0% found this document useful (0 votes)
15 views298 pages

Proceeding of 1st International Conference on Environmental Protection and Disaster RISKs

The 1st International Conference on Environmental Protection and Disaster Risks was held on September 29-30, 2020, in Sofia, Bulgaria, focusing on topics such as air pollution, biodiversity, and environmental monitoring. The proceedings include studies on the impacts of climate change, natural hazards, and the development of forecasting systems for environmental threats. The document outlines the objectives of a national research program aimed at ensuring a sustainable living environment in Bulgaria through comprehensive studies and the establishment of a National GeoInformation Center for better disaster management.
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
15 views298 pages

Proceeding of 1st International Conference on Environmental Protection and Disaster RISKs

The 1st International Conference on Environmental Protection and Disaster Risks was held on September 29-30, 2020, in Sofia, Bulgaria, focusing on topics such as air pollution, biodiversity, and environmental monitoring. The proceedings include studies on the impacts of climate change, natural hazards, and the development of forecasting systems for environmental threats. The document outlines the objectives of a national research program aimed at ensuring a sustainable living environment in Bulgaria through comprehensive studies and the establishment of a National GeoInformation Center for better disaster management.
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 298

Proceeding

of 1st International Conference


on Environmental Protection
and Disaster RISKs

Part One

29 – 30 September 2020
Sofia, Bulgaria

Az-buki National Publishing House

Sofia 2020
The 1st International Conference on Environmental Protection and
disaster RISKs took place on 29 – 30 September 2020 in Sofia, Bulgaria.

Part One of the proceeding book presents texts on the following topics:
Air Pollution, Climate and Health; Biodiversity; Informatics, Remote
Sensing, High Performance Computing and GIS for Environmental
Monitoring and Management.

In Part Two of the proceeding book will be presented texts on the


following topics: Natural Hazards and Risks; Water Resources, Human
Activities and Management.

Proceeding of 1st International Conference on Environmental


Protection and disaster RISKs – Pаrt One

Editors:
Georgi Gadzhev
Nina Dobrinkova

Cover Design: Ivo Hristov


Stylist-corrector: Anelia Vracheva
Preprint: Az-buki National Publishing House, www.azbuki.bg

Print: Alliance Print


Size: 70/100/16
Printed Quires: 18,5

ISBN 978-619-7065-38-1
е-ISBN 978-619-7065-39-8
National research program
Environmental protection and minimization of unfavorable events
and natural disasters hazards

The general Program objectives are carrying out fundamental and applied
studies aiming at ensuring sustainable, favorable and safe living environment for
the population of the Republic of Bulgaria

The specific Program objectives are Fundamental studies. Development and/


or adoption of methods and carrying out of reliable, comprehensive and detailed
studies of:
1. Spatial/temporal variability of the parameters and characteristics of the
atmosphere and their impact on quality of life, health risks and ecosystem status;
2. Quality of waters atmosphere and their impact on quality of life, health risks
and ecosystem status;
3. Quality of urban environment (buildings, transport and technical infrastructure,
green areas, public and open areas) and their impact on quality of life, health risks
and ecosystem status;
4. Recurrence and spatial distribution of extreme, unfavorable and catastrophic
natural phenomena (droughts, storms, hail, floods, fires, sea waves, soil erosion,
etc.) and their connection with atmospheric processes and climate changes; risk
assessment and evaluation of the possibilities for forecast, early warning and
prevention;
5. Geological environment and geological hazards - recurrence and spatial
distribution; risk assessment and evaluation of the possibilities for forecast, early
warning and prevention;
6. Hazards for biodiversity and ecosystems functions as a component of
sustainable and favorable living environment;
7. Reaction and adaptation of ecosystems to unfavorable and catastrophic events
in the geological past (global warming and cooling, collisions with space objects) –
a key to understanding the current and predicting the future changes.

Other specific Program objectives are applied studies – elaboration, on the bases
of the knowledge, generated by the fundamental studies of:
1. Systems for forecasting of unfavorable and catastrophic natural phenomena;
2. Early warning systems
3. Systems assisting the actions during catastrophic events and disasters;
4. Long-term strategies and measures for preventing and overcoming of
unfavorable and catastrophic phenomena;
5. Enriching the knowledge about mechanisms of biodiversity loses in Bulgaria
and the Black Sea in climate change conditions, their connections with local human
activities, possible pathways for mitigation and adaptation to negative consequences.
Enhancing the qualification of the scientists and developing of up-to date research
infrastructure, Dissemination of the scientific results, Transfer of knowledge to the
decision making bodies (elaboration of scientifically robust strategies and programs
for sustainable government) and Promoting multi-disciplinary partnership between
research teams from different scientific fields are also part of the general Program
objectives.

Expected results
Fundamental results:
1. Generation of new knowledge about the processes in the atmosphere,
hydrosphere, lithosphere and biosphere in regional and local scales, about their
interactions and their impact on quality of life, health and ecosystem status.
2. Accounting for the multi-scale nature of the processes.
3. Tracking of the main mechanisms and pathways for forming the characteristics
the atmosphere, hydrosphere, lithosphere and biosphere and respectively their
impact on quality of life, health and ecosystem status and the genesis of different
natural disasters.

4. Applied results:
5. Evaluation of the spatial/temporal variability (mapping) of the unfavorable
events risk: extreme seismic events with a big reccurrence period, extreme
temperatures, bad air and water quality, draughts, floods, storms, hail, icing, fires,
geological disasters, biodiversity loses, etc.
6. Systems for forecasting of unfavorable and catastrophic natural phenomena;
7. Early warning systems;
8. Early warning systems for ecosystem changes with elements of worsening of
the living environment or bio resource or other ecosystem services loss;
9. Systems assisting the actions during catastrophic events and disasters;
10. Long-term strategies and measures for preventing of unfavorable and
catastrophic phenomena, worsening of the living environment or bio resource or
other ecosystem services loss;
11. Strategic measures for emission reduction aiming at reduction of the health
risks and impact on ecosystems.
12. Elaboration of comments and recommendations for the National action plan
for adaptation to climate changes.
NATIONAL GEOINFORMATION CENTER

The NATIONAL GEOINFORMATION CENTER (NGIC) is a newly established


scientific infrastructure for cooperation and integration of human resources and
informational products and data on the geo-environment, their complex analysis and
creation of mathematical models. It is part of the National road map for scientific
infrastructure (2017-2023), adopted by the Council of Ministers of Bulgaria. The
partners joined in the research consortium are four institutes working in the field
of Earth observation: the National Institute of Geophysics, Geodesy and Geography
(NIGGG-BAS), the National Institute of Meteorology and Hydrology (NIMH),
the Institute of Oceanology (IO-BAS), the Geological Institute (GI-BAS), and two
institutes competent in ICT: the Institute of Mathematics and Informatics (IMI-BAS)
and the Institute of Information and Communication Technologies (IICT-BAS).
NGIC is a distributed scientific infrastructure with national territorial
coverage and integrates monitoring networks, observatories, observation stations,
laboratories, computing centers and other specialized equipment of the participating
partners. Existing technical resources include unique facilities, scientific equipment
and computer networks.
The creation of single, dynamic ICT-based network will ensure the development
of multidisciplinary, broad-spectrum of geo products to serve a wide range of
users - government, local authorities, business and the public, to solve important
national tasks for forecasting and prevention of natural and anthropogenic risks and
disasters.
The impact and benefits expected from the implementation of the project include:
■ Response to natural and anthropogenic disasters and accidents and creation of
prevention plans, including dangerous meteorological conditions and related health
problems;
■ Sustainable urban development; land use; design and implementation of large
infrastructure projects (gas pipelines, highways, water supply, etc.);
■ Improved management of water resources through a better understanding of
the water cycle, improving the management and protection of terrestrial, coastal
and marine ecosystems;
■ Analysis of environmental factors affecting human health and well-being, as
well as assessing, forecasting, mitigating and adapting to climate change;
■ Raised public awareness of natural disasters and industrial accidents through
new interactive products;
■ Support to the transfer of scientific results in operational technologies by
encouraging cooperation and collaboration between research communities.
■ Support to the research and development in key areas of Earth sciences,
contributing to advances in science and technology.
CONTENTS

AIR POLLUTION, CLIMATE AND HEALTH


8 Winter and Summer Variability and Trends of Modis Derived Sea Surface
Temperatures for the Western Black Sea / Nadezhda Valcheva, Violeta
Slabakova
21 Variations of Temperature over Bulgaria and their Connection with Solar
Cycles / Yavor Chapanov
33 Study of One Month High Polluted Pm Event for Sofia Region / Margret
Velizarova, Reneta Dimitrova
44 Satellite Data Assimilation of Air Quality Parameters in Bulgaria / Dimiter
Syrakov, Maria Prodanova, Emilia Georgieva
53 Preliminary Results for the Recurrence of Air Quality Index for the City of
Sofia from 2008 to 2019 / Georgi Gadzhev
65 Particulate Matter Characteristics and Atmospheric Boundary Layer Height
over Sofia / Plamen Savov, Nikolay Kolev, Ekaterina Batchvarova, Hristina
Kirova, Maria Kolarova
79 Sensitivity to the Wrf Model Configuration of the Wind Chill Index for Sofia
Region – Preliminary Results / Vladimir Ivanov, Reneta Dimitrova
90 Modelling of the Sulphur and Nitrogen Depositions over the Balkan
Peninsula by CMAQ and EMEP-MSC-W – Preliminary Results / Georgi
Gadzhev, Vladimir Ivanov
101 How Well Do the Air Quality Models EMEP and CAMS Reproduce
Particulate Matter Surface Concentrations in Bulgaria / Hristina Kirova,
Nadya Neykova, Emilia Georgieva
112 Cloud and Rain Water Chemical Composition at Peak Cherni Vrah, Bulgaria /
Elena Hristova, Blagorodka Veleva, Krum Velchev, Emilia Georgieva
123 Characteristics of Ultrafine PM Emitted During Gazification of Biomass /
Iliyana Naydenova, Tsvetelina Petrova, Ognyan Sandov, Ricardo Ferreira,
Rositsa Velichkova, Mario Costa
134 Black Sea Freezing and Winter Severity: Relation to the Temperature / Mirna
Matov, Elisaveta Peneva, Greta Georgieva
144 Assessment of the Future Climate over Southeast Europe Based on CMIP5
Ensemble of Climate Indices – Part One: Concept and Methods / Hristo
Chervenkov, Vladimir Ivanov, Georgi Gadzhev, Kostadin Ganev
157 Assessment of the Future Climate over Southeast Europe Based on CMIP5
Ensemble of Climate Indices – Part Two: Results and Discussion / Hristo
Chervenkov, Vladimir Ivanov, Georgi Gadzhev, Kostadin Ganev
170 Application of Backward Trajectories and Cluster Analysеs for Study of
Variations in PM10 Concentrations / Rozeta Neykova, Elena Hristova
180 Relationship between Particulate Matter and Health Indicators for Acute
Morbidity in Sofia / Petar Zhivkov, Aleksandar Simidchiev
194 Nocturnal Boundary-Layer Characteristic in a Costal Site Using Long-Term
Sodar Data / Damyan Barantiev, Ekaterina Batchvarova
205 Environmental Impact of Quarry Activities: the Case Study of a Gneiss
Deposit in the Region of Blagoevgrad / Radostina Angelova, Rositsa
Velichkova, Sylvia Alexandrova

BIODIVERSITY
216 Peniophora Proxima (Peniophoraceae, Basidiomycota): Newly Recorded
for the Bulgarian Mycota / Aneta Lambevska-Hristova, Svetlana Bancheva
221 Applicability of Control Region of Mitochondrial DNA for Assessment of
Turbot Populations Along the Bulgarian Black Sea Coast / Petya Ivanova,
Nina Dzhembekova, Ivan Atanassov, Krasimir Rusanov, Violin Raykov,
Ivelina Zlateva, Maria Yankova

INFORMATICS, REMOTE SENSING, HIGH


PERFORMANCE COMPUTING AND GIS
FOR ENVIRONMENTAL MONITORING AND MANAGEMENT
231 Open Source GIS Decision Support Tools for Wildfire and Flood Natural
Hazards with Embeded Interactive Training Resources / Stefan Stefanov,
Nina Dobrinkova, Emilia Tosheva
241 Methodology for Calculating the Parameters of Radio Paths with Ionospheric
Reflection / Rumiana Bojilova, Plamen Mukhtarov
249 GIS in Crises Management Use of Smart Phone GIS Application in the
Event of Natural Disaster / Aleksandar Petrovski, Nenad Taneski, Jugoslav
Ackoski
259 Framework for Wildfire Danger Prediction System / Nina Dobrinkova,
Maria Asenova
271 Evaluation of Sentinel-3a OLCI Ocean Color Products in the Western Black
Sea / Violeta Slabakova, Snejana Moncheva, Natalyia Slabakova, Nina
Dzhembekova
282 A Modelling System for Numerical Weather Prediction in the Bulgarian
Antarctic Base Area / Boriana Chtirkova, Elisaveta Peneva, Gergana
Georgieva
Proceeding of 1st International conference on Environmental protection and disaster RISKs

https://doi.org/10.48365/envr-2020.1.1 Air Pollution,


Climate and Health

WINTER AND SUMMER VARIABILITY


AND TRENDS OF MODIS DERIVED SEA SURFACE
TEMPERATURES FOR THE WESTERN BLACK SEA

Nadezhda Valcheva, Violeta Slabakova


Institute of Oceanology – Bulgarian Academy of Sciences (IO-BAS)

Abstract: The study investigates on sea surface temperature (SST) seasonal to


interannual variability in the western Black Sea using gridded fields of MODIS
derived SST monthly averages of 4 km horizontal resolution. More specifically,
winter and summer SSTs products were considered spanning a 17-year period
(2003-2019).To this end, seasonal climatic mean fields were obtained and observed
features were discussed from a hydrodynamic view point. Furthermore, time series
of winter and summer SST means were analysed to reveal the spatial variability
pattern. In terms of zonal flow, the winter SST increases eastward, while the
summer ones - from northwest to southeast. Both winter and summer SSTs exhibit
increasing long-term trends; there is almost uniform increase of the winter SSTs
of 0.6-0.7°C per decade for all zonal areas, while the summer ones tend to vary in
somewhat wider range – 0.5-0.7°C per decade. Besides, based on the normalized
SST anomalies, it was found out that during 2003-2019 SSTs have grown with a
lower rate compared to the two preceding decades. Finally, in attempt to link the
SST variability to global teleconnection patterns such as North Atlantic Oscillation
(NAO), a spectral analysis of seasonal time series was performed, which indicated
the existence of quasi-decadal scale of variation that roughly follows the NAO
positive and negative cycles.
Keywords: Black Sea, sea surface temperature, remote sensing, seasonal
anomaly, decadal variability

INTRODUCTION
Sea surface temperature (SST) is the temperature of the extremely thin inter-
face between the ocean and the overlying atmosphere, which controls heat and
gas exchange. SST is also the longest and most widely measured parameter in the
ocean (Emery, 2015). In the Black Sea, it influences the water circulation and eco-
logical condition of the basin and has an impact on the temperature of the Cold
Intermediate Layer (CIL). The temperature of the surface layer and SST respond
most quickly to the atmospheric impacts, which is the reason why SST experiences
8
Winter and Summer Variability and Trends of Modis...

a significant spatial and temporal (interannual, seasonal, and synoptic) variability


(Ginzburg et al., 2008). Therefore, it is important to investigate about the spatial
SST variability on intraannual (seasonal) and inter-annual scales as well as relevant
long-term trends.
In the last four decades, Black sea SST variability has become a subject of
increased scientific interest. The first studies on interannual variability of temper-
ature started to be published in the 1980s and 1990s after collecting large series of
quasi-regular measurements launched back in the 1950s.(Polonsky et al., 2013).
Some of the studies are based on hydrologic observations. For example, Polonsky
and Lovenkova (2004) analysed the long-term variability and trends in the upper
layer temperature during the cold half of the year using contact observation for the
period 1960–1990. They report typical cooling values of 1°C/10 yr is located in the
northern part of the studied standard section. The temporal trends and interannual
to interdecadal scale SST changes during the 20th century over contrasting shelf and
deep sea regions were presented in Shapiro et al. (2010).Analysis of the time se-
ries has revealed decoupling between the shelf and deep sea regions in response to
climate change. The deep sea experienced general cooling, which was particularly
strong from the beginning of the century until the late 1960s. The western shelf did
not exhibit a particular trend, except for the last 15-20 years when warming was
detected.
However, only the satellite infrared measurements can represent the SST of the
10 µm thick skin layer of the ocean, opposed the bulk SSTs measured by the buoys
and ships in the layer 1-5 m (Emery, 2015). Furthermore, due to their larger cover-
age, satellite datasets give the opportunity to investigate interannual SST variability
and to reveal long-term climatic trends not only on the basin wide scale but also
for smaller regions of particular dynamics (Ginzburg et al., 2004).For example,
night-time weekly AVHRR measurements were used to investigate seasonal and
interannual variability of the Black Sea SST during 1982-2000. Analysis of the
weekly SST means were used to calculate winter and summer SST anomalies as
well as a long-term trend. As a result a positive basin wide trend of about 0.09°C
yr-1 has revealed, which however was different for the western and eastern deep sea
regions – about 0.08 and 0.11°C yr-1, respectively (Ginzburg et al., 2004). Polonsky
et al. (2013) discovered a quasi-decadal SST variability with winter and summer
fluctuations differing in amplitudes and phases. The difference between the temper-
ature values from satellite and contact measurement data on average for all decades
did not exceed 0.3°C. In winter, the large-scale SST spatial distribution was char-
acterized by temperature increase from north to south (in the north-western shelf
with high horizontal temperature gradients) and from west to east in the deep water
region, while in summer, the increase was observed from northwest to southeast.
Furthermore, the long-term variability of the temperature fields in the Black
Sea can serve as a reliable indicator of regional climate change. During the 20th
9
Nadezhda Valcheva, Violeta Slabakova

century, the Black Sea surface temperature had been constantly increasing due to
climate change. Sakalii and Başusta (2018) analysed SST data covering the period
1982-2015 and predicted the Black Sea SST until 2100. They found out that SST
increased with a rate of 0.64°C per decade thus conforming to the worst climate
change scenario. By the end of 21stcentury, the relative increase of the average SST
was predicted to be 5.1°C.
Moreover, SST spatial variations can be used as an indicator for the large- and
mesoscale water dynamics. For example, SST data were used to explore the Black
Sea heat balance underpinning the relevance of the fundamental oceanographic
issue about how close the SST reproduces the air temperature (Esbensen and Reyn-
olds, 1981; Gulev, 1994). Stanev and Staneva (2001) demonstrated that large ar-
ea-dependent differences could exist between the SST and air temperature, creating
an important eddy-dominated component of the ocean-atmospheric heat exchange.
Finally, SST exerts an impact on the ocean primary productivity through both
physical effect on water stratification (which in turn affects nutrient availability)
and its biological effect on plankton metabolic rates. The rise of global mean SST
over the past century is linked with both decreases and increases in primary pro-
ductivity depending on the time period and the region (Belkin, 2016). Oguz (2005)
suggests that climate plays a crucial role in the Black Sea ecosystem state fluc-
tuations. During the 1980s and 1990s, the existing atmospheric and hydrological
processes, together with nutrient supply, led to an enhanced plankton productivity,
which ultimately resulted in an increased jellyfish production and a longer over-
fishing period. On the other hand, a decreasing trend in Chlorophyll a and primary
production has been observed worldwide as SST increases, suggesting that reduced
phytoplankton biomass in the Black Sea ecosystem, hence its recovery, is not solely
a factor of diminished nutrient loading (McQuatters-Gollop et al., 2008).
Therefore, the proposed study aim to assess the spatial SST variability on sea-
sonal and interannual scales as well as relevant long-term trends making use of
remote sensing data, and to detect variability cycles possibly related to global tele-
connection patterns.

DATA AND METHODS


Study domain encompasses the western Black Sea between 27.5-30.0°E and 42.0-
44.0°N (Fig. 1). It includes near-shore (coastal) and shelf (inner and outer), spreading
meridionally along the shore curve, and open (deep sea) areas. The SST data used in
this study represent monthly global binned MODIS/AQUA Level 3 standard mapped
night-time images (Brown & Minnett, 1999) at 4 km horizontal resolution spanning
the period 2003-2019. The data were obtained from Ocean Color website provided by
National Aeronautics and Space Administration (NASA). For extraction of the fixed
region of interest and highest quality data from each monthly composite SST images,
the SeaDAS v.7.5 (Fu et al., 1999) software was used.
10
Winter and Summer Variability and Trends of Modis...

Previously, Slabakova & Valcheva (2007) have performed validation of MODIS


SSTs against in situ (CTD) sea water temperature for the western Black Sea. Match-
up data were extracted for a period of four years (2003-2006). A strong correlation
(0.953) between the two data types was ascertain, while root-mean square errors
and biases for separate datasets (cruises) were found to vary in ranges 0.18-0.30˚C
and 0.278-0.450˚C, respectively. This indicates a good agreement with standards
established in EOS Data Product Handbook (2000).

Figure 1. Scheme of the study domain. Location of points, for which seasonal
SST time series were derived, are illustrated as coloured rings as follows: coastal
stations in blue, inner shelf ones in red, outer shelf in green, open sea in violet

In order to achieve the study aims, mean winter and summer climatic fields are
constructed by averaging the monthly gridded fields over the time-span of interest
(entire study period or separate years within it). Seasons are defined as three-month
periods: winter includes January, February and March, and summer - July, August
and September.
Furthermore, the SST normalized anomaly (referred hereafter as anomaly)
(˚C) is calculated using the following formulation:

(1)
11
Nadezhda Valcheva, Violeta Slabakova

where is the seasonal mean for each year, – climatic seasonal mean for
the entire 17-year period, – respective standard deviation. The overall anom-
aly is computed by averaging the seasonal normalized anomaly over appropriate
number of years. Spectral analysis was performed applying Fast Fourier Trans-
formation on the mean seasonal SST scalar time series using estimates of spectral
density as mean values of periodogram.

RESULTS AND DISCUSSION


Seasonal mean SST climatic fields, presented in Fig. 2, are very similar to those
derived from in-situ CTD bulk temperature data spanning the period 1995-2010
(Valcheva, 2014). More specifically, the winter climatic SST distribution reveals
the main features of typical seasonal dynamics, namely presence of narrow frontal
zone of sharp horizontal gradients between coastal and inner shelf waters, on one
hand, and outer shelf and deep basin, on the other, defined between 6.8 and 7.6°C
isotherms; colder coastal waters of temperatures lower than 6.2°C; and vast area of
quasi uniform open sea waters of temperatures about 8°C. The front is determined
by intensification of the large-scale circulation – the Rim current flow is strongly
pressed against the basin lateral boundaries – intense mixing of surface waters and
restricted water exchange with the northwestern shelf.
In summer, the relatively cooler waters in the northernmost part of the domain
are marked by temperatures lower than 24.1°C. Farther within the western shelf,
these waters are being transformed while their impact can be traced along the
entire shelf curve. The frontal zone is less discernable and considerably wider
with respect to winter, which is determined by the weakening and instability
of the large scale circulation (lower Rim current velocity) and existence of
quasi-stable (recurrent) anticyclonic eddies. With this respect, Burgas eddy is
visible on the SST climatic distribution on the coastal side of the Rim current
in both seasons marked by temperatures higher than 24.3°C in summer and
lower than 6°C in winter, that is being slightly warmer and cooler, respectively,
with respect to the surrounding waters. The same is valid for Kaliakra eddy that
contains cooler waters in both seasons – less than 6.4°C in winter and 24.0°C
in summer. Thus, the summer circulation pattern allows for horizontal water
exchange between the periphery of the large-scale gyres and the shelf. The
process of entrainment of warmer open sea waters (SST>24.5°C) in the shelf
and near-shore zone, which are being spread to the north, is particularly visible
in the middle and southern part of the study area. The summer SSTs increase
from northwest to southeast as discovered also by Polonsky et al. (2013). Spatial
temperature variability ranges in less than 0.7°C, i.e. three to four times smaller
compared to the winter (2.6°C).

12
Winter and Summer Variability and Trends of Modis...

Figure 2. Winter (left) and summer (right) mean SST (°C) climatic spatial
distribution over the western Black Sea for 2003-2019

With this respect, the analysis of seasonal time series derived for locations in
near-shore (coastal), inner and outer shelf, and open sea can be particularly useful.
Therefore, 23 points were selected normal to the general exposure of the shore (Fig.
1). In Table 1 seasonal SST minima and maxima for mentioned areas as well as the
maximum and average temperature differences are summarized. This allows for
more thorough examination of the SST spatial variability.

Table 1. SST minima / maxima and relevant temperature differences


for areas of particular surface dynamics
SST min (°C) SST max (°C) SST diff. max (°C) SST diff. average (°C)
Area
winter summer winter summer winter summer winter summer
Coastal 3.9 22.6 7.8 25.2 2.5 0.7 1.2 0.5
Inner shelf 4.2 22.9 7.9 25.2 2.5 0.7 1.0 0.4
Outer shelf 4.8 22.9 8.5 25.1 2.8 0.6 1.4 0.3
Open sea 6.5 22.8 9.2 25.2 0.9 0.9 0.4 0.4

In meridional direction, SST increases southward.This increase is more pronounced


in coastal and shelf areas, particularly in winter – about 2-3°C (1.0-1.4°C on the
average). In summer, this pattern is less discernible and the temperature difference
does not exceed 0.7°C (0.3-0.5°C on the average). Deep water SSTs do not differ
noticeably between northern and southern regions, in both seasons, as the maximum
difference is less than a degree and the average one is 0.4°C (Table 1, Fig. 3). In terms
of zonal flow, the winter SST increases eastward, which means that the cooling of
coastal waters is the most marked. The temperature differences between near-shore
13
Nadezhda Valcheva, Violeta Slabakova

and open sea regions experience significant fluctuations of 1.5-3.8°C. In summer, the
opposite takes place as the coastal waters were heated more than the offshore ones.
However, the difference is so minor (less than an order of magnitude compared to the
winter one) that it is not greater than a couple of tenths of a degree (Table 1, Fig. 3).
From hydrodynamic point of view, the open sea waters are part of a large scale
circulation cell, namely the western gyre, and the small spatial variability can be
attributed to the more intense mixing that takes place there. On the other hand,
the water flow in shelf and coastal areas is mainly governed by the mesoscale
circulation features, which defines the greater spatial variability observed. In
winter, it is still more uniform due to the intensified circulation, while in summer
the pattern may considerably differ between areas due to circulation instability and
presence of eddies. Interannual variability of zonal temperature differences is much

Figure 3. Time series of winter (A) and summer (B) SST climatic means obtained
for zonal stretchesfor2003-2019: coastal (C), inner shelf (S1), outer shelf (S2)
and deep sea (D). Respective linear trends are represented with dashed line
14
Winter and Summer Variability and Trends of Modis...

more pronounced than the meridional ones, in winter in particular. For this season,
the range is largest in the beginning – 2003-2004 (about 3°C) – and the middle of
the study period – 2010-2011 (about 3.5°C) – whereas it is the smallest in 2007
(0.5°C). Thus, 2010-2011 data sustain the hypothesis for decoupling between open
sea and periphery. The variability of summer zonal temperature differences appears
to be more uniform with a steady decreasing trend.
Several authors (e.g. Oguz et al., 2003; Stanev et al., 2014) suggested that the
erosion of the Black Sea upper layer stratification is mainly associated with the
atmospheric warming of the sea surface and winter SST can be used as indicator of
the cooling intensity (Piotukh et al., 2011). With this respect, the warmest winters
were those of 2007, 2014 and 2016 (SST maxima of 7.9-8.6°C), as well as the
winters of 2009-2010 for the open sea area with absolute maximum of 8.7°C (Fig.
3A). During such winters, the cooling of surface waters is insufficient leading to
limited ventilation of the upper layer and shrinking of cold intermediate layer. Such
phenomena were reported in 2011 and 2015-2016 (Stanev et al., 2019; Valcheva
et al., 2020). On the other hand, the coolest winters were in 2006, 2012 and 2017
when SSTs in near-shore waters dropped to 4.3-4.8°C and in open sea areas – to
6.8-7.7°C, which assisted the recovery of the vertical stratification. The warmest
summers took place in 2010, 2012 and 2016 when SSTs reached 25°C.This suggests
that periods of the SST rise and fall alter every 5-6 years. Moreover, the fluctuations
are synphase in all areas (especially well pronounced for summer, see Fig. 3B)
except for winters of 2008-2010. Besides, the winter SST fluctuations get smoother
in offshore direction.
Both winter and summer SSTs exhibit increasing long-term trends (Fig. 3).
Linear approximation predicts almost uniform increase of the winter SSTs of 0.06-
0.07°C yr-1for all zonal areas, while the summer ones tend to vary in somewhat
wider range – 0.05-0.07°C yr-1 – as the near-shore SSTs rise is more pronounced.
These values are in agreement with the findings of Sakalii and Başusta (2018) and,
having in mind those reported by Ginzburg et al. (2004), it seems that during 2003-
2019 SSTs have grown with a lower rate compared to the two preceding decades.
Furthermore, spatial distribution of the anomaly calculated using (1) averaged
over the study period is presented in Fig. 4. It can be seen that the summer anomalies
(ranging within ±0.6°C) are more than ten times greater than the winter ones
(roughly ±0.04°C). While the summer anomalies field is smoother and resembles the
mean climatic SST distribution, the winter one shows a more varied aspect. Thus,
the water flow from the northwestern shelf was marked by negative summer SSTs
anomalies at the same time that the warming took place mainly in the open sea area
(positive anomalies within the western gyre in particular). Through the process of
entrainment, these warmer waters had contributed to the positive anomaly observed
on the southern shelf and adjacent coastal waters. Nevertheless, small negative SST
anomaly is present within the middle part of the shelf (~- 0.2°C).
15
Nadezhda Valcheva, Violeta Slabakova

Figure 4. Winter (left) and summer (right) spatial distribution of the average
SST normalized anomaly (°C) over the western Black Sea for 2003-2019

As for the winter SST anomaly field, the negative anomaly can be ob-
served in deep sea area; there is neutral anomaly over the larger part of the
shelf and positive one over certain spans of the middle and mostly southern
shelf (especially in front of Burgas Bay). The slight cooling of the open sea
surface might have contributed to the reduced SST growth rate during the
last 15-20 years. As a whole, the result suggest that the trend of decoupling
between the basin periphery and deep sea region, detected by Shapiro et al.
(2010) and attributed to the climate change, is still present in first two de-
cades of the 21stcentury.
The cycle of SST interannual variability suggested above agrees with previous
studies (e.g. Oguz et al., 2006; Capet et al., 2012), which imply the existence
of teleconnection between Black Sea regional atmospheric condition and large
scale atmospheric pattern driven by the North Atlantic Oscillation (NAO).
More specifically, this refers to a sequence of cold and mild winter cycles with
approximately 5-year duration that roughly follows the NAO positive and
negative cycles. Similar fluctuations were detected for mixed layer temperature
(Valcheva, 2018) and cold intermediate layer thickness (Valcheva et al., 2020). In
attempt to confirm this finding with the SST dataset used in this study, a spectral
analysis of seasonal time series was performed. The result revealed that the
largest spectral density corresponds to the highest frequency fluctuations of two
years. However, the second peak indicates an8-year cycle of variability, which is
more discernable for the winter time series. Such quasi-decadal fluctuations agree
with the observed peak in 8-10 year band of the winter mean NAO index power
spectrum, which was found to enhance particularly over the second half of the
20th century (Polonsky et al., 2004).
16
Winter and Summer Variability and Trends of Modis...

CONCLUSIONS
The study dealt with the spatio-temporal variability and related long-term
trends of winter and summer sea surface temperature derived from MODIS remote
sensing data. Thus, climatic SST field reveals the main features of typical seasonal
hydrodynamics: in winter the temperatures vary between 6.2°C in near-shore zone
to about 8°C in the open sea presenting a marked narrow transition zone of strong
horizontal gradients, whereas in summer relatively cooler waters coming from
the northwestern shelf mixed with warmer ones (SST~24.5°C) as they flow in the
domain due to instability of the large scale circulation. It was confirmed existence
of quasi-stable anticyclonic mesoscale features, such as Burgas and Kaliakra eddies
on the coastal side of the Rim current, as well as the process of entrainment of open
sea waters in the shelf and near-shore zone, which are being spread to the north.
In summer, spatial temperature variability ranges in less than 0.7°C, i.e. three
to four times smaller compared to winter (2.6°C). In meridional direction, SST
increases southward. This increase is more pronounced in coastal and shelf
areas, particularly in winter – about 2-3°C. In terms of zonal flow, the winter
SST increases eastward, while the summer ones – from northwest to southeast.
Interannual variability of zonal temperature differences is much more pronounced
than the meridional ones, in winter in particular.
With respect to the increased atmospheric warming of the sea surface and use of
winter SST as indicator of the cooling intensity, warmest and coolest winters have
been identified –2007, 2014 2009-2010, and 2016 (7.9-8.7°C); and 2006, 2012
and 2017 (4.3-4.8°C). Both winter and summer SSTs exhibit increasing long-term
trends; there is almost uniform increase of the winter SSTs of 0.6-0.7°C per decade
for all zonal areas, while the summer ones tend to vary in somewhat wider range –
0.5-0.7°C per decade. The results suggest that during 2003-2019 SSTs have grown
with a lower rate compared to the two preceding decades.
Furthermore, spatial distribution of the averaged normalized anomalies reveals
that the summer anomalies (±0.6°C) are more than ten times greater than the winter
ones (±0.04°C). While the summer anomalies field is smoother and resembles the
mean climatic SST distribution, the winter one shows a more varied aspect. As a
whole, the outcome confirms the trend of decoupling between the basin periphery
and deep sea region. Finally, the spectral analysis of seasonal time series confirmed
the existence of a quasi-decadal SST variability cycle, which agrees with the
observed peak of the winter mean NAO index power spectrum.

ACKNOWLEDGMENTS
This work has been carried out in the framework of the National Science Program
"Environmental Protection and Reduction of Risks of Adverse Events and

17
Nadezhda Valcheva, Violeta Slabakova

Natural Disasters", approved by the Resolution of the Council of Ministers №


577/17.08.2018 and supported by the Ministry of Education and Science (MES)
of Bulgaria (Agreement № Д01-322/18.12.2019).

REFERENCES
Belkin, I. (2016). Chapter 5.2: Sea surface temperature trends in large
marine ecosystems. In IOC-UNESCO and UNEP (2016). Large Marine
Ecosystems: Status and Trends, United Nations Environment Programme,
Nairobi, 101-109.
Brown, O. & Minnett, P. J. (1999). Modis Infrared Sea Surface Temperature
Algorithm. Algorithm Theoretical Basis Document, Version 2.0, modis.
gsfc.nasa.gov/data/atbd/atbd_mod25.pdf.
Capet, A., Barth, A., Beckers, J.-M. & Grégoire, M. (2012). Interannual
variability of Black Sea's hydrodynamics and connection to atmospheric
patterns. Deep Sea Research Part II: Topical Studies in Oceanography,
77–80, 128-142.
Emery, W. J. (2015). Sea Surface Temperature. Encyclopedia of Atmospheric
Sciences 2nd Edition, Vol. 1, 136-143.
EOS Data Product Handbook Volume 2 (2000), Parkinson, C. L. &
Greenstone, R. (Eds). NASA Goddard Space Flight Greenbelt,
Maryland, 256.
Esbensen, S.K. & Reynolds, R.W. (1981). Estimating monthly averaged
air-sea transfers of heat and momentum using the bulk aerodynamic
method. J. Phys. Oceanogr. 11, 457–465.
Fu, G., Baith, K. & McClain, C. (1999). The SeaWiFS Data Analysis System.
Proceedings of the 4th Pacific Ocean Remote Sensing Conference, 73–79.
Ginzburg, A. I., Kostianoy, A. G. & Sheremet, N. A. (2004). Seasonal and
interannual variability of the Black Sea surface temperature as revealed
from satellite data (1982-2000). Journal Marine Systems. 52, 33–50.
Ginzburg, A. I., Kostianoy, A. G. & Sheremet, N. A. (2008). Sea surface
temperature variability. The Black Sea Environment. The Handbook of
Environmental Chemistry. Vol. 5: Water Pollution, Part 5/Q, Kostianoy
A. G. and Kosarev A. N. (Eds.). Springer-Verlag Berlin Heidelberg,
255–276.
Gulev, S.K. (1994). Influence of space-time averaging on the ocean-
atmosphere exchange estimates in the North Atlantic midlatitudes. J.
Phys. Oceanogr. 24, 1236–1255.
McQuatters-Gollop, A., Mee, L.D., Raitsos, D.E. & Shapiro, G.I. (2008).
Non-linearities, regime shifts and recovery: The recent influence of
climate on Black Sea chlorophyll. Journal of Marine Systems 74,
649–658.
18
Winter and Summer Variability and Trends of Modis...

Oguz, T. (2005). Black Sea ecosystem response to climatic teleconnections.


Oceanography, Vol. 18(2), 122-133.
Oguz, T., Cokacar, T., Malanotte-Rizzoli, P. & Ducklow, H.W. (2003).
Climatic warming and accompanying changes in the ecological regime
of the Black Sea during 1990s. Global Biogeochem. Cycles, 17, 1088.
Oguz, T., Dippner, J.W. & Kaymaz, Z. (2006). Climatic regulation of the
Black Sea hydrometeorological and ecological properties at interannual
to decadal time scales. J. Mar. Syst., 60, 235–254.
Piotukh, V. B., Zatsepin, A. G., Kazmin, A. S. & Yakubenko, V. G. (2011).
Impact of the winter cooling on the variability of the thermohaline
characteristics of the active layer in the Black Sea. Oceanology 51,
221–230.
Polonsky, A. B. & Lovenkova, E. A. (2004). Temperature and salinity
trends in the active layer of the Black Sea in the second half of the 20th
century and their possible causes. Izvestiya, Atmospheric and Oceanic
Physics, Vol. 40 (6), 747–755.
Polonsky, A.B., Basharin, D.V., Voskresenskaya, E.N. & Worley, S. (2004).
North Atlantic Oscillation: description, mechanisms, and influence on
the Eurasian climate. Phys. Oceanogr., 15 (2), 96-113.
Polonsky, A. B., Shokurova, I. G. & Belokopytov, V. N. (2013). Decadal
variability of temperature and salinity in the Black Sea. Journal of
Marine Hydrophysic, 16, 27-41. (In Russian).
Sakalii, A. & Başusta, N. (2018). Sea surface temperature change in
the Black Sea under climate change: A simulation of the sea surface
temperature up to 2100, Int. J Climatol.; 38, 4687-4698.
Shapiro, G. I., Aleynik, D. L. & Mee, L. D. (2010). Long term trends in the
sea surface temperature of the Black Sea. Ocean Sciences, 6,491–501.
Slabakova, V. & Valcheva, N. (2007). Estimation of feasibility of the Sea
Surface Temperature satellite data use through comparison against in situ
measurements for the western Black Sea Region. Proc. of the 3rd Int.
Scientific Conf. “Space Ecology Nanotechnology Safety”, 185-190.
Stanev, E.V. & Staneva, J.V. (2001). The sensitivity of the heat exchange
at sea surface to meso and sub-basin scale eddies: Model study for the
Black Sea.Dynamics of Atmospheres and Oceans 33, 163–189.
Stanev, E.V., He, Y., Staneva, J. & Yakushev, E. (2014). Mixing in
the Black Sea detected from the temporal and spatial variability of
oxygen and sulphide Argo float observations and numerical modelling,
Biogeosciences, 11 (20), 5707–5732.
Stanev, E. V., Peneva, E. & Chtirkova, B. (2019). Climate Change and
Regional Ocean Water Mass Disappearance:Case of the Black Sea. J.
Geophys. Res. C: Oceans 124.

19
Nadezhda Valcheva, Violeta Slabakova

Valcheva, N. (2014). Climatic distribution of summer and winter temperature


and salinity in the Western Black sea. Proc. of 12th Int. Conf. on “Marine
Sciences and Technologies - Black Sea’2014”, 189-195.
Valcheva, N. (2018). Interannual variability of the mixed layer summer sea
water temperature in the Western Black Sea. Proc. of 14th Int. Conf. on
“Marine Sciences and Technologies - Black Sea’2018”, 204-209.
Valcheva, N., Marinova, V., Peneva, E., Lima, L., Celiberti, S. & Masina, S.
(2020). Spatio-temporal variability of the Black Sea Cold Intermediate
Layer properties derived from in situ data. The Copernicus Marine
Environment Monitoring Service Ocean State Report. J. of Operational
Oceanography, 33-41.

 Nadezhda Valcheva
https://orcid.org/0000-0002-7690-8731
Institute of Oceanology
Bulgarian Academy of Sciences (IO-BAS)
Varna, Bulgaria
E-mail: [email protected]

 Violeta Slabakova
https://orcid.org/0000-0002-3089-0126
Institute of Oceanology
Bulgarian Academy of Sciences (IO-BAS)
Varna, Bulgaria
E-mail: [email protected]

20
Proceeding of 1st International conference on Environmental protection and disaster RISKs

https://doi.org/10.48365/envr-2020.1.2 Air Pollution,


Climate and Health

VARIATIONS OF TEMPERATURE OVER BULGARIA


AND THEIR CONNECTION WITH SOLAR CYCLES

Yavor Chapanov
Climate, Atmosphere and Water Research Institute –
Bulgarian Academy of Sciences (CAWRI-BAS)

Abstract. The Sun is the main source of energy for all Earth’s geosystems,
including climate, weather, mean sea level, winds, precipitation, and etc., mainly
through Total Solar Irradiance (TSI), whose variations during solar activity cause
various changes on the Earth surface. Climate processes, interactions between
atmosphere-and ocean system, various local, regional and global hydrological
cycles are the main mediator between solar activity and a number of geophysical
processes on the Earth surface. The temperature at the Earth surface is widely used
climate index, whose variations consist of significant seasonal oscillations, trend
and long-term cycles. The variations of temperature over Bulgaria is investigated
by means of several long time series of meteorological observations. The changes
of seasonal components of temperature and long-term oscillations are analyzed
in narrow frequency bands by means of the Partial Fourier Approximation. These
temperature variations are compared with the corresponding cycles of solar activity
in order to create long-term forecast models.
Keywords: solar activity; temperature; climate variations

INTRODUCTION
The solar activity affects terrestrial systems by means of direct radiation over
Earth surface, solar wind, and the solar magnetic field. The TSI cycles are the main
source of climate indices variations. The solar wind directly affects Earth magnetic
field, ionosphere and atmosphere. The variations of solar magnetic field modulate
solar wind and cosmic rays in the frame of the heliosphere. The cosmic rays (CR)
near Earth are modulated by Earth magnetic field variations, too. Recently a new
mechanism of climate modulation, based on cosmic rays variations, has been pro-
posed (Kilifarska et al., 2005, 2008; Kilifarska 2011; Velinov et al., 2005). This
mechanism is based on chain processes near tropopause by ozone production, tem-
perature variations, followed by vertical winds and water content change. The last
step of this chain affects surface temperature, because the atmospheric water is
one of the most powerful greenhouse gas. This model provides an explanation for
21
Yavor Chapanov

the cascade processes in which CR, whose total energy is relatively small, cause
climatic effects with much more energy. The solar activity cycles modulate CR
directly by the heliosphere and indirectly by the geomagnetic field changes, whose
effect is visible mainly at high latitudes. A significant part of cosmic rays consists
of charged solar particles, whose interaction with the Earth atmosphere is preceded
by focussing effect of geomagnetic field over polar regions, while the most ener-
getic galactic cosmic rays affect all Earth regions. The cosmic ray intensity exposes
antiphase variations to the TSI variations. The TSI is strong during solar activity
maximum, when the cosmic ray intensity has minimum and vice versa. So, during
TSI maxima, the warming processes occur on Earth surface, and during TSI min-
ima the thermal cycle amplitudes are amplified, due to cooling effects of cosmic
rays. The variations of Earth temperature are affected directly by TSI cycles and
indirectly by solar wind and solar magnetic field. These solar cycles are presented
by the indices of TSI, sunspot numbers (also known as Wolf’s number) and North-
South (N-S) solar asymmetry.
The knowledge of temperature variations and cycles is important in various
fields of human activity. An aspect of this knowledge is assessment of agromete-
orological conditions in national agricultural lands. The determination of statisti-
cally significant tendency of average annual air temperature is important for spring
crop growing (Georgieva, Shopova and Kazandjiev, 2019; Kazandjiev, Shopova,
Georgieva, 2018). The climate change and coming with them abiotic stressors, a
consequence of extreme weather conditions affect agriculture plant productivity
(Slavcheva-Sirakova et al., 2020; Shopova, Alexandrov and Todorova, 2019). The
goal of this study is to determine the thermal trends and changes of seasonal and
long-term components of temperature at several Bulgarian meteorological stations,
their connection with the cycles and harmonics of solar activity in order to improve
long term forecasts.

DATA
The used data consist of centennial time series of monthly temperature variations
from 4 Bulgarian stations in Vidin, Pleven, Razgrad, Dobrich and Sofia (Fig.1).
The solar data are presented by the Total Solar Irradiance (TSI) variations and
N-S solar asymmetry (Fig.2). The estimated TSI for the last 400 years is based on
the NRLTSI2 historical TSI reconstruction model (Kopp et al., 2016; Lean, 2010;
Coddington at al., 2016). The index Sa of N-S solar asymmetry is calculated by
formula
, (1)

where AN is the total area of the sunspots over Northern solar hemisphere and AS -
the total area of the sunspots over Southern solar hemisphere. The data of sunspot
area since 1874 are observed by the Royal Greenwich Observatory and merged after
22
Variations of Temperature Over Bulgaria and Their...

Figure 1. Monthly temperature data from stations Dobrich,


Pleven, Razgrad, Vidin and Sofia

1976 with the US Air Force (USAF) and the US National Oceanic and Atmospheric
Administration (NOAA) data by D. Hathaway.
23
Yavor Chapanov

Figure 2. Solar indices TSI and N-S solar asymmetry

METHODS
The time series spectra are determined by the well-known Fast Fourier
Transform (FFT). The periodical variations are derived from the data by means of
partial Fourier approximation (PFA) based on the Least-Squares (LS) estimation
of Fourier coefficients. The Partial Fourier approximation F ( t ) of time series is
given by
(2)

where P0 is the period of the first harmonic, t0 - the mean epoch of observations,
f0, f1, ak and bk are unknown coefficients and n is the number of harmonics of
the partial sum, which covers all oscillations with periods between P0/n and P0.
The application of the LS estimation of Fourier coefficients needs at least 2n+2
observations, so the number of harmonics n is chosen significantly smaller than
the number N of sampled data fi. The small number of harmonics n yields to LS
estimation of the coefficient errors. This estimation is the first essential difference
with the classical Fourier approximation. The second difference is the arbitrary
choice of the period of first harmonic P0, instead of the observational time span,
so the estimated frequencies may cover the desired set of real oscillations. This
method allows a flexible and easy separation of harmonic oscillations into different
frequency bands by the formula
(3)

where the desired frequencies wk are limited by the bandwidth


24
Variations of Temperature Over Bulgaria and Their...

2ðm1 2ðm2
≤ ωk ≤ (4)
P0 P0 ,
After estimating the Fourier coefficients, it is possible to identify a narrow
frequency zone presenting significant amplitude, and defining a given cycle.
Then this cycle can be reconstructed in time domain as the partial sum limited
to the corresponding frequency bandwidth. Doing this for terrestrial and solar
time series, we shall identify their respective cycles, isolate and compare the
common ones.
The used time series cover 105-year overlapped time interval for the period
1900.0 – 2005.0. Each time series consist of 1260 monthly data points. The PFA
performs estimation of 150 harmonics with the accuracy better than 0.07°C for
temperature; 0.02 for N-S solar asymmetry and 0.5mW/m2 for TSI.
The seasonal amplitude of temperature variations AT for each year is calculated
by
, (5)

where the Tmax is the maximal summer temperature of a given station, and Tmin – the
minimal winter temperature.

RESULTS
The main results of this study are described by analyses of FFT spectra; seasonal
and long-term variations of the temperature; thermal trends due to global warming;
influence of TSI and N-S solar asymmetry harmonics on temperature oscillations
and variations of seasonal components.

FFT spectra
The time series spectra are calculated by the Fast Fourier Transform (FFT) and
they are shown in Fig. 3. It is remarkable that the temperature spectra are almost
coherent for interannual oscillations, especially for the stations from North Bulgar-
ia. All time series have significant oscillations with periods 2.2, 2.5, 2.8, 3.4, 4.0,
5.2, 7.0 and 10.5 year, so a high correlation between the long-term variations of
temperature is expected.

Seasonal and long-term variations of temperature


The amplitudes of seasonal oscillations of the temperature, determined by
(5) are shown in Fig. 4. It is visible that these variations consist of a lot identical
interannual cycles and the time series are highly correlated. The seasonal
components of temperature have common long-term variations and trends. It is
remarkable significant positive linear trends after 1975.
25
Yavor Chapanov

Figure 3. Spectra of temperature variations

Figure 4. Seasonal amplitudes of temperature at stations Dobrich


(red), Pleven (blue), Razgrad (purple), Vidin (green) and Sofia (black)

The long-term oscillations of temperature are separated from the seasonal


components as superposition of oscillations with periods 2 – 104yr and 15 – 104yr,
determined by the Method of Partial Fourier Approximation (Fig.5). All time series
in Fig. 5 are highly correlated with common interannual and decadal cycles. The
26
Variations of Temperature Over Bulgaria and Their...

common tendency of temperature rise, associated with the global warming, is


presented after 1995 for the decadal oscillations with periodicity 15 – 104 years,
while in second case (periodicity 2 – 104 years) the positive tendency is after 1980.

Figure 5. Long-term variations of temperature at stations Dobrich (red),


Pleven (blue), Razgrad (purple), Vidin (green) and Sofia (black)

Thermal trends due to global warming


The time series from Fig. 5 are not suitable to determine the response of tem-
perature rise due to global warming, because the time series contain significant
interannual and long-term oscillations. The temperature rate from different time
series are shown in Table 1.

Table 1. Linear trends of temperature in different time series in [°C/yr]


Station Original data Long-term Long-term Seasonal amplitude
2-104yr 15-104yr after 1975
Vidin +0.005 +0.004 +0.009 +0.10
Pleven +0.011 +0.010 +0.015 +0.11
Razgrad –0.002 –0.002 +0.002 +0.13
Dobrich –0.003 –0.003 +0.001 +0.14
Sofia +0.009 +0.009 +0.011 +0.15

The original and long-term time series don’t match the temperature rate con-
nected with the global warming. Their rates of linear trends are almost zero or
negative. The only exception is the positive rate of seasonal amplitude after 1975,
whose value is between +0.010 and +0.15°C/yr (Fig. 4 and Table 1).
27
Yavor Chapanov

Influence of solar harmonics on temperature cycles


The influence of solar harmonics on temperature variations is often hidden,
because the major part of solar signals is absorbed by cooling process of evaporation,
especially in the case of decadal cycles. Nevertheless, some interannual cycles of
the temperature are sensitive to the solar harmonics.
The solar-temperature influence will be demonstrated on data from a single
station, because the oscillations are highly correlated with a lot of identical cycles
between different stations. The TSI harmonics from period bands 3.6 – 3.7yr; 4.3 –
4.4yr and 5.8 – 6.2yr have excellent agreement with the corresponding oscillation
of temperature from station Vidin (Fig.6).

Figure 6. Common interannual cycles of temperature


at station Vidin and TSI variations.

The harmonics of N-S solar asymmetry have good agreement with temperature
cycles with periods from bands 4.8-5.0yr; 8.7 – 9.5yr and 11.7 – 13.1yr (Fig.7).
These solar cycles are connected with the solar magnetic field variations and
corresponding influence on the heliosphere and galactic cosmic rays, whose effects
on temperature oscillations is explained by the model of Kilifarska.
28
Variations of Temperature Over Bulgaria and Their...

Figure 7. Common cycles of temperature at station Vidin


and N-S solar asymmetry variations

Influence of solar harmonics on seasonal variations of the temperature


The seasonal variations are also affected by the solar activity. The amplitudes
of seasonal components of temperature have significant interannual and decadal
oscillations from the interval 20°C – 30°C. A part of these oscillations can be
explained by the influence of solar harmonics. The TSI harmonics with periods 6.2
– 6.6yr and 7.5 – 81yr have good phase agreement with the amplitudes of seasonal
components of temperature at station Sofia (Fig.8). Common cycles with periods
17.2 – 20.6yr of seasonal amplitude of temperature at station Sofia and N-S solar
asymmetry with short phase reverse exist (Fig.9). The cumulative effects of these
cycles may explain at least 5°C of the observed amplitude of seasonal components.

Temperature variations driven by Schwabe –Wolf and Hale solar cycles


The classical 11-year Schwabe-Wolf and 22-year cycles of solar magnetic field
(Hale cycles) also affect temperature variations. The common 10.5-year cycles of
temperature and TSI variations are shown in Fig.10. The 10.5-year temperature
cycles from all stations have very small phase differences and the solar signal
delay is below 3yr. The 21-year cycles of temperature have relatively small phase
differences (Fig.11) and the delay of the TSI effects on the temperature is between
0 and 2 years for different stations. The cumulative effects of Schwabe –Wolf and
Hale solar cycles on temperature variations is about 0.7°C.
29
Yavor Chapanov

Figure 8. Common cycles of seasonal amplitude of temperature


at station Sofia and TSI variations

Figure 9. Common cycles of seasonal amplitude of temperature


at station Sofia and N-S solar asymmetry variations

CONCLUSIONS
The temperature variations tor the period 1900-2005 from 5 Bulgarian stations
are analysed and compared with some solar harmonics and cycles. The linear
trends, seasonal components and long-term oscillations of the temperature are de-
termined. Almost all of periodic components of observed temperature are highly
correlated, including the seasonal variations and long-terms. The linear trends of
the centennial time series of temperature don’t match the tendency of temperature
rise due to the global warming. The global warming signature is discovered in var-
iations of seasonal amplitudes after 1975, where the amplitude rates are between
0.10°C/yr and 0.15°C/yr. The 11-year Schwabe-Wolf and 22-year cycles of solar
magnetic field (Hale cycles) affect temperature variations with time delay between
0 and 3 years, where the common effect on temperature is about 0.7°C. The TSI
and N-S solar asymmetry harmonics have good agreement with the variations of
temperature and seasonal amplitudes in narrow frequency bands, whose periods are
between 3.6 and 20.6 years. These results may improve climatic models and some
long-term forecasts.
30
Variations of Temperature Over Bulgaria and Their...

Figure 10. Common 10.5-year cycles of temperature and TSI variations

Figure 11. Common 21-year cycles of temperature and TSI variations

ACKNOWLEDGEMENTS
This work has been carried out in the framework of the National Science Program
“Environmental Protection and Reduction of Risks of Adverse Events and
Natural Disasters”, approved by the Resolution of the Council of Ministers №
577/17.08.2018 and supported by the Ministry of Education and Science (MES)
of Bulgaria (Agreement № Д01-322/18.12.2019).

REFERENCES
Coddington, O., Lean, J.L., Pilewskie, P., Snow, M., Lindholm, D. (2015).
A solar irradiance climate data record. Bull. American Meteorological
Soc. 97(7), 1265 – 1282.
Georgieva, V., Shopova, N., Kazandjiev, V. (2019). Assessment of conditions
in South Bulgaria for spring crop growing using agrometeorological
indices. AIP Conf. Proc., Vol. 2075, Iss. 1, id.120014
31
Yavor Chapanov

Kazandjiev V., Shopova N., Georgieva V. (2018). Hydrothermal conditions


during vegetation season and spring crop growing in Plovdiv region.
Journal of Balkan Ecology, 21, 1, PublishScieSet-Eco Publisher, 2018,
23 – 38.
Kilifarska, N.A. and Haight, J.D. (2005). The impact of solar variability on
the middle atmosphere in present day and pre-industrial atmospheres. J.
Atmos. Solar Terr. Phys., 67, 3, 241 – 249.
Kilifarska, N.A., Tassev, Y.K. and Tomova, D.Y. (2008). Cosmic ray
showers and their relation to the stratospheric sudden warmings. Sun and
Geosphere, 3, 1, 10 – 17.
Kilifarska, N.A. (2011). Long-term variations in the stratospheric winter
time ozone variability – 22 year cycle. Comptes rendus de l’Académie
bulgare des Sciences, 64, 6, 867 – 874.
Kopp, G., Krivova, N., Lean, J., and Wu, C.J. (2016). The Impact of the
Revised Sunspot Record on Solar Irradiance Reconstructions, Solar
Physics. doi: 10.1007/s11207-016-0853-x
Lean, J.L. (2010). Cycles and Trends in Solar Irradiance and Climate, Wiley
Interdisciplinary Reviews, Climate Change 1. doi: 10.1002/wcc.018.
Shopova, N., Alexandrov, V., Todorova G. (2019). Cluster analysis of the
highest daily temperature amplitudes in some agricultural regions of
Southeastern Bulgaria, Proc. First Sci. Conf.“Climate, atmosphere and
water resources in the face of climate change” Sofia, 14 – 15 October
2019, 64 – 72.
Slavcheva-Sirakova, D., Shopova, N., Kostasdinov, K., Filipov, S.,
Velichkova, K. (2020). Climate analysis and effects of abiotic stress
on salad grounded in underground greenhouse and outdoor and effects
of organic fertilizers in the fight with stress factors. Proc. Conference
"Agriculture for Life, Life for Agriculture", Bucharest 4 – 6 June 2020.
Velinov, P.I.Y., Mateev L. and Kilifarska N.A., (2005). 3-D model for cosmic
ray planetary ionisation in the middle atmosphere. Ann. Geophys., 23, 9,
3043 – 3046.

 Yavor Chapanov
http://orcid.org/0000-0002-6159-3554
Climate, Atmosphere and Water Research Institute
Bulgarian Academy of Sciences (CAWRI-BAS)
Sofia, Bulgaria
E-mail: [email protected]

32
Proceeding of 1st International conference on Environmental protection and disaster RISKs

https://doi.org/10.48365/envr-2020.1.3 Air Pollution,


Climate and Health

STUDY OF ONE MONTH EVENT


OF HIGH PM POLLUTION FOR SOFIA REGION

Margret Velizarova1, Reneta Dimitrova1,2


1
“St. Kliment Ohridski” University of Sofia
2
National Institute of Geophysics, Geodesy and Geography –
Bulgarian Academy of Sciences (NIGGG –BAS)

Abstract: The habitants in Bulgarian capital city are unfortunately still exposed
to high levels of particulate matter (PM) despite a lot of efforts made by the
government and municipality during the last decades. A high resolution modelling
using ADMS-Urban system was carried out for prolonged one month (January
2014) event with elevated levels of PM concentration for Sofia region. Four main
emission sources (point industrial, domestic heating, roads and non-well defined
transport) were taken into account for this study. Although this type of modelling
accounts only for the local sources and has some deficiency, it can be very useful
tool for estimation of the contribution of each source type to the concentration field
and also allows to simulate different scenarios for future emission reduction that can
help the authorities in decision making.
Keywords: Air quality modelling, PM pollution in Sofia, ADMS-Urban model.

INTRODUCTION
Air pollution causes great harm to European citizens’ health and about 400 000 people
die prematurely due to excessive air pollutants such as dust particles, nitrogen dioxide
and ozone each year (EU Special Report, 2018). Around one-quarter of Europeans
living in urban areas were exposed to air pollutant levels exceeding some EU air quality
standards and up to 96% of EU citizens living in urban areas were exposed to levels of
air pollutants considered by the WHO as damaging to health (EEA, 2018). Lost years
of healthy life in some EU Member States are similar to countries often associated with
poor air quality, such as China and India and unfortunately Bulgaria is the leader in this
bad statistics with almost 2.5 lost years of healthy life from ambient air pollution per
hundred inhabitants (EU Special Report, 2018).
The air quality in big cities is substantially affected by anthropogenic pollutant
emissions and weather patterns. Most harmful for the Bulgarian big cities like Sofia
and Plovdiv is particulate matter (PM) pollution during the winter according to
the national reports on the state and protection of the environment for 2010-2014.
33
Margret Velizarova, Reneta Dimitrova

Plenty of work in the field of the regional air quality modelling has been made in
Bulgaria in the last two decades (Gadzhev et al., 2014a, b; Georgieva et al., 2015,
2017; Syrakov et al, 2015, 2016), most of the simulations are based on the US
EPA WRF (Skamarock et al., 2008) - CMAQ (Byun and Schere, 2006) system.
Advantages of these models include the ability to use temporally and spatially
varying meteorology and model chemical reactions, but due to used parametrizations
of physical processes however, these models have restrictions of grid resolution
(finest is usually 1 km). ADMS-Urban system (CERC, 2017) is a city-scale model
that resolves concentration fields explicitly representing the near-field features of
the dispersion of emissions with very high resolution (meters). Furthermore it can
be nested within the regional model CMAQ and this approach accounts for the full
range of temporal and spatial scales (Stoker et al., 2012), which will be subject of
our future work.
A lot of efforts were made by the government and Sofia municipality during
the last decades, but unfortunately the capital city habitants are still exposed to
high levels of PM10 (defined according European Union DIRECTIVE 2008/50/
EC; EEA, 2008). High resolution modelling is carried out for prolonged one
month event with elevated levels of PM concentration for Sofia region. The
case study was selected based on observations with persistent elevated levels
of PM in January 2014. The main goal is to estimate the abilities of the ADMS-
Urban to capture the concentration pattern using available detailed emission
inventory developed for Sofia and fill the missing details of concentration of
the pollutants.

METHODOLOGY AND MODEL SET-UP


ADMS-Urban system
ADMS-Urban is a comprehensive system for modelling air quality and it is
being used across the world for air quality management and assessment studies
of complex situations in urban areas, cities, towns and close to motorways, roads
and large industrial areas (Stoker et al., 2012; Hood et al., 2018; Biggart et al.,
2020). The “local” Gaussian type model with integrated street canyon model
is nested within a trajectory model so that significant areas may be considered.
ADMS-Urban is significantly more advanced than most of the other air Gaussian
type dispersion models in that it incorporates the latest understanding of the
boundary layer structure, using advanced algorithms for the height-dependence of
wind speed, turbulence and stability to produce improved predictions. The model
also takes account of the impacts of street canyons on dispersion, turbulence and
mixing induced by traffic, and includes a photochemical model for NOx and ozone.
Predicting pollutant concentrations from an urban area is a complex modelling
problem. ADMS-Urban has been developed with a number of features to simplify
the modelling process and help users.
34
Study of one Month Event of High Pmpollution...

Model set-up
The domain in this study covers Sofia city and suburbs, approximately 38 km by
32 km, with grid resolution of 50 m (Figures 1a-d). Spatial data entered into ADMS-
Urban should be in a Cartesian coordinate system measured in metres in one of the
available geographic projections. Different number of vertical levels can be exploited,
but high number increases the computational time significantly. We have used only
2 levels for this study as we consider only surface concentration. All different types
of sources – point (industrial stacks), roads, domestic heating and not well-defined
transport are mapped into the selected domain. We have used in this study detailed
emission inventory developed for Sofia municipality, published and described in
details in Sofia municipality reports (2017; 2019). The emissions required very
detailed shape description for line and area sources, while point sources are defined
by their geographical coordinates transferred in the study domain. All sources were
described into details and their locations are presented in Figure 1(a-d).

Figure 1. Map for simulations with all types of emission sources included –
(a) point (industrial stacks), (b) roads, (c) domestic heating and (d) not well-
defined transport; and (e) wind rose at Druzhba site for January 2014

One month of simulations for January 2014 was selected based on 10-years of record
of hourly concentrations of PM10 taken from the official air quality network (EEA,
http://eea.government.bg/kav/), for the stations located in Sofia (Druzhba, Nadezhda,
Pavlovo, Hipodruma, Orlov most). In addition, concentrations of PM2.5 at Hipodruma
35
Margret Velizarova, Reneta Dimitrova

station were used. The mountain site Kopitoto, located near Sofia, was used as a
background station in this study. The meteorological conditions (measured also at each
site) were prescribed from Druzhba site, as it consists full record of hourly data with
a few missing records only, which were reconstructed based on interpolation method
integrated into the model, taking the available values from the nearest hours. The
meteorological variables used in this study are hourly data for wind speed (minimum
0.75 ms-1 is required) and direction, temperature, relative humidity, incoming solar
radiation (when data are missing for cloud cover). The local meteorological conditions
for the simulated period were low wind velocity (less than 2 ms-1) and predominant
south-eastern and north-eastern flow. Wind rose for January 2014 is shown in Figure
1e. ADMS-Urban was run daily providing 24-h average concentration outputs for the
entire domain and prescribed locations, corresponding to the measurement sites.
ADMS-Urban allows a lot of additional options that we applied in this study.
Sofia valley has a complex topography that affects the pollution transport and this
option was included with the model runs. The effect of complex terrain is modelled
by changing the plume trajectory and dispersion to account for disturbances in
Monin-Obukhov length. Turbulence is considered using three sets corresponding to
unstable (convective), near neutral and stable conditions, based on Monin-Obukhov
length and corresponding to different diffusivity coefficients. Loss of the airborne
concentration of pollutant occurs by dry and wet deposition. Wet deposition did
not apply in this study as there was no precipitation data for this period. The plume
rise module predicts the trajectory, enhanced spread and inversion penetration
of a buoyant jet or plume, given the conditions at the source and in the external
environment. This module was applied to the industrial point sources requiring
additional data for the temperature, flow velocity and geometry of the stack (height
and diameter). The basic street canyon module was used for the road sources which
required additional information such as the coordinates of each road segment and
averaged parameters for road width and canyon height. Only primary pollutants
were modelled, chemical reactions were not included in this study.

RESULTS AND DISCUSSION


The study period of high polluted PM event for Sofia region covers one
month. Maps with the average concentration of PM10 and PM2.5 for January 2014
are shown in Figure 2. These maps present contribution of all main sources with
added background concentration calculated for each day from the measurements
at Kopitoto site. Main polluted areas are located near the busy roads and cover the
villages surrounding the city, where the domestic heating emissions are coming
from wood and coal burning. Note, that emission inventory for PM10 and PM2.5
differ, as for PM10 the basic year is 2014 (the modelling year), for the PM2.5 2017 is
taken as a basic year (Sofia municipality report, 2017; 2019). The PM2.5 inventory
covers several additional sources in Lozenetc, Malinova Dolina, Gorna Bania,
36
Study of one Month Event of High Pmpollution...

Ovtscha Kupel, Suhodol, Vitosha, Bakston and Manastirski Livadi, which are
included for PM2.5 , but not for PM10 simulations.

Figure 2. Maps of the average concentration of PM10 (a) and PM2. 5


(b) from all sources with added background concentration for January 2014

Estimated contribution of each separate source to the total concentration field,


with the impact (presented in %) are shown for all main four sources (Figures 3 –
6). Two more sources – from construction and repair activities, and landfills and
quarries, were included and tested in simulations, but their contribution was less
than 1% and they are excluded from the analysis.
Figure 3 shows that the pollution related to domestic heating is higher in the
villages around Sofia, especially those, located in the proximity to the mountain
slopes north (Stara Planina) and south (Vitosha and Plana mountains), more likely
due to the effect of complex topography that affects the local circulation. For these
areas the contribution of domestic heating is between 80 to 100 %.

Figure 3. Contribution of domestic heating to averaged concentration fields


for PM10 (a) and PM2.5 (b) from all sources for January 2014
37
Margret Velizarova, Reneta Dimitrova

Figure 4 presents the pollution due to the main busy roads with most significant
contribution to central city area and ring road around Sofia. Note, that the newly developed
north tangent from the ring road is not included into simulations, as the last available
developed emission inventory, used in this study, covers previous periods (2014 for PM10
and 2017 for PM2.5). The most polluted areas with contribution between 80 -100 % due
to traffic are nearby boulevards “Tsarigradsko shose”, “Slivnitsa”, “Todor Aleksandrov”,
“Botevgradsko shose”, suburbs Gorublene – Pancharevo, the south ring road in proximity
of Vitosha mountain, as well as the north-western ring road, which gives connection
between big residential areas – Nadezhda, Obelya, Lyulin and “Struma” motorway.

Figure 4. Contribution of main roads to averaged concentration fields for PM10


(a) and PM2.5 (b) from all sources for January 2014

Figure 5 shows contribution of pollution due to not well-defined transport,


considered in this study as a volume source with vertical extension of 1m. These
sources correspond to the residential areas and the emission is calculated based on

Figure 5. Contribution of not well-defined transport sources to averaged


concentration fields for PM10 (a) and PM2.5 (b) from all sources for January 2014
38
Study of one Month Event of High Pmpollution...

population living in the area and registered vehicles. This source provides between
50 – 80% of the total concentration in the described big residential regions and the
Sofia airport area.
The impact of the industrial point sources is much lower (less than 50%) in
comparison to the other sources (Figure 6). The main contribution they have in the
northern part of the Sofia valley where a cluster of objects is located in the industrial
city area (in proximity to Nadezhda residential area), power plants “Iztok”, “Sofia”,
“Lyulin” and big industrial sources in Novi Iskar.

Figure 6. Contribution of point industrial sources to averaged concentration fields


for PM10 (a) and PM2.5 (b) from all sources for January 2014

Direct comparison at selected locations (observational sites) shows that simulated


concentrations are much lower than observed for some periods (especially 3-8, 18-
19 January 2014), which is not surprizing, as the modelling considers only primary
pollutants and these periods correspond to very high stable conditions favourable for
long living pollution in the atmosphere and possible secondary pollution formation
(Figure 7). Also this study do not take into account for possible transport coming
from sources located out of the study area. The background concentration, taken
at the Kopitoto station, cannot provide information on additional transport due to
local circulation and for some cases is not representative enough.
The evaluation of the performance of WRF-CMAQ air quality modelling
system in Bulgaria also shows significant underestimation of PM10 concentration
especially during the winter period (from October to March), when modelled
values are 2-4 times lower than observed ones (Georgieva et al., 2015, Syrakov
et al., 2015), which is in agreement with the presented study. This deficiency
in air quality modelling in regional and local scale pointed out for a possible
problem with the emission inventory for the Sofia region and more efforts
needed in this direction.
39
Margret Velizarova, Reneta Dimitrova

Figure 7. Comparison of simulated 24-h concentration


(with and without background concentration taking account)
at all available measurement sites for January, 2014

Figure 8. Contribution of different sources to concentration at locations


of the air quality stations for January, 2014

In spite of this deficiency, simulations are useful for estimation of the contribution
of the local sources at the selected receptors (Figure 8). It is obvious that stations
Pavlovo, Hipodruma and Orlov most mostly account for the traffic pollution (more
40
Study of one Month Event of High Pmpollution...

than 50%). Nadezhda and Druzhba stations are located in a residential areas and
high contribution (~50%) have the not well-defined area transport and only less
than 2% due to the point industrial sources. The domestic heating produces almost
double concentration for PM2.5 in comparison with PM10 (see Hipodruma site). The
contribution of domestic heating to PM10 concentration is also visible at all sites,
but predominant at Nadezhda and Druzhba, most likely because they are located in
the suburbs in proximity to satellite villages around Sofia.

CONCLUSIONS
In spite of efforts made by the Sofia municipality, the air quality is still poor, espe-
cially related to PM (Sofia municipality report, 2017; 2019). The most favourable for
smog formation are meteorological conditions during the winter, due to formation of
prolonged inversions and low turbulence under calm conditions related to non-dynamic
(anticyclonic) weather. One month with very high PM pollution registered by the of-
ficial air quality stations was simulated by ADMS-model. Although this type of model-
ling account only for the local sources, it can be very useful tool for estimation of the
contribution of each source type to the concentration field, and it allows to simulate
different scenarios that can help the authorities in the decision making. Only primary
pollutants (PM10 and PM2.5) were modelled (chemical transformations are not included)
with meteorological hourly data measured at Druzhba station, which can be comment
as a deficiency of the study. But a lot of additional options available with the sophisti-
cated Gaussian model (ADMS-Urban) allow to make advanced estimation of the local
pollution, much better than using other models from this type for air quality assessment.

The main findings in this study are:


I) Modelling of PM10 and PM2.5 concentrations with ADMS-Urban shows very
high levels (200 µgm-3) at the villages around Sofia, where air quality measurements
are not available at present.
II) Contribution of transport (from the main busy roads and area non well-
defined transport) is much higher than input from other sources.
III) Domestic heating is significantly important in the suburban satellite villages
around Sofia and its contribution to the central city area is low, less than 20% for
PM10 and less than 30% for PM2.5.
IV) Contribution of industrial point sources is very low, less than 5% in the area
in close proximity to the main source for the simulated case, but keep in mind that
the pattern can be different under strong winds when the pollution can affect areas
on larger distances.
V) Direct comparison with the measurement shows that local primary emissions
contribute only partly to the measured concentrations and the rest is more likely due
to several possible reasons – underestimation of the emissions, secondary pollutants
and transport from the sources outside of the domain.
41
Margret Velizarova, Reneta Dimitrova

VI) This work has pointed out the need for updated and reliable detailed
emission inventory for Sofia, including all different types of sources and pollutants.
Developed new atmospheric air quality management program for Sofia municipality
and future work on this topic can contribute significantly to improve air quality
modelling for the region.

ACKNOWLEDGMENTS
This work has been carried out in the framework of the National Science Program
"Environmental Protection and Reduction of Risks of Adverse Events and
Natural Disasters", approved by the Resolution of the Council of Ministers
№ 577/17.08.2018 and supported by the Ministry of Education and Science (MES)
of Bulgaria (Agreement № Д01-322/18.12.2019).
We acknowledge the provided access to the e-infrastructure of the NCDSC – part of
the Bulgarian National Roadmap on RIs, with the financial support by the Grant
No D01-221/03.12.2018.

REFERENCES
Biggart, M., Stocke, J., Doherty, R.M., et al. (2020). Street-scale air quality
modelling for Beijing during a winter 2016 measurement campaign. Atmos
Chem Phys, 20(5), 2755-2780.
Byun, D. & Schere, K.L. (2006). Review of the Governing Equations, Computational
Algorithms, and Other Components of the Models-3 Community Multiscale Air
Quality (CMAQ) Modeling System, Applied Mechanics Reviews, 59, 51-77.
CERC, (2017). CERC_ADMS-Urban4.1.1_User_Guide.pdf.
European Court of Auditors, (2018). Air pollution: Our health still insufficiently
protected. Report N°23.
European Environment Agency (2008). Directive 2008/50/EC of the European
Parliament and of the Council of 21 May 2008 on ambient air quality and cleaner
air for Europe. OJ L 152, 11.6.2008, 1–44.
European Environment Agency (2018). Environmental indicator report 2018, EEA
No 19/2018, Luxembourg.
Gadzhev, G., Ganev, K., Miloshev, N., Syrakov, D. & Prodanova, M., (2014a).
Some basic facts about the atmospheric composition in Bulgaria - Grid computing
simulations, 8353 LNCS, 484-490.
Gadzhev, G., Ganev, K., Miloshev, N., Syrakov, D. & Prodanova, M., (2014b).
Analysis of the processes which form the air pollution pattern over Bulgaria,
8353 LNCS, 390-396.
Georgieva, E., Syrakov, D., Prodanova, M., Etropolska, I. & Slavov, K. (2015).
Evaluating the performance of WRF-CMAQ air quality modelling system
in Bulgaria by means of the DELTA tool. IJEP, 57(3-4), DOI: 10.1504/
IJEP.2015.074512.
42
Study of one Month Event of High Pmpollution...

Georgieva, I., Gadzhev, G., Ganev. K., Melas, D. & Wang, T. (2017). High
Performance Computing Simulations of the Atmospheric Composition in
Bulgaria and the City of Sofia, Cybernetics and Information Technologies, 17(5),
DOI: 10.1515/cait-2017-0053.
Hood, C., MacKenzie, I., Stocker, J., Johnson, K., Carruthers, D., Vieno, M. &
Doherty, R. (2018). Air quality simulations for London using a coupled regional-
to-local modelling system. Atmos Chem Phys, 18, 11221-11245.
Skamarock, W.C., Klemp, J.B., Dudhia, J., et al. (2008) A Description of the
Advanced Research WRF Version 3. NCAR Technical Note, NCAR/TN-475+STR
Sofia municipality, (2017). Atmospheric air quality management program of capital
municipality for the period 2015-2020 – reduction of emissions and reach of the
established norms for fine particles PM10. (in Bulgarian).
Sofia municipality, (2019). Program for supplementation of the program
Atmospheric air quality management program of capital municipality for the
period 2015-2020 - reduction of emissions and achievement of the established
norms for fine particles PM10, by indicators fine particles with size up to 2.5
microns and polycyclic aromatic hydrocarbons. (in Bulgarian).
Stocker, J., Hood, C., Carruthers, D. & McHugh, C. (2012). ADMS-Urban:
developments in modelling dispersion from the city scale to the local scale.
IJEP, 50(1/2/3/4), 308-316.
Syrakov, D., Prodanova, M. & Georgieva, E. (2015). Performance of the Bulgarian
WRF-CMAQ modelling system for three subdomains in Europe. Física de la
Tierra, 27, 137-153.
Syrakov, D., Prodanova, M., Georgieva, E., Etropolska & I., Slavov, K. (2016).
Simulation of European air quality by WRF–CMAQ models using AQMEII-2
infrastructure. J Comput Appl Math, 293, 232–24.

 Margret Velizarova
Department of Meteorology and Geophysics
Faculty of Physics
“St. Kliment Ohridski” University of Sofia
Sofia, Bulgaria
E-mail: [email protected]

 Reneta Dimitrova
http://orcid.org/0000-0002-5931-8713
Department of Meteorology and Geophysics
Faculty of Physics
“St. Kliment Ohridski” University of Sofia
National Institute of Geophysics, Geodesy and Geography
Bulgarian Academy of Sciences
Sofia, Bulgaria
E-mail: [email protected]

43
Proceeding of 1st International conference on Environmental protection and disaster RISKs

https://doi.org/10.48365/envr-2020.1.4 Air Pollution,


Climate and Health

SATELLITE DATA ASSIMILATION


OF AIR QUALITY PARAMETERS IN BULGARIA

Dimiter Syrakov1,2, Maria Prodanova1,2, Emilia Georgieva1


1
National Institute of Meteorology and Hydrology (NIMH)
2
National Institute of Geophysics, Geodesy and Geography –
Bulgarian Academy of Sciences (NIGGG-BAS)

Abstract The operational Bulgarian Chemical Weather Forecast System


(BgCWFS) was modified and applied for assimilation of satellite retrieved
atmospheric chemistry parameters - Aerosol Optical Depth (AOD) and columnar
values of NO2 and SO2. The work outlines the methodology based on calculation
of correction factors between model estimated and satellite derived parameters.
Two model simulations for August 2017 and February 2019 are made – Model Run
(without satellite data assimilation – mod-run) and Satellite Run (with satellite data
assimilation – sat-run) for all 5 domains of BgCWFS. In the paper the effect of the
assimilation is demonstrated stressing mainly on the difference between sat- and
mod-runs. The comparison of sat-run results with observations and with third party
modelling results are object of other publications.
Keywords: Modelling air quality, satellite retrieve data, GOME-2, MetOP
satellites, data assimilation

INTRODUCTION
Satellite derived air pollution data are nowadays increasingly used in
combination with comprehensive chemical transport models (CTM) for better
description of the atmospheric composition and for improved forecast of
pollutants concentrations at ground level (Benedetti et al., 2009, Park et al., 2011,
Gadzhev et al., 2015). In the last few decades, major efforts are put on improving
modeled aerosol parameters, as aerosols play significant role in the Earth’s
energy budget and climate. Aerosols in the atmosphere are product of complex
interactions between sources and chemical transformations at different scales,
determining their high variability in space and time (Boucher, 2015). Satellites
provide information for AOD that is a measure of the aerosol loading in the entire
atmospheric column. They could provide information for sources, typically not
included in the models, such as dust storms and wild fires. Together with aerosol
parameters (AOD, AAI etc.) some gaseous parameters are retrieved. In the article
44
Satellite Data Assimilation of Air Quality...

the effects of similar assimilation of two gaseous parameters – columnar values


of NO2 and SO2 – is presented.
The purpose of this work is to establish a methodology for satellite data
assimilation in the current Bulgarian Chemical Weather Forecast System and to
discuss some preliminary results from simulations for two one-month periods. Up
to now, satellite data are not used in relation to air quality in Bulgaria. This work is
the first attempt to benefit from available satellite data for better understanding and
simulating air quality parameters in the country.

THE MODELING SYSTEM


The Bulgarian Chemical Weather Forecast System (Syrakov et al., 2013, 2014) is
based on the state of the art WRF-CMAQ modelling chain. WRF v3.6 (Skamarock
& Klemp, 2008) is used as meteorological pre-processor to the CMAQ model, v4.6
(Byun & Schere, 2006). The nesting capabilities of both models are used; they are
run over five nested domains – Europe (EU, 81 km horizontal resolution), Balkan
Peninsula (BP, 27 km), Bulgaria (BG, 9 km), Sofia district (SD, 3 km) and Sofia
city (SC, 1km). WRF is fed by the forecast/reanalyse production of the US NCEP
(National Centres for Environmental Prediction). As far as this data is global, it
provides the initial and boundary meteorological conditions for WRF as well. The
chemical boundary conditions over the mother domain (Europe) are set according
to the climatic profiles, embedded in the CMAQ software. The presumption is that
the possible errors decrease inside the domain because of the continuous action
of pollution sources. The boundary conditions for each of the other domains are
determined from the senior one.
The physical parameterisations selected for WRF are the well-known and widely
used Kain-Fritsch scheme for the cumulus parameterization, YSU for PBL, WSM6
for microphysics, RRTM for radiation, Noah for land-surface model, RADM for wet
deposition of gases and particulates.. The exploited version of CMAQ uses CBIV as gas-
phase chemical mechanism and ISORROPIA 1.7 for inorganic aerosol thermodynamics/
partitioning. The emissions are based on the inventory, provided by TNO (The Netherlands
Organization for Applied Scientific Research) for 2010 for Europe (Kuenen et al., 2014).
For Bulgaria, national emission inventories for 2015 are used.

SATELLITE DATA ASSIMILATION DESCRIPTION


AOD calculations
AOD is not routinely calculated in BgCWFS as well as in the most of the
CTMs. Different algorithms are described in the literature. Results from testing
some of them for calculating AOD at 550 nm are reported in (Syrakov et al., 2019).
Finally, the FlexAOD software (Curci, 2012, Curchi et al., 2015) is chosen for
post-processing of CMAQ estimated profiles of aerosols species. The algorithm
is based on the Mie theory and the tool has a version consistent with the CMAQ
45
Dimiter Syrakov, Maria Prodanova1, Emilia Georgieva

output. It is well known that models usually underestimate the particulate matter
concentrations. The assimilation of satellite retrieved AOD is supposed to improve
this shortcoming.

Assimilation technology
For BgCWFS we consider aerosol data provided by GOME-2 instrument on
board of MetOpA, MetOpB and MetOpC satellites. An objective analysis scheme
(called further ANALIZ) is used for combining model estimated and satellite
retrieved data (AOD and columnar densities of NO2 and SO2). The object analysis
scheme is original one, based on calculation of autocorrelation functions and
application of spline interpolation.
The analyzed data is used to calculate the correction factors between model
estimated and satellite assimilated parameters – different factors for AOD, NO2
and SO2. The gridded values of the correction factors are then used to correct
the concentrations of different pollutants (Fig.1). The correction factor of AOD is
applied to all particle variables and is one and the same for all levels. The correction
factor of NO2 and SO2 is applied to the respective variable profiles. In such a way
new fields for the satellite passage hour are produced serving as initial condition for
further CMAQ calculations.

Figure1. Scheme for use of satellite retrieved AOD in BgCWFS


for the satellite overpass hour Hs.

BgCWFS is modified in a way to organize pseudo-operational (real time)


satellite data assimilation. The main data stream for one day is as follows:
1. Performs full-day CMAQ integration with initial conditions from the previous
day. Concentrations for the whole day are obtained – CONC(00÷24)
2. Prepares new initial condition (IC) for satellite passage hour Hs performing:
a. Prepares the satellite data as input to ANALIZ (for Hs).
46
Satellite Data Assimilation of Air Quality...

b. Extracts and saves CONC(00÷Hs-1).


c. Extracts and saves CONC(Hs).
d. Prepares input for FlexAOD (for the whole day).
e. Runs FlexAOD.
f. Prepares AOD data as input to ANALIZ (for Hs).
g. Prepares NO2 and SO2 columnar data as input to ANALIZ (for Hs).
h. Runs ANALIZ (for AOD, NO2 and SO2 for Hs).
i. Calculates Correction Factor (CF) for AOD, NO2 and SO2 (for Hs).
j. Modifies PM, NO2 and SO2 fields by CF - new IC for Hs.
3. Performs a part-day CMAQ integration for h=Hs÷24 – CONC(Hs÷24).
4. Concatenates CONC(00÷Hs-1) with CONC(Hs÷24:00) – new CONC(00÷24)
The assimilation is taking place in the first three model domains of BgCWFS
(EU, BP and BG). The remaining two domains (SD and SC) obtain satellite
influence via their boundary conditions, calculated from the senior domains. This
is made because of the specific characteristic resolution of GOME-2 data (about
40 km). Such assimilation is made for first time, worldwide. According to the
numerous publications, satellite data assimilation has been applied for one domain,
only. Another novelty is the simultaneous assimilation of AOD, NO2 and SO2.

RESULTS
As far as only limited amount of pollutants are monitored and the measurements
are made at the ground, mainly, a post-processing of obtained data is performed
consisting in extracting the surface values of several key pollutants as well as
producing compositions as PM10, PM2.5, AOD and columnar values of NO2
(NO2_C) and SO2 (SO2_C). This is made for both sat- and mod-runs, results saved
as so called ARCH-files (ARCHs and ARCHm, respectively).
The detailed investigation of the results of the satellite data assimilation is object
of another publication. Here, only effects of the assimilation will be demonstrated
exploiting the difference “sat-mod” of several variables, for different days and
domains. The main behavior of this difference is as follows:
● In first hours of the day, the differences remaining from the previous day are
relatively small, with positive and negative values.
● At hour Hs (satellite overpass fixed to 09:00 UTC) a disturbance appears like
a simultaneous area source of pollution (similar to fire or other area accident)
● The shape and the structure of this disturbance are quite irregular and it
appears approximately over the territory covered by satellite measurements (but
not over the whole area).
● Similarly to pollution distribution, the disturbances evolves with time and
move as if due to atmospheric circulation.
● Usually, its maximal values decrease and for some variables become small or
disappear by the end of the day.
47
Dimiter Syrakov, Maria Prodanova1, Emilia Georgieva

● Often, the disturbances remain to the next day, especially for the bigger
modelling domains.
An example of this behavior is shown on Fig. 2 (satellite overpass hour Hs=9:00
UTC).

Figure 2. Evolution of the disturbance in the AOD “sat-mod” difference caused


by satellite data assimilation over domain BP (Balkan Peninsula), 27 km
resolution, h=0, 8, 9(Hs), 10, 14, on 22. August 2017

Of course, the details of this common behavior of the “sat–mod” difference vary
very much with the different days, different variables, different domains, and for
different seasons, as well.

Figure 3. Maps of dAOD fields on 01.08.2017 at time step h=Hs


for the first 3 BgCWFS domains

The AOD-differences (dAOD) by the two versions of BgCWFS at the moment


Hs more or less follow the satellite data coverage. This can be seen in Fig. 3, where
dAOD is shown in the first three BgCWFS domains for 01.08.2017. One can
notice that the three fields are quite consistent. More or less similar is the spatial
distribution of dPM10. Note that PM10 is directly influenced by the ratio AODsat/
AODmod (not the differences but the ratio) at the same hour.
The satellite data for NO2_C cover bigger areas than for the AOD ones. The
coverage by SO2_C data is smaller, but it is still bigger than for AOD. The dNO2
behavior is slightly different than dAOD. Up to the moment Hs of a certain day
48
Satellite Data Assimilation of Air Quality...

the differences “sat-mod” are around zero. At Hs, fragmentary spots appear with
positive and negative values, predominantly in the first two modelling domains.
At h=Hs+1, the positive spots practically disappear and the negative ones decrease
their maxima. Possible reason for this behavior is the high reactivity of NO2. Fig.
4 shows the time evolution of the dNO2 spatial distribution in domain Bulgaria.

Figure 4. Maps of dNO2 for the domain BG on 03.08.2017


at different time steps, h=8, 9 (Hs), 10, 11

Fig. 5 shows the time evolution of dSO2 fields for domain Bulgaria on
07.02.2019. The decrease of the differences with the time is smaller than for August
2017; small positive differences remains for the next day.

Figure 5. Maps of dSO2 for domain BG on 07.02.2019


at different time steps, h=0, 9 (Hs), 10, 23

Figure 6. Maps of dO3 for 3.08.2017 at different time steps:


h=0, 9 (Hs), 10, 11, 18.

Interesting fact is the appearance of high values of dSO2 around the upper
border of the domain. This is example of influence of the senior domain - satellite
data assimilation from domain BP is transferred to domain BG via the boundary
conditions. Such an effect is observed for other days and pollutants.
49
Dimiter Syrakov, Maria Prodanova1, Emilia Georgieva

The final example for the effect of satellite data assimilation is the evolution of
ozone surface concentrations. Like the other parameters directly influenced by the
satellite data assimilation (PM, NO2, SO2) the dO3 behavior is also characterized
by “explosion” (higher values with respect to previous hour), but at time Hs+1,
i.e. one hour later than satellite passage. This delay is due to the fact that ozone
is a secondary pollutant and some time is necessary for reactions (mainly with
already changed NO2) leading to this deviation from the usual daily evolution of
the changes “sat-mod”. The spots with differences dO3 are quite irregular on the
maps, mainly positive, and their decrease lasts much longer time, often continuing
during the next day. Increase of spot maxima during the following one or two hours
after Hs is also possible. In Fig. 6, one can see the time evolution of dO3 during
the first hours of the day (with remains from the previous day), the appearance of a
new disturbance at h=Hs+1=10:00 UTC due to assimilation of satellite NO2 data,
the increase of its intensity and further decrease several hours later.

CONCLUSIONS
Satellite retrieved atmospheric chemistry data were assimilated for the first
time in the Bulgarian Chemical Weather Forecast System. The off-line version of
the modelling system was run over five nested domains for two months (August
2017 and February 2019) in two modes – without satellite data (mod-run), and
with satellite data assimilation for AOD and columnar NO2 and SO2 (sat-run). The
preliminary analysis of the results in this work was focussed mainly on the evolution
of the fields of the “sat-mod” differences of several parameters (AOD, PM10, NO2
and SO2). The main behaviour of these differences is demonstrated. The deviations
are also outlined and commented. Further investigations are required to understand
the advantages and the weaknesses of the BgCWFS-sat modelling system, using
both comparisons with observations and model inter-comparisons to data from air
quality networks.

ACKNOWLEDGMENTS
This study was carried out in the framework of the project SIDUAQ, funded by the
European Space Agency (ESA) through contract No. 4000124150/18/NL/SC.
Deep gratitude is due to ESA services providing data from GOME-2 instrument
on board of MetOp-A, B and C satellites.

REFERENCES
Benedetti A., J.-J. Morcrette, O. Boucher, A. Dethof, R. J. Engelen, M.
Fisher, H. Flentjes, N. Huneeus, L. Jones, J. W. Kaiser, S. Kinne, A. Man-
gold, M. Razinger, A. J. Simmons, M. Suttie, and the GEMS-AER team
50
Satellite Data Assimilation of Air Quality...

(2009): Aerosol analysis and forecast in the ECMWF Integrated Forecast


System. Part II: Data assimilation. J. Geophys. Res., 114, D13205.
Boucher, O. (2015). Atmospheric Aerosols: Properties and Climate Im-
pacts, Springer, Dordrecht. ISBN 9789401796484.
Byun, D., & Schere, K. L. (2006) Review of the governing equations, com-
putational algorithms and other components of the Models-3 Communi-
ty Multiscale Air Quality (CMAQ) modeling system. Appl. Mech. Rev.,
59, 51–77
Curci, G. (2012). FlexAOD: A Chemistry-transport Model Post-processing
Tool for A Flexible Calculation of Aerosol Optical Properties. In: Pro-
ceedings of the 9-th International Symposium on Tropospheric Profiling,
L'Aquila, Italy, September 2012, ISBN: 978-90-815839-4-7
Curci, G., Hogrefe, C., Bianconi, R., Im, U., Balzarini, A., Baró, R., Brun-
ner, D., Forkel, R., Giordano, L., Hirtl, M., Honzak, L., Jiménez-Guer-
rero, P., Knote, C., Langer, M., Makar, P. A.,Pirovano, G., Pérez, J. L.,
San José, R., Syrakov, D., Tuccella, P., Werhahn, J., Wolke, R., Žabkar,
Zhang, R. J., and Galmarini, S. (2015): Uncertainties of simulated aero-
sol optical properties induced by assumptions on aerosol physical and
chemical properties: An AQMEII-2 perspective. Atmos. Environ. , 115,
541–552
Gadzhev, G., Ganev, K., Miloshev, N., 2015) Numerical study of the at-
mospheric composition climate of Bulgaria - Validation of the computer
simulation results, International Journal of Environment and Pollution,
57 (3-4), 189-201.
Kuenen, J.J.P., Visschedijk, A.J.H., Jozwicka, M.,Denier van der Gon,
H.A.C. (2014). TNO-MACC_II emission inventory; a multi-year (2003–
2009) consistent high-resolution European emission inventory for air
quality modelling. Atmos. Chem. Phys.,14 (20), 10963–10976.
Park R. S., Song C. H., Han K. M., Park M. E., Lee S.-S., Kim S.-B., Shi-
mizu A. (2011). A study on the aerosol optical properties over East Asia
using a combination of CMAQ-simulated aerosol optical properties and
remote-sensing data via a data assimilation technique. Atmos. Chem.
Phys., 11 (23), 12275–12296.
Skamarock, W.C. & Klemp, J.B. (2008): A time-split non-hydrostatic atmo-
spheric model. J. Comput. Phys., 227 (7), 3465–3485.
Syrakov, D., Prodanova, M., Slavov, K., Etropolska, I., Ganev, K., Milo-
shev, N., Ljubenov, T. (2013). Bulgarian System for Air Pollution Fore-
cast. J. Int. Sci. Publ.: Ecol. Saf., 7 (1), 325-334.
Syrakov, D., Prodanova, M., Etropolska, I., Slavov, K., Ganev, K., Mi-
loshev, N., Ljubenov, T. (2014). A Multi- Domain Operational Chem-
ical Weather Forecast System. In: Lirkov I. et al. (eds.) Large-Scale

51
Dimiter Syrakov, Maria Prodanova1, Emilia Georgieva

Scientific Computing, LNCS 8353, 413–420. Springer-Verlag Berlin


Heidelberg.
Syrakov D., Prodanova M., Georgieva E., Dimitrova M., Spassova T., Ata-
nassov D., Veleva B., Nedkov R. (2019). Aerosol optical depth calcula-
tions using the Bulgarian Chemical Weather Forecast System. Bulg. J.
Meteorol. Hydrol. 22 (3).

 Dimiter Syrakov
https://orcid.org/0000-0001-5057-8490
National Institute of Meteorology and Hydrology
National Institute of Geophysics, Geodesy and Geography – Bulgarian Academy of Sciences
Sofia, Bulgaria
E-mail: [email protected]

 Maria Prodanova
https://orcid.org/0000-0002-7395-9314
National Institute of Meteorology and Hydrology
National Institute of Geophysics, Geodesy and Geography – Bulgarian Academy of Sciences
Sofia, Bulgaria
E-mail: [email protected]

 Emilia Georgieva
https://orcid.org/0000-0002-8466-4976
National Institute of Meteorology and Hydrology
Sofia, Bulgaria
E-mail: [email protected]

52
Proceeding of 1st International conference on Environmental protection and disaster RISKs

https://doi.org/10.48365/envr-2020.1.5 Air Pollution,


Climate and Health

PRELIMINARY RESULTS FOR THE RECURRENCE


OF AIR QUALITY INDEX FOR THE CITY
OF SOFIA FROM 2008 TO 2019

Georgi Gadzhev
National Institute of Geophysics, Geodesy and Geography –
Bulgarian Academy of Sciences (NIGGG-BAS)

Abstract: The living environment of human beings and, obviously, the


atmospheric composition has a great impact for the quality of life and human health.
Air Quality (AQ) is a key element of European citizens' wellbeing and quality of
life. The objectives of the present work are to conduct reliable, comprehensive and
detailed studies of the impact of lower atmosphere composition on the quality of
life and health risks for the population in the city of Sofia. The performed numerical
simulations with the US EPA Models-3 system are for 12 years from 2008-
2019 and calculated on five domains: Europe, Balkan Peninsula, Bulgaria, Sofia
Municipality and Sofia City with increasing space resolution - from 81 km (Europe)
to 1 km (Sofia City). The System is based on the well-known models WRF (Meso-
Meteorological Model) and US EPA dispersion model CMAQ (Chemical Transport
Model). As emission input the TNO data is used for the two biggest domains. For
the 3 Bulgarian domains the current emission inventory prepared by Bulgarian
environmental authorities is exploited.
Keywords: Air Quality Indices, air quality, quality of life, health risks.

INTRODUCTION
The Air Quality is a key element for the well-being and quality of life of
human beings. According to the World Health Organization, air pollution severely
affects the health of European citizens. There is increasing evidence of adverse
effects of air pollution on the respiratory and the cardiovascular system as a result
of both acute and chronic exposure. In particular, a significant reduction of life
expectancy by a year or more is assumed to be linked to long-term exposure
to high air concentrations of particulate matter (PM). There is considerable
concern about impaired and detrimental air quality conditions over many
areas in Europe, especially in urbanized areas, despite 30 years of legislation
and emission reductions. Current legislation, e.g. the Ozone daughter directive
53
Georgi Gadzhev

2002/3/EC (European Parliament, 2002), requires informing the public on AQ,


assessing air pollutant concentrations throughout the whole territory of Member
States and indicating exceedances of limit and target values, forecasting potential
exceedances and assessing possible emergency measures to abate exceedances.
For this purpose, modeling tools must be used in parallel with air pollution
measurements. There are also different kind indexes for the well-being and
quality of life of the humans, such as UV, Heat and Wind chill -Indexes, but they
are not subject of interest in present study (Bojilova, Mukhtarov, Miloshev 2020,
Evtimov, Ivanov 2008, Ivanov, Evtimov 2014, Chervenkov, Slavov, Ivanov 2019).
The goals of reliable air quality studies are the efficient control and protection of
population exposure as well as possible emission abatement measures. In recent
years the concept of “chemical weather” arises and in many countries respective
forecast systems are being developed along with the usual meteorological weather
forecasts (Sofiev at al., 2006, Poupkou et al., 2008, Monteiro et al., 2005, San
Jose et al., 2006, Mukhtarov, Bojilova 2017, Bojilova, Mukhtarov 2019). Air
pollution easily crosses national borders. It would be cost-effective and beneficial
for citizens, society and decision-makers that national chemical weather forecast
and information systems were networked across Europe.

MODELING TOOLS
The present study is based on air quality simulations with US EPA Models-3 air
quality modeling system, consisting of 3 models:
• CMAQ v.4.6 - Community Multi-scale Air Quality model, http://www.
cmaq-model.org /, (Denis et al. 1996, Byun, Ching 1999, Byun, Schere 2006), the
Chemical Transport Model (CTM);
• WRF v.3.2.1 - Weather Research and Forecasting Model, http://www.
wrf-model.org/ , (Skamarock et al. 2007), the meteorological pre-processor
to CMAQ. The Weather Research and Forecasting (WRF) Model is a next
generation meso-scale numerical weather prediction system designed to serve
both operational forecasting and atmospheric research needs. It is an evolutionary
successor to the MM5 model. The creation and further development of WRF is
due to the collaborative efforts of several US institutions like NCAR, NOAA,
NCEP and others. The WRF is a fully compressible and non-hydrostatic model
with terrain-following hydrostatic pressure coordinate. The grid staggering is
the Arakawa-C type. One can find more info on http://www.wrf-model.org/
index.php;
• SMOKE v.2.4 - Sparse Matrix Operator Kernel Emissions Modelling
System, http://www.smoke-model.org/ , (Coats, Houyoux 1996, Houyoux,
Vukovich 1999, CEP 2003), the emission pre-processor to CMAQ. CMAQ
demands its emission input in specific format reflecting the time evolution of
all pollutants accounted for by the chemical mechanism used (CB-IV in this
54
Preliminary Results for the Recurrence of Air Quality...

case). Emission inventories are used as row data for anthropogenic emission
processing. The inventories are made on annual basis for big territories; many
pollutants are estimated as groups (VOC and PM2.5 for instance). Preparation
of emission input to a Chemical Transport Model requires emission processing.
Such emission processing component in EPA Models-3 system is SMOKE but
it is partly used, here, because it’s quite strong relation to US emission sources
specifics. In this study SMOKE is used only for calculating biogenic (BgS)
emissions and for merging Area sources (AS), Large point sources (LPS) and
BgS-files into a CMAQ emission input file. The area source emissions and the
large point source emissions are prepared by the interface programs AEmis and
PEmis.
In the System, WRF model is driven by the NCEP data in GRIB-2 format with
space resolution of 1°×1° and 6-hour time resolution. TNO inventory for 2005
(Denier van der Gon et al., 2010) is exploited partly for Bulgaria domain, TNO
being the Netherlands’s Organization for Applied Scientific Research. For Bulgaria
itself and for the other Bulgarian domains, the National inventory for 2010 as
provided by Bulgarian Executive Environmental Agency is used. That means TNO
inventory is used only for the territories outside Bulgaria in the mother CMAQ’s
domain.

Figure 1. Five computational domains (CMAQ domains are nested in WRF ones)

The nesting capabilities of WRF and CMAQ are used to downscale the
simulations from European region to the Sofia city area. The resolution of the
mother domain (Europe) is 81 km, big enough as to correspond to met-data space
resolution. Four other domains are nested in it and in each other – Balkan Peninsula
(27km resolution), Bulgaria (9 km), Sofia municipality (3 km) and Sofia city (1
km) as shown in Fig. 1. The climatic data is used for chemical boundary conditions
following the presumption that the errors introduced by this assumption will
decrease quickly to the centre of the domain due to the continuous acting of the
pollution sources. All other domains receive their boundary conditions from the
previous domain in the hierarchy.
The post-processing program XtrCON extracts part of the pollutants for archiving
and further handling. Only surface values of the most important pollutants are saved
55
Georgi Gadzhev

- 8 gases and 11 aerosols (including PM10 and PM2.5). Part of these pollutants
is more or less monitored and they are referred in the European legislation with
the respective thresholds. For the moment it presents 4 main pollutants - Ozone,
NO2, SO2 and PM10 which are used to calculate the Air Quality Indices (AQI).
Calculation of the Air Quality (AQ) impact on human health and quality of life
in Sofia city is the objective of the present study. The impact is calculated in the
terms of the so called AQI – an integral characteristic directly measuring the effects
of AQ on human health. The calculations are made on the basis of long term AQ
simulations, which make it possible to reveal the climate of AQI spatial/temporal
distribution and behaviour.
The AQI is defined as a measure of air pollution seen in the context of its impact
on human health. It provides an integrated assessment of the impact of the whole
range of pollutants on human health and is calculated based on the concentration of
various pollutants obtained from measurements or numerical modeling. The index
is defined in several segments (EPA, 2009), each of which is a linear function of the
concentration of each considered pollutant:

(1)

where:
I = the AQI,
C = the pollutant concentration,
– the concentration breakpoint that is ≤ C,
– the concentration breakpoint that is ≥ C,
– the index breakpoint corresponding to ,
– the index breakpoint corresponding to .

In that calculation the index falls in one of the ranges of the dimensionless scale.
In each range index values are associated with an intuitive colour code, a linguistic
description and a health description.
Pretty often in order to evaluate the air quality situation in European cities, all
detailed measurements are transformed into a single relative figure: the Common
Air Quality Index (CAQI) and this index have 5 levels using a scale from 0 (very
low) to > 100 (very high). The index is based on 3 pollutants of major concern
in Europe: PM10, NO2, O3 and will be able to take into account to 3 additional
pollutants (CO, PM2.5 and SO2).
One of the most commonly used air quality index is the UK Daily Air Quality
Index (Leeuw, F. de, Mol, W., 2005), also used in Bulgaria (Etropolska et al. 2010),
(Syrakov et al, 2012, 2013, 2014a, 2014b, 2015), (Georgieva, I., 2014), (Georgieva
et al. 2015), (Georgieva, I. and Ivanov, V., 2017, 2018), (Ivanov, V. and Georgieva,
I., 2017) and (Gadzhev 2018).
56
Preliminary Results for the Recurrence of Air Quality...

RESULTS
Annual recurrence of AQI in “Low”, “Moderate” and “High” bands over
territory of Sofia city from 2015 to 2019 and average recurrence for the whole
period 2008-2019 (08-19):
Figure 2 demonstrate the spatial and diurnal variation of the annual recurrence of
Low band for the chosen hours 04:00, 12:00 and 18:00UTC for the chosen periods.
Here we have to mention that in the Low range the air is most clean, so high recurrence
values mean more cases with clean air (red colour) and lower recurrence values mean
(blue colour), less cases with clean air (worse AQ status). What can be noticed is: the
recurrence in Low range is different for all years at 04:00UTC.

04:00 12:00 18:00

Figure 2. Annual recurrence of AQI in “Low” band over territory of Sofia city from
2015 to 2019 and average recurrence for the whole period 2008-2019 (08-19)
57
Georgi Gadzhev

In 2015, 2018 and 08-19 at 12:00 UTC the recurrence in Low band is bigger
than this in 2016, 2017 and 2019. While at 18:00 UTC the higher recurrence is in
2015, 2018, 2019 and 08-19. The high polluted areas are the city centre at 04:00 and
two spots (Kostinbrod and Sofia airport) in 2019 and they are very well displayed
in figure.

04:00 12:00 18:00

Figure 3. Annual recurrence of AQI in “Moderate” band over territory of Sofia city
from 2015 to 2019 and average recurrence for the whole period 2008-2019 (08-19)
58
Preliminary Results for the Recurrence of Air Quality...

04:00 12:00 18:00

Figure 4. Annual recurrence of AQI in “High” band over territory of Sofia city from
2015 to 2019 and average recurrence for the whole period 2008-2019 (08-19)
59
Georgi Gadzhev

In Figure 3 and Figure 4 (Moderate and High ranges) - high recurrence values
means less favourable and respectively bad AQ status. It can be seen that most areas
with high recurrence of cases with lower AQ status are at noon in the whole domain
and mostly over the Vitosha Mountain in other hours. This is due to the intensive
O3 photo-chemistry reactions, the higher NO2 concentrations lead to production of
higher O3 concentration. The major NO2 sources in the city are the surface sources
(road transport). It can be seen at 18:00 UTC in 2016 and 2017 where over the
road network it leads to decreasing of O3 concentration and with that improving
the AQ status. Average for the 2008-2019 in Moderate band at 18:00 UTC it can
be also noticed about 10-20% recurrence with not so good AQ status over Vitosha
Mountain. Higher values over the Vitosha Mountain in the night and the afternoon
are due to the higher concentration of O3 in mountain areas and intensive ozone
transport from higher levels (intensive turbulence during midday). The behavior
of the surface ozone is complex. The O3 in Bulgaria is to a great extent due to
transport from abroad and above (Gadzhev et al. 2013 and Kaleyna et al. 2013a,
2013b, 2014). This is the reason why the O3 concentrations early in the morning
are smaller (less intensive transport from higher levels), and higher at noon and
afternoon (more turbulent atmosphere (more intensive transport from higher levels)
and O3 photochemistry).
The high recurrence of cases in Figure 4 with most polluted air (High band)
appears again in the city centre almost at all hours in all years. In the city centre
can be observed more than 20% "High" pollution in the night and 10% at the day.
Bad AQ status from the High band almost never disappears. In 2018 and mostly
in 2019 it’s make impression that the cases with more worse AQ status are much
higher than the other years, as the recurrence around the Kostinbrod, TPP Sofia and
Sofia airport can reach about 70% at the different time of the day.
At the plots for average for the 2008-2019 in High band can be seen that at all
hours there are places with worst AQ status where the recurrence is almost 20% of
all happenеd cases.

CONCLUSIONS
The simulations for Sofia city show that the air quality status of Sofia is not so
good (evaluated with a spatial resolution of 1km).
AQ status falls mostly in Low and Moderate bands, but the recurrence of cases
with High pollution is close to 20% mostly at the city center.
The recurrence of cases in Low and Moderate bands has been different for dif-
ferent years.
The pollution in the city is probably due to the surface sources like road trans-
port and also the TPPs in the city and Sofia airport.
Apart from these general features the climatic behavior of the AQI probabili-
ties is rather complex with significant spatial, seasonal and diurnal variability. The
60
Preliminary Results for the Recurrence of Air Quality...

areas with slightly worse AQ status are not necessarily linked to the big pollution
sources. Wide rural and even mountain regions can also have significant probability
for AQI from the Moderate range.
The hot spot in Sofia city, where index with higher impact (High band) is the
city center. The recurrence in High band is relatively high - about 20 % in the morn-
ing and 10% in the afternoon.

ACKNOWLEDGMENTS
Bulgarian National Science Fund (grant DN-04/2/13.12.2016).
Special thanks are due to US EPA and US NCEP for providing free-of-charge
data and software and to the Netherlands Organization for Applied Scientific
research (TNO) for providing the high-resolution European anthropogenic
emission inventory.

REFERENCES
Bojilova R., Mukhtarov P., Miloshev N., (2020) Climatology of the Index
of the Biologically Active Ultraviolet Radiation for Sofia. An Empirical
Forecast Model for Predicting the UV-Index, Comptes rendus de
l’Acad´emie bulgare des Sciences, 73, 4, 2020, 531-538
Bojilova R., Mukhtarov P., (2019) Response of the electron density profiles
to geomagnetic disturbances in January 2005, Stud. Geophys. Geod., 63,
3, 436-444
Byun, D., Ching, J., (1999), Science Algorithms of the EPA Models-3
Community Multiscale Air Quality (CMAQ) Modeling System. EPA
Report 600/R-99/030, Washington DC.
Byun, D., Schere, K.L., (2006), Review of the governing equations,
computational algorithms, and other components of the models-3
community multiscale air quality (CMAQ) modeling system. Appl.
Mech. Rev. 59, 51–77
CEP, (2003), Sparse Matrix Operator Kernel Emission (SMOKE) Modeling
System, University of North Carolina, Carolina Environmental Programs
- CEP, Research Triangle Park, North Carolina.
Chervenkov H., Slavov K., Ivanov V., (2019) STARDEX and ETCCDI
climate indices based on E-OBS and CARPATCLIM: Part one:
General description, Lecture Notes in Computer Science (including
subseries Lecture Notes in Artificial Intelligence and Lecture Notes in
Bioinformatics), 11189 LNCS, pp. 360-367
Coats, C.J.,Jr., and Houyoux, M.R., (1996), Fast Emissions Modeling With the
Sparse Matrix Operator Kernel Emissions Modeling System, The Emissions
61
Georgi Gadzhev

Inventory: Key to Planning, Permits, Compliance, and Reporting, Air and


Waste Management Association, New Orleans, September
Denier van der Gon, H., Visschedijk, A., van de Brugh, H., Droge, R.,
(2010), A high resolution European emission data base for the year
2005, TNO-report TNO-034-UT-2010-01895 RPT-ML, Apeldoorn, The
Netherlands
Dennis,R.L.,Byun,D.W.,Novak,J.H.,Galluppi,K.J.,Coats,C.J.,andVouk,M
.A., (1996), The Next Generation of Integrated Air Quality Modeling:
EPA’s Models-3, Atmospheric Environment 30, 1925–1938.
EMEP/CORINAIR (2002), Atmospheric emission inventory guidebook,
third edition, European Environmental Agency.
EPA (2009). Technical assistance document for the reporting of daily
air quality - the Air Quality Index (AQI). EPA- 454/B-09-001, US
Environmental Protection Agency, Research Triangle Park, North
Carolina, Office of Air Quality Planning and Standards, Research
Triangle Park, North Carolina 27711.EU (2007à).
Etropolska, I., Prodanova, M., Syrakov, D., Ganev, K., Miloshev, N.,
Slavov, K., (2010), Lecture Notes in Computer Science (including
subseries Lecture Notes in Artificial Intelligence and Lecture Notes in
Bioinformatics), 6046 LNCS, 2011, 141-149
European Parliament (2002), DIRECTIVE 2002/3/EC of 12 February
2002 relating to ozone in ambient air, Official Journal of the European
Communities (9.3.2002) L67, 14–30
Evtimov S., Ivanov V., (2008) Correspondence analysis of the atmospheric
phenomena in Bulgaria, Comptes Rendus de L'Academie Bulgare des
Sciences, 61(4), 443-450
Gadzhev, G., (2018) Recurrence of air quality for the city of Sofia for 2013
and 2014, Bulgarian Geophysical Journal, 2018, 41, 46-58
Gadzhev, G., Ganev, K., Miloshev, N., Syrakov, D., Prodanova, M.,
(2013) Numerical Study of the Atmospheric Composition in Bulgaria,
Computers and Mathematics with Applications, 65, 402-422.
Georgieva, I., (2014) Study of the air quality index climate for Bulgaria,
Proc. of the international conference on numerical methods for scientific
computations and advanced applications, May 19-22, 2014, Bansko,
ISBN978-954-91700-7-8, 39-42.
Georgieva I., G.Gadzhev, K. Ganev, M. Prodanova, D. Syrakov, N.
Miloshev (2015) Numerical study of the air quality in the city of Sofia
– some preliminary results, International Journal of Environment and
pollution, 57, Nos. 3/4, 162-174.
Georgieva, I., Ivanov, V., (2017), Impact of the Air Pollution on the Quality
of Life and Health Risks in Bulgaria, HARMO 18 - 18th International

62
Preliminary Results for the Recurrence of Air Quality...

Conference on Harmonisation within Atmospheric Dispersion Modelling


for Regulatory Purposes, Proceedings, 647-652
Georgieva, I., Ivanov, V., (2018), Computer simulations of the impact of
air pollution on the quality of life and health risks in Bulgaria, Int. J.
Environment and Pollution, 64, Nos. 1/3, 2018, 35-46
Houyoux, M.R., and Vukovich, J.M., (1999), Updates to the Sparse Matrix
Operator Kernel Emission (SMOKE) Modeling System and Integration
with Models-3, The Emission Inventory: Regional Strategies for the
Future, Raleigh, NC, Air and Waste Management Association
Ivanov V., Evtimov S., (2014), Wind chill hazard in Bulgaria during 2003-
2012 period, Comptes Rendus de L'Academie Bulgare des Sciences,
67(11), pp. 1521-1530
Ivanov, V. and Georgieva, I., (2017) Air quality index evaluations for Sofia city,
17th IEEE International Conference on Smart Technologies, EUROCON
2017 - Conference Proceedings, Article number 8011246, Pages 920-925
Kaleyna P., Pl. Muhtarov, N. Miloshev, (2013a) Condition of The
Stratospheric and Mesospheric Ozone Layer over Bulgaria for the
Period 1996-2012: Part 1 – Total Ozone Content, Seasonal Variations,
Bulgarian Geophysical Journal, 39, National Institute Of Geophysics,
Geodesy And Geography, Bulgarian Academy Of Sciences
Kaleyna P., Pl. Muhtarov, N. Miloshev, (2013b) Condition of The
Stratospheric and Mesospheric Ozone Layer over Bulgaria for the
Period 1996-2012: Part 2 – Total Ozone Content, Short term variations,
Bulgarian Geophysical Journal, 39, National Institute Of Geophysics,
Geodesy And Geography, Bulgarian Academy Of Sciences
Kaleyna P., Mukhtarov P., Miloshev N., (2014) Seasonal Variations of the
Total Column Ozone over Bulgaria in the Period 1996–2012, Comptes
rendus de l’Acad´emie bulgare des Sciences, 67, 7.
Leeuw, F. de, Mol, W., (2005), Air Quality and Air Quality Indices: a world
apart. ETC/ACC Technical Paper 2005/5
Monteiro, A., Lopes, M., Miranda, A.I., Borrego, C. and Vautard, R., (2005),
Air pollution forecast in Portugal: a demand from the new air quality
framework directive, Internat. J. Environment and Pollution, 25, 4–15
Mukhtarov P., Bojilova R., (2019) INFLUENCE OF SOLAR AND
GEOMAGNETIC ACTIVITY ON THE IONOSPHERE OVER
BULGARIA, Comptes rendus de l’Acad´emie bulgare des Sciences, 70,
9, 1289-1296
Poupkou, A., I. Kioutsioukis, I. Lisaridis, K. Markakis, D. Melas, C.
Zerefos and Giannaros, T., (2008b) Air quality forecasting for Europe,
the Balkans and Athens, 3rd Environmental conference of Macedonia,
14-17 March 2008, Thessaloniki, Greece.

63
Georgi Gadzhev

San Jose, R., J. Perez and R. Gonzalez, (2006), Air Quality Real-Time
Operational Forecasting System for Europe: an application of MM5-
CMAQ-EMIMO modeling system, in Brebia, C. (Ed.), “Air Pollution
XIV”, WIT Press, Southampton, Boston, 75
Skamarock et al., (2007), A description of the Advanced Research WRF
Version 2.
Sofiev, M., Siljamo, P., Valkama, I., (2006), A dispersion modeling system
SILAM and its evaluation against ETEX data. Atmos. Environ. 40, 674–
685
Syrakov, D., Etropolska, I., Prodanova, M., Ganev, K., Miloshev, N.,
Slavov, K., (2012), Operational Pollution Forecast for the Region of
Bulgaria, American Institute of Physics, Conf. Proc. 1487.
Syrakov, D., Etropolska, I., Prodanova, M., Slavov, K., Ganev, K.,
Miloshev, N., Ljubenov T., (2013), Downscaling of Bulgarian Chemical
Weather Forecast from Bulgaria region to Sofia city, American Institute
of Physics, Conf. Proc. 1561, 120-132.
Syrakov, D., Prodanova, M., Etropolska, I., Slavov, K., Ganev, K., Miloshev,
N., Ljubenov T., (2014a), A multy-domain operational chemical
weather forecast system, Lecture Notes in Computer Science (including
subseries Lecture Notes in Artificial Intelligence and Lecture Notes in
Bioinformatics), 8353 LNCS, 2014, 413-420
Syrakov, D., Prodanova, M., Georgieva, E., Etropolska, I., Slavov, K., (2014b),
Impact of NOx emissions on air quality simulations with the Bulgarian
WRF-CMAQ modelling system, HARMO 2014 - 16th International
Conference on Harmonisation within Atmospheric Dispersion Modelling
for Regulatory Purposes, Proceedings 2014, 187-190
Syrakov, D., Prodanova, M., Georgieva, E., Etropolska, I., Slavov, K.,
(2015), Impact of NOx emissions on air quality simulations with the
Bulgarian WRF-CMAQ modelling system, International Journal of
Environment and Pollution, 57, Issue 3-4, 2015, 285-296

 Georgi Gadzhev
http://orcid.org/0000-0002-6159-3554
National Institute of Geophysics, Geodesy and Geography
Bulgarian Academy of Sciences
Sofia, Bulgaria
E-mail: [email protected]

64
Proceeding of 1st International conference on Environmental protection and disaster RISKs

https://doi.org/10.48365/envr-2020.1.6 Air Pollution,


Climate and Health

PARTICULATE MATTER CHARACTERISTICS


AND ATMOSPHERIC BOUNDARY LAYER HEIGHT
OVER SOFIA

Plamen Savov1, Nikolay Kolev1, 2,


Ekaterina Batchvarova3, Hristina Kirova4, Maria Kolarova4
1
Department of Physics, University of Mining and Geology “St. Ivan Rilski”
2
Institute of Electronics – Bulgarian Academy of Sciences (IE-BAS)
3
Climate, Atmosphere and Water Research Institute –
Bulgarian Academy of Sciences (CAWRI-BAS)
4
National Institute of Meteorology and Hydrology (NIMH)

Abstract: The paper presents results from aerosol experimental campaigns


performed in the urban environment of Sofia. Laser Particle Counter (PLC) data of
2 summer days (June, 7 and 8) and 4 winter days (December, 18, 19, 20, 21) of
2019 are discussed. Aerosol particle concentrations (number/l and mass µg/m3) in
channels 0-2.5 µm and 2.5-10 µm are measured at the open green area near Pliska (at
30 m distance from boulevard Tsarigradsko shose). The combined effect of the daily
development of atmospheric boundary layer height, meteorological parameters, and
hourly variations in the concentrations of the aerosol fractions is discussed. WRF-
GDAS and HYSPLIT models are used for determination of the atmospheric boundary
layer height and to follow the transport of air masses. BSC-DREAM dust model is
used as an additional source of information to assess the long-range intrusion of dust.
Keywords: air quality, aerosol concentration, atmospheric boundary layer (ABL),
particle number concentrations, WRF-(GDAS) model, HYSPLIT, BSC-DREAM

INTRODUCTION
The Atmospheric Boundary Layer (ABL) height determines the volume in which
different gaseous and aerosol pollutants are mixed due to turbulent processes within the
atmosphere. To model the ABL height in urban environments is additional challenge
due to the specific physical and chemical characteristics which play important role
for pollutant dispersion, climate comfort, and weather forecasting (Chen et al., 2011;
Batchvarova and Gryning, 2006; Batchvarova et al., 2006; Batchvarova et al., 2011;
Rotach et al., 2005; Avolio et al., 2017). Specific studies on the city of Sofia including
remote sensing measurements are presented by Kirova&Batchvarova, 2017; Kolev
et al., 2016; Savov et al., 2016; Kolev et al., 2019. The combination of models and
particle counter measurements provides comprehensive information on both the
65
P. Savov, N. Kolev, E. Batchvarova, H. Kirova, M. Kolarova

aerosols concentration characteristics in the urban atmosphere as well as about the


vertical structure of aerosol layers and meteorological parameters determining the
transport of air mass (Ngan et al., 2015; Stein et al.,. 2015; Lin Su et al., 2015).

METHODOLOGY AND EQUIPMENT


Experimental site and instruments
A two-channel BQ20 (TROTEC, Germany) laser particle counter (LPC) with
channel 1 (0 - 2.5 µm) and channel 2 (2.5 - 10 µm) denoted further in the paper
as PM2.5 and PM10, respectively, was used to measure instantly the number of
particles and their mass with time step 10 or 15 minutes. The sampling rate is
0.9 l/min. The accuracy of the devices is in the range of 15-20%. Measurements
were performed near bus stop Pliska (in the green area in front of the building
of Philosophy Faculty of Sofia University), 30 m North of Blvd. “Tsarigradsko
shose”, the largest boulevard in Sofia, one of the roads of heaviest traffic. It runs
11 km from west to east connecting the center of city with Trakiya highway (A1).
Only cars and busses are allowed with speed limit of 80 km/h.
Meteorological data were obtained from an automatic weather station located at
Sofia airport with time resolution 30 minutes.

Experimental days
Measurements were performed during two typical summer days (7-8 June 2019),
characterized by relatively calm anticyclone weather, and four winter days of (18
– 21 December 2019) characterised with high level of pollution and occurrence of
fog in the morning hours. Particular behaviour of the aerosol concentrations was
noted on December 21, a day with foehn wind.

Application of models
The Weather Research and Forecasting Model (WRF) is a numerical weather
forecasting and atmospheric simulation system designed for both research and
operational applications. The WRF model is used for a wide range of meteorological
applications from tens to thousands of kilometres. The purpose of applying the
WRF-GDAS model is to produce new atmospheric analyses using historical data
(available from 2004 to present) and to analyze current atmospheric conditions by
using Global Data Assimilation System (GDAS).
The Hybrid Single-Particle Lagrangian Integrated Trajectory model (HYSPLIT,
http://www.arl.noaa.gov/ ready/hysplit4.htm), a Lagrangian dispersion model,
has been coupled (online) to the Weather Research and Forecasting (WRF)
meteorological model in such a way that the HYSPLIT calculation is run as part of
the WRF-ARW prediction calculation (Ngan et al., 2015; Lin Su et al., 2015) . The
embedded HYSPLIT includes dispersion, trajectories, deposition (dry and wet),
etc. (Chen et al., 2011; Stein et al., 2015).
66
Particulate Matter Characteristics...

The updated BSC-DREAM8b and the new NMMB/BSC-Dust models (https://


ess.bsc.es/bsc-dust-daily-forecast), developed in the Earth Sciences Department to
simulate and/or predict the atmospheric cycle of mineral dust at BSC were used to
assess the long-range transport of dust over the area of Sofia.

RESULTS AND DISCUSSIONS


Summer 2019
Anticyclone synoptic conditions prevailed on 7 and 8 June 2019 causing sunny
weather and development of cumulus clouds in the afternoon mostly on the first
day. Measured maximal temperatures were 25-30 oC and minimal - 13 oC.

(a) (b)
Figure 1. Daily variations in PM2.5 and PM10 on 7 June 2019,
measured as (a) particles per liter (N/l) and (b) concentration mass (µg/m3)

The number of PM10 changes from 20 N/l in the morning hours to 10 N/l in the
afternoon, and of PM2.5 from 250 N/l to 100 N/l (Fig. 1a). The corresponding mass
concentrations for PM10 are from 25-30 µg/m3 to 5 µg/m3 in the afternoon and for
PM2.5 from 10-12 µg/m3 to 5 µg/m3 (Fig.1b).
WRF-GDAS model forecast for Boundary Layer Depth (Zi) over Sofia is 1200
– 1300 m (Fig. 2a). The model suggests prevailing wind from N (1-4 m/s) from 9
to 15 LT. HYSPLIT Backward trajectories of 1000, 2000 and 3000 m ending in
Sofia at 12UTC on June, 7 2019 are shown in Fig. 2b. Transport of air masses from
Morocco and Sahara desert at altitude of 3 km can be noted.
Dust concentrations of 20 µg/m3 forecasted by BSC Dust model at 4 km height
do not influenced the level of particulates measured in the urban surface layer.
NMMB/BSC Dust (Dust Forecast at 06UTC Friday 07 June) gives a higher con-
centration of about 20 µg/m3 in the layer between 3 and 6 km height forecasted over
Sofia (Fig. 3a). LON-Height cross-section and LAT-Height cross-section present
67
P. Savov, N. Kolev, E. Batchvarova, H. Kirova, M. Kolarova

the distribution of dust concentration across North Africa and Europe (Fig. 3b). The
dashed line indicates the position of Sofia.

(a) (b)
Figure 2. Model results for Sofia (LAT 42.65; LON 23.38)
on 7 June (a) GDAS Stability plot and (b) HYSPLIT Backward trajectories

(a) (b)
Figure 3. NMMB/BSC Dust Forecast Concentrations (µg/m3)
at 06 UTC on 7 June (a) vertical profile and (b) LON&LAT-Height cross-section
68
Particulate Matter Characteristics...

Measurements on 8 June show that the number of PM10 is around 20 N/l with
max of 25 N/l from 9:30 to 10:30 LT and for PM2.5 the numbers are 220 in the
early morning hours and increase to 300-350 N/l from 9:30 to 10:30 LT, then fall
to 175 N/l after 12 LT (Fig. 4 a). The corresponding mass concentrations for PM10
are from 15 to 30 µg/m3. For PM2.5 the concentrations are around 10 µg/m3 before
12 LT and around 5 µg/m3 in the early afternoon (Fig. 4 b).

(a) (b)
Figure 4. Daily variations in PM2.5 and PM10 on 8 June 2019,
measured as (a) particles per liter (N/l) and (b) concentration mass (µg/m3)

WRF-GDAS model forecast for ABL Height for Sofia is 1800-1900 m.


Pasquill Stability Class is from stable (E) in the morning to unstable (B-C) at
12 UTC (Fig. 5a). GDAS wind speed for Sofia on July 8 is 7 m/s in the morning
hours and 1-4 m/s after 11 LT from NW. The HYSPLIT Backward trajectories of
1000, 1500 and 2000 m ending in Sofia at 12UTC on June, 8 2019 are shown in
Fig. 5b. The transport of air masses at the tree levels is from different areas. The
2000-metre trajectory starts from Algeria. NMMB/BSC Dust (Dust Forecast at
06UTC Sat 08 June) shows a higher concentration of about 20 µg/m3 in the layer
below 2 km, Fig 6a. This suggests that long-range transport of dust may influence
the surface PM concentrations in Sofia given the fact that ABL height is about
1900 m. The LON&LAT-Height cross-sections present different distribution of
dust concentration compared to the previous day (Fig 6b).
During these summer days, the measured concentrations were lower than the
sanitary norms. The height of the convective ABL reached 1350 and 1900 m, on 7
and 8 June respectively. On 7 June (Friday) the morning rush hour concentration
peak is at 10 LT and is strongly pronounced (Fig. 1). On 8 June (Saturday), high
concentrations remain between 9 and 12 LT. Secondary increase is noted at 13 and
69
P. Savov, N. Kolev, E. Batchvarova, H. Kirova, M. Kolarova

(a) (b)

Figure 5. Model results for Sofia on 8 June (a) GDAS Stability plot
and (b) HYSPLIT Backward trajectories

(a) (b)

Figure 6. NMMB/BSC Dust Forecast Concentrations (µg/m3) at 06 UTC


on 8 June (a) vertical profile and (b) LON&LAT-Height cross-section

14 LT. Summary of measured and modelled parameters for the summer days is
given in Table 1.
70
Particulate Matter Characteristics...

Table 1. PM10 and PM2.5 concentrations and meteorological


parameters (modelled and observed)
Day of PM10 mass PM10 number PM2.5 conc. PM2.5 ABLH Wind dir/ Tmax /
experi- min-max min-max min-max (µg/ min-max min-max speed (m/s) Tmin
ment (µg/m3) (N/l) m3) (N/l) (m) (model) (oC)
7 June 5-30 5-20 2-12 50-250 250-1350 N / 1-4 25-30 / 13
8 June 15-30 10-30 5-10 100-350 200-1900 NW / 1-4 25-30 / 13

It can be noted that the measured particle number is higher on Saturday, 8 June,
despite the expected lower traffic contribution. Furthermore, the ratio highest/
lowest values is smaller on that day (PM2.5: 350/100 N/l), while on 7 June the
ratio is 250/50 N/l. Possible reason for these differences is the long-range transport
contribution (Fig. 4 and Fig. 6).

Winter 2019
Anticyclone synoptic circulation prevailed in the period 18-20 December 2019
causing low wind speed from SW-S, morning fog conditions in valleys and low
stratus clouds. Measured maximal temperatures were 10 oC and minimal - 0 oC.
No precipitation and no snow cover were measured in Sofia. Change of weather to
cyclone circulation starts from 20 December, leading to feohn wind from S-SW in
Sofia reaching 16 m/s, gusts up to 23 m/s and abnormally high temperatures of 15 oC.
Measurements on 18 December were performed from 8 to 14 LT under heavy
fog (observed relative humidity of 100 %) conditions and show number of PM10
of 200 N/l until 10 LST and lower values (100 N/l) in the afternoon. Registration
for PM2.5 shows 4500 N/l in the early morning hours, maximum of 5000-5500 N/l

(a) (b)
Figure 7. Daily variations in PM2.5 and PM10 on 18 December 2019,
measured as (a) particles per liter (N/l) and (b) concentration mass (µg/m3)
71
P. Savov, N. Kolev, E. Batchvarova, H. Kirova, M. Kolarova

around 10 LT (related to morning rush hours and fog) and gradual decrease to 2500-
2000 N/l after 12 LT (Fig. 7a). The corresponding mass concentrations for PM10
(Fig. 7b) is from 220 µg/m3 at 8 LT, with max of 250-260 µg/m3 at 9:30-10:30 LT
and decrease to 100 µg/m3 in the early afternoon. The PM2.5 concentrations show
similar behavior: around 140 µg/m3 before 12 LT with, 160-170 µg/m3 at 10:30 LT
and decrease to 60 µg/m3 in the early afternoon. Summarized information of the
winter experimental days is given in Table 2.
GDAS model ABL height of maximum 200 m and stable to neutral stratification
over Sofia (Fig. 8a), constant SW-S wind of 1-4 m/s during the day. The 96-hour back-
ward trajectories of height 200, 500 and 1000 m show constant SW flow (Fig. 8b).

Table 2. PM10 and PM2.5 concentrations and meteorological


parameters (modelled and observed)
Day of PM10 mass PM10 number PM2.5 mass PM2.5 ABLH Wind Tmax /
experi- min-max min-max min-max number min-max dir/ speed Tmin
ment (µg/m3) (N/l) (µg/m3) min-max (N/l) (m) (m/s) (model) (oC)
18 Dec. 80-260 100-200 50-170 2000-5500 50-180 / fog SW-S / 1-4 10 / 0
19 Dec. 150-600 100-400 100-375 3000-11000 50-150 / fog S-SE / 1-4 10 / 0
20 Dec. 70-120 50-100 30-70 1100-1700 450-150 S / 1-4 10 / 0
21 Dec. 10-100 20-90 10-40 300-1400 Strong wind S-SW/16 obs 10-15 / > 0

(a) (b)
Figure 8. Model results for Sofia on 18 December (a) GDAS Stability
plot and (b) HYSPLIT Backward trajectories
72
Particulate Matter Characteristics...

Measurements on 19 December were performed from 7 to 12 LT. The stable


stratification and fog persisted leading to double increase of all maximal particle
number and mass concentrations compared to the previous day (Fig. 9 and Table 2).

(a) (b)
Figure 9. Daily variations in PM2.5 and PM10 on 19 December 2019,
measured as (a) particles per liter (N/l) and (b) concentration mass (µg/m3)

GDAS ABL height over Sofia is below 200 m , stratification is stable to very stable
(Fig. 10a), and constant S-SE wind of 1-4 m/s during the day. The 96-hour backward
trajectories of height 200, 500 and 1000 m show SW flow from Sicily (Fig. 10b).

(a) (b)

Figure 10. Model results for Sofia on 19 December (a) GDAS Stability plot
and (b) HYSPLIT Backward trajectories
73
P. Savov, N. Kolev, E. Batchvarova, H. Kirova, M. Kolarova

On 20 December the meteorological conditions start to change to multi-


centered low pressure structure over the Balkan Peninsula. The transport of
warm air masses from SW causes increase of temperatures. NMMB/BSC Dust
Forecast model indicates dust concentration of 50 µg/m3 at 4000 m, showing
long-range transport, which remains far above the ABL (maximal of 450 m)
over Sofia.
The number and mass concentration measurements on 20 December show
distinct decrease (two times in values) and different from the previous days
changes with time. PM2.5 mass concentration starts from 70 µg/m3 in the early
morning (8:30 LT), a minimum of 30 µg/m3 is recorded between 10 and 11 LT,
related to the destruction of the fog. Higher values (80 µg/m3) are recorded
again between 11:30 and 12:30 LT, related with increased traffic of cars leaving
the capital for the following 6 days of Christmas. For PM10 the concentrations
are with similar behavior to PM2.5 but with higher values starting from 100 µg/
m3 in the morning (8:30 LT), diminishing to 70 (µg/m3) and growing again to
120 µg/m3 after 11 LT. After 12:30 LT the concentrations slowly decrease (Fig.
11 and Table 2).

(a) (b)
Figure 11. Daily variations in PM2.5 and PM10 on 20 December 2019,
measured as (a) particles per liter (N/l) and (b) concentration mass (µg/m3)

On 21 December the cyclone formed over Italy moves to NE above the


Hungarian Plane. The southwesterly wind speed increases. Passing over Vitosha
Mountain this flow is observed as feohn in Sofia reaching wind speed of 16 m/s at
Sofia airport and gusts of 23 m/s.
The measured PM10 and PM2.5 concentrations show peculiarities related to
the new meteorological situation starting with low morning concentrations from
7 – 10 LT (PM2.5: 10-20 µg/m3 and for PM10: 30 – 40 µg/m3). Very rapidly the
74
Particulate Matter Characteristics...

concentrations increase twice to 50 µg/m3 and 90 µg/m3, correspondingly (Fig.


12b). After 10:30 LT the concentrations decrease. This behavior is possibly related
to short time wind gust situation during the peak period when large amounts of dust
are lifted from surface. It is interesting to note that PM10 number concentration
increased twice and PM2.5 three times for the period 10-10:30 LT (Fig. 12a). The
peak cannot be explained with high transport traffic, as it is Saturday of a 6-day
holiday period.

(a) (b)
Figure 12. Daily variations in PM2.5 and PM10 on 21 December 2019,
measured as (a) particles per liter (N/l) and (b) concentration mass (µg/m3)

The last 2 days of the winter experimental campaign are peculiar in view of
particle concentrations and show the dynamic interplay between meteorology and
sources for the air quality in the city.

CONCLUSIONS
This study presents experimental results concerning the daily distribution of
the aerosol particles in the urban environment and the correlation to the synoptic
situations, meteorological parameters and ABL height. The analysis is based on
2 summer and 4 winter days of aerosol particles (concentration and number)
measurements with laser aerosol particle counters.
In summer, the ABL is high and the observed concentration of aerosol particles
is under the sanitary norms and changes two times from maximal to minimal values.
The maximum is related to intensive transport traffic along the Tsarigradsko shose
in the morning hours.
In winter, the ABL height is low the observed concentrations of aerosol particles
are higher than the norms for 18 and 19 December. The maximal values are 3 times
higher than the minimal and are related to fog conditions and intensive transport
75
P. Savov, N. Kolev, E. Batchvarova, H. Kirova, M. Kolarova

traffic along the Tsarigradsko shose in the morning hours. On 20 and 21 December
the concentrations show peculiar changes with time related to Christmas holidays
and rare meteorological conditions.

ACKNOWLEDGMENTS
This work is carried out in the framework of the National Science Program
"Environmental Protection and Reduction of Risks of Adverse Events and
Natural Disasters", approved by the Resolution of the Council of Ministers
№ 577/17.08.2018 and supported by the Ministry of Education and Sci-
ence (MES) of Bulgaria (Agreement № D01-230/06.12.2018). The Contri-
butions of Kolarova and Kirova is supported by the project DN4/7 (Study
of the PBL structure and dynamics over complex terrain and urban area),
funded by the National Science Fund of Bulgaria. The authors gratefully
acknowledge the NOAA Air Resources Laboratory (ARL) for the provi-
sion of the WRF-GDAS Model used in this publication; the NOAA Air
Resources Laboratory (ARL) for the provision of the HYSPLIT transport
and dispersion model. Some data and images used in the publication are
from the (NMMB/BSC-Dust or BSC-DREAM8b) model, operated by the
Barcelona Supercomputing Center (http://www.bsc.es/ess/bsc-dust-daily-
forecast/).

REFERENCES
Avolio, E., Federico, S., Miglietta, M., Lo Feudo, T., Calidonna, C.,
& Sempreviva, A. (2017). Sensitivity analysis of WRF model PBL
schemes in simulating boundary layer variables in southern Italy: An
experimental campaign. Atmospheric Research, 192, 58–71.
Batchvarova, E. & Gryning, S.-E. (2006). Progress in Urban Dispersion
Studies. Theor. Appl. Climatol. 84, No. 1-3, 57-67.
Batchvarova, E., Gryning, S.-E., Rotach, M. & Christen, A. (2006).
Comparison of modelled aggregated turbulent fluxes and measured
turbulent fluxes at different heights in an urban area. In: Air pollution
modeling and its application XVII, Borrego, C. and Norman, A.
(Eds.), Kluwer Academic/Plenum Publishers (NATO Challenges of
Modern Society series) 363-370.
Batchvarova E., Kolarova M., Veleva B., Neykov N., Neitchev P.,
Videnov P., Gamanov A., & Barantiev D. (2011). The atmospheric
boundary layer – parameterizations, observations and applications.
(2011) Bul. J. Meteo & Hydro, 16/1, 41-53.
Chen, F., Kusaka, H., Bornstein, R., Ching, J., Grimmon, C.S.B.,
Grossman-Clarke, S., Loridan, T., Manning, K.W., Martilli, A.,
76
Particulate Matter Characteristics...

Miao, S., Sailor, D., Salamanca, F.P., Taha, H., Tewari, M., Wang,
X., Wyszogrodzki, A.A., & Zhang, C. (2011). The integrated WRF/
urban modelling system: development, evaluation, and applications
to urban environmental problems. Int. J.Climatol. 31 (2), 273–288.
Kirova, H. & Batchvarova, E. (2017). Mesoscale simulation of
meteorological profiles during the Sofia Experiment 2003.
International Journal of Environment and Pollution, 61, 2, 2017, 134-
147. DOI:10.1504/IJEP.10006760
Kolev, N., Savov, P., Evgenieva, Ts., Miloshev, N., Petkov, D., &
Donev, E. (2016). Summer measurements of atmospheric boundary
layer (ABL), aerosol optical depth (AOD) and water vapour content
(WVC) over Sofia (Bulgaria) 2010–2014. Compt. rend. Acad. bulg.
Sci., 69, 4, 421-429.
Kolev, N., Savov, P., Evgenieva, Ts., Miloshev, N., Gueorguiev,
O., Batchvarova, E., Kolarova, M. Danchovski, V., Ivanov, D. &
Petkov, D. (2019). Investigation of the atmospheric boundary layer
and optical characteristics of the atmospheric aerosols over Sofia in
summer 2016, 10th Jubilee International Conference of the Balkan
Physical Union, AIP Conference Proceedings 2075, 120004 (2019)
https://doi.org/10.1063/1.5091262, Published Online: 26 February
2019.
Lin Su, Yuan, Z., Hung Fung, J. C., & Hon Lau, A. K. (2015). A
comparison of HYSPLIT backward trajectories generated from two
GDAS datasets, Science of the Total Environment, 506-507, 527-
37.
Ngan, F., Stein, A., & Draxler, R. (2015). Inline coupling of WRF–
HYSPLIT: Model development and evaluation using tracer
experiments. J. Appl. Meteor. Climatol., 54, 1162–1176
Rotach, M. W., Vogt, R., Bernhofer, C., Batchvarova, E., Christen, A.,
Clappier, A., Feddersen, B., Gryning, S.-E., Martucci, G., Mayer,
H., Mitev, V., Oke, T. R., Parlow, E., Richner, H., Roth, Roulet,
M., Y.-A., Ruffieux, D., Salmond, J. A., Schatzmann, M. & Voogt,
J. A. (2005). BUBBLE – an Urban Boundary Layer Meteorology
Project, Theor. Appl. Climatol. 81, 231–261, DOI 10.1007/s00704-
004-0117-9.
Savov, P., Kolev, N., Evgenieva, Ts., Vatzkitcheva, M., & Danchovs-
ki, V. (2016). Correlations between particle number concentrations,
boundary layer height, meteorological parameters and urban environ-
ments. Compt. rend. Acad. bulg. Sci., 69, 1, 19-24.
Stein, A., Draxler, R., Rolph, G., Stunder, B., Cohen, M., & Ngan, F.
(2015). NOAA’S HYSPLIT Atmospheric transport and dispersion

77
P. Savov, N. Kolev, E. Batchvarova, H. Kirova, M. Kolarova

modeling system. Bulletin of the American Meteorological Society,


96, 12 (DEC. 2015), 2059-2078. Published by: American Meteoro-
logical Society.

 Plamen Savov
Department of Physics
University of Mining and Geology “St. Ivan Rilski”
Sofia, Bulgaria
E-mail: [email protected]

 Nikolay Kolev
Department of Physics
University of Mining and Geology “St. Ivan Rilski”
Institute of Electronics
Bulgarian Academy of Sciences
Sofia, Bulgaria
E-mail: [email protected]

 Ekaterina Batchvarova
Climate, Atmosphere and Water Research Institute
Bulgarian Academy of Sciences
Sofia, Bulgaria
E-mail: [email protected]

 Hristina Kirova
National Institute of Meteorology and Hydrology
Sofia, Bulgaria
E-mail: [email protected]

 Maria Kolarova
National Institute of Meteorology and Hydrology
Sofia, Bulgaria
E-mail: [email protected]

78
Proceeding of 1st International conference on Environmental protection and disaster RISKs

https://doi.org/10.48365/envr-2020.1.7 Air Pollution,


Climate and Health

SENSITIVITY TO THE WRF MODEL


CONFIGURATION OF THE WIND CHILL INDEX
FOR SOFIA REGION – PRELIMINARY RESULTS

Vladimir Ivanov1, Reneta Dimitrova1,2


National Institute of Geophysics, Geodesy and Geography –
1

Bulgarian Academy of Sciences (NIGGG-BAS)


2
Faculty of Physics, Department of Meteorology and Geophysics,
“St. Kliment Ohridski” University of Sofia

Abstract: The thermal comfort is one of the main issues nowadays. The
implications from the modelled future climate projections also put a question of
the expected thermal environmental conditions. The objective of this paper is a
winter human discomfort study for Sofia and its surroundings. Data from numerical
simulations with Weather Research and Forecasting (WRF) model were used to
calculate a Wind Chill index, which describes the deviation from the winter thermal
comfort. This index depends on air temperature and wind, and it is categorized in
a 6-grade scale, which describes how far or close are the environmental conditions
from the human thermal comfort. Numerical experiments with combination
of different parameterization schemes for atmospheric boundary layer and
microphysical processes were carried out. Model performance for the temperature
and wind speed were used for estimation of the best model options for calculation
of the Wind Chill index.
Keywords: Wind chill index modelling, WRF model performance estimation

INTRODUCTION
The thermal comfort is one of the main issues nowadays. The implications
from the modelled future climate projections also put a question of the expected
thermal environmental conditions (Stocker et al., 2013) and related air pollution
(Gadzhev et al., 2014a, Gadzhev et al., 2014b, Gadzhev et al., 2012, Gadzhev et
al., 2011). The population of the cities increase more and more nowadays, which
entails the growth of the cities through the gradual amplification of urbanizing. That
intensify the urban heat island effect (Rizwan, Dennis & Chunho, 2008), change
the surroundings environment modifying the local circulations (Hidalgo, Masson,
& Gimeno, 2010), lead to implications in the precipitation distribution (Collier,
2006), etc. Therefore, the consequences from the increasing or decreasing of the
79
Vladimir Ivanov, Reneta Dimitrova

air temperatures become of bigger importance. The deviation from the thermal
comfort is expressed as a number called Wind Chill index, which depends on air
temperature, wind and/or relative humidity, and possibly from other parameters. It
is categorized in ordinal scale for deviation of the environmental conditions from
the thermal comfort for human beings (Cheng, Niu, & Gao, 2012). The thermal
discomfort for the territory of Bulgaria and the Balkan Peninsula is considered
in several studies. Some of them deal with observations (Ivanov & Evtimov,
2014a; Ivanov & Evtimov, 2014b). Their results show that the deviations from
the thermal comfort in Bulgaria could be significant from the human perspective.
Others, estimated by outputs from the regional climate model also show that the
Balkan Peninsula is subjected to heat-related illness (Ivanov, Gadzhev, Ganev &
Chervenkov, 2020).
The objective of this paper is a winter human discomfort study for Sofia and its
surroundings. The Weather Research and Forecasting Model (WRF) model was
used for simulation of temperature and wind fields. Different parameterization
schemes for atmospheric boundary layer and microphysical processes were used
for estimation of the best model options for calculation of the Wind Chill index.

METHODOLOGY
Temperature and wind fields modelling
WRF is a mesoscale numerical simulation system for research and operational
forecasting of the atmospheric environment (WRFv3.9; Skamarock et al 2008).
Five nested domains in Lambert projection, with D1 at 9 km, D2 at 3 km, and three
at 1 km horizontal resolution (D3, D4, D5) and with hourly output, were selected.
The map with the modelling domains is shown in Figure 1. The bigger domain
D1covers the north and central parts of the Balkan Peninsula, the inner domain D2 -
mainly the territory of Bulgaria, the innermost domains: D3 - the Sofia valley, D4 -
Plovdiv region, D5 - Varna region. The study considers only the innermost domain
D3, which includes geographically the city of Sofia and its surroundings with com-
plex terrain. The model was implemented with 50 pressure-based terrain-following
vertical levels from the surface to 50 hPa. The initial and boundary conditions were
derived from the 0.25-degree NCEP Final Operational Model Global Tropospheric
Analyses (http://rda.ucar.edu/datasets/ds083.2/) datasets available every 6 hours.
Data assimilation (fdda model option) was used for the outermost domains D1 for
all vertical levels and for domain D2 for the first 10 model levels above the ground.
The atmospheric physics options are responsible for the including of the sub-grid
atmospheric and surface processes, which cannot be solved explicitly by the model
- the planetary boundary layer (PBL), microphysical processes, convection, short-
wave, and longwave radiation processes, land-air interaction. The WRF physics
package included: the new version of Radiative Transfer Model - RRTMG param-
eterization (Iacono et al 2008) - for longwave and shortwave radiation to compute
80
Sensitivity to the Wrf Model Configuration...

Figure 1. Modelling domains with an enlarged view


of the innermost domain D4 and Sofia city
radiation at every 10 minutes; Noah land surface model (Chen, Dudhia 2001); and
Kain cumulus parameterization, for D1 domain only (Kain 2004). For this specific
study, the most important was to calculate properly surface temperature and wind
speed to ensure reliable calculation of the Wind Chill index. The model sensitivity
was evaluated using three PBL schemes -Yonsei University scheme (YSU; Hong
et al 2006), Mellor-Yamada Nakanishi and Niino Level 2.5 (MYNN2.5; Nakanishi,
Niino 2006), the Quasi-Normal Scale Elimination scheme (QNSE; Sukoriansky et
al., 2005), and two microphysical schemes – Thompson (Thompson et al 2008),
and Lin (Lin et al., 1983). The YSU PBL is a non-local K scheme, first-order;
the MYNN2.5 - turbulent kinetic energy, second-order scheme with level 2.5, and
QNSE uses a self-consistent, quasi-normal scale elimination algorithm and spec-
tral space representation. The Lin microphysics is a single-moment sophisticated
scheme that has ice, snow, and graupel processes, suitable for real-data high-
resolution simulations; Thompson also is a single-moment for cloud water, ice,
snow, rain, graupel, hail, but calculate also the rain number concentration. The
selections were made after laborious inspection process of various schemes in pre-
vious studies for the same domain (Egova et al. 2017, Vladimirov et al. 2018). Six
experiments were performed in this study with all possible combinations of op-
tions, described later by the names of the using schemes.

Wind Chill index calculation


The focus of the study is on the winter bio-meteorological conditions, where the
air temperature and the wind speed play the most important role in determining the
degree of discomfort. For that purpose, a typical winter month (January 2016) was

81
Vladimir Ivanov, Reneta Dimitrova

selected for simulations. The cold temperatures and bigger wind speeds makes the
human body cooling stronger, deteriorating the thermoregulation system, and pos-
sibly the tone and the health. For describing the cooling power of the wind in cold
weather, we use the Wind Chill index (Osczevski, Bluestein, 2005), given in tempera-
ture dimension. It is the temperature that the human individual feel in calm weather,
with heat losses from the body equal to ones for the given air temperature and wind
speed. The body reacts to different combinations of the air temperature (at 2m) and
wind speed (at 10m), which is categorized in a six-grade linguistic scale (Table 1).

Table 1 Wind Chill index severity categorization


Wind Chill Wind Chill
Environment Risk
Temperature category
0°C ÷ -9°C Low Risk Slight increase in discomfort
-10°C ÷ -27°C Moderate Risk Increased discomfort, with risk of hypothermia and frostbite
-28°C ÷ -39°C High Risk Exposed skin can freeze in 10 to 30 minutes
-40°C ÷ -47°C Very High Risk Exposed skin can freeze in 5 to 10 minutes
-48°C ÷ -54°C Severe Risk Exposed skin can freeze in 2 to 5 minutes
< -55°C Extreme Risk Exposed skin can freeze in less than 2 minutes

The wind chill temperature is calculated using simple relation:

(1)
The “T” is the air temperature at 2 m in C°. The “V” is wind speed at 10 m in
kmh-1. The Wind Chill index has temperature dimension, so we will call it wind
chill temperature (WCT).

WRF MODEL VALIDATION


The output from all described above model configurations was validated against
observations in winter conditions. Data for air temperature and wind speed from
the stations of the National Environment Agency at four locations in the Sofia city
(Druzhba, Nadezhda, Pavlovo, Hipodruma) and the mountain site Kopitoto were
used. Standard statistics - the Mean Bias (MB) and the Root Mean Square Error
(RMSE) were used as criteria for model performance.
The calculated MB are given in Table 2, and the RMSE in Table 3. The MB of
the temperature for Druzhba, Hipodruma, and Pavlovo is lower for the MYNN2.5_
Thompson configuration. The best model options for simulation of the tempera-
ture at Nadezhda is MYNN2.5_Lin, and at Kopitoto is YSU_Thompson. The wind
speed MB for Druzhba has minor value for the MYNN2_Lin simulation. The best
simulation for the wind speed at Hipodruma is MYNN2.5_Thompson, for Nadezh-
da it is QNSE_Lin, and for Kopitoto it is the YSU_Thompson. The wind speed at
Pavlovo is in the best agreement with observations for the MYNN2.5_Lin sim-
82
Sensitivity to the Wrf Model Configuration...

ulation. Overall, the results of the model validation by the MB suggest that the
MYNN2.5_Thompson simulates temperature the best. On the other hand, the wind
speed is simulated with the lowest MB by the MYNN2.5_Lin configuration.

Table 2 Mean Bias of the model configurations of the temperature


at 2 m (T) and wind speed at 10 m (WS)
Stations
Model configuration Druzhba Hipodruma Kopitoto Nadezhda Pavlovo
T WS T WS T WS T WS T WS
MYNN2.5_Lin -0.54 0.25 -0.96 1.10 -4.16 -5.58 -0.09 1.13 -1.07 0.84
MYNN2.5_Thompson 0.05 0.37 -0.68 1.13 -3.95 -5.47 0.40 1.17 -0.66 0.86
QNSE_Lin -3.06 0.37 -3.61 1.14 -4.02 -3.63 -2.93 1.09 -3.00 0.87
QNSE_Thompson -1.02 0.65 -1.82 1.27 -3.82 -3.63 -1.19 1.29 -1.73 1.02
YSU_Lin -0.84 0.54 -1.62 1.32 -3.60 -2.40 -0.59 1.43 -1.68 0.86
YSU_Thompson -0.21 0.56 -1.13 1.36 -3.40 -2.37 -0.23 1.48 -1.10 0.91

The smallest model RMSE for the temperature at Druzhba is MYNN2.5_


Thompson, at Hipodruma is MYNN2.5_Lin and at Kopitoto is QNSE_Thompson.
The model configuration that provides the lowest temperature RMSE at Nadejda is
MYNN2.5_Thompson and at Pavlovo is the YSU_Thompson. The smallest RMSE
of the wind speed at Druzhba and Hipodruma is the MYNN2_Lin. The QNSE_
Thompson provides the smallest RMSE at the Kopitoto site. The model set-up that
gives the best results at Nadejda and Pavlovo sites is the MYNN2_Lin. These results
of the RMSE lead to the same conclusions as for the MB. Therefore, the MYNN2_
Thompson model set-up simulates the air temperature with smaller error, and the
MYNN2_Lin set-up simulates in the best way the wind speed. In general, all configu-
rations underestimate the temperature (most significantly at mountain station Kopi-
toto) and overestimate the wind speed with approximately 1 ms-1 or less.
Table 3. Root Mean Square Error of the model configurations
of the temperature at 2 m (T) and wind speed at 10 m (WS)
Stations
Model configuration Druzhba Hipodruma Kopitototo Nadezhda Pavlovo
T WS T WS T WS T WS T WS
MYNN2.5_Lin 3.02 1.04 3.21 1.46 4.60 12.03 3.07 1.52 3.40 1.35
MYNN2.5_Thompson 2.93 1.07 3.35 1.50 4.35 12.00 2.89 1.58 3.36 1.43
QNSE_Lin 4.99 1.66 5.44 1.92 3.94 11.77 4.93 1.84 4.51 1.78
QNSE_Thompson 3.11 1.76 3.67 2.01 3.70 11.70 3.32 1.97 3.31 1.86
YSU_Lin 3.16 1.33 3.54 1.79 3.80 12.73 3.21 1.88 3.60 1.50
YSU_Thompson 3.05 1.38 3.57 1.86 3.54 12.63 3.04 1.97 3.25 1.51

83
Vladimir Ivanov, Reneta Dimitrova

Overall, the model validation shows that the configurations with MYNN2.5
PBL scheme and Thompson microphysics have better behaviour in comparison
to the others, concerning the two meteorological parameters. Their results are
superior at Druzhba and Hipodruma stations, while the other model configurations
give better results for the other stations. Moreover, the model configuration using
the Thompson scheme for microphysical parameterization presents better results
than Lin. The Sofia city is located in complex topography with mountain Vitosha
nearby, and most likely, these differences result from the local modifications of the
large-scale weather. Therefore, reasons could be the roughness and/or topography
characteristics surrounding the city, proximity to the mountain areas, terrain height,
and the station exposure to the solar radiation. Note that structures are not presented
explicitly in these types of models, and shadow as well as the building wake effects
can affect the temperature and wind fields significantly.

WIND CHILL MODELLING


The results from the Wind Chill modelling show that different cases belong
mainly to the categories – Low Risk, Moderate Risk, and High Risk, and mostly to
the first two. The number of Low Risk cases in the different domain locations for
the period under consideration is shown in Figure 2. There are between 210 and 360
Low Risk cases in large area of the domain. The different districts of the Sofia city
itself are characterized by 210 to 300 cases, except for the areas nearby the Vitosha
Mountain. The number of Low Risk cases in the north-western, north-eastern, and
south-eastern areas of the domain is higher. There are, however, some differences
between the model configurations. The number for the north-eastern near-city areas

Figure 2. Number of cases for Low Risk conditions from WRF


simulations with different model options for January 2016
84
Sensitivity to the Wrf Model Configuration...

of the domain for MYNN2.5_Lin, YSU_Lin, and YSU_Thompson, is smaller in


comparison with the other model configurations. The QNSE_Thompson has a spot
in that area, with about 30 more Low Risk cases than in the surrounding area. The
Low Risk count distribution in the Vitosha Mountain is more diverse. It is layered
from above 390 in the lowest terrain heights to below 150 cases at the highest ones.
The figure shows also, that the north-west and western slopes have more Low Risk
cases than the neighbour ones.

Figure 3. Number of cases for Moderate Risk conditions from WRF


simulations with different model options for January 2016

Figure 3 shows the spatial distribution of the Moderate Risk conditions counts in
the domain area. Here, the number of cases in the northeastern and the southeastern
parts are between 240 and 330. The other areas except Vitosha Mountain have between
around 120 and 240 cases with Moderate Risk conditions. There are more than 240
cases in the Vitosha Mountain, and over large area, they are above 390. The Moderate
Risk cases spatial distribution has also some distinctive characteristics. The QNSE_
Lin model configuration appears to have more Moderate Risk cases than anyone else
does. The spatial distribution of that one, QNSE_Thompson and YSU_Thompson
is more homogenous than in the other model configurations. The Vitosha Mountain,
however, is an exception with above 390 cases for all model configurations.
The plots for the High Risk cases (Figure 4) is very similar for all model con-
figurations, except the QNSE_Lin, which shows some unique spatial features in
the eastern northwestern parts of the domain, similar to the Vitosha Mountain. The
entire domain area has up to 10 cases. Only for the area of the Vitosha Mountain the
number of cases increases from 10 to about 90 at the elevated areas.
85
Vladimir Ivanov, Reneta Dimitrova

Figure 4. Number of cases for High Risk conditions from WRF


simulations with different model options for January 2016

CONCLUSION
As was noted previously, the model configurations with MYNN2.5 show
better results than the other (QNSE or YSU) PBL schemes. It is very likely
due that Sofia city is located in a valley surrounded by mountains and highland
terrain and some PBL schemes do not succeed to simulate the turbulence re-
gime in the low atmospheric levels. The local turbulent kinetic energy scheme
MYNN2.5 is more suitable for stable winter conditions. Furthermore, the mod-
el configuration with Thompson microphysics gives a little better results than
Lin microphysics.
The spatial distribution of the number of Moderate Risk cases of the Wind
Chill index calculated from the WRF output for the region of Sofia in the winter
depends on the model configuration. Mainly, the type of PBL parameterization
scheme ensures that dependency. The model simulations with the MYNN2.5
PBL scheme have a more heterogeneous spatial pattern of the Wind Chill index
than the other ones for the Moderate Risk categories. The Low Risk distribution
also varies between the model simulations. There are a few High Risk cases,
which does not imply significant differences between the model configurations.
An exception from the last statement is the Vitosha Mountain, where a high
number of High Risk wind chill cases are calculated. This study can recom-
mend using MYNN2.5 PBL and Thompson microphysics for Wind Chill mod-
elling at the Sofia region. However, a more comprehensive study is needed to
strengthen these preliminary conclusions.
86
Sensitivity to the Wrf Model Configuration...

ACKNOWLEDGMENTS
This work has been carried out in the framework of the National Science Program "Envi-
ronmental Protection and Reduction of Risks of Adverse Events and Natural Disasters",
approved by the Resolution of the Council of Ministers № 577/17.08.2018 and sup-
ported by the Ministry of Education and Science (MES) of Bulgaria (Agreement №
Д01-322/18.12.2019).
We want to acknowledge the Bulgarian National Science Fund (grant DN-04/2/13.12.2016).
Special thanks to the Faculty of Physics at the University of Sofia “St. Kl. Ohridski” for
the access to the supercomputing facilities (NESTUM cluster http://hpc-lab.sofiatech.bg/
home/, funded by the ERDF Project BG161PO003-1.2.05-0001-C0001) at Sofia Tech
Park, Sofia, Bulgaria and to US NCEP for providing free-of-charge data and software.

REFERENCES
Chen, S. H., Sun W.Y., (2002) A one-dimensional time dependent cloud model. J.
Meteor. Soc. Japan., 80(1), 99–118.
Chen F., Dudhia J (2001) Coupling an advanced land surface–hydrology model with
the Penn State–NCAR MM5 modeling system. Part I: Model implementation
and sensitivity. Mon Wea Rev 129: 569–585
Cheng, Y., Niu, J., & Gao, N. Thermal comfort models: A review and numerical
investigation. (2012) Building and Environment, 47(1), 13–22.
Collier C. G. (2006) The impact of urban areas on weather. Q. J. R. Meteorol. Soc.,
132, 1–25.
Gadzhev, G., Ganev, K., Miloshev, N., Syrakov, D. & Prodanova, M., (2014a) Some
basic facts about the atmospheric composition in Bulgaria - Grid computing
simulations, Lecture Notes in Computer Science (including subseries Lecture
Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 8353
LNCS, 484-490.
Gadzhev, G., Ganev, K., Miloshev, N., Syrakov, D. & Prodanova, M., (2014b)
Analysis of the processes which form the air pollution pattern over Bulgaria,
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial
Intelligence and Lecture Notes in Bioinformatics), 8353 LNCS, 390-396.
Gadzhev, G., Ganev, K., Syrakov, D., Miloshev, N. & Prodanova, M., (2012)
Contribution of biogenic emissions to the atmospheric composition of the Balkan
Region and Bulgaria, International Journal of Environment and Pollution, 50
(1-4), 130-139.
Gadzhev, G., Syrakov, D., Ganev, K., Brandiyska, A., Miloshev, N., Georgiev,
G. & Prodanova, M., (2011a) Atmospheric composition of the Balkan region
and Bulgaria. Study of the contribution of biogenic emissions, AIP Conference
Proceedings, 1404, 200-209.
Hidalgo, J., Masson V., & Gimeno, L. (2010) Scaling the Daytime Urban Heat
Island and Urban-Breeze Circulation. J. Appl. Meteor. Climatol., 49, 889–901.
87
Vladimir Ivanov, Reneta Dimitrova

Hong, S., Noh Y. & Dudhia J. (2006), A new vertical diffusion package with an
explicit treatment of entrainment processes. Mon Weather Rev 134(9):2318–2341.
Iacono M. J., Delamere J.S., Mlawer E.J., Shephard M.W., Clough S.A. & Collins
W.D. (2008) Radiative forcing by long-lived greenhouse gases: Calculations
with the AER radiative transfer models. J Geophys Res Atmos 113:2-9.
Ivanov,V., Gadzhev G., Ganev K. & Chervenkov H. (2020) Sensitivity of the Simulated
Heat Risk in Southeastern Europe to the RegCM Model Configuration—
Preliminary Results. LNCS, 11958, 340-347
Ivanov V., Evtimov, S. (2014a) Wind chill hazard in Bulgaria during
2003-2012 period. Comptes Rendus de L'Academie Bulgare des Sciences,
67(11), 1521-1530.
Ivanov V., Evtimov, S. (2014b) Heat Risks in Bulgaria during 2003-2012 period.
Bulgarian Geophysical Journal, 40, 3-12.
Kain J.S. (2004) The Kain–Fritsch convective parameterization: An update. J Appl
Meteorol 43:170–181
Lin C, Zhang Z, Pu Z, Wang F (2017) Numerical simulations of an advection
fog event over Shanghai Pudong International Airport with the WRF model. J
Meteorol Res 31(5):874–889.
Nakanishi, M., Niino, H. (2006) An improved Mellor–Yamada level 3 model:
its numerical stability and application to a regional prediction of advecting
fog. Bound. Layer Meteor., 119, 397–407.
Nakanishi, M., Niino, H. (2009) Development of an improved turbulence closure
model for the atmospheric boundary layer. J. Meteor. Soc. Japan, 87, 895–912.
Niu, G.–Y., Yang Z-L., Mitchell K. E., Chen F., Ek, M. B., Barlage M., Kumar A.,
Manning K, Niyogi D., Rosero E., Tewari M. & Xia Y. (2011) The community
Noah land surface model with multiparameterization options (Noah–MP): 1.
Model description and evaluation with local–scale measurements. J. Geophys.
Res., 116, D12109.
Olson, J. B., Kenyon, J. S., Angevine, W.M., Brown, J. M., Pagowski, M. & Sušelj
K. (2019) A Description of the MYNN-EDMF Scheme and the Coupling to
Other Components in WRF–ARW. NOAA Technical Memorandum OAR
GSD, 61, 37.
Osczevski R., Bluestein, M. (2005) The New Wind Chill Equivalent Temperature
Chart. Bull. Amer. Meteor. Soc., 86 (10), 1453–1458.
Rizwan, A. M., Dennis, L. Y., & Chunho, L. I. U. (2008) A review on the generation,
determination and mitigation of Urban Heat Island. Journal of Environmental
Sciences, 20(1), 120-128.
Skamarock W C, Klemp JB, Dudhia J., et al. (2008) A Description of the Advanced
Research WRF Version 3. NCAR Technical Note, NCAR/TN-475+STR.
Song–You, H.,Noh, Y., Dudhia, J. (2006) A new vertical diffusion package with an
explicit treatment of entrainment processes. Mon. Wea. Rev., 134, 2318–2341.

88
Sensitivity to the Wrf Model Configuration...

Stocker, T.F., Qin D., Plattner G.-K., Alexander L.V. et al. (2013) Technical
Summary. In: Climate Change 2013: The Physical Science Basis. Contribution
of Working Group I to the Fifth Assessment Report of the Intergovernmental
Panel on Climate Change [Stocker, T.F., D. Qin, G.-K. Plattner, M. Tignor, S.K.
Allen, J. Boschung, A. Nauels, Y. Xia, V. Bex and P.M. Midgley (eds.)]. United
Kingdom and New York: Cambridge University Press, Cambridge.
Sukoriansky, S., B. Galperin & V. Perov, (2005), Application of a new spectral
theory of stably stratified turbulence to the atmospheric boundary layer over sea
ice. Boundary-Layer Meteorol 117(2):231–257.
Thompson, G., Field, P. R., Rasmussen, R. M. & Hall, W. D. (2008) Explicit
Forecasts of Winter Precipitation Using an Improved Bulk Microphysics
Scheme. Part II: Implementation of a New Snow Parameterization. Mon. Wea.
Rev., 136, 5095–5115.
Vladimirov E., Dimitrova R. & Danchovski V., 2018: Sensitivity of WRF model
results to topography and land cover: study for the Sofia region, Annuaire de
l’Université de Sofia “St. Kliment Ohridski”, Faculté de Physique, 111:87 -101.
Yang, Z.–L., Niu G.–Y., Mitchell K. E., Chen F., Ek M. B., Barlage M., Longuevergne
L., Manning K., Niyogi D., Tewari M., & Xia Y. (2011) The community
Noah land surface model with multiparameterization options (Noah–MP): 2.
Evaluation over global river basins. J. Geophys. Res., 116, D12110.

 Vladimir Ivanov
https://orcid.org/0000-0001-9768-1049
National Institute of Geophysics, Geodesy and Geography
Bulgarian Academy of Sciences
Sofia, Bulgaria
E-mail: [email protected]

 Reneta Dimitrova
http://orcid.org/0000-0002-5931-8713
National Institute of Geophysics, Geodesy and Geography
Bulgarian Academy of Sciences
Department of Meteorology and Geophysics
Faculty of Physics
“St. Kliment Ohridski” University of Sofia
Sofia, Bulgaria
E-mail: [email protected]

89
Proceeding of 1st International conference on Environmental protection and disaster RISKs

https://doi.org/10.48365/envr-2020.1.8 Air Pollution,


Climate and Health

MODELLING OF THE SULPHUR AND NITROGEN


DEPOSITIONS OVER THE BALKAN PENINSULA
BY CMAQ AND EMEP-MSC-W – PRELIMINARY
RESULTS

Georgi Gadzhev, Vladimir Ivanov


National Institute of Geophysics, Geodesy and Geography –
Bulgarian Academy of Sciences (NIGGG-BAS)

Abstract. The air quality US EPA models-3 system consisting of SMOKE -


emission model and pre-processor, MM5 – meteorological driver, and CMAQ –
chemical-transport model, is used in many studies of the air quality in the Balkan
Peninsula, and in particular Bulgaria. It runs in different model resolutions,
depending on the domain, from European to city scale. The EMEP-MSC-W model
is another chemical transport model, widely used in air quality modelling. Two of
the processes involved in the concentration change of some pollutant are the dry
and wet depositions. The air quality modelling capability depends on many factors,
for example, meteorology and emissions. We study the differences in the simulation
of the wet and dry depositions for nitrogen and sulphur compounds, between the
CMAQ and the EMEP-MSC-W model for a period of 8 years.
Keywords: modelling; CMAQ; EMEP; pollution; composition; air quality

INTRODUCTION
The air pollution nowadays forces many countries to take actions for mitigating
its adverse effects on human health. Therefore, we need a lot of information,
which is increasing in recent years. There are already more direct and indirect
data connected to the air quality from different surface-based and satellite-based
observing systems. However, we need to understand the different processes involved
in the creation, transportation, and transformation of the air pollutant species, which
help us to understand their distribution at different spatial and temporal scales.
The research community performs these tasks by air quality models systems, with
chemical transport models as the main component. We use one of these systems
with the chemical transport model CMAQ, for modeling the air quality in the
Balkan Peninsula. Previous results from air pollution modelling for the Balkan
Peninsula and Bulgaria are published in a lot of research works (Gadzhev et al.
2014, Georgieva 2014, Syrakov et al. 2015, Kaleyna, Mukhtarov, Miloshev 2013a,
90
Modelling of the Sulphur and Nitrogen Depositions...

2013b, 2014). The air composition is formed by several processes, which involve
dry deposition, wet deposition, horizontal and vertical advection, horizontal and
vertical diffusion, emission, chemical transformation, aerosol processes, and
aqueous chemistry (Gadzhev, Ganev, Mukhtarov 2020, Gadzhev et al. 2011). They
interact in between and determine the air composition at different scales. There
are some studies with CMAQ of the dry and wet deposition and their influence on
the precipitation for Bulgaria (Syrakov et al. 2019a, 2019b, Georgieva et al. 2017,
2019) for different periods up to two years. Another chemical transport model –
EMEP-MSC-W is also widely used for air quality studies in Europe (Simpson et al.
2012). Our aim is to make a preliminary study of the model comparison between
long-term high-resolution simulations with the CMAQ and the EMEP-MSC-W
simulations of the Nitrogen (N) and Sulphur (S) dry and wet deposition processes
in the Balkan Peninsula for a long-term period and high spatial resolution.

METHODOLOGY
The study is based on air quality simulations with two chemical transport models
over the Balkan Peninsula from 2000 to the 2007 year. One of these simulations is
performed with the US EPA Models-3 system, which includes CMAQ (Community
Multiscale Air Quality) model (Denis et al. 1996, Byun, Ching 1999, Byun, Schere
2006), SMOKE (Sparse Matrix Operator Kernel Emissions Modelling System)
(Coats, Houyoux 1996, Houyoux, Vukovich 1999, CEP 2003) and the regional
mesoscale meteorological model MM5. The CMAQ is a numerical chemical
transport model for modelling the different processes and their contribution
involved in changing the surface and airborne gases and aerosols. That model needs
three kinds of input information – initial and boundary conditions, meteorology,
and emissions.
We use the regional mesoscale numerical model MM5 for modelling the weather
and climate conditions (Dudhia 1993, Grell, Dudhia, Stauffer) over the Balkan
Peninsula. It is a non-hydrostatic high-resolution model, providing the needed raw
meteorological output for further processing. We use the nesting capabilities of the
MM5, where the output from each outer domain excluding the last one, is used as
input for the smaller one. The first and the bigger one (D1) is the European domain
with background information, provided from the NCEP Global Analysis Data with
1º x 1º (~81 x 81km) horizontal resolution. Our research work is concentrated on
the domain D3 geographically limited to the Balkan Peninsula and some adjacent
territories. The output from the MM5 model, need to be reprocessed to the right
format for ingesting in the CMAQ. For that purpose, we use the Meteorological
– Chemistry Interface Processor – MCIP, which prepares all meteorological input
information CMAQ needs.
The emissions from the large source sources and area sources for the whole
domain excluding Bulgaria and some adjacent territories are ingested from the TNO
91
Georgi Gadzhev, Vladimir Ivanov

high-resolution emission inventory with spatial resolution 0.25°x0.125° (Denier


van der Gon et al., 2010) in a longitude-latitude grid, reprocessed from the 50-
km grid of the EMEP (European Monitoring and Evaluation Programme) database.
The emissions for Bulgaria are from the National Emission Inventory. The CMAQ
needs also from biogenic emissions. They are provided from the emission pre-
processor SMOKE. The input information is provided from the TNO emissions, the
MCIP output, and the land-use database.
The CMAQ model accounts for the following processes with a different
contribution to the changing of the concentration field for each pollutant: horizontal
diffusion (HDIF); horizontal advection (HADV); vertical diffusion (VDIF);
vertical advection (VADV); dry deposition (DRYDEP); emissions (EMISS);
chemical transformations (CHEM); aerosol processes (AERO); cloud processes
(CLOUD). The solution of the transport and transformation equations gives the
mean concentration change of i-th pollutant in the first model layer from time t to
time t + ∆t.
It is presented as a sum of the contribution of the former processes:

We focus on the dry and wet depositions modelled by the CMAQ in this study.
The N deposition contains the contribution from NO2 (Nitrogen dioxide), NO
(Nitrogen oxide), NO3 (Nitrogen trioxide), N2O5 (Dinitrogen pentoxide), HNO3
(Nitric acid), HONO (Nitrous acid), ANH4J (Accumulation-mode ammonium
mass), ANH4I (Aitken-mode ammonium mass), ANO3J (Accumulation-mode nitrate
mass), ANO3I (Aitken-mode aerosol nitrate mass) and NH3 (Ammonia):

The S deposition contains the contribution from SO2 (Sulphur dioxide), SULF
(Sulphate aerosols), ASO4J (Accumulation-mode aerosol sulphate mass), and ASO4I
(Aitken-mode aerosol sulphate mass):

The CMAQ deposition output is in 1-hour frequency. Therefore, we sum up the


hourly values of the N and S components for every day of the simulation, finding
the daily deposition values.
92
Modelling of the Sulphur and Nitrogen Depositions...

The second model used for comparison with the previous one is with the
Meteorological Synthesizing Centre-West (MSC-W) of the European Monitoring
and Evaluation Programme (EMEP). It is a chemical transport model (Simpson et
al., 2012), a key tool involving in the European air pollution policy assessments.
In the beginning, the model covers the whole of Europe with a resolution of about
50 km x 50 km, with vertical levels up to the tropopause (100 hPa). The model
has changed over the years, adding different features, and currently, his horizontal
resolution ranging from 5 km to 1 degree with 20 vertical levels. In our study, we
use a grid size 0.1° x 0.1°. The EMEP-MSC-W model runs with meteorological
fields from the numerical weather prediction system ECMWF-IFS Cycle36r1.
The model output is with daily frequency, so we do not need to do further post-
processing.
For comparison of the models, we use two kind of error characteristics. The first
is Normalised Mean Bias noted as (NMB):

,
and the second is the Mean Bias (MB):

The notions in these equations are i - ith value, M - the output form CMAQ,
E - the output from EMEP-MSC-W. The results are revealed with the multiyear
averaged values of the NMB for each grid point and the annual spatial-averaged
values of the bias of the CMAQ output.

RESULTS
The results are given for the N depositions and for the S depositions, separated
in dry component, wet component, and total (dry+wet) component. The multiyear
average of the S dry, wet, and total depositions (figure 1) reveals the following
features. There is a difference between the CMAQ and the EMEP-MSC-W model
due to the difference in the emission inventories. We can clearly note the missing of
some of the S sources in one model, but not in the other. We can see from the sum
of dry and wet deposition shown on the figure, that the TPP Bobov dol, the TPP
Pernik, the Sofia city, the town of Devnia, the Bucharest city and the Istanbul city
are noticeable in the CMAQ model output, but not in the EMEP-MSC-W output.
On the other hand, Zlatna Panega and Southern Italy sources show up in the EMEP
output, but not in the CMAQ one. The influence of the input meteorological data
and the meteorological driver for the models have a considerable effect mainly on
the wet deposition modelling capability. The wet deposition in the EMEP-MSC-W
model has smaller spatial gradients, more intensive and local maximums on larger
93
Georgi Gadzhev, Vladimir Ivanov

areas around the corresponding sources. On the other hand, the wet deposition in
the CMAQ has smaller values and bigger local spatial gradients.

Figure 1. CMAQ (upper row) multiyear average sulphur dry deposition


(left plot), wet deposition (middle plot) and dry + wet (right plot).
EMEP-MSC-W (middle row) multiyear average Sulphur dry deposition
(left plot), wet deposition (middle plot) and dry + wet (right plot). Normalized
mean bias [%] of the CMAQ model (lowest row) for sulphur dry deposition
(left plot), wet deposition (middle plot) and dry + wet (right plot) in comparison
to the EMEP-MSC-W.
94
Modelling of the Sulphur and Nitrogen Depositions...

The influence of the meteorological conditions and the orography is more notable
for figure 1 where the normalized mean biases are shown. The normalized mean bias
of the dry deposition over the land areas reaches 50 % and more and only above some
of the sources is negative. The normalized mean bias of the sum of the dry and wet
depositions has a similar spatial structure with one of the wet depositions for the two
models but is more complex. The CMAQ and EMEP-MSC-W simulate the annual
area-averaged dry plus wet deposition in a quite similar way from 2000 to 2007, as is
shown in table 1. Although more or less different in particular years, they are close.
The bias from 2000 to 2003 is negative (figure 2), which is easy to suggest from the
area-averaged total depositions and the CMAQ bias for the whole period.

Figure 2. Annual area averaged Sulphur total (dry + wet) depositions


and bias of the CMAQ model

Table 1. Annual area averaged and multiyear area averaged (YA)


Sulphur total (dry + wet) depositions
EMEP dry + wet average S CMAQ dry + wet average S Bias of dry+ wet S
Year
(mg/m2) (mg/m2) (mg/m2)
2000 3.6684 2.6886 -0.9798
2001 3.4655 3.2473 -0.2182
2002 3.8176 2.8822 -0.9355
2003 3.2111 3.1706 -0.0405
2004 2.9467 3.0579 0.1112
2005 2.6357 3.2199 0.5842
2006 2.5942 2.965 0.3708
2007 2.3126 2.7426 0.4299
YA 2.9426 3.0188 0.0762

The result for the N depositions is shown in figure 3. They have different spatial
and temporal features from the S ones. As is seen in figure 3, the model difference
95
Georgi Gadzhev, Vladimir Ivanov

between the dry depositions appears in the almost homogeneous distribution for the
CMAQ, in contrast to the clearly outlined sources in the EMEP-MSC-W. The input
meteorological data and the meteorological driver for the models exert substantial
influence mostly on the wet deposition modelling. The spatial gradient of the mean
EMEP-MSC-W wet deposition is smaller than the CMAQ one. However, the CMAQ
mean wet deposition is smaller and with bigger local spatial gradients following the
orography features. The results for the normalized mean biases of the dry, wet, and
sum of the dry and wet depositions (figure 3) suggest a substantial influence of the

Figure 3. CMAQ (upper row) multiyear average Nitrogen dry deposition


(left plot), wet deposition (middle plot) and dry + wet (right plot). EMEP-MSC-W
(middle row) multiyear average nitrogen dry deposition (left plot), wet deposition
(middle plot) and dry + wet (right plot). ). Normalized mean bias [%] of the
CMAQ model (lowest row) for nitrogen dry deposition (left plot), wet deposition
(middle plot) and dry + wet (right plot) in comparison to the EMEP-MSC-W
96
Modelling of the Sulphur and Nitrogen Depositions...

meteorological input and the orography on the spatial distribution of the mean wet
deposition. The normalized mean bias of the mean dry deposition reaches 50% not
only on the land however, is negative in some places. The normalized mean bias of
the sum of the dry and wet depositions has a similar, but a more complex structure
with the one of the wet deposition, because of the influence of the dry deposition.
The data in table 2 and figure 4 suggest that the simulated annual area-averaged
total nitrogen depositions by the CMAQ and EMEP-MSC-W models pretty much
the same, although the CMAQ value is smaller in 2002. The results for the multiyear
area-averaged total nitrogen depositions are very similar.

Figure 4. Annual area averaged Nitrogen total (dry + wet) depositions


and bias of the CMAQ model.

Table 2. Annual area averaged and multiyear area averaged (YA)


Nitrogen total (dry + wet) depositions
EMEP dry + wet average N CMAQ dry + wet average N Bias of dry+ wet N
Year
(mg/m2) (mg/m2) (mg/m2)
2000 2.1538 2.1827 0.0289
2001 2.3273 2.5671 0.2398
2002 2.2701 2.1798 -0.0903
2003 2.1185 2.285 0.1665
2004 2.2455 2.4198 0.1743
2005 2.3357 2.451 0.1153
2006 2.1773 2.2341 0.0568
2007 2.1989 2.2235 0.0246
YA 2.2331 2.3286 0.0955

CONCLUSION
The results suggest that the Nitrogen annual area-averaged total depositions
are represented more similarly by the two models, than the Sulphur ones. There
is a large orography influence on the sum of dry and wet deposition for both
groups of chemical species. The current research suggests that the orography
97
Georgi Gadzhev, Vladimir Ivanov

and meteorology exert substantial influence on the total Nitrogen and Sulphur
depositions.

ACKNOWLEDGMENTS
This work has been carried out in the framework of the National Science Pro-
gram "Environmental Protection and Reduction of Risks of Adverse Events and
Natural Disasters", approved by the Resolution of the Council of Ministers №
577/17.08.2018 and supported by the Ministry of Education and Science (MES)
of Bulgaria (Agreement № Д01-322/18.12.2019).
Bulgarian National Science Fund (grant DN-04/2/13.12.2016).
Special thanks are due to US EPA and US NCEP for providing free-of-charge data
and software and to the Netherlands Organization for Applied Scientific re-
search (TNO) for providing the high-resolution European anthropogenic emis-
sion inventory

REFERENCES
Byun, D., Ching, J., (1999), Science Algorithms of the EPA Models-3
Community Multiscale Air Quality (CMAQ) Modeling System. EPA Report
600/R-99/030, Washington DC.
Byun, D., Schere, K.L., (2006), Review of the governing equations,
computational algorithms, and other components of the models-3
community multiscale air quality (CMAQ) modeling system. Appl. Mech.
Rev. 59, 51 – 77.
CEP, (2003), Sparse Matrix Operator Kernel Emission (SMOKE) Modeling
System, University of North Carolina, Carolina Environmental Programs -
CEP, Research Triangle Park, North Carolina.
Coats, C.J.,Jr., and Houyoux, M.R., (1996), Fast Emissions Modeling With the
Sparse Matrix Operator Kernel Emissions Modeling System, The Emissions
Inventory: Key to Planning, Permits, Compliance, and Reporting, Air and
Waste Management Association, New Orleans, September
Denier van der Gon, H., Visschedijk, A., van de Brugh, H., Droge, R., (2010),
A high resolution European emission data base for the year 2005, TNO-
report TNO-034-UT-2010-01895 RPT-ML, Apeldoorn, The Netherlands.
Dennis,R.L.,Byun,D.W.,Novak,J.H.,Galluppi,K.J.,Coats,C.J.,andVouk,M
.A., (1996), The Next Generation of Integrated Air Quality Modeling:
EPA’s Models-3, Atmospheric Environment 30, 1925 – 1938.
Dudhia, J. (1993) A nonhydrostatic version of the Penn State-NCAR
mesoscale model: validation tests and simulation of an Atlantic cyclone
and cold front, Monthly Weather Review, 121 (5), 1493 – 1513.
98
Modelling of the Sulphur and Nitrogen Depositions...

Gadzhev G., Ganev K., Mukhtarov P., (2020) Statistical Moments of the
Vertical Distribution of Air Pollution over Bulgaria, Lecture Notes
in Computer Science (including subseries Lecture Notes in Artificial
Intelligence and Lecture Notes in Bioinformatics), 11958 LNCS, 213 –
219.
Gadzhev, G., Ganev, K., Miloshev, N., Syrakov, D., Prodanova, M. (2013)
Numerical Study of the Atmospheric Composition in Bulgaria, Computers
and Mathematics with Applications, 65, 402 – 422.
Gadzhev G., Syrakov, D., Ganev, K., Brandiyska, A., Miloshev, N., Georgiev,
G., Prodanova, M. (2011) Atmospheric composition of the Balkan region
and Bulgaria. Study of the contribution of biogenic emissions, AIP
Conference Proceedings, 1404, 200 – 209.
Georgieva E., Syrakov D., Hristova E., Prodanova M., Gospodinov I. (2019)
Comparison of EMEP and WRF_CMAQ modelling results for deposition
estimates in Bulgaria for 2016 and 2017, 19th International Conference on
Harmonisation within Atmospheric Dispersion Modelling for Regulatory
Purposes, Harmo 2019.
Georgieva E., Hristova E., Syrakov D., Prodanova M., Batchvarova
E. (2017) Preliminary evaluation of CMAQ modelled wet deposition
of sulphur and nitrogen over Bulgaria. HARMO 2017 – 18th International
Conference on Harmonisation within Atmospheric Dispersion Modelling
for Regulatory Purposes, Proceedings 2017 – October, 51 – 55.
Georgieva, I. (2014) Study of the air quality index climate for Bulgaria,
Proc. of the international conference on numerical methods for scientific
computations and advanced applications, May 19 – 22, 2014, Bansko,
ISBN978-954-91700-7-8, 39. – 42.
Grell, G.A., Dudhia, J., Stauffer, D.R. A description of the Fifth Generation
Penn State/NCAR Mesoscale Model (MM5) NCAR Technical Note, NCAR
TN-398-STR, 138.
Houyoux, M.R., and Vukovich, J.M. (1999) Updates to the Sparse Matrix
Operator Kernel Emission (SMOKE) Modeling System and Integration
with Models-3, The Emission Inventory: Regional Strategies for the
Future, Raleigh, NC, Air and Waste Management Association.
Kaleyna P., Mukhtarov P., Miloshev N. (2014) Seasonal Variations of the
Total Column Ozone over Bulgaria in the Period 1996–2012, Comptes
rendus de l’Acad´emie bulgare des Sciences, 67, 7.
Kaleyna P., Pl. Muhtarov, N. Miloshev, (2013a) Condition of The Stratospheric
and Mesospheric Ozone Layer over Bulgaria for the Period 1996 – 2012:
Part 1 – Total Ozone Content, Seasonal Variations, Bulgarian Geophysical
Journal, 39, National Institute Of Geophysics, Geodesy And Geography,
Bulgarian Academy Of Sciences.

99
Georgi Gadzhev, Vladimir Ivanov

Kaleyna P., Pl. Muhtarov, N. Miloshev, (2013b) Condition of The Stratospheric


and Mesospheric Ozone Layer over Bulgaria for the Period 1996 – 2012:
Part 2 – Total Ozone Content, Short term variations, Bulgarian Geophysical
Journal, 39, National Institute Of Geophysics, Geodesy And Geography,
Bulgarian Academy Of Sciences.
Leeuw, F. de, Mol, W., (2005), Air Quality and Air Quality Indices: a world
apart. ETC/ACC Technical Paper 2005/5.
Simpson, D., Benedictow, A., Berge, H., Bergström, R., Emberson, L. D.,
Fagerli, H., Flechard, C. R., Hayman G. D., Gauss M., Jonson J. E., Jenkin
M. E., Nyíri A., Richter C., Semeena V. S., Tsyro S., Tuovinen J.-P.,
Valdebenito A & Wind P. (2012). The EMEP MSC-W chemical transport
model – technical description. Atmos. Chem. Phys., 12, 7825 – 7865.
Syrakov D., Prodanova M., Georgieva E., Hristova E., (2019a) Applying
WRF-CMAQ models for assessment of sulphur and nitrogen deposition in
Bulgaria for the years 2016 and 2017, International Journal of Environment
and Pollution, 66(1-3), 162-186
Syrakov D., Georgieva E., Prodanova M., Hristova E., Gospodinov I., Slavov
K., Veleva B. (2019b) Application of WRF-CMAQ model system for
analysis of sulfur and nitrogen deposition over Bulgaria, Lecture Notes
in Computer Science (including subseries Lecture Notes in Artificial
Intelligence and Lecture Notes in Bioinformatics), 11189 LNCS, 474 –
482.
Syrakov, D., Prodanova, M., Georgieva, E., Etropolska, I., Slavov, K. (2015),
Impact of NOx emissions on air quality simulations with the Bulgarian
WRF-CMAQ modelling system, International Journal of Environment
and Pollution, 57, Issue 3-4, 2015, 285 – 296.

 Georgi Gadzhev
http://orcid.org/0000-0002-6159-3554
National Institute of Geophysics, Geodesy and Geography
Bulgarian Academy of Sciences
Sofia, Bulgaria
E-mail: [email protected]

 Vladimir Ivanov
https://orcid.org/0000-0001-9768-1049
National Institute of Geophysics, Geodesy and Geography
Bulgarian Academy of Sciences
Sofia, Bulgaria
E-mail: [email protected]

100
Proceeding of 1st International conference on Environmental protection and disaster RISKs

https://doi.org/10.48365/envr-2020.1.9 Air Pollution,


Climate and Health

HOW WELL DO THE AIR QUALITY MODELS EMEP


AND CAMS REPRODUCE PARTICULATE MATTER
SURFACE CONCENTRATIONS IN BULGARIA

Hristina Kirova, Nadya Neykova, Emilia Georgieva


National Institute of Meteorology and Hydrology (NIMH)

Abstract: The main objective of this study is to evaluate the performance of


some well-known and widely used operational air quality modelling systems (EMEP-
MSC-W, and the models at the Copernicus Atmosphere Monitoring Service (CAMS))
for simulations of particulate matter in Bulgaria. The analysis is focussed on a summer
month (August 2017) and includes comparison of model to observations from 24
regular air quality background stations, as well as model inter-comparison. Along with
statistical indicators various graphs are used - box plots, kernel density estimations,
and scatter plots. The spatial distribution is discussed upon maps of monthly mean
concentrations and comparison of domain averaged model concentrations. The spatial
distribution of PM10 is similar only over the Black sea. The highest domain mean
surface concentrations are simulated by CAMS global – by about 12% for PM10 and
by about 14% for PM2.5 higher than the other models.
Keywords: chemical transport models, PM10, PM2.5, regular air quality observations,
model validation, Bulgaria

INTRODUCTION
Particulate matter (PM) are constantly studied because of their effects on the
human health and on the environment. While monitoring facilities for surface level
concentrations are growing in number and type, the information from observational
data is still limited in time and space. Chemical transport models (CTM) has been
recognized as valuable tool not only for air quality assessment, but also for policy
support in determining abatement measures (McMurry et al., 2004, Miranda et al.,
2015). The PM modelling, however, still represents a challenging task, as their
concentrations are influenced not only by different emission sources, but also by
atmospheric processes and chemical transformation mechanisms taking place
over various spatial scales. Some international initiatives for evaluation of CTMs
in the last years (e.g. AQMEII (Air Quality Modelling Evaluation International
Initiatives), FAIRMODE (Forum for AIR quality MODeling)) have allowed better
understanding of the weaknesses and strengths of the models and have contributed
101
Hristina Kirova, Nadya Neykova, Emilia Georgieva

to their improvement and further development. Nowadays, CTMs are the backbone
of many comprehensive air quality forecasting systems designed for different scales
– from country, to European and global ones (e.g. Schaap et al., 2008, Sofiev et al.,
2008, Mailler et al., 2017).
At the National Institute of Meteorology and Hydrology (NIMH) a chemical
weather forecasting system has been also set up, BgCWFS (Syrakov et al. 2013,
Etropolska et al., 2011). The system was evaluated on European scale in the
framework of AQMEII (Brunner et al., 2015, Curci et al., 2015, Im et al., 2015).
On national scale, BgCWFS results with 9 km spatial resolution, showed
underestimation for PM10 (Georgieva et al., 2015). Exceedances of PM limit
values are often observed at many sites in Bulgaria (Naydenova et al., 2018, EEA
Report, 2019), and there is a public and expert interest in possibilities of operational
modelling systems to predict surface PM concentrations.
For this study we have chosen to look at freely available results of three well-
known and widely used operational air quality modelling systems for a case
study in Bulgaria. The systems differ in their input data, emissions handling,
parameterisation schemes, chemical mechanisms etc., but it is believed that they
capture the main characteristics of the surface PM distribution. In brief, results
from the following systems are used here:
EMEP MSC-W (denoted further as EMEP) – the model of the Meteorological
Synthesizing Centre-West (MSC-W) of the European Monitoring and Evaluation
Programme) (Simpson et al., 2012). The EMEP model is one of the major instru-
ments which is applied for decision and policy making not only because of its
coverage but also because its output includes photooxidants, inorganic and organic
aerosols and depositions. This is the model used by the European Environmental
Agency for annual reporting on the air quality status in Europe. The model domain
covers an extended European region with a horizontal resolution of 0.1° x 0.1° and
20 vertical layers (the lowest with a height of approximately 50 meters). Results for
the year 2017 from version rv4_33 are used.
The second system is the regional (European) ensemble air quality forecast-
ing system at the Copernicus Atmosphere Monitoring Service (CAMS) (denoted
further as CAMS-ENS). Its main characteristics are: coverage – Europe, spatial
and temporal resolution: 0.1° and 1 h, vertical levels up to 5000 m. The Ensemble
forecast is the median of the forecasts from 9 different state-of-the-art atmospheric
modelling systems (CHIMERE, EMEP, EURAD-IM, DEHM, GEM-AQ, LOTOS-
EUROS, MATCH, MOCAGE, SILAM), ECMWF Copernicus Report, 2019. All
CTM’s use the same emissions data, the same meteorological driver and the same
boundary conditions. The regional CAMS system combines model data and in-situ
observations to provide air quality forecasts.
The third system is the global CAMS (CAMS-ECMWF). It provides opera-
tional forecast for atmospheric chemistry parameters globally with horizontal reso-
102
How Well do the Air Quality Models Emep and Cams Reproduce...

lution of 40 km on 60 vertical levels going up to 0.1 hPa, the temporal resolution


is 3 h. CAMS – ECMWF assimilates data from satellite observations. Thus means
emissions by dust storms and wildfires are taken into account.
The period for the case study is one month - August 2017. This is a typical
summer month, characterized by high temperatures and low precipitations. The
prevailing weather is of anticyclonic type (except 5 days with weak pressure field
and 10 days with cyclonic type). Massive thunderstorms were registered on 4 days.
The number of days at NIMH stations with rainfall more than 1 mm was between
1 and 4, and with rainfall above 10 mm was between 0 and 3 (National Institute
of Meteorology and Hydrology - Monthly Bulletin). High temperatures and low
amount of precipitation are favorable conditions for wildfires which occurred at the
beginning of the studied period and in the second part of the month in southern Bul-
garia. Saharan dust intrusions towards the country, detected by data from GOME2
instrument on MetOP satellites, were identified in about 14 days throughout the
month. As during summer the domestic heating - one of the major emission sources
for PM - is missing, the selected month is assumed to be influenced by natural
aerosols, which CAMS modelling systems can treat through the data assimilation.
The main goal of this work is to evaluate the performance of EMEP, CAMS-
ENS and CAMS-ECMWF for PM10 and PM2.5 surface concentrations in Bulgaria
during the selected period, comparing model results to observations and performing
model intercomparison.

METHODS
To check the performance of the models, two main aspects were considered:
a) Model to observation analysis
The modelled daily PM values are compared to the daily observed ones in each
station with available observations of PM10 (24 stations) and PM2.5 (8 stations)
August 2017. The stations are part of the air quality monitoring network managed
by the Bulgarian Executive Environment Agency (ExEA), Fig.1.

Figure 1. Map of the stations


with available observations
of PM10 and PM2.5 for August
2017. The 3 stations selected
for graphs are marked with red
color: Sofia Hipodruma, Varna
– SOU Angel Kanchev, and
Rozhen
103
Hristina Kirova, Nadya Neykova, Emilia Georgieva

The performance of the EMEP and CAMS-EN Smodels is evaluated through


box plots, kernel density estimations (kde) of the probability density functions
(pdf), scatter plots and statistical indicators. Here we show as example the plots for
3 selected stations located in different environment: one urban - Sofia Hipodruma
(BG0050A), one at the Black Sea coast-Varna – SOU Angel Kanchev (BG0075A),
and one of rural/mountain type - Rozhen (BG0053R, 1720 m asl), Fig. 1.
The daily values for all stations are grouped and shown on one scatter plot for
each model. The statistical indicators, averaged over all stations, include the mean
observed and modelled concentrations, the mean bias error (MBE), the root mean
square error (RMSE), the correlation coefficient (Corr), the fractional gross error
(FGE), and the normalized mean bias (NMB).
b) Model - to - model comparison
For this analysis we use qualitative comparison - maps for mean monthly
concentrations of particulate matters (PM10 and PM2.5). As a quantitative measure
the values of the domain mean concentrations, as simulated by the different systems
for the month, аre calculated.

RESULTS
The comparison of model data from EMEP and CAMS-ENS to observed PM10
concentrations is visualized in Fig.2 for the selected three stations. The graphs pre-
senting the kernel density estimations show that PM10 concentrations estimated by
both models have similar distributions. The box plots indicate that both models
have smaller means and dispersions than the observed data, but CAMS-ENS model
has smaller dispersion than the EMEP model. For the rural remote station Rozhen
the models have better resemblance to the observed data than for the urban stations,
as expected because this station is not influenced by local emissions typical for the
cities.
Both models underestimate the observed PM10 concentrations. This can also
be seen from the statistical indicators of modelled versus observed PM10 averaged
over all 24 stations presented in Table 1 and on the scatter plots for all 24 stations
presented in Fig. 3. The statistical indicators show very similar results for the two
models with underestimation on average by about 50%. The scatter plot indicates
that the EMEP model has more often overestimation of some daily values than
CAMS-ENS. Overall we can conclude that EMEP model underestimates the
observed PM10 concentrations to a lesser extent than CAMS-ENS model, but the
correlation is better for CAMS-ENS.

104
How Well do the Air Quality Models Emep and Cams Reproduce...

A B C

Figure 2. Observed vs modelled PM10 in August 2017:


a) Sofia – Hipodruma, b) Varna, c) Rozhen: Box plots on 1st row
and corresponding kernel density estimations on 2nd row

Table1.Statistical indicators of modelled versus observed PM10


concentrations averaged over 24 stations
PM10 Mean_OBS Mean_MOD MBE RMSE
Cor FGE NMB[%]
(Nstations=24) [µgm-3] [µgm-3] [µgm-3] [µgm-3]
EMEP 25.68 12.90 -12.78 15.82 0.47 0.79 -49.77
CAMS-ENS 25.68 10.99 -14.69 16.11 0.56 0.80 -57.22

The distributions of the modelled and observed PM2.5 concentrations


are presented in the box plots and kernel density estimation plots, in Fig.
4. Compared to the results for PM10, the difference between modelled and
observed PM2.5 concentrations is smaller. The modelled PM2.5 concentrations
by both models are similarly distributed. Both models have smaller means and
dispersion than the observed data but CAMS-ENS model has smaller dispersion
than the EMEP model. As with PM10 the performance is better for the rural
remote station Rozhen.
Both models underestimate the observed PM2.5 concentrations, as indicated by the
statistical indicators in Table 2 and the scatter plots in Fig. 5. The underestimation,
105
Hristina Kirova, Nadya Neykova, Emilia Georgieva

A B
Figure 3. Scatter plots for daily PM10 at 24 stations in Bulgaria in August 2017:
a) EMEP and b) CAMS-ENS

on average by about 32%, is less than the underestimation for PM10. To note that the
number of stations for PM2.5 is only 8.

A B C
Figure 4. Observed vs modelled PM2.5 in August 2017: a) Sofia –
Hipodruma, b) Varna, c) Rozhen: Box plots on 1st row
and corresponding kernel density estimations on 2nd row
106
How Well do the Air Quality Models Emep and Cams Reproduce...

Table2. Statistical indicators of modelled versus observed PM2.5


concentrations averaged over 8 stations
PM2.5 Mean_OBS Mean_MOD MBE RMSE
Corr FGE NMB[%]
(Nstations=8) [µgm-3] [µgm-3] [µgm-3] [µgm-3]
EMEP 13.55 9.34 -4.21 7.26 0.48 0.58 -31.09
CAMS-ENS 13.55 8.82 -4.73 5.93 0.66 0.45 -34.90

A B
Figure 5. Scatter plots for daily PM2.5 at 8 stations in Bulgaria in August 2017:
a) EMEP and b) CAMS- ENS

The model to model comparison is based on results for Bulgaria obtained by


the three model systems – EMEP, CAMS-ENS and CAMS-ECMWF for August
2017. The spatial distribution of mean monthly concentrations of PM10 and PM2.5
is shown in Fig. 6, the averaged for the domain monthly values are provided in
Table 3. All models simulate higher concentrations in the eastern part of the domain
(Black Sea). EMEP results over land indicate some hot spots in correspondence
with big cities. Higher concentrations in the Lower Danube plain (north-western
part of the domain) are simulated by CAMS-ECMWF. This might be a consequence
of emission sources that are not included in the other models, but this system is
producing due to satellite data assimilation. The mean monthly map for the Aerosol
Optical Depth, as retrieved by MODIS Terra satellite, Fig. 7, also indicates higher
aerosol loading north of the country. The character of the spatial distribution by the
models is maintained in the maps for PM2.5. The domain mean surface concentra-
tions for PM10 and PM2.5 are with 9 % and 12.5 % higher for CAMS-ECMWF than
for EMEP (Table 3). The lowest values are simulated by CAMS-ENS: with 14 %
and 15.6 % lower than for CAMS-ECMWF.
107
Hristina Kirova, Nadya Neykova, Emilia Georgieva

Figure 6. Monthly mean PM10 (first row) and PM2.5 (second row)
concentrations (µgm-3) for domain Bulgaria (August 2017): EMEP (left),
CAMS-ENS (middle) and CAMS-ECMWF (right)

Table 3. Domain mean surface concentrations [µgm-3] for August 2017


EMEP CAMS-ENS CAMS-ECMWF
PM10 12 11.3 13.2
PM2.5 8.4 8.1 9.6

Figure 7. Mean monthly (August, 2017) map of combined dark target


and deep blue AOD at 550 nm for land and ocean
108
How Well do the Air Quality Models Emep and Cams Reproduce...

CONCLUSIONS
In this study EMEP and CAMS-ENS simulated concentrations for PM10 and
PM2.5 were compared to data from regular air quality stations in Bulgaria for a
test period of one month (August 2017). Both models underestimate the observed
concentrations, on monthly basis by about 50 % for PM10 and by about 32 % for
PM2.5. The models perform better at the rural remote (mountain) site Rozhen than
for the urban background stations indicating that the outputs could be used for
indicative values of background PMs concentrations. EMEP model underestimated
the observed PM10 and PM2.5 values to a lesser extent than CAMS-ENS. The spatial
distribution showed higher values in the eastern part of domain Bulgaria and hot
spots over main cities. Interestingly, the CAMS-ECMWF model, that has coarser
grid resolution, indicates higher surface concentrations north of the country over
the Lower Danube plain. This might be due to sources not accounted for in the
other models, e.g. fires or mineral dust. As this system assimilates satellite data,
it could forecast influence of such events on surface PM concentrations. Further
analysis are ongoing in this direction.
Future plans foresee to perform similar analysis for a winter month when surface
concentrations of PM are higher due to higher anthropogenic emissions (traffic,
domestic heating, and energy production from thermal power plants) and when
the conditions for dispersion of atmospheric pollutants are worse due to thermal
inversions. Results by BgCWFS will be also included in the analysis.

ACKNOWLEDGMENTS
The work presented here was carried out in the frame of the project SIDIAQ“
Satellite Information Downscaled to Urban Air Quality in Bulgaria”, funded
by the Government of Bulgaria through the ESA Contract No.4000124150/18/
NL/SC under the PECS (Plan for Еurореаn Cooperating States). Тhе view
expressed herein саn in nо way bе taken to reflect the official opinion of the
Еurореаn Space Agency.
The European Centre for Medium-Range Weather Forecasts, the Copernicus
Atmosphere Monitoring Service and the Meteorological Synthesizing Centre-
West of the European Monitoring and Evaluation Programme and Met Norway
are acknowledged for providing analyzed and forecasted data on meteorological
parameters and atmospheric chemistry. Someanalyses and visualizations used in
this study were produced with the Giovanni online data system, developed and
maintained by the NASA GES DISC.

REFERENCES
Brunner, D., Savage, N., Jorba, O., Eder, B., Giordano, L., Badia, A.,
Balzarini, A., Baró, R., Bianconi, R., Chemel, C., Curci, G., Forkel, R.,
Jiménez-Guerrero, P., Hirtl, M., Hodzic, A., Honzak, L., Im, U., Knote,
109
Hristina Kirova, Nadya Neykova, Emilia Georgieva

C., Makar, P., Manders-Groot, A., van Meijgaard, E., Neal, L., Pérez, J.
L., Pirovano, G., San Jose, R., Schröder, W., Sokhi, R. S., Syrakov, D.,
Torian, A., Tuccella, A., Werhahn, J., Wolke, R., Yahya, K., Zabkar, R.,
Zhang, Y., Hogrefe, C. andGalmarini, S.. (2015) Comparative analysis
of meteorological performance of coupled chemistry-meteorology
models in the context of AQMEII phase 2,Atmospheric Environment,
115, 470-498.
EEA (2019) Air Quality in Europe – 2019 Report, EEA Report 10/2019,
ISSN 1977-8449, doi:10.2800/822355.
ECMWF Copernicus Report (2019) CAMS Regional Air Quality User
Guide, issued Meteo-France, https://www.regional.atmosphere.
copernicus.eu/doc/USER_GUIDE_dataServer.pdf
Etropolska, I., Prodanova, M., Syrakov, D., Ganev, K., Miloshev, N. and
Slavov, K. (2011) Bulgarian Operative System for Chemical Weather
Forecast. In: Dimov, I., Dimova, S., Kolkovska, N. (eds) Numerical
Methods and Applications. NMA 2010. Lecture Notes in Computer
Science, vol 6046. Springer, Berlin, Heidelberg, 141-149.
Georgieva, E., Syrakov, D., Prodanova, M., Etropolska, I. and Slavov, K.,
(2015) Evaluating the performance of WRF-CMAQ air quality model-
ling system in Bulgaria by means of the DELTA tool, International Jour-
nal of Environment and Pollution, Vol. 57, No. 3/4, 272-284.
Curci, G., Hogrefe, C., Bianconi, R., Im, U., Balzarini, A., Baró, R.,
Brunner, D., Forkel, R., Giordano, L., Hirtl, M., Honzak, L., Jiménez
Guerrero, P., Knote, C., Langer, M., Makar, P., Pirovano, G., Perez, J. L.,
San José, R., Syrakov, D., Tuccella, P., Werhahn, J., Wolke, R., Žabkar,
R. and Zhang, Y. (2015) Uncertainties of simulated aerosol optical
properties induced by assumptions on aerosol physical and chemical
properties: an AQMEII-2 perspective,Atmospheric Environment, 115,
Elsevier, 541-552.
MaillerS., Menut, L., Khvorostyanov, D.,Valari, M., Couvidat, F., Siour,
G., Turquety, S., Briant, R., Tuccella, P., Bessagnet, B., Colette, A.,
Letinois, L., and F. Meleux, F., (2017) CHIMERE-2017: from urban
to hemispheric chemistry-transport modelingGeosci. Model Dev., 10,
2397-2423, https://doi.org/10.5194/gmd-10-2397-2017.
McMurry, P., Shepherd, M.F and Vickery J.S. (2004) Particulate Matter
Science for Policy Makers: A NARISTO Assessment, Cambridge
University Press.
Miranda, A., Silveira, C., Ferreira,J., Monteiro, A., Lopes, D., Relvas, H.,
Borrego, C.and Roebeling, P. (2015) Current air quality plans in Europe
designed to support air quality management policies, Atmospheric
Pollution Research 6, 434-443.

110
How Well do the Air Quality Models Emep and Cams Reproduce...

Naydenova, I., Petrova, T,Velichkova and R., Simova I. (2018) PM10


Exceedance In Bulgaria, CBUInternational Conference on Innovations
in Science and Education, March 21-23, 2018, Prague, Czech Republic.
Schaap, M., Timmermans, R. M., Roemer, M., Boersen, G., Builtjes, P. J.,
Sauter, F. J., Velders, G. J., and Beck, J. P. (2008). The LOTOS EUROS
model: description, validation and latest developments, Int. J. Environ.
Pollut., 32, 270–290.
Simpson, D., Benedictow, A., Berge, H., Bergström, R., Emberson, L. D.,
Fagerli, H., Flechard, C. R., Hayman, G. D., Gauss, M., Jonson, J. E.,
Jenkin, M. E., Nyíri, A., Richter, C., Semeena, V. S., Tsyro, S., Tuovinen,
J.-P., Valdebenito, Á., and Wind, P. (2012). The EMEP MSC-W chemical
transport model – technical description, Atmos. Chem. Phys., 12, 7825–
7865, https://doi.org/10.5194/acp-12-7825-2012.
Sofiev, M., Galperin, M., and Genikhovich, E. (2008). A Construction
andEvaluation of Eulerian Dynamic Core for the Air Quality
andEmergency Modelling System SILAM, in: Air Pollution Modeling
and Its Application XIX, Springer Netherlands, Dordrecht,699–701.
Syrakov, D., Prodanova, M., Slavov, K., Etropolska, I., Ganev, K., Miloshev,
N. and Ljubenov, T. (2013): Bulgarian System for Air Pollution Forecast,
Journal of International Scientific Publications ECOLOGY & SAFETY,
Volume 7, Part 1, ISSN: 1313-2563, pp.325-334.

 Hristina Kirova
https://orcid.org/0000-0002-1764-2443
National Institute of Meteorology and Hydrology
Sofia, Bulgaria
E-mail: [email protected]

 Nadya Neykova
https://orcid.org/0000-0003-1892-2081
National Institute of Meteorology and Hydrology
Sofia, Bulgaria
E-mail: [email protected]

 Emilia Georgieva
https://orcid.org/0000-0002-8466-4976
National Institute of Meteorology and Hydrology
Sofia, Bulgaria
E-mail: [email protected]

111
Proceeding of 1st International conference on Environmental protection and disaster RISKs

https://doi.org/10.48365/envr-2020.1.10 Air Pollution,


Climate and Health

CLOUD AND RAIN WATER CHEMICAL


COMPOSITION AT PEAK CHERNI VRAH, BULGARIA

Elena Hristova, Blagorodka Veleva, Krum Velchev,


Emilia Georgieva
National Institute of Meteorology and Hydrology (NIMH)

Abstract. The purpose of the work is to present and discuss newly obtained
data for the chemical composition of cloud water (CW) and rain water (RW) at a
high-elevation site in Bulgaria. Sampling of CW and RW was organized in 2017
and 2018 during field experiments at Cherni Vrah, the highest peak in Vitosha
Mountain. Passive collectors designed and constructed at NIMN were used. All
collected samples (118) were analyzed for acidity (pH), conductivity (EC), main
anions - SO42-, NO3-, Cl-, ammonium ions (NH4+), macro and micro elements (Na,
K, Mg, Ca, Fe, Si, Zn, Cu). The average pH values for both types of samples were
in the acidity range (<5.0). The values of EC varied from 5 to 89.2 µS.cm-1 for RW
and from 0.7 to 202 µS.cm-1 for CW. The ion composition was dominated by NH4+,
Ca, nssSO42- and NO3-, which made up more than 63% of the total ionic content for
RW and 75% for CW. The relative contribution of the major compounds to the CW
and RW composition is presented and discussed. The effect of long-range transport
processes is studied for some selected periods of 2018 using HYSPLIT air mass
backward trajectory analysis.
Keywords: Cloud water; rain water; chemical composition; acidity; backward
trajectories

INTRODUCTION
The atmosphere is an important environment in which different gaseous and
aerosol species are transported. Cloud water (CW) and rain water (RW) play
important roles in removing particles and dissolved gaseous pollutants from the
atmosphere. They also scavenge sulphur dioxide (SO2), nitrogen oxides (NOx),
and other atmospheric pollutants, which can affect their acidity and chemical
composition and cause ecological damage to ecosystems (Seinfeld and Pandis,
2006, Gioda et al., 2013). The chemical composition of CW and RW depends on
pollutants emitted by sources of anthropogenic and natural origin, the dynamical
processes in the atmosphere and the chemical reactions that occur during both local
and long-range transport. Nitrates (NO3-), sulphates (SO42-) and other ions such
112
Cloud and Rain Water Chemical Composition at Peak...

as ammonium (NH4+), chloride (Cl), magnesium (Mg), calcium (Ca), potassium


(K), and sodium (Na) are commonly present in the aqueous phase. Some of these
species originate from natural sources such as sea spray, soils and forest fires.
Other species such as ammonium come from anthropogenic activities, including
agricultural fertilizers and biomass burning (Hůnová, et al., 2017, Mihajlidi-Zelić
et., 2006).
Several authors worldwide have explored trends in the chemical composition
of cloud and rain water since 1990 (Weathers, et al., 1988, Anderson, et al., 2006,
Aleksic et al., 2009, Gioda et al., 2013 Schwab et al., 2016). At high-elevation
environments, cloud and fog liquid water have generally been recognized as being
more acidic than rainwater. In Bulgaria, precipitation chemistry in forest mountain
areas was analysed occasionally during field campaigns related to depositions and
critical loads to ecosystems (Zeller et al., 1998, Ignatova & Damyanova, 2012,
Ignatova & Fikova, 2007). A few studies have focussed on analysis of meteorological
conditions and the influence of air mass transport on the chemical composition of
precipitations at peaks Cherni Vrah and Moussala (Iordanova & Blaskova, 2011).
In the last years, NIMH is conducting research activities on atmospheric
depositions in Bulgaria including both numerical simulations and observational
campaigns (Georgieva et al., 2015, Georgieva et al., 2018, Hristova et al, 2016,
Hristova, 2017, Hristova &Veleva, 2015). The observational campaigns are
organized in areas that could be adversely impacted by the depositions of acidifying
and eutrophying compounds, and thus, be exposed to environmental risks (mountain
and nature protected areas). The work presented here is part of these recent activities.
The scope of this study is to compare and discuss newly obtained data for the
chemical composition of CW and RW at a high-elevation site in Bulgaria – Cherni
Vrah (ChV), Vitosha Mountain. Another objective is to analyse the effect of long-
range transport processes on the chemical composition that will be briefly outlined
for selected case periods.

METHODS
Site description
Vitosha Mountain is the first national park in Bulgaria and in the Balkan
Peninsula. It is located on the outskirts of Sofia, with an area of 270.79 km2. Cherni
vrah is the highest peak where the Meteorological station of NIMH is situated
(42.6167 N, 23.2667 E, 2286 m asl), Fig.1. The CW and RW samples were collected
at this meteorological station from June 2017 to November 2018.
The cloud water is sampled by using a passive collector designed and constructed
in NIMH (Fig.2a). All construction is made from plexiglas and sampling elements
are made from fishing lines with length in total 180 m: 100m - ø1mm 80 m –
ø0.5mm. The cloud droplets impact on the vertical strings, combine to larger drops,
run down the strings, and drip into a 500 ml polyethylene (PE) bottle. The CW
113
Elena Hristova, Blagorodka Veleva, Krum Velchev, Emilia Georgieva

Figure 1. Map of Vitosha Mountain and synoptic station Cherni Vrah

samples were collected on an event basis, resulting in different time duration of


typically about several hours.

a) b)

Figure 2. Samplers for (a) cloud water and (b) rain water

Manual (bulk) rain water sampler is used for sampling of rainwater samples
(Fig.2b). The material of the collector is polyethylene terephthalate funnel of 20
cm in diameter and bottle with 5l capacity. The bulk sampler is washed every day
with deionized water (<1µS.cm-1) to avoid dry deposition. The RW samples were
taken on daily basis.

Chemical analysis of cloud and rain water samples


All 40 cloud and 78 rainwater samples were analysed for acidity (pH), electro
conductivity (EC), Cl-, SO42-, NO3-, Ca, Mg, K, Na, Fe, Si, Zn, Cu, NH4+. Acidity
and electro conductivity were measured at the moment of sampling by a portable
pH-meter and conductivity meter. The pH meter was calibrated before each
measurement using standard buffer solutions of pH 4.00 and 7.01. The conductivity
meter was periodically calibrated against KCl standard solutions. Chemical analysis
are performed in certified laboratory by Ion Chromatograph (ICS 1100, DIONEX)
114
Cloud and Rain Water Chemical Composition at Peak...

for SO42-, NO3-, Cl-, ICP OES (Vista MPX CCD Simultaneous, VARIAN) for Ca,
Mg, K, Na, Fe, Si, Zn, Cu and Spectrophotometer S-20 for NH4+.

Table 1. Detection limits (DL) for all analysed elements


Ca K Mg Na Fe Si Zn
DL, mg.l -1
0.10 0.10 0.10 0.002 0.1 0.002 0.1 0.005 0.10 0.005 0.05

The detection limits for all analysed elements are presented in Table 1. The
concentration of nss_SO42- has been estimated by correction based on assumption
that sodium is a sea salt tracer: [nss_SO42-] = [SO42-]-(0.25× [Na]), according to
WMO GAW, 2004.

Trajectory analysis for long range effects


The origin of the air masses back-ward analysis was conducted using the Hybrid
Single Particle Lagrangian Integrated Trajectory (HYSPLIT), (Stein et al, 2015,
Rolph et al., 2017). Three different arrival heights were used: 500, 1000, 1500 m
a.g.l (above ground level). The calculations were made for 72h using meteorological
input from NCEP GDAS with horizontal resolution of 0.5° x 0.5°.

RESULTS AND DISDUSIONS


Physico-chemical parameters
The CW and RW pH values in this study ranged from 3.5 to 5.7 and the
conductivity values from 0.7 to 202 μS cm-1. The distribution of the relative pH and
EC frequency for both types of samples are presented in Figure 3. This frequency
analysis shows that 100 % of the cloud and 98% of the rain samples have pH value
in the acidity range (Figure 3a). Тhe highest percentage of pH values are in the
range of 4.0 – 4.5 for both CW and RW (58% and 56%). Only 1.5% of the collected
rain water samples are in the neutral range (5.5-6.0), there are not cloud water
sample with pH value higher than 5.3. Around 10% of CW and RW samples have
pH in the slightly acidic range (5.0 – 5.5), and 28% of RW pH and 18% of CW are
in the 4.5 -5.0. In the very acidic range (3.5 – 4.0) are 13% of the CW pH values
and 5% of the RW pH values. The frequency analysis of electrical conductivity
presented in Figure 3b shows 57% of RW EC values and 33% of CW EC values in
the range 0.07 – 20 μS cm-1.
The percentage of EC values in the range 20 – 40 μS cm-1 is very close for both
type of samples (23% and 25%). The percentage with EC values in the range > 60
μS cm-1 is very different: 3% for RW and 31% for CW. Cloud water samples have
higher conductivities (0.7 – 202 μS cm-1) than rain water (4.7 – 89 μS cm-1) due to
dilution factors in the rain, i.e., cloud droplets have lower water content than rain
droplets; therefore the ionic concentrations were higher in cloud water (Gioda et
115
Elena Hristova, Blagorodka Veleva, Krum Velchev, Emilia Georgieva

a) b)
Figure 3. Frequency of pH (a) and EC (b) for cloud and rain water samples

al., 2013). The average EC values for CW and RW are 56 μS cm-1 and 23 μS cm-1,
respectively.
The CW and RW pH parameter is the result of acid-base reactions in the cloud
droplets. Sulphates and nitrates are the main ions that increase the concentration
of H + ion in rainwater, while NH4+, Ca (usually in the form of CaCO3), Mg, K are
the main neutralizing ions. The total ionic content (TIC) of cloud and rain water
samples is ranged from 1.1 to 68 mg l−1 and from 4 to 90 mg l−1, respectively. The
median of TIC in CW samples is 16.5 mg l−1 and for RW samples is 6.2 mg l−1. As

Figure 4. Frequency distribution of Total Ionic Content in cloud


and rain water samples
116
Cloud and Rain Water Chemical Composition at Peak...

shown by the frequency distribution of TIC in Fig. 4, a fraction of cloud and rain
samples have concentrations between 1 and 20 mg l−1 (89% of the RW and 62% of
the CW). Тhe percentage of samples in the concentration range 20 – 40 mg l−1 is
higher for the CW than for the RW (25% and 6%, respectively). Only 2.6% of the
CW samples have TIC in the range 80-100 mg l−1.
The variation in the concentrations of all studied elements is seen on the Box
Plot presented in Fig.5. The ion composition of RW and CW was dominated by
NH4+, Ca, nss_SO42- and NO3-, which made up more than 63% and 75% of the total
ionic content. As expected, concentrations of analyzed elements are higher in cloud
water than in rain water samples.

a) b)
Figure 5. Concentrations of the studied elements
in (a) cloud water and (b) rain water samples

The concentrations of the main acidifying ions - SO42- for the study period ranged
from 0.6 to 39mg.l-1 for CW and from 0.3 to 20.3mg.l-1 for RW. Concentrations of
NO3- vary from 0.15 to 23.1mg.l-1 for CW and from 0.18 to 15.8mg.l-1 for RW.
NH4+ ion concentrations for CW and RW samples are ranged from 0.01mg.l-1 to
8.3mg.l-1 and from 0.03mg.l-1 to 3.9mg.l-1. The lowest variations in concentrations
were observed for Fe, Cu and Zn. Their concentration is ranged from 0.005 to
1.8mg.l-1. High variation in Cl and Na concentrations are observed in RW samples.
The obtained Cl concentration ranged from 0.05mg.l-1 to 27mg.l-1 and for Na from
01 mg.l-1 to 15.6mg.l-1.

Long range transport effects for some selected periods


The origin of air masses was examined by using back-trajectories from the
model HYSPLIT (Stein et al, 2015, Rolph et al., 2017) for two periods: 19 – 20
March 2018 and 30 June – 4 July 2018.
The synoptic situation during the period 19 – 20 March 2018 is characterised
by Saharan outbreak towards the Balkans, associated with coloured rain and orange
117
Elena Hristova, Blagorodka Veleva, Krum Velchev, Emilia Georgieva

snow in many parts of Eastern Europe. At ChV the arriving air masses were from
South (S)-Southwest (SW) (Fig. 6) on 19.03 (rain water sample) and from West (W)
on 20.03 (cloud water sample) (Fig. 6). From all analysed samples collected in 2018,
the largest concentrations of Cl and Na were obtained in the rain water samples on 19
March. The TIC of the rain sample is 64.1mg.l-1 with 42% contribution of Cl, 24%
of Na, 3% ss_SO42- (sea salt SO42-) and only 6% of the sulphates from anthropogenic
source (nss_SO42-) (Fig.6). A very large difference in the SO42- concentrations be-
tween RW and CW is observed. The TIC of the cloud water sample on 20 March
2018 is 3.9 mg.l-1 containing 28% nss_SO42-, following by 28.3% Ca, 15.9% Cl and
11.1% K. These results indicated aged air masses with sea salt aerosols (Cl and Na)
and mineral dust (Ca and Si) associated with Saharan origin.

Figure 6. Back-trajectories and contribution of different elements in rain water


and cloud waters samples for the period 19 – 20 March 2018

Тhe synoptic situation for the second period, 30 June – 4 July 2018, is character-
ized by the influence of the slowly moving Mediterranean cyclone “Nefeli” cross-
ing the country from south to northeast.
The atmospheric conditions in the first part of the period were highly unstable
with heavy rains and thunderstorms in many places in Bulgaria, while the end of
the period was marked by increased surface pressure and occasional convective
precipitations (Monthly bulletin, 06 & 07, 2018). The TIC in rain sample collected
on 30 June (48.5 mg.l-1) is higher than this for cloud water sample collected on 3
July (36.7 mg.l-1). The trajectory analysis shows that on 30 of June the transport
of air masses to the ChV is from North while they are from W, NW on 3 of July
(Fig.7). Generally, nss_SO42- was found to be the dominant ion in both samples: RW
(36%) and CW (43%). The contribution of NO3- and NH4+ ions in the CW sample
118
Cloud and Rain Water Chemical Composition at Peak...

Figure 7. Back-trajectories and contribution of different elements


in rain water (30 Jun) and cloud water (3 Jul) samples

(23% and 12%) are higher than in the RW sample (16% and 5%). The contribution
of Ca is three times higher in the RW sample than observed in the CW sample. For
this selected period the TIC is consisted mainly of nss_SO42-, NO3- NH4+ and Ca
(RW-83% and CW-88%).

CONCLUSIONS
New results for the chemical composition of cloud water (CW) and rain wa-
ter (RW) at The high-elevation site Cherni Vruh were presented. The results were
based on 40 cloud and 78 rainwater samples collected and analysed in the pe-
riod June 2017 – November 2018. The comparison of the cloud and rain water
presents systematic differences concerning the pH, the electric conductivity, and
119
Elena Hristova, Blagorodka Veleva, Krum Velchev, Emilia Georgieva

concentrations of most elements. The frequency analysis showed that 100 % of the
cloud and 98% of the rain samples have pH value in the acidity range (<5.0).
Mean concentrations were generally higher in cloud water than in rainwater
samples. The ion composition of RW and CW was dominated by NH4+, Ca, nss_
SO42- and NO3- accounting for more than 63% and 75% of the total ionic content.
The obtained concentrations of nss_SO42- in the cloud water samples are two times
higher than those reported for various mountain sites by Marinoni et al., 2004,
Schwab et al., 2016 and Gioda et al., 2013. The mean concentrations of NO3- and
NH4+ in the CW samples are lower than ones derived in Marinoni et al., 2004 and
Gioda et al., 2013. The study of the origin of the air masses contributed to better
understanding of variations in the chemical composition and concentration levels
in two specific cases in 2018.

ACKNOWLEDGMENTS
This study was performed with the financial support from the Bulgarian National
Science Fund trough contract N. DN-04/4-15.12.2016. We acknowledge also to
the NOAA Air Resources Laboratory (ARL) for the provision of the HYSPLIT
model and READY website (http://www.ready.noaa.gov).

REFERENCES
Aleksic N, K. Roya, G. Sistla, J. Dukett. N. Houck. P. Casson. (2009).
Analysis of cloud and precipitation chemistry at Whiteface Mountain.
NY. Atmospheric Environment, 43, 2709 – 2716.
Anderson, J.B., Baumgardner, R.E., Grenville, S.E., (2006). Trends in
cloud water sulfate and nitrate as measured at two mountain sites in
the Eastern United States: regional contributions and temporal changes
compared with regional changes in emissions, 1986–1999. Atmospheric
Environment, 40, 4423 – 4437.
Fowler. D., Pilegaard. K., Sutton. M.A., et al., (2009). Atmospheric
composition change: ecosystems and atmosphere exchange. Atmospheric
Environment, 43, 5192 – 5263.
Hristova E., Veleva B, (2015), Experimental study of the chemical
composition of precipitation and PM10 in Sofia in 2014, Bulgarian
Journal of Meteorology and Hydrology, 20, 5, 23 – 32, (in Bulgarian).
Hristova E., Georgieva E., Velchev K., Kirova H., Oruc I., Nikolov V.,
Syrakov D., Prodanova M., Batchvarova E., Veleva B., Petrov A.,
Neikova R., Branzov H., Kolarova M., Etropolski E. (2016). Composition
of precipitations in the coastal area of Southeast Bulgaria, 3rd National
Congress on Physical Sciences, Section: Physics of Earth, Atmosphere
120
Cloud and Rain Water Chemical Composition at Peak...

and Space, 29 Sep. – 2 Oct. 2016, Sofia, CD, Heron-Press Sofia, 2016,
ISBN:978-954-580-364-2.
Hristova E., Chemical composition of precipitation in urban area. (2017).
Bulgarian Journal of Meteorology and Hydrology, 22, 1 – 2, 41 – 49, (12).
Hůnová, P. Kurfürsta, V. Stráníka, M. Modlíka. (2017). Nitrogen deposition
to forest ecosystems with focus on its different forms. Nitrogen deposition
to forest ecosystems with focus on its different forms, Science of the
Total Environment, 575, 791 – 79.
Gioda A., Mayol-Bracero O. L., Scatena F. N., Weathers K. C., Mateus
V. L., McDowell W. H.. (2013). Chemical constituents in clouds and
rainwater in the Puerto Rican rainforest: Potential sources and seasonal
drivers, Atmospheric Environment, 68, 208 – 220
Ignatova N. and Fikova R.. (2007). Precipitation chemistry in Sweet
Chestnut forests (Castanea sativa Mill.) in Belasitza Mountain.
Sustainable management of Sweet Chestnut ecosystems-CAST Bul.
book with reports from I-st Balkan regional workshop. 2-5 November.
Blagoevgrad. Bulgaria. publishing house Petekston. 2007. ISBN:
978-954-457-046-0
Iordanova L., Blaskova S. (2011). Impact of atmospheric precipitations on
the surface water chemistry in the high mountain area of the Bulgarian
Danube basin In: Proceedings of the XXV-th conference of the Danube
countries on hydrological forecasting and hydrological bases of water
management. Budapest. Hungary. Session 7, 6:822 – 830.
Marinoni, Laj P., Sellegri K., Mailhot G. (2004). Cloud chemistry at the
Puy de Dˆome: variability and relationships with environmental factors,
Atmospheric Chemistry and Physics, 4, 715 – 728.
Mihajlidi-Zelić A., Deršek-Timotić I., Relić D., Popović A., Đorđević D.
(2006). Contribution of marine and continental aerosols to the content of
major ions in the precipitation of the central Mediterranean, Science of
the Total Environment, 370, 441 – 451.
Monthly hydro metrological bulletin, (2018), NIMH, Sofia June, July (in
Bulgarian). ISSN 1314-894X
Schwab, J.J., Casson, P., Brandt, R., Husain, L., Dutkewicz, V., Wolfe,
D., Demerjian, K.L., Civerolo, K.L., Rattigan, O.V., Felton, H.D. and
Dukett, J.E. (2016). Atmospheric Chemistry Measurements at Whiteface
Mountain, NY: Cloud Water Chemistry, Precipitation Chemistry, and
Particulate Matter. Aerosol Air Qual. Res. 16: 841 – 854. https://doi.
org/10.4209/aaqr.2015.05.0344
Stein, A.F., Draxler, R.R, Rolph, G.D., Stunder, B.J.B., Cohen, M.D.,
and Ngan, F. (2015). NOAA's HYSPLIT atmospheric transport and
dispersion modeling system, Bull. Amer. Meteor. Soc., 96, 2059 – 2077.

121
Elena Hristova, Blagorodka Veleva, Krum Velchev, Emilia Georgieva

Rolph, G., Stein, A., and Stunder, B. (2017). Real-time Environmental


Applications and Display sYstem: READY. Environmental Modelling &
Software, 95, 210 – 228.
Zeller K., Bojinov Ch., Donev E., Nikolov N.. (1998). Bulgarian Rila
Mountain Forest Ecosystems Study Site: Site Description and SO4²-.
NO3- Deposition, In: Proceedings of the International Symposium on Air
Pollution and Climate Change Effects on Forest Ecosystems, February
5 – 9. 1996. Riverside. CA Gen. Tech. Rep. PSW-GTR-166. Albany.
CA: U.S. Department of Agriculture, Forest Service, Pacific Southwest
Research Station, 332 p.
Weathers, K.C., Likens, G.E., Bormann, F.H., Bicknell, S.H., Bormann,
B.T., Daube Jr., B.C., Eaton, J.S., Galloway, J.N., Keene, W.C., Kimball,
K.D., McDowell, W.H., Siccama, T.G., Smiley, D., Tarrant, R. (1988).
Cloud water chemistry from ten sites in North America, Environmental
Science & Technology, 22, 1018 – 1025.
WMO GAW Report No. 160. (2004). Manual for the GAW precipitation
chemistry programme, Edited by Mary A. Allan Prepared by GAW
Precipitation Chemistry Science Advisory Group.

 Elena Hristova
https://orcid.org/0000-0002-5681-4375
National Institute of Meteorology and Hydrology
Sofia, Bulgaria
E-mail: [email protected]

 Blagorodka Veleva
https://orcid.org/0000-0003-2848-5559
National Institute of Meteorology and Hydrology
Sofia, Bulgaria
E-mail: [email protected]

 Krum Velchev
National Institute of Meteorology and Hydrology
Sofia, Bulgaria
E-mail: [email protected]

 Emilia Georgieva
https://orcid.org/0000-0002-8466-4976
National Institute of Meteorology and Hydrology
Sofia, Bulgaria
E-mail: [email protected]

122
Proceeding of 1st International conference on Environmental protection and disaster RISKs

https://doi.org/10.48365/envr-2020.1.11 Air Pollution,


Climate and Health

CHARACTERISTICS OF ULTRAFINE PM EMITTED


DURING GASIFICATION OF BIOMASS

Iliyana Naydenova1, Tsvetelina Petrova1, Ognyan Sandov1,


Ricardo Ferreira2, Rositsa Velichkova3, Mario Costa2
Technical University of Sofia, College of Energy and Electronics
1
2
Universidade de Lisboa, Instituto Superior Técnico,
Mechanical Engineering Department, IDMEC – Lisboa (Portugal)
3
Technical University of Sofia, Faculty of Power Engineering and Power Machines

Abstract:
The present work aims at characterising ultrafine particulate matter (PM) that
was emitted during biomass gasification in a drop tube furnace (DTF). Three
different types of agricultural residue were gasified using the following gasifying
agents: O2/N2 and O2/N2/CO2. The PM in the flue gases were sampled, using 13 stage
impactor. Particulates, having aerodynamic diameter of 1 and 0.65 µm (PM1 and
PM0.65) had the highest mass fraction among all other particulates with aerodynamic
diameter up to 10 µm. Thus, the effects of biomass and gasifying agent on particles’
mass distribution were studied. Besides, PM1 and PM0.65 were characterised using
X-ray fluorescence (XRF) analysis. The PM elemental composition confirmed that
volatile ash compounds, such as K and Cl are typical constituents of the ultrafine
particulates, when biomass from agricultural residue (colza and sunflower husks)
was gasified.
Keywords: particulate matter, biomass, gasification.

INTRODUCTION
Currently in Bulgaria, the air quality investigations are mostly focused on
the atmospheric air pollution. Certain attention is drawn on the particulate mat-
ter (precisely PM10 and PM2.5), namely the exceedance of their norms (Directive
2008/50/EC), PM chemical composition and source apportionment. Chuturkova
(Chuturkova, 2015) assessed particulate matter (РМ10 and РМ2.5) in atmospheric
air for the period 2007-2014. Two regions were examined (urbanized and indus-
trial) involving three monitoring stations: city background, transport-oriented,
and industrial-oriented. The author reports the average annual PM10 and PM2.5
concentrations, the average monthly PM10 concentrations, the maximum PM10
concentrations (μg/m3) and the number of exceedings in the investigated pe-
123
I. Naydenova, Ts. Petrova, O. Sandov, R. Ferreira, R. Velichkova, M. Costa

riod. Vlaknenski et al. (Vlaknenski, 2015) traced the background air pollution
with PM10 in three medium-sized urban areas of Central North Bulgaria during
2007-2010. Veleva et al. (Veleva, 2014a and Veleva, 2014b) measured the daily
concentration of PM10 in a period of four seasons in Bulgaria, as well as the
chemical composition of the collected samples through EDXRF analysis. The
authors identify more than 23 chemical elements in the structure of the collected
PM10. Veleva et al. (Veleva, 2015) organized six experimental campaigns in Sofia
within the winter and the summer periods of 2012, 2013 and 2014. The authors
report the PM10 mass concentration and the elemental composition of more than
20 elements (P, S, Cl, K, Ca, Ti, V, Cr, Mn, Fe, Ni, Cu, Zn Br, Rb, Sr, Y, Zr, Cd,
Sn, Sb, I, Ba, Pb), which show significant variations in maximum and mean con-
centrations.
Antova et al. (Antova, 2019) collected and identified the main air pollutants in
indoor air. The authors studied the concentration of PM10 and PM2.5, ozone, carbon
monoxide, carbon dioxide, formaldehyde, nitrogen dioxide, volatile organic com-
pounds, moisture and mold presence in school classrooms.
The pronounced interest on PM10 and PM2.5 concentration in the urban atmo-
sphere is challenged by the European and the national regulations, straggling to
find efficient solution for various insufficiently controlled processes, such as com-
bustion in terms of residential heating and transport (IIR, 2019, Nikolaev, 2017
and Naydenova, 2018). Gvero et al. (Gvero, 2018) measure the daily average PM10
concentrations (μg m–3) in one part of Banja Luka city area, with dominant private
households, and determine the PM chemical composition. Juda-Rezler et al. (Juda-
Rezler, 2020) studied the PM2.5 for one calendar year and gave the seasonal concen-
tration of PM2.5, 19 trace elements in PM2.5, as well as PM2.5 sources apportionment
– residential combustion, exhaust traffic emissions, non-exhaust traffic emissions.
Zalakeviciute et al. (Zalakeviciutea, 2020) identified 28 different elements in the
structure of PM10, as well as the existence of ions, such as SO42-, NO3- and NH4+.
Pateraki et al. (Pateraki, 2020) chemically characterized PM2.5 and PM1, sampled
from Greater Athens Area. The authors identified 20 different PM-bound polycyclic
aromatic hydrocarbons (PAH). Chernishev et al. (Chernishev, 2018) investigated
the particle size distribution and the chemical composition, e.g. the fraction of par-
ticles with structured carbon (crystalline phase state) in the sampled PM10, which
originated from the exhausts of two-wheeled vehicles that are typically used in the
territory of Vladivostok, Russia.
Recently, the variety of processes for solid biofuels utilization gained signifi-
cant attention due to the increase of energy demands and shortage of fossil fuels,
the main source of energy in the planet (Villetta, 2017). However, the flue gases of
such processes are typical source of pollutants, including char and PM of different
size and chemical composition. The ultrafine PM contain significant amount of soot
particles known to be harmful to the environment and the human beings. They can
124
Characteristics of Ultrafine PM Emitted...

penetrate easily into the respiratory system, causing lung malfunction and prema-
ture death (Göktepe, 2016 and Naydenova, 2007).
Thus, detailed research is needed about the emission modes and factors of ultra-
fine PM, as well as on their chemical composition and behavior.
Gasification is one of the possible utilization ways to convert solid biofuels into
valuable products. Generally, biomass gasification is a thermo-chemical conversion
in a low oxygen environment (using different gasifying agents - usually air, oxy-
gen or steam), producing syngas. It is a mixture mainly made of carbon monoxide
(CO), hydrogen (H2), methane (CH4) and other hydrocarbons. These substances are
flammable and combustible (Villetta, 2017). According to Basu and Duarte (Basu,
2010 and Duarte, 2018) the gasification converts solid or liquid carbonaceous fuels
into a synthesis gas or syngas that can be used to produce chemical feedstock, liq-
uid fuels, power or gaseous fuels.
The present work was focused on the ultrafine PM emitted during biomass
gasification, which was carried out in a drop tube furnace (DTF). Three different
types of biomass were utilized with two types of gasifying agents: O2/N2 and O2/N2/
CO2. The effects of biomass and gasifying agent on the particles’ mass concentration
were investigated, along with the chemical characterisation of the PM1 and PM0.65,
using XRF analysis. The work is still in progress and herein only some preliminary
results are presented.

METHODS USED
In the present experiment, PM with different sizes was collected during biomass
gasification in DTF, which is a laboratory-scale reactor, suitable for pulverised-fuel
conversion. Detailed description of the applied experimental set up is provided by
Duarte and Duarte et al. (Duarte, 2018 and Duarte, 2019) and Adánez-Rubio et.al
(Adánez-Rubio, 2020). Duarte et al. (Duarte, 2019) conduct gasification experi-
ments of wheat straw. The authors examine the influence of the operating tempera-
ture of the DTF between 900 and 1200 ºC on the formation of syngas and soot.
The authors obtained that higher operating temperatures result in higher CO and
H2 yields, higher H2/CO volume ratio, which leads to a higher carbon conversion
efficiency but lower yields of CH4 and CO2. The formation of soot increases with
temperature increase and reaches a maximum at 1000ºC, above which it starts to
decrease. Adánez-Rubio et.al (Adánez-Rubio, 2020) investigate the pig manure
gasification at different temperatures (900-1200 ºC) and different gasifying agents
(N2/O2, N2/O2/CO2, N2/O2/H2O). The results show that the temperature increase
leads to a higher production of soot particles, and the use of water decreases the
soot production.
The present work was focused on the chemical characterization of ultrafine PM
that was sampled from the flue gases during biomass gasification. The experiment
was carried out at 1000ºC, expecting to have high soot yield, which would highly
125
I. Naydenova, Ts. Petrova, O. Sandov, R. Ferreira, R. Velichkova, M. Costa

affect the yield of ultrafine particulates. The selected types of biomass are woody
biomass (softwood and bark) and agriculture residue (colza and sunflower husks).
The present study was focused on the Bulgarian alternative biofuels that are easily
accessible in large amount annually as a result of food and energy industries. For
instance, sunflower husks and softwood are typically used for solid biofuels produc-
tion. Recent technologies allow co-combustion of biomass with coal in pulverised
coal boilers. On the other hand, the softwood is widely used for residential heating
purposes (as solid biofuel pellets or it is directly burnt in stoves at some regions in
Bulgaria).
Prior to the experiment, the biomass was characterized through proximate, ulti-
mate and ash analyses. The results are presented elsewhere (Naydenova, 2020). The
biomass was initially ground and sieved to assure fuel particle size below 1mm.
The biomass flow was kept at 15 g/h throughout all experiments, and the
excess air ratio (𝜆) was constant, at a value of 0.4. Two different gasifying
agents were used in the following proportions: (1) a mixture of 1% of O2 and
99% of N2 and (2) a mixture of 1% O2, 94% of N2 and 5% of CO2. The nitrogen
was used as a carrier gas. The chosen operating conditions were based on the
works of Duarte, Duarte et al. and Adánez-Rubio et.al. (Duarte, 2018, Duarte,
2019 and Adánez-Rubio, 2020), as well as on the requirements of Dekati PM
sampling system (Dekati, 2013).
During biomass gasification, the PM samples were collected with 13 stage im-
pactor, which assures D50% values (μm) for 13 cutdiameters (cut-off sizes), be-
tween 10 μm and 30 nm, (Dekati, 2010). The present work aims at characterizing
only the ultrafine particulates, namely the PM with D50% = 0.65 μm (PM0.65) and
D50% = 1 μm (PM1). For simplicity, further in the text, the above described particu-
lates are entitled as PM0.65 and PM1.
The PM elemental composition was obtained using XRF analysis, which was
carried out with Fischerscope X-Ray System XDАL. This method for qualitative
analysis was chosen as a primary step of PM characterization, because it is non-
destructive and allows identifying relatively wide range of chemical elements with
atomic numbers from 19 (potassium) to 92 (uranium) in the ppm range.

RESULTS AND DISCUSSION


In the present experiment, the ultrafine particulates (PM0.65 and PM1) constituted
the major portion of the PM, collected in the 13-stage cascade. Similar tendency
is observed in an independent study, described elsewhere (Naydenova, 2020), on
the combustion of five different types of biomass (cherry stones and pellets from
softwood, sunflower husk, rape, wheat straw) in fluidized bed conditions. The pres-
ently collected ultrafine particulates were subject to further research. The relative
mass concentration (wt. %) of both PM1 and PM0.65 among all collected PM in the
13-stage cascade was obtained:
126
Characteristics of Ultrafine PM Emitted...

When O2/N2 was used as gasifying agent: 71.63% for sunflower husks, 50.63%
for softwood and 35.55% for colza;
When O2/N2/CO2 was used as gasifying agent: 46.18% for sunflower husks,
43.72% for softwood and 68.46% for colza.

Table 1. Mass of the sampled PM


Mass (g) wt.%
Biomass type
PM0.65 PM1 PM0.65 PM1
O2/N2 – gasifying agent
Colza 0.00160 0.00038 28.73 6.82
Softwood 0.00518 0.00367 29.63 21.00
Sunflower husks 0.00607 0.00562 37.19 34.44
O2/N2/CO2 – gasifying agent
Colza 0.00571 0.00673 31.43 37.04
Softwood 0.00355 0.00735 14.24 29.48
Sunflower husks 0.00340 0.01324 9.44 36.75

Furthermore, the particle mass distribution PM1 and PM0.65 were investigated.
Thus, the effect of the biomass type and the gasifying agent on the PM mass and
wt. % distribution is presented in Table 1 and Fig. 1. In the case of O2/N2 - the
following relations were observed for the mass concentration of PM0.65 and PM1
(Fig. 1):
- Colza: the mass concentration of PM0.65 was about 4 times higher than the one,
measured for PM1;
- Softwood: PM0.65 was 29 % higher than PM1;
- Sunflower husks: PM0.65 and PM1 were with almost equal masses.
In the case of O2/N2/CO2 - the following relations were obtained (Fig. 1):
- Colza: PM0.65 was about 17% higher than PM1;
- Softwood: PM0.65 was twice timed less than PM1;
- Sunflower husks: PM0.65 was about 4 times less than PM1.
Besides, the gasification of colza resulted in an almost equal mass of PM0.65
among all other PM, sampled in the 13-stage cascade, for both gasifying agents.
On the other hand, a significant difference in PM1 mass concentration was ob-
served: 6.82 % in O2/N2 and 37.04 % in O2/N2/CO2 atmosphere. When softwood
was gasified, the fraction of PM0.65 among all other PM was 29.63 % in O2/N2
and 14.24 % in O2/N2/CO2, whereas the PM1 was 21.00 % in O2/N2 and 29.48%
in O2/N2/CO2 atmosphere. In the case of sunflower husks, the mass concentration
of PM0.65 was 37.19 % in O2/N2 and 9.44 % in O2/N2/CO2 atmosphere, whereas in
O2/N2/CO2 atmosphere the PM1 showed higher concentration for all investigated
biomass types.
127
I. Naydenova, Ts. Petrova, O. Sandov, R. Ferreira, R. Velichkova, M. Costa

Figure 1. The percentage of PM0.65 and PM1 relative to the total amount
of particulate matter

Fig. 2 shows a comparison of the emitted PM0.65 and PM1 depending on the bio-
mass used. When O2/N2 was the gasifying atmosphere, the lowest mass of PM0.65
and PM1 was detected in the colza gasification, whereas the highest mass - for the
sunflower husks. Slight variations were observed for the PM0.65 mass during bio-
mass gasification in O2/N2/CO2 atmosphere: softwood (0.00355 g), sunflower husks
(0.0034 g) and colza (0.00673 g). Distinctly higher mass of PM1 was obtained,
when sunflower husks (0.01324 g) were processed.

128
Characteristics of Ultrafine PM Emitted...

Figure 2. The mass of PM0.65 and PM1

Table 2 presents the elemental composition of the investigated PM0.65-1, resulting


from the XRF analysis.

Table 2. Elemental composition of the examined PM


Gasifying agent
Biomass type O2/N2/CO2 O2/N2
PM0.65 PM1 PM0.65 PM1
Colza Fe Fe Fe Fe
K K K K
Cl
Cu Cu
Mn Mn
Zn Zn
Softwood Fe Fe Fe Fe
Mn Mn Mn Mn
Zn Zn Zn
Sr Sr Sr Sr
Zr Zr
Cd
Cu Cu
Sb
As
Sunflower husks Fe Fe Fe Fe
Zn Zn Zn Zn
Ca
Mn Mn
K K

As expected, the ultrafine particulates that were emitted during gasification of


agricultural residue (colza and sunflower husks) contained volatile ash compounds,
129
I. Naydenova, Ts. Petrova, O. Sandov, R. Ferreira, R. Velichkova, M. Costa

such as K and Cl. These results correspond with the results from the ultimate and
the ash analyses of the same feedstock (Naydenova. 2020) and with the observa-
tions of Obernberger et al. (Obernberger, 2006). The authors concluded that the ash,
containing elements like K, Na, P and Cl is typically part of the fly ash, whereas
elements like Ca, Mg and Si are normally found in the bottom ash. The PM0.65-1
from woody biomass (softwood) contained greater variety of heavy metals and Sr
was obtained independently of the gasifying agent.
The combined effects of biomass and gasifying agent on the PM chemical com-
position can be summarized as follows: (a) colza - Fe and K were found in all
filters, independently of the gasifying agent. However, Cl was detected only in the
PM0.65 (O2/N2/CO2), whereas the elements: Cu, Mn and Zn were obtained in the
rest of the filters; (b) softwood – Fe, Mn and Sr were the dominant elements in both
PM0.65 and PM1, accompanied by unequally distributed Zn, Zr, Cd, Cu, Sb and As in
some of the filters; (c) sunflower husks - Fe and Zn were detectable in all samples,
while Ca, Mn and K – only in some of the filters. According to Argyropoulos et al.
(Argyropoulos, 2012), the biomass burning is a typical source of K, Ca, Mn and Sr.

CONCLUSIONS
The main goal of the present study was to characterise the ultrafine particulates
(PM1 and PM0.65), obtained from the flue gases during biomass gasification in
DTF. The PM elemental composition was experimentally measured applying XRF
technique. Thus, the effect of biomass and gasifying agent on the particulates’ mass
distribution and elemental composition was investigated.
When biomass was processed in O2/N2, the relative mass concentration of the
PM collected on both filters, PM0.65 and PM1, over the total mass of all other sized
PM increased in the following order: 35.55 % (colza), 50.63 % (softwood), 71.63 %
(sunflower husks). Nevertheless, when O2/N2/CO2 was the gasifying agent: 43.72 %
(softwood), 46.18 % (sunflower husks), 68.46 % (colza).
The elemental composition of the ultrafine particulates sampled during gasifi-
cation of agricultural residue (colza and sunflower husks) confirmed the existence
of volatile ash compounds, such as K and Cl, which were missing in the PM1-0.65,
sampled during gasification of woody biomass. This result corresponds well with
preliminary published observations (Obernberger, 2006) that elements, such as K
and Cl, play a significant role in the process of ultrafine particulates formation
through the nucleation mechanism.

ACKNOWLEDGMENTS
The PM analysis was funded through the Internal Competition of Technical
University of Sofia - 2020 for the scientific project “Perspective leaders”, Contract
№ 201ПР0017-10/26.05.2020. The authors would also like to acknowledge the
financial support from the National Science Program “Environmental Protection

130
Characteristics of Ultrafine PM Emitted...

and Reduction of Risks of Adverse Events and Natural Disasters”, approved by


the Resolution of the Council of Ministers No 577/17.08.2018 and supported
by the Ministry of Education and Science of Bulgaria (Agreement No D01-
230/06.12.2018 and Agreement No Д01-322/18.12.2019).

REFERENCES
Adánez-Rubio, I., Ferreira, R., Rio, T., Alzueta, M.U., Costa, M. (2020). Soot
and char formation in the gasification of pig manure in a drop tube reactor,
Fuel, 281, 118738, doi:10.1016/j.fuel.2020.118738.
Antova T. G., Panev T. I., Tzoneva M. Tz., Sidjimov M. A., Lukanova R. Т.
(2019). Indoor air quality assessment (pilot study), Journal of International
Scientific Publications, Ecology & Safety ISSN 1314-7234, 13, pp.77-85.
Argyropoulos G., Manoli E., Kouras A., Samara C. (2012). Concentrations and
source apportionment of PM10 and associated major and trace elements in the
Rhodes Island, Greece, Science of the Total Environment, 432, 12–22.
Basu P. (2010). Biomass gasification and pyrolysis: Practical design and
theory. 1st ed., Elsevier Inc.
Bulgaria’s Informative Inventory Report 2019 (IIR, 2019). Submission under
the UNECE Convention on Long-Range Transboundary Air Pollution,
Centre on Emission Inventories and Projections (CEIP), Umweltbundesamt,
Vienna, Austria.
Chernyshev V.V., Zakharenko A.M., Ugay S.M., Hien T.T., Hai L.H., Kholodov
A.S., Burykina T.I., Stratidakis A.K., Mezhuev Ya. O., Tsatsakis A.M.,
Golokhvast K.S. (2018). Morphologic and chemical composition of particu-
late matter in motorcycle engine exhaust, Toxicology Reports, 5, 224–230.
Chuturkova R. (2015). Particulate matter air pollution (PM10 and PM2.5) in
urban and industrial areas, Journal Science Education Innovation, 5, 13-32,
Association Scientific and Applied Research International Journal, ISSN
1314-9784.
Dekati (2013). Dekati DLPI Low Pressure Impactor, https://www.particleinstru-
ments.com/sites/default/files/pdfs/DLPI%20Brochure%206%2C2013.pdf
Directive 2008/50/EC of the European Parliament and of the Council of 21
May 2008 on ambient air quality and cleaner air for Europe, https://eur-
lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:02008L0050-
20150918&from=EN.
Duarte F.S. (2018), Biomass Gasification in a Drop Tube Furnace. Thesis to
obtain the Master of Science Degree in Mechanical Engineering, Instituto
Superior Técnico, Lisboa.
Duarte F., Ferreiro I., Costa M. (2019). Soot formation during the gasification of
wheat straw in a drop tube furnace. 5th International Conference WASTES: So-
lutions, Treatments and Opportunities, Almada, Portugal, 4th – 6th September.

131
I. Naydenova, Ts. Petrova, O. Sandov, R. Ferreira, R. Velichkova, M. Costa

Göktepe B., Umeki K., Gebart R. (2016). Does distance among biomass par-
ticles affect soot formation in an entrained flow gasification process? Fuel
processing Technology, 141, 99–105.
Gvero P., Radić R., Kotur M., Kardaš D. (2018). Urban air pollution caused
by the emission of PM10 from the small household devices and abatement
measures, Thermal Science, 152-152, DOI: 10.2298/TSCI180119152G.
Juda-Rezler K., Reizer M., Maciejewska K., Błaszczak B., Klejnowski K.
(2020). Characterization of atmospheric PM2.5 sources at a Central European
urban background site, Science of the Total Environment, 713, 136729.
Naydenova I. (2007). PhD Thesis, Soot Formation Modeling during Hydrocar-
bon Pyrolysis and Oxidation behind Shock Waves. Ruprecht-Karls-Universi-
t¨at Heidelberg Interdisziplin¨ares Zentrum f¨ur Wissenschaftliches Rechnen
Naydenova I., Petrova T., Velichkova R., Simova I. (2018). PM10 exceedance in Bul-
garia, CBU International conference proceedings, 2018, March 21-23, Prague,
Czech Republic, 6, 1129-1138. http://dx.doi.org/10.12955/cbup.v6.1305.
Naydenova I., Sandov O., Wesenauer F., Laminger T., Winter F. (2020)
Pollutants formation during single particle combustion of biomass under
fluidized bed conditions: An experimental study, Fuel, 278, 117958.
Nikolaev A, Konidari P. (2017). Development and assessment of renewable
energy policy scenarios by 2030 for Bulgaria. Renew Energy, 111, 792–802.
https://doi.org/10.1016/j.renene.2017.05.007.
Obernberger I, Brunner T, Baernthaler G. (2006). Chemical properties of solid
biofuels - significance and impact. Biomass Bioenergy, 30, 973–82. https://
doi.org/10.1016/j.biombioe.2006.06.011.
Pateraki St., Asimakopoulos D.N., Maggos Th., Assimakopoulos V.D., Bou-
giatioti A., Bairachtari K., Vasilakos Ch., Mihalopoulos N. (2020). Chemical
characterization, sources and potential health risk of PM2.5 and PM1 pollution
across the Greater Athens Area, Chemosphere, 241, 125026.
Veleva B., Hristova E., Nikolova E., Kolarova M., Valcheva R. (2014a). Sea-
sonal variation of PM10 elemental composition in urban environment, Jour-
nal of International Scientific Publications: Ecology and Safety 8, ISSN
1314-7234 (Online), http://www.scientific-publications.net.
Veleva B., Hristova E., Nikolova E., Kolarova M., Valcheva R. (2014b). Ele-
mental composition of air particulate (PM10) in Sofia by EDXRF techniques,
Journal of Chemical Technology and Metallurgy, 49 (2), 163-169.
Veleva B., Hristova E., Nikolova E., Kolarova M., Valcheva R. (2015). Experimental
study on elemental composition of PM10 in Sofia, 2012-2014, Science & Technolo-
gies, V (2), Nautical & Environmental studies, DOI: 10.13140/RG.2.1.4570.8562.
Villetta M., Costa M., Massarotti N. (2017). Modeling approaches to biomass
gasification: A review with emphasis on the stoichiometric method, Renew-
able sustainable energy reviews, 74, 71-88.

132
Characteristics of Ultrafine PM Emitted...

Vlaknenski T., Stoychev P., Chuturkova R. (2015). Background air pollution


by fine particles in urban areas of central northern Bulgaria, Environmental
technology, 1 (22), 82-90.
Zalakeviciutea R., Rybarczyk Y., Granda-Albuja M. G., Suarez M. V. D., Ale-
xandrino K. (2020). Chemical characterization of urban PM10 in the Tropical
Andes, Atmospheric Pollution Research, 11, 343–356.

 Iliyana Naydenova
College of Energy and Electronics
Technical University of Sofia
Sofia, Bulgaria
E-mail: [email protected]

 Tsvetelina Petrova
https://orcid.org/0000-0001-8482-4792
College of Energy and Electronics
Technical University of Sofia
Sofia, Bulgaria
E-mail: [email protected]

 Ognyan Sandov
https://orcid.org/0000-0001-7697-669X
College of Energy and Electronics
Technical University of Sofia
Sofia, Bulgaria
E-mail: [email protected]

 Ricardo Ferreira
Mechanical Engineering Department
Instituto Superior Técnico
Universidade de Lisboa
Lisboa, Portugal
E-mail: [email protected]

 Rositsa Velichkova
http://orcid.org/0000-0003-3757-8685
Faculty of Power Engineering and Power Machines
Technical University of Sofia
Sofia, Bulgaria
E-mail: [email protected]

 Mario Costa
Mechanical Engineering Department
Instituto Superior Técnico
Universidade de Lisboa
Lisboa, Portugal

133
Proceeding of 1st International conference on Environmental protection and disaster RISKs

https://doi.org/10.48365/envr-2020.1.12 Air Pollution,


Climate and Health

BLACK SEA FREEZING AND WINTER SEVERITY:


RELATION TO THE TEMPERATURE

Mirna Matov, Elisaveta Peneva, Greta Georgieva


Faculty of Physics, “St. Kliment Ohridski” University of Sofia

Abstract: Black Sea freezing is observed regularly in the northern part and near
the Kerch Strait. The reason is the relatively shallow northwestern shelf part and
the river inflow of the three major European rivers Danube, Dniepr and Dniestr,
carrying large amount of fresh water to this part of the Black Sea. The observed
global warming in the latest decades makes these episodes less intense. In 2012
and 2017 the sea ice extended unusually south. The aim of this study is to fill the
gap of recent analysis on the Northern Black Sea areas covered regularly with ice.
Based on the coastal meteorological stations measurements, we investigate the
interannual-to-decadal variability of the thermal regime.The relation with the sea-
ice cover based on satellite observations is studied. We define various categories of
winter conditions depending on the number of cold days.
Keywords: climate change, Black Sea, winter severity

INTRODUCTION
The Black Sea is a large deep water basin on the border between European and
Asian continents lying in the continental mid-latitude climate zone. It is an impor-
tant climatic factor for all borderline countries (Bulgaria, Romania, Ukraine, Rus-
sia, Georgia and Turkey). The open plane in north direction enables the propagation
of the Siberian High influence in winter and thus severe winter conditions could oc-
cur. From the other side, the Mediterranean Sea influence is significant through the
Mediterranean cyclones passing frequently the area, which leads to mild weather
conditions.
Black Sea freezing is observed regularly in the northern part and near the Kerch
Straits and occasionally spread during cold winters to south reaching Romanian
coast (Simonov and Altman, 1991). Since 1972 moderate freezing was observed,
but February 2012 was extremely cold and the Black sea ice covered area reached
Constanta in Romania (Figure 1). 2017 was also anomalously cold in this area.
There is not a thorough recent study of the areas covered regularly with ice in
the Northern Black Sea, and our aim is to fill this gap and to check to what extent
the sea ice occurrence is important to be taken into account from numerical models.
134
Black Sea Freezing and Winter Severity: Relation...

The other objective is to study the interannual-to-decadal variability of the thermal


regime and winter severity in the region in relation to the freezing.

Figure 1. Romanian coast near Constanta in February 2012

DATA USED IN THE STUDY


In order to reveal the interannual variability during the last decades we use three
sources of data:

Figure 2. Locations of he meteorological stations used in the study

1) Historical archives and literature material on the Black Sea ice extend and
temperature from the Black Sea book (Simonov and Altman, 1991)
2) Data for 2m air temperature coming from 9 coastal meteorological stations
along the Black Sea coast - Burgas, Varna, Odessa, Herson, Mariupol, Batumi,
Trabzon, Sinop and Istanbul (Figure 2). The sea temperature in Burgas and Varna
is also used. The data source is the regular SYNOP telegrams distributed by the
Global Surface Summary of the Day – GSOD (data.nodc.noaa.gov). The data are
decoded from SYNOP code and cover the period since 1980 till present.
135
Mirna Matov, Elisaveta Peneva, Greta Georgieva

3) Data for the ice extent are taken from the US National Snow and Ice Data
Center product Multisensor Analyzed Sea Ice Extent - Northern Hemisphere
(MASIE-NH). This provides daily sea ice data from January 1st, 2006 to the present
(MASIE, Technical References).

ANALYSIS OF THE WINTER TEMPERATURE VARIATIONS

Figure 3. Interannual variability of the winter temperatures (Jan, Feb, Mar)


in Burgas, Varna, Bartumi, Odesa, Herson, Mariupol, Trabzon, Sinop
and Istanbul. The grey columns in the background refer to the data availability,
the periods with no or insufficient data are not taken in the analysis.

We have used data from the NOAA data set Global Surface Summary of the
Day – GSOD (data.nodc.noaa.gov). The mean winter temperature is calculated for
the 9 meteorological stations along the Black Sea coast – Burgas and Varna in
Bulgaria, Odessa, Herson and Mariupol in Ukraine, Batumi in Georgia and Istanbul,
Trabzon and Sinop in Turkey, as the mean value for the months January, February
and march). The graphs are presented in Figure 3 and show significant interannual
variability. As expected the north-coast stations (Odesa, Herson and Mariupol) are
much colder than the south ones (Istanbul, Sinop and Trabzon), and the one eastern
(Batumi) is similar to the south ones. Burgas and Varna winter temperature on the
west coast is in between. Overall, the variations in the 9 stations are very similar
and in phase, suggesting that the winter conditions are uniform in the entire area.
Interesting fact to notice also is the slight tendency for milder winter.
136
Black Sea Freezing and Winter Severity: Relation...

In order to categorize how cold was the winter, we have followed Simonov and
Altman (1991) approach: we define a Winter Severity Index (WSI) as the sum of
negative daily temperature values (frozen days) with opposite sign for the months
December to March for each of the 9 stations. Figure 4 show that in Odesa, Mariupol
and Herson every winter the temperature falls regularly below zero, on the contrary
- in Istanbul, Sinop and Trabzon very rare. Burgas and Varna on the west coast are
similar to the north but Batumi on the east is like the southern stations. The most
indicative for the winter intensity are the north cities and thus in the further analysis
we have considered these three locations.

Figure 4. Winter severity index for Burgas, Varna, Bartumi, Odesa, Herson, Mari-
upol, Trabzon, Sinop and Istanbul. The grey columns in the background refer to
the data availability, the periods with no or insufficient data are not taken in the
analysis.

WINTER INTENSITY IN THE PERIOD 1925-2020.


The curves in Figure 5 represent the WSI variations for the historical data (Si-
monov and Altman 1991) and our calculated WSI from the synoptic measurements
in Odesa, Herson and Mariupol for the period 1926-2020. It is seen that the curves
evolve very similar. Mariupol is the northern-most city and as consequence the
winter severity is most pronounced there. Odesa and Herson are similar in latitude
but the west position of Odessa is the reason for less values of WSI. Only severe
winter cause significant WSI along the southern coast (Figure 4).
Such winter was for example the 1984/1985, which is similar to the coldest
from the record . In the past decade the relatively cold winters were the 2002/2003,
137
Mirna Matov, Elisaveta Peneva, Greta Georgieva

2005/2006, 2011/2012 and


2016/2017. An extremely
cold winter was 1953/1954.
The analysis shows that
the “cold” weather in win-
ter decreases from north to
south and from west to east:
the freezing days are almost
absent in the Trabzon sta-
tion. Another interesting
observation is the negative
trend of WSI after 2000. Figure 5. WSI of the historical data for
Also, the peaks in the last the northern coast of the Black Sea and
three decades generally de- in Odesa, Herson and Mariupol.
crease in height. This defin-
itively reflects the warming trend in this period.

SEA ICE COVER VARIABILITY.


Data for the ice extent are taken from the US National Snow and Ice Data Cen-
ter product Multi-sensor Analyzed Sea Ice Extent - Northern Hemisphere (MA-
SIE-NH). This provides daily sea ice data from January 1st, 2006 to the present.
The satellites used for this research are ALOS, AQUA, DMSP, ENVISAT, GOES,
MSG, and RADARSAT-2. Parameters measured by these sensors include sea ice,
ice edges, ice extent, and ice growth/melt. Spatial coverage in the study is the entire
Northern Hemisphere specified as N 90° to N 0° and E 180° to W -180°. Raster
data pixels (grid cells) are 4 km x 4 km thus 16 km2 each (MASIE, Technical Ref-
erences).
The available geotiff image in polar stereographic projection (Figure 6) is pre-
processed to Mercator projection and using the mask for the Black Sea pixels (Fig-
ure 6), only the “sea” pixels with ice are counted. Thus the average and maximal
sea ice extent for the four months in the winter season (December to March) are
calculated. One can note that the mask contains also the major rivers estuaries.

138
Black Sea Freezing and Winter Severity: Relation...

Figure 6.Figure illustrating the Figure 7. Mask for the Black


MASIE-NH product; The geotiff image sea pixels
is processed to obtain Mercator projection
and only the Black Sea area is cut.

The results for the daily sea-ice cover in the winters of 2006-2020 period are
shown in Figure 8. It is obvious that freezing occurs every year with exception of
the winters 2006/2007 and latest one 2019/2020. Maximal ice extent is reached in
February 2012 but for relatively short period. The ice coverage in 2006 and 2010
for example is less than that in 2012, but stays for a longer period of time.

Figure 8. Daily sea ice extent variability in the period 2006-2020


from MASIE-NH.

Figure 9 shows a combined plot of historical data from 1950-1985 for sea ice
extend (Simonov and Altman, 1991) and our calculations from MAISIE-NH in
the period 2006-2020, processed as described above. Nevertheless global warming
trend, it is clear that the freezing occurs rather regularly which leads to the conclu-
sion that it is result from synoptic situations rather than the seasonal averages.
139
Mirna Matov, Elisaveta Peneva, Greta Georgieva

Figure 9. Combined historical data from 1950-1985 from sea ice extent
and data from MAISIE-NH in the period 2006-2020

SEA SURFACE TEMPERATURE VARIATION


The available data from synoptic daily measurements of the sea surface
temperature (SST) in Burgas and Varna meteorological stations allow us to calculate
the winter months mean SST for the period 2000-2020 (Figure 10). The winter
months are taken from January to March. The two curves show rather similar in-
phase behavior: in Varna the winter SST is generally lower but not always. The
comparison of SST with the sea-ice cover together with the winter air temperature
and WSI variation reveals that the different sources of data confirm the anomalous
cold and warm years: the winters in 2003, 2012 and 2012 were especially cold and
on the contrary, the winters in 2001, 2007 and 2020 were very warm.

WINTER CLASSIFICATION
The variation in the ice extent is in agreement with the variations of the WSI and SST.
This allows us to make a classification of the winter in the Black Sea region using the
sea-ice extent and the threshold values for the winter intensity as defined in the Table 1.
In Table 2 the winters in the period from 1926 to 2020 are listed and the type according to
the classification in Table 1 is given. Note that since 2003 no “cold” winter conditions are
observed, and the relative colder moderate winters are classified as Moderate*.
From the winter classification it could be noted that “cold winter” conditions
usually last only one season and the following winter is either mild or moderate.
Rarely there are two cold winters in sequence. In this region the very cold synoptic
situations are due to the Siberian High expanding to the west and reaching the Cen-
140
Black Sea Freezing and Winter Severity: Relation...

Figure 10. Winter mean SST measured near the coast in the meteorological
stations in Varna and Burgas.

tral Europe, thus the next step will consider variations of the Siberian High in order
to understand the climate feedback mechanism in the Black Sea region.

Table 1. Winter classification


WSI value Winter type

[0, 200] Mild


[200, 400] Moderate
[400, -] Cold

Table 2. Winter type from 1926 to 2020


Year Winter type Year Winter type Year Winter type
1926 Moderate 1960 Moderate 1991 Moderate
1927 Moderate 1961 Mild 1992 Moderate
1928 Cold 1962 Mild 1993 Moderate
1929 Cold 1963 Cold 1994 Mild
1930 Mild 1964 Moderate 1995 Moderate
1931 Moderate 1965 Moderate 1996 Cold
1932 Cold 1966 Mild 1997 Cold
1933 Moderate 1967 Moderate 1998 Mild
1934 Moderate 1968 Moderate 1999 Mild
1935 Cold 1969 Moderate 2000 Mild
1936 Mild 1970 Mild 2001 Mild
1937 Moderate 1971 Mild 2002 Moderate
1938 Mild 1972 Moderate 2003 Cold
1939 Mild 1973 Moderate 2004 Mild

141
Mirna Matov, Elisaveta Peneva, Greta Georgieva

1940 Moderate 1974 Mild 2005


1941 Moderate 1975 Mild 2006 Mild
1945 Moderate 1976 Moderate 2007 Moderate*
1946 Moderate 1977 Mild 2008 Mild
1947 Cold 1978 Moderate 2009 Mild
1948 Mild 1979 Moderate 2010 Mild
1949 Moderate 1980 Moderate 2011 Moderate
1950 Moderate 1981 Mild 2012 Moderate
1951 Mild 1982 Mild 2013 Moderate*
1952 Mild 1983 Mild 2014 Mild
1953 Mild 1984 Mild 2015 Mild
1954 Cold 1985 Cold 2016 Mild
1955 Mild 1986 Moderate 2017 Mild
1956 Moderate 1987 Cold 2018 Moderate
1957 Moderate 1988 Moderate 2019 Mild
1958 Mild 1989 Mild 2020 Mild
1959 Mild 1990 Mild Mild

CONCLUSIONS AND FUTURE PLANS


We have used synoptic observations from 9 coastal Black Sea meteorological
stations in order to analyse the interannual-to decadal variations in the mean winter
temperature . Our study shows that the variations are very similar and in phase,
suggesting that the winter conditions are uniform in the entire area. There is a slight
tendency for milder winter in the last 4 decades in all 9 stations.
Along the north coast (Odesa, Mariupol and Herson) every winter the temperature
falls below zero, and along the south coast (Istanbul, Sinop, Trabzon and Batumi)
only the severe winters cause negative temperatures. For example - 1985/1986 was
extremely cold and in the past 20 years there are 4 very cold winters - 2002/2003,
2005/2006, 2011/2012 and 2016/2017.
The sum of daily negative temperatures with opposite sign for the winter season
is used to define the winter intensity and is Winter Severity Index (WSI). It presents
a negative trend after 2000 reflecting the global warming tendency.
Satellite observations are used to estimate the sea ice cover area and it is found
that in the north part freezing is observed rather regularly. The winters in 2007 and
2020 are the unique with no ice observed, and maximal area covered with ice is
reached in February 2012 but for relatively short period. The ice cover in 2006 and
2010 for example is less than that in 2012, but stays for a longer period of time.
The conclusion is that the freezing results from extremely cold synoptic situations
rather than mean seasonal extreme.
The variation of the winter season ice extent is in agreement with the variation
of the WSI, estimated from the temperature, as well as the measurements of SST in
Burgas and Varna. These data time-series were used to classify the winters’ severity
in the period 1926-2020 as cold, moderate and mild.
142
Black Sea Freezing and Winter Severity: Relation...

Since 2003 no real “cold” winter conditions are observed, but 2006, 2012 and
2017 are cool. On the other “warm” end are the winters in 2001, 2007 and 2020.
The winter 2019/2020 is anomalous warm; the mildest from the analysed period.
Future analysis is foreseen to investigate the impact of the Siberian High, as well as
the “buffer” role of the Black Sea.

ACKNOWLEDGMENTS
This work has been carried out in the framework of the National Science Program
"Environmental Protection and Reduction of Risks of Adverse Events and
Natural Disasters", approved by the Resolution of the Council of Ministers №
577/17.08.2018 and supported by the Ministry of Education and Science (MES)
of Bulgaria (Agreement № Д01-322/18.12.2019).

REFERENCES
National Ice Center and National Snow and Ice Data Center. Compiled by
F. Fetterer, M. Savoie, S. Helfrich, and P. Clemente-Colón. 2010, up-
dated daily. Multisensor Analyzed Sea Ice Extent - Northern Hemisphere
(MASIE-NH), Version 1. Boulder, Colorado USA. NSIDC: National
Snow and Ice Data Center. doi: https://doi.org/10.7265/N5GT5K3K.
[May 6th 2019].
Simonov A. and Altman E (eds) (1991), Hydrometeorology and Hydro-
chemistry of the USSR Seas. Vol. IV. Black Sea. Issue 1. St. Petersburg,
Gidrometeoizdat (in Russian)

 Mirna Matov
http://orcid.org/0000-0002-9900-0638
Faculty of Physics
“St. Kliment Ohridski” University of Sofia
Sofia, Bulgaria
E-mail: [email protected]

 Elisaveta Peneva
http://orcid.org/0000-0003-1325-685X
Faculty of Physics
“St. Kliment Ohridski” University of Sofia
Sofia, Bulgaria
E-mail: [email protected]

 Greta Georgieva
Faculty of Physics
“St. Kliment Ohridski” University of Sofia
Sofia, Bulgaria
E-mail: [email protected]

143
Proceeding of 1st International conference on Environmental protection and disaster RISKs

https://doi.org/10.48365/envr-2020.1.13 Air Pollution,


Climate and Health

ASSESSMENT OF THE FUTURE CLIMATE


OVER SOUTHEAST EUROPE BASED
ON ENSEMBLE OF CLIMATE INDICES –
PART ONE: CONCEPT AND METHODS

Hristo Chervenkov1, Vladimir Ivanov2,


Georgi Gadzhev2, Kostadin Ganev2
1
National Institute of Meteorology and Hydrology (NIMH)
2
National Institute of Geophysics, Geodesy and Geography –
Bulgarian Academy of Sciences (NIGGG-BAS)

Abstract: Nowadays there is a strong degree of agreement that the climate


change is the defining challenge of our time. The analysis based on climate indices
is probably the most widely used non-parametric approach for quantification of
extreme climate events. This study which consist of two parts, is dedicated on the
assessment of the spatial patterns and the temporal evolution of 6 temperature-
based and 3-precipitation based indices in projected future climate over south-
east Europe up to the end of the century. The annual means of the daily mean
and extreme temperatures are also analysed in consistent manner. The indices are
computed from the bias-corrected output of 5 CMIP5 global models, forced with all
4 RCP emission scenarios. The multi model ensemble medians of the temperature-
based indices shows considerable warming which is consistent with the increase of
the mean temperatures and is statistically significant in most cases. The revealed
changes of the precipitation-based indices are more complex when compared with
temperature changes.
Keywords: climate indices, CMIP5 ensemble, RCPs, future climate, South-east
Europe

INTRODUCTION
There is a strong degree of agreement that the climate change is the defining
challenge of our time. It will exert influence on the ecosystems, on all branches of
the international economy, and on the quality of life. The globally averaged surface
temperature of the Earth increased 0.85°C over the 1880 to 2012 period. It is
extremely likely that the observed warming of the climate system was caused by
the increased anthropogenic emission of greenhouse gases (IPCC, 2013). However,
immediate damages to humans and their properties as well as to ecosystems are not
144
Assessment of the Future Climate Over Southeast Europe...

obviously caused by gradual changes in temperature or precipitation but mainly by


so-called extreme climate events (Sillmann & Röckner, 2007). The rare occurrence of
extremes makes it necessary to investigate long data records to determine significant
changes in their frequency and intensity. To this end, global circulation models
(GCMs) or, more generally, coupled atmosphere-ocean general circulation models
(CAOGCM) are physically consistent way to simulate past, present, and future
climate states inclusive extreme events. Regional Climate Models (RCM) applied
with higher spatial resolution over a limited area and driven by GCMs can provide
more appropriate information on such smaller scales supporting more detailed impact
and adaptation assessment and planning (Rummukainen, M., 2010; Belda et al.,
2015; Giorgi & Gutowski, 2015). Therefore RCMs have an important role to play
by providing projections with much greater detail and more accurate representation
of localized extreme events (Gadzhev et al., 2021). The Working Group on Coupled
Modelling (WGCM) established the Coupled Model Intercomparison Project (CMIP)
as a standard experimental protocol for studying the output of CAOGCMs. CMIP
provides a community-based infrastructure in support of climate model diagnosis,
validation, intercomparison, documentation and data access (Taylor et al., 2009). The
main aim of the fifth phase of CMIP, CMIP5, is to study the climate and climate
change in the past, present and future, using a set of simulations with different climate
simulators in various spatial and temporal scales (Taylor et al., 2012).
The Mediterranean region lies in a transition zone between the arid climate of
North Africa and the temperate and rainy climate of central Europe and it is affected
by interactions between mid-latitude and tropical processes. Because of these
features, even relatively minor modifications of the general circulation, e.g. shifts
in the location of mid-latitude storm tracks or sub-tropical high pressure cells, can
lead to substantial changes in the Mediterranean climate (Giorgi & Lionello, 2008).
In addition to planetary scale processes and teleconnections, the climate of the
Mediterranean is affected by local processes induced by the complex physiography
of the region and the presence of a large body of water (the Mediterranean Sea).
Several projects and, consequently, many studies are dedicated in the recent
decades on the climate projections over Europe and the Mediterranean basin. The
CMIP5 projections generally agree on warming in all seasons in Europe during this
century, while precipitation projections are more variable across different parts of
Europe and seasons (Belda et al., 2015). Despite of the overall agreement for general
reduction of the precipitation amount in the middle and at the end of the 21th century,
there are still many differences in the magnitude of the expected changes, annual
and seasonal variability and areal distributions (Dai, 2013; Giorgi & Lionello, 2008;
Orlowsky & Seneviratne, 2012; Ulbrich et al., 2006). Part of the problems could be
attributed to the models deficiencies of the precipitation simulation. The model es-
timations includes also non-negligible uncertainties which results, in particular, in a
less spatially coherent pattern of change, bigger inter-model spread and a lower lev-
145
Hristo Chervenkov, Vladimir Ivanov, Georgi Gadzhev, Kostadin Ganev

el of statistical significance when compared with temperature changes (Alexander


et al., 2006). In Sillmann et al. (2013a) is demonstrated that GCMs underestimate
observed precipitation magnitudes, although CMIP5 models show an improvement
compared to CMIP3. The results of Orlowsky & Seneviratne (2012) show that de-
spite the uncertainty in other regions, droughts have increased in the Mediterranean
and are projected to increase further, emphasizing the need for proactive adapta-
tion planning. Central and Eastern Europe is a region where precipitation changes
remain also still uncertain (Belda et al., 2015). Although regional climate change
amplitudes of temperature and precipitation in Europe follow global trends, they
can be also affected by changes in the large-scale circulation and regional feedback
processes (Kjellstrom et al., 2011). In the comprehensive study of Stagge et al.
(2015) is used of the RCMs outcome from CORDEX (the Coordinated Regional
Climate Downscaling Experiment – see Kotlarski et al., 2014 for details), forced
with CMIP5 climate projections, to quantify the projected change in meteorological
drought for Europe during the 21th century, revealing increasing projected drought
throughout the Mediterranean, including the eastern Mediterranean. In agreement
with previous studies, in Sillmann & Röckner (2007) is evidenced a considerable
intensification of heat and water stress in the region.
The free worldwide exchange of methods, software and especially data is from
essential importance for the expert community of the geophysical sciences. Such
services are provided either from the primary vendors (institutions, organisations or
projects) or from single point access portals as the Copernicus Data Store (CDS).
They are reliable source for elaboration of objective climatologies, both in regional
(Birsan et al., 2014; Chervenkov et al., 2019; Chervenkov & Slavov, 2019; Spinoni
et al., 2018) and global (Sillmann et al., 2013a, 2013b; Orlowsky & Seneviratne,
2012) scale for recent and projected future climate as well as for development of
various custom-tailored applications. Our working group uses also this possibility
in optimal way. The present article, which is the first part of more common work,
describes shortly 5 collected and/or implemented by the authors data bases in our
general effort to describe the historical, near past and recent as well as the projected
future climate over south-east (SE) Europe most concisely and comprehensively.
The work is organized as follows. The CMIP5 scenarios and the used models
are described in Section 1. Section 2 is dedicated on the used methods and the con-
sidered data bases. Section 3 and Section 4 are the core of the present study. Due
to their importance, the first one, ‘Results for the mean temperatures’, is dedicated
solely on these variables. In the second one are described the results for the other
indices. The concluding remarks are in the last section.

1. CMIP5 SCENARIOS AND USED MODELS


The CMIP5 experiment uses new emission scenarios called representative
concentration pathways (RCP) (Moss et al., 2010) to assess the interactions
146
Assessment of the Future Climate Over Southeast Europe...

between the human activities on the one hand and the environment on the
other hand, and their evolution. Unlike the CMIP3 scenarios, the RCPs are
mitigation scenarios that assume policy actions will be taken to achieve certain
emission targets. For CMIP5, four RCPs have been formulated: RCP2.6,
RCP4.5, RCP6.0 and RCP8.5. They are based on a range of projections of
future population growth, technological development, and societal responses.
The labels for the RCPs provide a rough estimate of the radiative forcing
in the year 2100 (relative to pre-industrial conditions). RCP2.6 represents
mitigation scenarios that aim to limit the increase of global mean temperature
to <2°C. Different than other RCPs and earlier CMIP3 scenarios, RCP2.6
has a peak in greenhouse gases (GHG) concentration arround 2050 and then
declines at a moderate rate. Under RCP4.5, GHG-emissions will peak around
the early 2050s and then stabilize, causing a CO2 equivalent of about 650
parts per million and a temperature increase of approximately 1.8–2.0°C in
2100, compared to the control period of 1986–2005. RCP8.5, on the other
hand, predicts a continuous rise of GHG emissions until 2100, causing a CO2
equivalent larger than 1370 ppm and a global average temperature increase
close to 4°C (Spinoni et al., 2018).
The CMIP3 and CMIP5 model output are available from the data archives of
the Program for Climate Model Diagnosis and Intercomparison (PCMDI) and the
Earth System Grid data distribution portal (ESG). To generate the considered in this
study indices for historical and future time periods, bias-corrected climate datasets
provided through Inter Sectoral Impact Model Intercomparison Project (ISIMIP
1), Fast Track simulation round have been used. For each simulation round a set
of gridded bias-corrected climate variables have been produced to be used as input
data for running impact models. These climate datasets contain daily-resolution,
bias-corrected climate data from 5 CMIP5 GCMs according Table 1 covering the
period 1950-2099 (historical run up to 2005), downscaled to a 0.5°×0.5° lat-lon
grid. They cover the global land area.

Table 1. Main Features of the Considered Models


Spat. Resolution
Model Acronym Institution
(Lon×Lat~Lev.)
GFDL-ESM2M Geophysical Fluid Dynamics Laboratory, USA 144×90L24
HadGEM2-ES Met Office Hadley Centre, UK 192×145L40
IPSL-CM5A-LR Institut Pierre-Simon Laplace, France 96×96L39
MIROC-ESM-CHEM AORI, NIES, JAMSTEC, Japan 128×64L80(T42)
NorESM1-M Norwegian Climate Centre, Norway 144×96L26

Note that the models used in this study differ from the models in Chervenkov
& Slavov (2020a, 2020b, 2021). ISIMIP offers significantly fewer models, than the
147
Hristo Chervenkov, Vladimir Ivanov, Georgi Gadzhev, Kostadin Ganev

applied in Sillmann et al. (2013b) and Orlowsky & Seneviratne (2012) but these
studies considers only CMIP5 RCP2.6, RCP4.5 and RCP8.5 (i.e. not RCP6.0) and
CMIP3 SRES A2 scenario, correspondingly.

2. METHODS
There are various methods to characterize extreme events, but the computation
and analysis of climate indices (CIs) based on daily temperature and precipitation
data is probably the most widely used non-parametric approach (Sillmann
& Röckner, 2007). The modern sets of such indices, among which the most
widely used is the collection of the Expert Team on Climate Change Detection
and Indices (ETCCDI, Zhang et al., 2011), are statistically robust, cover a wide
range of climate conditions, and have a high signal-to-noise ratio (Alexander &
Arblaster, 2009). They are used in several projects on climate change with focus
on different spatial scales, from planetary to continental, regional, national or
local scale, as prevailing indicators of changes of the extreme events (Birsan
et al., 2014). Subsequently, the number of publications on this topic is very
large (Alexander et al., 2006; Frich et al., 2002; Kiktev et al., 2003; Klein Tank
& Können, 2003; Moberg et al., 2006 and many others). Our group has also
previous, partially project-driven, experience in CI-based analysis of historical
(Chervenkov & Slavov 2020a, 2020b), near past and present (Chervenkov et al.,
2019; Chervenkov & Slavov, 2019; Chervenkov & Slavov, 2021; Malcheva et
al., 2016) and projected future regional climate (Gadzhev et al., 2021). The free
availability of databases of Cis or other climate indicators, with focus on different
spatial and temporal scales, facilitates any assessment which includes these
parameters. In Table 2 are listed the main features of some gridded databases
recently used in our group. It is worth to emphasize that the information from
these sources is rarely suitable for direct implementation in the tasks of the
regional climatology. Thus, the data from these sources have to be essentially
post-processed in order to fit to the specific needs.

The present study is based entirely on the data from ISIMIP Fast Track. Although
this project is intended to be a collection of agroclimatic indicators datasets,
most of the indices in scope, inclusive all considered in this study, are based on
the ETCCDI definitions which makes them universal. Agricultural indicators in
ISIMIP Fast Track have been pre-calculated for this complete matrix of 5 GCMs×4
RCPs combinations. In addition, as a proxy for historical observations, the “Watch
Forcing Data methodology applied to ERA-Interim (WFDEI)” (Weedon et al.,
2014) were used to generate observational historical Agroclimatic indicators. This
dataset is available at the same spatial resolution of ISIMIP climate datasets, covers
the time range of 1979 to 2013 and its 30 year long part 1981-2010 is used in the
study as reference for the current climate.
148
Assessment of the Future Climate Over Southeast Europe...

Table 2. Mean Features of the Used Data Bases


Spat. Coverage/ Time Span, Basic Reference/
Acronym Main Content Institution
Resolution Scenario(s) Access
4 data sets of SPI-1, SPI-3,
1900-2010, 1901- Chervenkov et al.
SPI-6, SPI12 based on UDEL/
Global, 0.5°×0.5°; 2013, NIMH, (2016);
SPI DB GEOG/CCR v3.02, GPCC
1.5°×1.5° 1851-2011, Bulgaria ftp://xeo.cfd.meteo.
v7.0, NOAA-CIRES 20CR
1900-2010 bg/SPI/
v2c, ECMWF ERA20C
1961-1990
Central&SE Belda et al. (2015);
CECILIA 2021-2050 CECILIA
152 CI based on RCMs Europe, http://cecilia.
DB 2071-2100 SRES project
0.1°×0.1° dmi.dk
A1B
Chervenkov et al.
E-OBS&
(2019);
STARDEX&ETCCDI CARPATCLIM
1951-2016 NIMH, https://repo.
ClimData Cis based on E-OBS domains;
1961-2010 Bulgaria vi-seem.eu/
&CARPATCLIM 0.25°×0.25°,
handle/21.15102/
0.1°×0.1°
VISEEM-343.
Canadian Sillmann et al.
1850-2100; Centre for (2013a); http://
ETCCDI Cis based on
EIA Global, varios res. CMIP5 RCP2.6, Climate www.cccma.ec.gc.
CMIP5-GCMs
RCP4.5,RCP8.5 Modelling& ca/data/climdex/
Analysis climdex.shtml
1951-2099; CDS
ISIMIP 26 Cis based on CMIP5- CMIP5 RCP2.6, documentation;
Global, 0.5°×0.5° ISIMIP 1
Fast Track GCMs RCP4.5, RCP6.0, https://www.isimip.
RCP8.5 org/protocol/

After the download from the CDS the ISIMIP-datasets are significantly post-
processed. The most essential stages are:
● The datasets for each model and RCP which are downloadable in 30-years
time slices are merged in common data streams for 2011-2099
● The indices with equal temporal resolution are joined in common
netCDF4 files
● Multi-model (MM) ensemble quantities as multi-model mean (MMM), MM
25-, 50- and 75-percentile which are often refereed as lower quartile, median
and upper quartile and traditionally noted as X25, X50 and X75 are computed.
● Due to storage constrains only a spatial subset over Europe is preserved.
All netCDF manipulations are performed with the powerful tool Climate Data
Operators (cdo). Additionally, for the current study only, all of the considered indices
are aggregated in time on annual basis. The aggregation method depend on the
indicator, e.g. min, max, sum, mean. The magnitude of the trend in time as well as
its statistical significance are estimated individually for all grid cells and separately
for each scenario by means of the Theil-Sen slope estimator (TSE) and the Mann-
149
Hristo Chervenkov, Vladimir Ivanov, Georgi Gadzhev, Kostadin Ganev

Kendall (MK) test correspondingly. Thus far, the study is constrained over SE Europe
only as will be shown in the next section and with details in the next part.

3. RESULTS FOR THE MEAN TEMPERATURES


The annual means of the daily minimum, mean and maximum temperature
(noted traditionally TN, TG and TX) are important variables, providing information
on the current state as well as on the long-term climate variability and change and
thus are studied separately from the other considered indices. We analyse the spatial
patterns of the multiyear means of the CMIP5 projections of TN, TG and TX as
well as the spatial patterns for the reference period. In order to intercompare them,
the ensemble median (MMX50) of the listed in Table 1 models for all 4 scenarios is
superimposed to the median for the reference period, as shown on Fig. 1.

Figure 1. MMX50 of the multiyear means of the TN (first row)


TG (second row) and TX (third row) for the reference period (1981-2010)
in the first column and multiyear means for 2070-2099 for RCP2.6, RCP4.5,
RCP6.0 and RCP8.5 in the second, third, fourth and fifth column correspondingly.
The absolute changes of the RCP2.6, RCP4.5, RCP6.0 and RCP8.5 relative
to the reference period are shown in the sixth seventh eighth and ninth column
correspondingly. The units are oC.

Figure 1 shows for all variables gradual increase of the projected changes from
RCP2.6 to RCP8.5, i.e. proportional to the radiative forcing. The changes are simi-
lar in magnitude for all parameters for fixed scenario RCP2.6-RCP6.0 and have
not clear spatial structure. For RCP8.5 the changes for TG and TX are exceeding
6oC and are somewhat bigger than the changes for TN. All changes are statistically
significant at the 5% level.
The area-weighted regional averages over land of TN, TG and TX are shown
on Figure 2.
150
Assessment of the Future Climate Over Southeast Europe...

Figure 2. Area-weighted regional averages (index and unit according subplot


title) for the reference (solid black line) and simulated by the CMIP5 ensemble for
the RCP2.6 (blue), RCP4.5 (green), RCP6.0 (yellow) and RCP8.5
(red). Solid lines indicate the ensemble median (i.e. the 50th quantile)
and the shading, respectively the thin lines, indicates the interquartile
ensemble spread (25th and 75th quantiles).

Figure 2 shows relatively smooth (in comparison with the other Ci’s as will be
shown further) but steady increase of the temperatures with apparent difference
between scenarios in the second half of the century.
The importance of assessing trends in climate extremes is often emphasized
(e.g. Klein Tank & Können, 2003; Meehl et al., 2000; Moberg et al., 2006). The

Figure 3. Trend slopes (unit: oC/10 years) of the TN (first row) TG


(second row) and TX (third row) of the MMM (left pane) and MMX50
(right pane) for the scenarios according the subtitles in the first row.
151
Hristo Chervenkov, Vladimir Ivanov, Georgi Gadzhev, Kostadin Ganev

main reason is that extreme weather conditions related to temperature, precipitation,


storms or other aspects of climate, can cause loss of life, severe damage and large
economic and societal losses. Thus, the trend assessment is essential part of the
recent climate studies (Alexander et al., 2006; Frich et al., 2002; Sillmann et al.,
2013b; and many others). Figure 3 shows the slope of the estimated by means of
TSE linear trend of the MMM and MMX50 of the TN, TG and TX for the whole
future period 2011-2099.
The most apparent result of analysis of Figure 3 is that the fields of the slopes for
MMM and MMX50 are practically identical for each corresponding variable and
scenario. As expected, the lowest values are for the scenario with the modest forcing
(RCP2.6) and the highest – for the scenario with the strongest forcing (RCP8.5).

CONCLUSION TO PART ONE


The referenced in Table 2 projects and initiatives contributes to the availability
of a valuable sets of spatially and temporally representative data to prepare relevant
climate change studies in the corresponding domains. These datasets are reliable
sources of various climate indicators, which can be presented as continuous, both in
space and time, digital maps. The positive consequences to the end user community
in exploration of the single point access data portal Copernicus Climate Data Store
are manifold, but the free access of data sets standardised file formats via unified
transfer protocols seems most significant.
Relevant outcome of the presented part one of the study, is the clearly expressed
warming signal in the field of the mean temperatures. It is spatially dominating over the
domain and everywhere statistically significant. The amount of warming by scenario
generally ranges from high to low as follows: RCP8.5, RCP6.0, RCP4.5 and RCP2.6
which shows principal proportion of the temperature increase to the radiative forcing.
Part two of this study is dedicated on the analysis of the spatial patterns and the
time evolution of the other considered indices.

ACKNOWLEDGMENTS
The authors thank the anonymous reviewers for their insightful suggestions. This
work has been carried out in the framework of the National Science Program
"Environmental Protection and Reduction of Risks of Adverse Events and
Natural Disasters", approved by the Resolution of the Council of Ministers №
577/17.08.2018 and supported by the Ministry of Education and Science (MES)
of Bulgaria (Agreement № Д01-322/18.12.2019).
This work has been accomplished with the financial support by the Grant No
BG05M2OP001-1.001-0003, financed by the Science and Education for Smart
Growth Operational Program (2014-2020) and co-financed by the European
Union through the European structural and Investment funds.

152
Assessment of the Future Climate Over Southeast Europe...

REFERENCES
Alexander, L. V., et al. (2006) Global observed changes in daily climate
extremes of temperature and precipitation, J. Geophys. Res., 111,
D05109, doi:10.1029/2005JD006290
Alexander, L. V. and Arblaster, J. M. (2009) Assessing trends in observed
and modelled climate extremes over Australia in relation to future
projections. Int. J. Climatol., 29: 417-435. doi:10.1002/joc.1730
Belda, M., Skalák, P., Farda, A., et al. (2015) CECILIA Regional Climate
Simulations for Future Climate: Analysis of Climate Change Signal,
Advances in Meteorology, vol. 2015, Article ID 354727, 13 pages, 2015
doi:10.1155/2015/354727
Birsan, M. V., Dumitrescu, A., Micu, D.M., Cheval, S. (2014) Changes in
annual temperature extremes in the carpathians since AD 1961. Natural
Hazards 74(3), 1899-1910 doi:10.1007/s11069-014-1290-5
Chervenkov, H., Tsonevsky, I., Slavov, K. (2016) Presentation of Four
Centennial-long Global Gridded Datasets of the Standardized Precipitation
Index, International Journal of Environmental & Agriculture Research
(IJOEAR) ISSN: 2454-1850 Vol. 2, Issue-3, March-2016 pp. 93-105
Chervenkov, H., Slavov, K., Ivanov, V., (2019) STARDEX and ETCCDI
Climate Indices Based on E-OBS and CARPATCLIM; Part One: General
Description, in G. Nikolov et al. (Eds.): NMA 2018, LNCS 11189, pp.
360-367, doi: 10.1007/978-3-030-10692-8 40
Chervenkov, H., Slavov, K. (2019) STARDEX and ETCCDI Climate
Indices Based on E-OBS and CARPATCLIM; Part Two: ClimData in
Use, in G. Nikolov et al. (Eds.): NMA 2018, LNCS 11189, 368-374,
(2019) DOI: 10.1007/978-3-030-10692-8 41
Chervenkov, H. Slavov, K. (2020a) Historical Climate Assessment of
Temperature-based ETCCDI Climate Indices Derived from CMIP5
Simulations, Compt. rend. Acad. bulg. Sci., 73, No 6, 784-790, DOI:
10.7546/CRABS.2020.06.05
Chervenkov, H. Slavov, K. (2020b) Historical Climate Assessment of
Precipitation-based ETCCDI Climate Indices Derived from CMIP5
Simulations, Compt. rend. Acad. bulg. Sci., 73, No 7, 942-948, DOI:
10.7546/CRABS.2020.07.06
Chervenkov H., Slavov K. (2021) ETCCDI Climate Indices for Assessment
of the Recent Climate over Southeast Europe. In: Dimov I., Fidanova S.
(eds) Advances in High Performance Computing. HPC 2019. Studies in
Computational Intelligence, vol 902. Springer, Cham. doi: 10.1007/978-
3-030-55347-0_34
Dai, A., (2013) Increasing drought under global warming in observations
and models, Nature Clim. Change, 3: 52-58

153
Hristo Chervenkov, Vladimir Ivanov, Georgi Gadzhev, Kostadin Ganev

Frich, P., Alexander, L.V., Della-Marta, P.M., Gleason, B.E., Haylock,


M.R., Tank, A.M., & Peterson, T.C. (2002). Observed coherent changes
in climatic extremes during the second half of the twentieth century.,
Climate Research, Vol. 19, No. 3, pp. 193-212, doi:10.3354/cr019193
Gadzhev G., Ivanov V., Valcheva R., Ganev K., Chervenkov H. (2021) HPC
Simulations of the Present and Projected Future Climate of the Balkan
Region. In: Dimov I., Fidanova S. (eds) Advances in High Performance
Computing. HPC 2019. Studies in Computational Intelligence, vol 902.
Springer, Cham. doi: 10.1007/978-3-030-55347-0_20
Giorgi, F., Gutowski, W. J. Jr. (2015) Regional Dynamical Downscaling and
the CORDEX Initiative, Annu. Rev. Environ. Resour. 2015. 40:467-490
Giorgi, F., Lionello, P. (2008) Climate change projections for the
Mediterranean region, Global and Planetary Change, 63(2-3): 90-104
IPCC (Intergovernmental Panel on Climate Change), 2013: The physical
science basis. In (eds: Stocker, T.F., Qin, D., Plattner, G-K., Tignor, M.,
Allen, S.K., Boschung, J., Nauels, A., Xia, Y., Bex, V., Midgley, P.M.,)
Contribution of Working Group I to the Fifth Assessment Report of the
Intergovernmental Panel on Climate Change. Cambridge University
Press, Cambridge.
Kiktev, D., Sexton, D. M. H. Alexander, L., Folland, C. K. (2003) Comparison
of Modeled and Observed Trends in Indices of Daily Climate Extremes,
Journal of Climate, Vol. 16, No. 22, pp. 3560-3571 doi:10.1175/1520-
0442(2003)016%3C3560:COMAOT%3E2.0.CO;2
Kjellstrom, E., Nikulin, G., Hansson, U.,Strandberg, G., Ullerstig, A.(2011)
21st century changes in the European climate: uncertainties derived from
an ensemble of regional climate model simulations, Tellus, Series A:
Dynamic Meteorology and Oceanography, vol. 63 No.1 pp24-40
Klein Tank, A.M. and G.P. Können, (2003) Trends in Indices of Daily
Temperature and Precipitation Extremes in Europe, 1946-99. J. Climate,
vol. 16, 3665-3680,doi: 10.1175/1520-0442(2003)016<3665:
TIIODT>2.0.CO;2
Kotlarski, S., K. Keuler, O. B. Christensen, A. Colette, M. Dèquè, A.
Gobiet, K. Goergen, D. Jacob, D. Lüthi, E. van Meijgaard, G. Nikulin,
C. Schär, C. Teichmann, R. Vautard, K. Warrach-Sagi, and V. Wulfmeyer
(2014) Regional climate modeling on European scales: a joint standard
evaluation of the EURO-CORDEX RCM ensemble. Geosci. Model
Dev., 7, 1297-1333 doi:10.5194/gmd-7-1297-2014
Malcheva, K., Chervenkov, H., Marinova, T., (2016) Winter Severity
Assessment on the Basis of Measured and Reanalysis data. In: 16th
International Multidisciplinary Scientific GeoConference SGEM 2016.
SGEM, pp. 719-726

154
Assessment of the Future Climate Over Southeast Europe...

Meehl, G. A., Zwiers, F., Evans, J., Knutson, T., Mearns, L., & Whetton, P.
(2000). Trends in extreme weather and climate events: Issues related to
modeling extremes in projections of future climate change. Bulletin of
the American Meteorological Society, 81(3), 427-436. doi:10.1175/1520-
0477(2000)081<0427:TIEWAC>2.3.CO;2
Moberg, A., et al. (2006) Indices for daily temperature and precipitation
extremes in Europe analyzed for the period 1901–2000, J. Geophys.
Res., 111, D22106, doi:10.1029/2006JD007103.
Moss, R. H. et al. (2010) The next generation of scenarios for climate
change research and assessment, Nature, Vol. 463, No. 7282, 747-756,
doi:10.1038/nature08823
Orlowsky, B. and Seneviratne, S. I. (2012) Elusive drought: uncertainty in
observed trends and short- and long-term CMIP5 projections, Hydrol.
Earth Syst. Sci. 17(5):1765-1781 doi:10.3929/ethz-b-000073994.
Rummukainen, M. (2010) State-of-the-art with regional climate models.
WIREs: Clim. Change 1:82-96
Sillmann, J., Röckner, E., (2007) Indices for extreme events in projections
of anthropogenic climate change. Climatic Change 86 (1-2), pp. 83-104.
Sillmann, J., V. V. Kharin, X. Zhang, F. W. Zwiers, and D. Bronaugh
(2013a) Climate extremes indices in the CMIP5 multimodel ensemble:
Part 1. Model evaluation in the present climate, J. Geophys. Res. Atmos.,
118, pp. 1716-1733,doi: 10.1002/jgrd.50203
Sillmann, J., V. V. Kharin, F. W. Zwiers, X. Zhang, and D. Bronaugh (2013b)
Climate extremes indices in the CMIP5 multimodel ensemble: Part 2.
Future climate projections, J. Geophys. Res. Atmos., 118, pp. 2473-2493
doi: 10.1002/jgrd.50188.
Spinoni, J., Vogt, J.V., Barbosa, P., Dosio, A., McCormick, N., Bigano, A.
and Füssel, H.‐M., (2018) Changes of heating and cooling degree‐days in
Europe from 1981 to 2100. Int. J. Climatol, 38: e191-e208. doi:10.1002/
joc.5362
Stagge, J.H., Rizzi, J., Tallaksen, L. M., and Stahl, K. (2015) Future
meteorological drought: projections of regional climate models for
Europe, Technical Report No. 25, Future Meteorological Drought
Projections of Regional Climate, DROUGHT-RSPI Project, University
of Oslo, Oslo, 23 pp.
Taylor, K. E., Stouffer, R. J., Meehl, G. A. (2009) A summary of the CMIP5
experiment design, PCDMI Rep., 33 pp., 2009 [Available online at http://
cmip-pcmdi.llnl.gov/cmip5/docs/Taylor_CMIP5_design.pdf.]
Taylor, K. E., Stouffer, R. J., Meehl, G. A., (2012) An Overview of CMIP5
and the Experiment Design Bull. Amer. Meteor. Soc., 93, 485-498, 2012
doi:10.1175/BAMS-D-11-00094.1

155
Hristo Chervenkov, Vladimir Ivanov, Georgi Gadzhev, Kostadin Ganev

Ulbrich, U., et al., (2006) The Mediterranean climate change under global
warming. In: Lionello, P., Malanotte-Rizzoli, P., Boscolo, R. (Eds.), R.
(Eds.), Mediterranean Climate Variability. Elsevier, Amsterdam, pp.
398-415.
Weedon, G.P., Balsamo, G., Bellouin, N., Gomes, S., Best, M.J. and
Viterbo, P., (2014) The WFDEI meteorological forcing data set: WATCH
Forcing Data methodology applied to ERA‐Interim reanalysis data.
Water Resources Research, 50(9), pp.7505-7514.
Zhang, X., Alexander, L., Hegerl, G.C., Jones, P., Tank, A.K., Peterson,
T.C., Trewin, B. and Zwiers, F.W. (2011), Indices for monitoring changes
in extremes based on daily temperature and precipitation data. WIREs
Clim Change, 2: 851-870. doi:10.1002/wcc.147

 Hristo Chervenkov
http://orcid.org/0000-0002-7658-3041
National Institute of Meteorology and Hydrology
Sofia, Bulgaria
E-mail: [email protected]

 Vladimir Ivanov
https://orcid.org/0000-0001-9768-1049
National Institute of Geophysics, Geodesy and Geography
Bulgarian Academy of Sciences
Sofia, Bulgaria
E-mail: [email protected]

 Georgi Gadzhev
http://orcid.org/0000-0002-6159-3554
National Institute of Geophysics, Geodesy and Geography
Bulgarian Academy of Sciences
Sofia, Bulgaria
E-mail: [email protected]

 Kostadin Ganev
National Institute of Geophysics, Geodesy and Geography
Bulgarian Academy of Sciences
Sofia, Bulgaria
E-mail: [email protected]

156
Proceeding of 1st International conference on Environmental protection and disaster RISKs

https://doi.org/10.48365/envr-2020.1.14 Air Pollution,


Climate and Health

ASSESSMENT OF THE FUTURE CLIMATE


OVER SOUTHEAST EUROPE BASED
ON ENSEMBLE OF CLIMATE INDICES –
PART TWO: RESULTS AND DISCUSSION

Hristo Chervenkov1, Vladimir Ivanov2,


Georgi Gadzhev2, Kostadin Ganev2
1
National Institute of Meteorology and Hydrology (NIMH)
2
National Institute of Geophysics, Geodesy and Geography –
Bulgarian Academy of Sciences (NIGGG-BAS)

Abstract: Nowadays there is a strong degree of agreement that the climate


change is the defining challenge of our time. The analysis based on climate indices
is probably the most widely used non-parametric approach for quantification of
extreme climate events. This study which consist of two parts, is dedicated on the
assessment of the spatial patterns and the temporal evolution of 6 temperature-
based and 3-precipitation based indices in projected future climate over south-
east Europe up to the end of the century. The annual means of the daily mean
and extreme temperatures are also analysed in consistent manner. The indices are
computed from the bias-corrected output of 5 CMIP5 global models, forced with all
4 RCP emission scenarios. The multi model ensemble medians of the temperature-
based indices shows considerable warming which is consistent with the increase of
the mean temperatures and is statistically significant in most cases. The revealed
changes of the precipitation-based indices are more complex when compared with
temperature changes.
Keywords: climate indices, CMIP5 ensemble, RCPs, future climate, South-east
Europe

INTRODUCTION TO PART TWO


This part is dedicated on the analysis of the considered indices except the
mean temperatures. The study is continuation of our scientific work, documented
in suite of publications (Chervenkov et al., 2019; Chervenkov & Slavov, 2019,
2020a, 2020b, 2020c, 2021) which first part is described in Chervenkov et al.,
(2021). As in Chervenkov & Slavov, (2021), the present study is inspired from
the comprehensive study of Sillman et al., (2013b) and fits in the same conceptual
framework. Beside the different models, considered indices and significantly finer
157
Hristo Chervenkov, Vladimir Ivanov, Georgi Gadzhev, Kostadin Ganev

grid spacing, this study differs, however, from Chervenkov & Slavov (2021) in other
two substantial aspects: first and foremost, the input data (i.e. daily temperatures
and precipitation sums) are bias-corrected, as described by Hempel et al., (2013),
prior the computation of the indices. Although some criticism exists, the general
view in the expert community is that the bias-corrected climate change signal is
more reliable compared with the uncorrected one and thus is more suitable for
impact assessments (Chervenkov & Spiridonov 2020, 2021). Second, the study is
focused on the multi-model statistics (mean, MMX25, MMX50, MMX75) rather
than the simulation output of the individual models. This is modern common
approach, adopted in many recent studies (Orlowsky & Seneviratne, 2012; Sillman
et al., 2013a, 2013b).

4. RESULTS
4.1. TEMPERATURE INDICES
4.1.1. ABSOLUTE AND THRESHOLD INDICES
We start our analysis with a comparison of the spatial patterns of the multiyear
means of the CMIP5 projections of the extreme temperatures, TNn and TXx with
their counterparts for the reference period. In order to intercompare them, the
ensemble median (MMX50) for all 4 scenarios is superimposed to the median for
the reference period, as shown on Fig. 1. for the annual extreme temperatures

Figure 1. MMX50 of the multiyear means of the TNn (first row) and TXx
(second row) for the reference period (1981-2010) in the first column
and multiyear means for 2070-2099 for RCP2.6, RCP4.5, RCP6.0 and RCP8.5
in the second, third, fourth and fifth column correspondingly. The absolute
changes of the RCP2.6, RCP4.5, RCP6.0 and RCP8.5 relative to the reference
period are shown in the sixth seventh eighth and ninth column correspondingly.
Stippling indicates grid points with changes that are not significant
at the 5% significance level. The units are oC.

Figure 1 shows gradual increase of the projected changes of the both indices
from RCP2.6 to RCP8.5, i.e. proportional to the radiative forcing. The greatest
158
Assessment of the Future Climate Over Southeast Europe Based...

changes in TNn, exceeding 9oC, are simulated in RCP8.5 over the northern
half of the domain. The most apparent difference between the spatial patterns
of the changes of the TNn and TXx is the stronger temperature increase for
TNn. This difference is well expressed for all four scenarios. Alexander et al.,
(2006) documents analysis of global data base of historical records, revealing
such asymmetric warming. According to their global study (for data since 1951),
changes in daily maximum temperatures are less marked, implying that our world
in many places has become less cold rather than hotter. Other studies, however,
based solely on assimilated European data (see Moberg et al., 2006 and references
therein) suggest that this conclusion is not representative for Europe if the entire
twentieth century is considered. An overall warming is observed also in Moberg
et al., (2006), but they find only a small difference, or no difference at all, between
average trends in daily minimum and maximum temperatures when they average
trends for 75 stations across Europe. According to the future in the CMIP5
projections, Sillmann et al., (2013b) outlines the differences in the changes of
the daily extreme temperatures. In particular, TNn increases more strongly in
higher latitudes of the Northern Hemisphere. It is worth to emphasize also that
our previous study Chervenkov & Slavov (2021), which is conceptually similar
to the present one as underlined above, do not detects clear enough and rigour
evidences for ’warming asymmetry’ between the indices, based on the minimum
temperature from the one hand and these, based on maximum temperature, from
the other hand.
The widely used threshold indices tropical nights (TR) and frost days (FD),
both based on the daily minimum temperature, have limited applicability over the
domain. In the climatological study of the Carpathian region Birsan et al., (2014) is
stated that the changes in the occurrence of TR are substantial only in low-elevation
areas (below 800 m.), located outside the Carpathian Mountains range, which are
particularly exposed to persistent and intense warm spells in summer. Generally, TR
are not characteristic to the climate of the mountain regions, which are significant

Figure 2. Same as Figure 1 but for the FD (first row)


and TR (second row). The units are days.
159
Hristo Chervenkov, Vladimir Ivanov, Georgi Gadzhev, Kostadin Ganev

part of the domain. Similarly, the FD are not very meaningful for maritime climate
– this is valid for the southern half of the model region and especially for the areas
along the fragmented coastline. Nevertheless, at least for methodological reasons,
analysis of these indices have to be performed as shown on Figure 2 in the case for
their spatial patterns.
Consistent with the changes of the minimum temperature, the fields of the
threshold indices shows progressive (i.e. from RCP2.6 to RCP8.5) decrease of FD
and, contrary to FD, increase of the TR. The vertical gradient of the FD is well
expressed especially along the main Carpathian ridge. The increase of the TR under
RCP2.6, RCP4.5 even RCP6.0 is, over the bigger part of the domain, relatively
small and, which is more important, statistically not significant.
The area-weighted regional averages over land of TNn, TXx, FD and TR, which
will be called area-averages (AA) for sake of brevity henceforth, are shown on
Figure 3.

Figure 3. AA of the considered absolute and threshold indices


(index and unit according subplot title) for the reference (solid black line)
and simulated by the CMIP5 ensemble for the RCP2.6 (blue), RCP4.5 (green),
RCP6.0 (yellow) and RCP8.5 (red). Solid lines indicate the ensemble median (i.e.
the 50th quantile) and the shading, respectively the thin lines, indicates
the interquartile ensemble spread (25th and 75th quantiles).

The overall tendencies, commented above, are markedly expressed for AAs
of all indices. In the recent study Chervenkov & Slavov, (2020a), among other
problems, are analysed the trends of five temperature-based indices, including TNn
and TXx. The trend estimation of the AAs over the domain is based on the gridded
160
Assessment of the Future Climate Over Southeast Europe Based...

data base of the HadEX2 project and is for the period 1900-2010. The study reveals
statistically significant increasing trend for the both parameters with bigger lapse
for TNn. These conclusions agrees with the present results, suggesting that the
projected changes are natural continuation of the already detected changes in the
near past and present climate. The analysis of Figure 3 shows also that the CMIP5
interquartile model spreads in the four RCPs practicality remain overlapping for
TXx, FD and TR until the middle and for TNn until the end of the 21st century.

4.1.2. DURATION AND PERCENTILE INDICES


The Cold and Warm Spell Duration Indices (CSDI & WSDI) are most
frequently used indicators for cold and heat waves respectively (Alexander et al.,
2006; Alexander & Arblaster, 2009; Sillmann &, Röckner, 2007 and many others)
inclusive in climatological studies of the considered region (Birsan et al., 2014).
Hence these indices are calculated using a percentile based, rather than fixed value,
threshold, they could be also considered as percentile indices. The spatial patterns
of the CSDI and WSDI for the control period as well as for the projected with the
CMIP5 future are shown on Figure 4.

Figure 4. Same as Figure 1 but for the CSDI (first row) and WSDI
(second row). The units are days.

Regarding the CSDI, the most obvious result on Figure 4 is that the small on
magnitude and uniformly distributed during the reference period index practically
disappears in the future even under the scenario with the weakest radiative forcing
(RCP2.6). Subsequently, CSDI remains near zero constant under the other three
scenarios. The change of the WSDI is very expressive both in magnitude and spatial
extent. The differences from RCP to RCP are significant, especially for RCP8.5
compared with others. The absolute increase of WSDI relative to the reference is
drastic: more than three months practically over the whole domain. This result,
rather embarrassing indeed, is in principal agreement with Sillmann et al., (2013b).
The temporal evolution of the considered duration indices is depicted on Figure 5.
161
Hristo Chervenkov, Vladimir Ivanov, Georgi Gadzhev, Kostadin Ganev

Figure 5. Same as Figure 2 but for the CSDI (first row) and WSDI
(second row). The units are days.

Figure 5 confirms the outcomes from the analysis of the spatial patterns of the
CSDI and WSDI. Interesting specifics of the dynamics of the AA of the CSDI is
considerable changes in both directions for relatively short time during the reference
period. This is consequence of single abnormally hot and cold years in this time
span documented also in Birsan et al., (2014). The AA of the WSDI during the
whole reference period remains with practically negligible values almost constant.
In contrast, the dynamics in the projected future demonstrates steady increase,
especially for RCP8.5. Consequently, the ensemble median for this scenario is over
150 days around the 2080’s.

4.2. PRECIPITATION INDICES


As in many other places of the world, in contrast to the projected changes in
the temperature indices, where there is a general agreement on the sign of change
independent of the region considered, changes in the precipitation indices over the
considered region are less consistent in this regard Sillmann et al., (2013b), Cher-
venkov et al., (2021).
Our analysis is focused on the indices Annual Precipitation Sum, Heavy Precipi-
tation Days and Consecutive Dry Days, noted RR, RR10mm and CDD correspond-
ingly. These indices are used as key parameters in many studies of present (for
example Sillmann and Röckner, 2007) and projected future climate (see Sillmann
et al., 2013b and citation therein).
Figure 6 provides a more detailed regional picture of the spatial patterns of
the considered precipitation indices for the control period as well as for the future
projections. Traditionally, the changes in precipitation sum RR relative to the 1981-
2010 reference period are expressed in percentage terms.
First and foremost, Figure 6 demonstrates the complex nature of the expected
precipitation changes. Although the total precipitation amount (the first row on
Figure 6) shows clear reduction tendency, especially over the southeastern part
of the domain, there is no big difference, both in magnitude and spatial distribu-
tions, in the relative changes in the scenarios RCP2.6-RCP6.0. Second, which
162
Assessment of the Future Climate Over Southeast Europe Based...

Figure 6. Same as Figure 1 but for the RR (first row), RR10mm


(second row) and CDD (third row). Relative instead of absolute changes of the
RR are considered. The units of the RR are mm and of the RR10mm and CDD as
well as their changes – days. The relative changes of the RR are expressed in %.

is most important, these changes are not statistically significant at the 5% level.
This result agrees with Sillmann et al., (2013b). Similar is the overall picture
with the days with heavy rain distribution: general reduction, approximately up
to a week over the bigger part of the domain, but without substantial difference
from scenario to scenario and without statistical significance over wide areas
for all RCPs, except RCP8.5. It is notable that this result is somewhat different
that the outcomes in Sillmann et al. (2013b): there is shown a small (generally
2-4 days) increase of R10mm over the Balkan Peninsula for all scenarios except
RCP8.5 and decrease in the latter of about 2-4 days. It have to be emphasized,
however, that in this study is noted that the models disagree even on the sign
of change in the total precipitation and R10mm over the Mediterranean. Our
recent experiments with the RCM RegCM driven by the GCM HadGEM2-ES
(Gadzhev et al., 2021) shows prevailing positive change (i.e. increase) for all
seasons except for the summer and on an annual basis. The projected increase
is roughly 25–35% for RCP2.6 and 35–45% for RCP8.5. The expected pre-
cipitation reduction in the summer reaches values of 35–45% for RCP8.5 over
Bulgaria and Romania.
The spatial patterns of the CDD, both in the present and projected future climate
is also complex. Most apparent is the well expressed gradient form southeast to
northwest. The contrary tendencies in the future, increase of the CDD in southeast
and decrease in northwest will leads to strengthening of this contrast. The analysis,
163
Hristo Chervenkov, Vladimir Ivanov, Georgi Gadzhev, Kostadin Ganev

performed in Malcheva et al., (2016) which is based on historical records and on


the climate reanalysis ERA20C, outlines the drying tendency over SE Bulgaria and
the neighboring territories in Greece and Turkey. In Sillmann et al., (2013b) is not-
ed that in the Mediterranean, the increases in CDD are accompanied by increases
in the index R95p (very wet days, not considered in the present study) suggesting
that dry spells in these regions become longer, but that precipitation may be more
extreme when it occurs. Our previous analysis, based on HadEX2 (Chervenkov &
Slavov, 2020a), demonstrates also simultaneous increase of the AAs of the CDD
and R95p. It have to be kept in mind, however, that the applicability of HadEX2 in
regional climate studies, especially for precipitation-related parameters, is disput-
able. Possible intensification of the extreme precipitation events in generally dryer
climate is also discussed in Giorgi & Lionello, (2008).
It is worth to emphasize, that the projected changes under RCP2.6 and RCP4.5,
are not significant over the bigger part of the domain which also indicates the com-
plexity of the phenomena.
The temporal evolution of the considered precipitation indices is shown on
Figure 7.

Figure 7. Same as Figure 2 but for the RR (first row) and R10mm
(second row) and CDD (third row). The units are according the subplots titles.

As most remarkable on Figure 7 appears the fact that the evolution lines of RR
and R10mm and in smaller extent these for CDD, respectively their interquartile
ensemble spreads for all scenarios, essentially overlaps for the bigger part of the
time span 2011-2099. Such dynamics is expectable, keeping in mind the relatively
small changes from RCP to RCP commented above. The absence, at least appar-
ent, of big outliers is also remarkable. The inter-model coherency is demonstrated
also in agreement between the models in the simulation of isolated wet years in the
2040s and 2070s in the evolution of the CDD.
164
Assessment of the Future Climate Over Southeast Europe Based...

SUMMARY AND CONCLUSION


Based on the availability of new sources of information, which represent the
state of the art global climate change simulations in the frame of the CMIP5 project
and are free accessible from the Copernicus Data Store, we present an updated
assessment of future climate change projections over south-east Europe.
In the present study 6 temperature-based and 3 precipitation-based indices,
together with the annual means of the daily minimum, mean and maximum
temperature are analysed systematically. The indices are calculated in in consistent
manner in the frames of the Global Agriculture project and the assessment
covers the recent climate (1981-2010) as well as CMIP5 multimodel ensemble
projections of the 21st century for all four RCP scenarios. The spatial patterns and
temporal evolution of changes presented in this work are in principal agreement
with previous studies based on GCM output data (Orlowsky & Seneviratne 2012;
Sillmann et al., 2013b; Chervenkov & Slavov, 2021), RCM simulations Belda et
al., 2015; Gadzhev et al., 2021) or such based on combined analysis of GCMs
and RCMs (Giorgi & Lionello 2008; Ulbrich et al., 2006). The results of this
study are also coherent with the consolidated outcomes from all Assessment Re-
ports (AR) of the Intergovernmental Panel on Climate Change (IPCC) (see, for
example, IPCC, 2007) concerning the expected long-term regional changes. How-
ever, the present results are not directly comparable, at least not quantitatively,
to theirs, due to different factors. These factors includes, among others, different
time spans and model ensemble members and variations in the applied methodol-
ogy of estimation. Nevertheless, the most general and important conclusion of the
study is the distinct warming, expressed in the spatial patterns and time evolution
of all of the considered thermal indices. The climate change of the considered
temperature-based indices is consistent with the tendencies of the annual means
of the daily mean and extreme temperatures. The warming dominates practically
over the whole domain and is statistically significant over its essential part in the
most cases. The revealed patterns of climate change intensify gradually with the
increasing radiative forcing in the considered scenarios, which also agrees gener-
ally with the outcome of the prevailing number of the recent studies. The signifi-
cantly finer grid spacing of 0.5o than this in Sillmann et al. (2013b), Chervenkov
& Slavov, (2021), were it is 1.5o, leads to representation (although still not clear
enough) of such structures as the vertical gradients of some indices especially
along the main Carpathian ridge and the typical spatial pattern of the CDD. Both
of them are poorly resolved in the cited above studies. Beside the revealed spatial
details which are well known benefit of the increased resolution, this study dif-
fers from Chervenkov & Slavov, (2021) in other substantial aspect regarding the
results: the clearly demonstrated ‘warming asymmetry’ manifested in the fields of
the TNn and TXx. Generally, the revealed warming is, evidently, continuation of
already detected tendency in the historical records of the twentieth century over
165
Hristo Chervenkov, Vladimir Ivanov, Georgi Gadzhev, Kostadin Ganev

the region (Chervenkov & Slavov, 2019; Chervenkov & Slavov, 2020; Malcheva
et al., 2016; Moberg et al., 2006).
Concerning the precipitation-based indices, the study confirms the complexity of
the expected precipitation-related changes and their inherent ambiguity. The latter
is clearly evidenced by the lower level of statistical significance for the scenarios
RCP2.6-RCP6.0 when compared with temperature changes. It is worth emphasizing
that the projected precipitation reduction over the SE part of the domain and increase
of the CDD could amplify the negative impact of the expected hotter climate.
The study could be continued in many directions. Key moments as, for example,
seasonal variations and detailed regional specifics, could be focal point of further
works. The 10-daily temporal resolution of some indices gives unique possibility
even for sub-sesonal analysis. Other way is to utilize more actively the output of
RCMs, as demonstrated in Belda et al., (2015) and Gadzhev et al., (2021). Such
studies are methodologically reliable scientific basis of various impact studies and
the development of adaptation strategies.

ACKNOWLEDGMENTS
The authors thank the anonymous reviewers for their insightful suggestions. Hence
this study is entirely based on free available data and software, the authors
would like to express their deep gratitude to the primary CMIP5 model output
vendors as well as all other organizations and institutions (MPI-M, UNI-DATA,
Copernicus Data Store), which provides free of charge software and data.
Without their innovative data services and tools this work would be not possible.
This work has been carried out in the framework of the National Science Program
"Environmental Protection and Reduction of Risks of Adverse Events and
Natural Disasters", approved by the Resolution of the Council of Ministers №
577/17.08.2018 and supported by the Ministry of Education and Science (MES)
of Bulgaria (Agreement № Д01-322/18.12.2019).
This work has been accomplished with the financial support by the Grant No
BG05M2OP001-1.001-0003, financed by the Science and Education for Smart
Growth Operational Program (2014-2020) and co-financed by the European
Union through the European structural and Investment funds.

REFERENCES
Alexander, L. V., et al. (2006) Global observed changes in daily climate
extremes of temperature and precipitation, J. Geophys. Res., 111,
D05109, doi:10.1029/2005JD006290
Alexander, L. V. and Arblaster, J. M. (2009) Assessing trends in observed
and modelled climate extremes over Australia in relation to future
projections. Int. J. Climatol., 29: 417-435. doi:10.1002/joc.1730
166
Assessment of the Future Climate Over Southeast Europe Based...

Belda, M., Skalák, P., Farda, A., et al. (2015) CECILIA Regional Climate
Simulations for Future Climate: Analysis of Climate Change Signal,
Advances in Meteorology, vol. 2015, Article ID 354727, 13 pages, 2015
doi:10.1155/2015/354727
Birsan, M. V., Dumitrescu, A., Micu, D.M., Cheval, S. (2014) Changes in
annual temperature extremes in the carpathians since AD 1961. Natural
Hazards 74(3), 1899-1910 doi:10.1007/s11069-014-1290-5
Chervenkov, H., Slavov, K., Ivanov, V., (2019) STARDEX and ETCCDI
Climate Indices Based on E-OBS and CARPATCLIM; Part One: General
Description , in G. Nikolov et al. (Eds.): NMA 2018, LNCS 11189, pp.
360-367, doi: 10.1007/978-3-030-10692-8 40
Chervenkov, H., Slavov, K. (2019) STARDEX and ETCCDI Climate
Indices Based on E-OBS and CARPATCLIM; Part Two: ClimData in
Use, in G. Nikolov et al. (Eds.): NMA 2018, LNCS 11189, 368-374,
(2019) DOI: 10.1007/978-3-030-10692-8 41
Chervenkov H., Spiridonov V. (2020) Bias Correcting of Selected ETCCDI
Climate Indices for Projected Future Climate, In: Lirkov I., Margenov S.
(eds) LSSC 2019, LNCS 11958, pp. 292-299, doi:10.1007/978-3-030-
41032-2_33
Chervenkov H., Slavov K. (2020a) Historical Climate Assessment of
Temperature-based ETCCDI Climate Indices Derived from CMIP5
Simulations, C. R. Acad. Bulg. Sci. 73, No 6, 784-790, DOI: 10.7546/
CRABS.2020.06.05
Chervenkov, H. Slavov, K. (2020b) Historical Climate Assessment of
Precipitation-based ETCCDI Climate Indices Derived from CMIP5
Simulations, Compt. rend. Acad. bulg. Sci., 73, No 7, 942-948,
DOI:10.7546/CRABS.2020.07.06
Chervenkov H., Slavov K. (2020c) ETCCDI Thermal Climate Indices in the
CMIP5 Future Climate Projections over Southeast Europe, Proceedings
of the 14th Annual Meeting of the Bulgarian Section of SIAM, Studies
in Computational Intelligence (in press)
Chervenkov H., Slavov K. (2021) ETCCDI Climate Indices for Assessment
of the Recent Climate over Southeast Europe. In: Dimov I., Fidanova
S. (eds) Advances in High Performance Computing. HPC 2019. Studies
in Computational Intelligence, vol 902, 398-412, Springer, Cham. doi:
10.1007/978-3-030-55347-0_34
Chervenkov H., Spiridonov V. (2021) Sensitivity of Selected ETCCDI
Climate Indices from the Calculation Method for Projected Future
Climate. In: Dimov I., Fidanova S. (eds) Advances in High Performance
Computing. HPC 2019. Studies in Computational Intelligence, vol 902,
413-427 Springer, Cham. doi:10.1007/978-3-030-55347-0_35

167
Hristo Chervenkov, Vladimir Ivanov, Georgi Gadzhev, Kostadin Ganev

Chervenkov, H., Ivanov V., Gadzhev G., Ganev K. (2021) Assessment of


the future over South-east Europe based on CMIP5 ensemble of climate
indices – Part one: General Concept and Methods, Proceedings of the
1st International conference on ENVIROnmental protection and disaster
RISKs, 02-04 October 2020, Sofia, Bulgaria (in press)
Gadzhev G., Ivanov V., Valcheva R., Ganev K., Chervenkov H. (2021)
HPC Simulations of the Present and Projected Future Climate of the
Balkan Region. In: Dimov I., Fidanova S. (eds) Advances in High
Performance Computing. HPC 2019. Studies in Computational
Intelligence, vol 902, 234-248 Springer, Cham. https://doi.
org/10.1007/978-3-030-55347-0_20
Giorgi, F., Lionello, P. (2008) Climate change projections for the
Mediterranean region, Global and Planetary Change, 63(2-3): 90-104
Hempel, S., Frieler, K., Warszawski, L., Schewe, J. and Piontek, F., (2013)
A trend-preserving bias correction–the ISI-MIP approach. Earth System
Dynamics, 4(2), 219-236.
Intergovernmental Panel on Climate Change (IPCC) (2007) Climate
Change 2007: The Physical Science Basis. Contribution of Working
Group I to the Fourth Assessment Report of the Intergovernmental Panel
on Climate Change, edited by S. Solomon et al., Cambridge Univ. Press,
Cambridge, U. K.
Malcheva, K., Chervenkov, H., Marinova, T., (2016) Winter Severity
Assessment on the Basis of Measured and Reanalysis data. In: 16th
International Multidisciplinary Scientific GeoConference SGEM 2016.
SGEM, pp. 719-726
Moberg, A., et al. (2006), Indices for daily temperature and precipitation
extremes in Europe analyzed for the period 1901–2000, J. Geophys.
Res., 111, D22106, doi:10.1029/2006JD007103.
Orlowsky, B. and Seneviratne, S. I. (2012) Elusive drought: uncertainty in
observed trends and short- and long-term CMIP5 projections, Hydrol.
Earth Syst. Sci. 17(5):1765-1781 doi:10.3929/ethz-b-000073994.
Sillmann, J., Röckner, E., (2007) Indices for extreme events in projections
of anthropogenic climate change. Climatic Change 86 (1-2), pp. 83-104.
Sillmann, J., V. V. Kharin, X. Zhang, F. W. Zwiers, and D. Bronaugh
(2013a) Climate extremes indices in the CMIP5 multimodel ensemble:
Part 1. Model evaluation in the present climate, J. Geophys. Res. Atmos.,
118, pp. 1716-1733,doi: 10.1002/jgrd.50203
Sillmann, J., V. V. Kharin, F. W. Zwiers, X. Zhang, and D. Bronaugh (2013b)
Climate extremes indices in the CMIP5 multimodel ensemble: Part 2.
Future climate projections, J. Geophys. Res. Atmos., 118, pp. 2473-2493
doi: 10.1002/jgrd.50188.

168
Assessment of the Future Climate Over Southeast Europe Based...

Ulbrich, U., et al., (2006) The Mediterranean climate change under global
warming. In: Lionello, P., Malanotte-Rizzoli, P., Boscolo, R. (Eds.), R.
(Eds.), Mediterranean Climate Variability. Elsevier, Amsterdam, pp.
398-415.

 Hristo Chervenkov
http://orcid.org/0000-0002-7658-3041
National Institute of Meteorology and Hydrology
Sofia, Bulgaria
E-mail: [email protected]

 Vladimir Ivanov
https://orcid.org/0000-0001-9768-1049
National Institute of Geophysics, Geodesy and Geography
Bulgarian Academy of Sciences
Sofia, Bulgaria
E-mail: [email protected]

 Georgi Gadzhev
http://orcid.org/0000-0002-6159-3554
National Institute of Geophysics, Geodesy and Geography
Bulgarian Academy of Sciences
Sofia, Bulgaria
E-mail: [email protected]

 Kostadin Ganev
National Institute of Geophysics, Geodesy and Geography
Bulgarian Academy of Sciences
Sofia, Bulgaria
E-mail: [email protected]

169
Proceeding of 1st International conference on Environmental protection and disaster RISKs

https://doi.org/10.48365/envr-2020.1.15 Air Pollution,


Climate and Health

APPLICATION OF BACKWARD TRAJECTORIES


AND CLUSTER ANALYSЕS FOR STUDY
OF VARIATIONS IN PM10 CONCENTRATIONS

Rozeta Neykova, Elena Hristova


National Institute of Meteorology and Hydrology (NIMH)

Abstract: This study is on estimation of the influence of the air mass origin
on particulate matter (PM10) levels in the city of Plovdiv, Bulgaria during 2019.
The HYSPLIT (Hybrid Single-Particle Lagrangian Integrated Trajectory) and the
statistical software package “Openair” in R are applied. The cluster analysis statistical
method is used to group the back-trajectories into 5 sets according to the angle,
direction and speed of the air mass. The Potential Source Contribution Function
(PSCF) and Concentration Weighted Trajectory (CWT) statistical methods are also
applied in order to identify source areas of pollutants. The PM10 concentrations
in two air quality stations in Plovdiv and some meteorological elements are also
presented and discussed. The influence of the clusters on PM10 levels is explained.
Keywords: Air quality, PM10, back-trajectories, cluster analysis, PSCF

INTRODUCTION
Air pollution in urban area is one of the main risk factors concerning the human
health in medium and high- developed countries according to the World Health Orga-
nization (WHO, 2019). Major problem with the air quality in some European coun-
tries is the particulate matter (PM) air pollution (EEA Rep. 2018). The major problem
in Bulgarian cities is also air pollution with PM. In our country there is still a problem
with the regulated number of exceedances of PM10 (ЕЕA, 2018). 78.6 % of the
Bulgarian population is exposed to exceedance of the limit values of PM10, while this
percent for Europe is about 21% to 41% (MoEW Report, 2019).The harmful effect of
PM depends on its size (TSP, PM10, PM2.5, PM1), the concentration and the chemical
composition. It is easy for PM to be inhaled into the respiratory tract, where they
can cause inflammatory processes and diseases of the respiratory and cardiovascular
systems (Lim et al., 2012; Straif et al, 2013). Despite the plans for improving the air
quality and reduction of the emissions in regional and European scales, the decrease
of the PM concentration during the last years is still unsatisfactory (ЕЕА, 2018). The
PM particles are emitted in the atmosphere by many anthropogenic sources such as
170
Application of Backward Trajectories and Cluster...

energetic, industry, road transport, burning of fuels, incineration of waste. Significant


contribution according to the area can be the resuspension and soil erosion, sea spray
and volcanic eruption and etc. (Seinfeld and Pandis, 2006).
The application of Trajectory Statistical Methods as a supplementary tool
can be very useful in this type of studies aimed to characterize sources and
formation processes of particulate matter (PM). Characterization of wind-flow
circulations can help us to identify specific meteorological situations that favour
external (transboundary) contributions on ambient levels of atmospheric PM. The
best way for interpreting atmospheric pollutant measurements at a receptor site
using back-trajectories is by performing the analysis of large sets of them using
specific Trajectory Statistical Methods (TSM). The use of TSM deploying large
trajectory ensembles can significantly reduce the trajectory uncertainty generated
by interpolation and truncation processes, low temporal or spatial resolution of
wind data, or an inappropriate selection of the starting heights. Characterizing the
synoptic meteorological situations which give rise to low and high concentration
levels of PM at a receptor site can help us identify potential source regions of PM
which can be transported towards the sampling site.
The main objective is to study the influence of the origin of air mass at a synoptic
scale on PM levels at urban air quality monitoring stations in the city of Plovdiv for
the period 01.01.2019 – 31.12.2019.

2. METHODOLOGY
2.1. The study area
Plovdiv is city in southern Bulgaria placed in the southern part of the Plain of
Plovdiv on the banks of the Maritsa River. The Sredna Gora mountain range rise to
the northwest, the Chirpan Heights to the east, and the Rhodope mountains to the
south. Plovdiv is the second populated city in Bulgaria, with population 347 851
up to 31.12.2019 according National statistical institute of Republic of Bulgaria
(https://www.nsi.bg/en/content/6710/population-towns-and-sex. The city has been
struggling with poor air quality for years, which mainly results from its location
and limited ventilation and also from cumulative impacts of local, regional and
transboundary emissions. In Plovdiv there are two official air quality stations, part
of the National air quality network, named Plovdiv - Kamenitsa (42.142889° N,
24.765239° E) and Plovdiv - zh.k. Trakia (42.141186° N, 24.787952° E). Both
sites are in urban area, but the Kamenitsa is Background type while zh.k. Trakia is
Traffic (Figure 1).

2.2. PM10 data


The PM10 hourly concentration levels for 2019 obtained at two air quality
monitoring sites on the Executive Environment Agency are used in the statistical
analysis.
171
Rozeta Neykova, Elena Hristova

Figure 1. Map of Plovdiv with urban air quality stations –


Kamenitsa (square) and zh.k. Trakia (triangle)

2.3. Back-trajectories
A daily meteorological analysis was carried out for the study period to identify
the different source regions of the air masses influencing the study area. 3-days
back-trajectories at 04:00, 12:00 and 18:00 UTC and at 3 different heights above
the starting point located at ground level were computed on a daily basis by using
the NOAA (National Oceanic and Atmospheric Administration) HYSPLIT 4.0
model (Air Resources Laboratory 2017) (Stein et al. 2015). The weekly archived
data GDAS (Global Data Assimilation System) with resolution 1º were used as
input. The best procedure to illustrate the vertical structure of the atmosphere is to
run trajectories at several heights above the point of interest. 500, 1500 and 2000 m
AGL was chosen in this study.

2.4. Trajectory statistical methods (TSM)


Trajectory statistical methods were applied to the HYSPLIT back-trajectories
with the software package “Openair” in R (http://www.opair-project.org/PDF/
OpenAir_NewsLetter_Issue14.pdf). Firstly, on the air mass back-trajectories data
was performed cluster analysis (CA). CA is a statistical method used to examine
data and group it into sets of similar data known as clusters. It is a useful method for
172
Application of Backward Trajectories and Cluster...

organizing large data sets into smaller, similar groups. Trajectory coordinates are
used as the clustering variables. CA can be used to classify the air mass origins that
arrive at a site (Dorling et al., 1992; Brankov et al., 1998; Salvador et al., 2008),
but CA does not provide any information on the geographical location of potential
source regions. This information can be obtained by using Potential Source Con-
tribution Function.
Potential Source Contribution Function (PSCF) calculates the probability that
a source is located at latitude 𝑖 and longitude j (Fleming et al.,2012; Pekney et
al.,2006). The basis of PSCF is that if a source is located at (𝑖,𝑗), an air parcel back
trajectory passing through that location indicates that material from the source can
be collected and transported along the trajectory to the receptor site. PSCF solves

(1)
where nij is the number of times that the trajectories passed through the cell (𝑖,𝑗) and
𝑚ij is the number of times that a source concentration was high when the trajecto-
ries passed through the cell (i,j).
The criterion for determining 𝑚ij is controlled by percentile, which is by default
90. Note also that cells with few data have a weighting factor applied to reduce their
effect (The openair Project newsletter Issue 14, February 2013).
Finally, the potential source regions of particles were evaluated by the Con-
centration Weighted Trajectory (CWT) statistical method, available in the above
mentioned “Openair” package computing a concentration field to identify source
areas of pollutants (Seibert et al. 1994). For each cell of the grid domain a weighted
mean concentration of the pollutant species under study is calculated. The weight
for each concentration value of the pollutant’s time series is the time spent in that
grid cell by the associated trajectory. Areas with high CWT values in the concentra-
tion field indicated that, on average, air parcels residing over them resulted in high
concentrations of the atmospheric pollutant at the receptor site. Thus, these concen-
tration fields show those potential source areas whose emissions can be transported
to the measurement site by prevailing synoptic winds (López, V., et al., 2019)

3. RESULTS AND DISCUSSION


3.1. Particulate matter levels
Time series of PM10 hourly concentration levels obtained at two air quality
monitoring stations (Trakia and Kamenitsa) during the study period is presented
in Figure 2. There is a good relation between the cases of increasing values at
the two air quality stations. The exceedance of the 24-h limit value (50 µg.m-3) is
observed in both air quality stations during 2019. 93 days are with exceedance in
the Trakia station and 66 days for the Kamenitsa station. There are some days with
exceedance only in Trakia station (27 days). Most of those days are during the cold
period of the year.
173
Rozeta Neykova, Elena Hristova

Figure 2. Annual course of the daily mean PM10 values


in Trakia and Kamenitsa air quality stations

There are some days with PM10 concentration above 50 µg.m-3 during the warm
period of 2019.
Very high differences in the PM10 concentration between Trakia and Kamenitsa
are observed on 11 April, 03 May and 08 August. Probably these cases of exceedance
are due to local sources of air pollution.
The PM10 concentrations for the periods with exceedance are between 54.3 and
194.5 µg.m-3 at the Trakia station and from 50.2 to 169.6 µg.m-3 at the Kamenitsa
station. 24 days with fog are registered and 60 with rain. 90% of cases with rain are
with 0 mm amount. The daily averaged wind speed is ranged from 0.1 to 4 m.s-1.

3.2. TSM results


A total of 3285 backward trajectories were obtained ending at 04:00, 12:00 and
18:00 UTC hours in Trakia station (42.141186° N, 24.787952° E) for each day of
the period 1.01.-31.12.2019, at 3 different heights (500, 1500, 2000 m agl). They
have been clustered using “Openair” to establish the different air-flow patterns.
Five typical meteorological synoptic situations (clusters) were obtained. The
comparison of obtained clusters, PSCF and weighted mean PM10 concentration for
three different highs are presented in Figure 3.
Back-trajectories with the origin at 500 m agl were grouped into 5 different
origins of air mass (Figure 3). Cluster 1 (C1) indicated a 30% air mass movement
from the north western (central Europe). A flow movement from the north was
represented by Cluster 2 (C2). Cluster 3 (C3) is characterized by air mass coming
from the south west (18.3%). Cluster 4 (C4) contained short back-trajectories of
174
Application of Backward Trajectories and Cluster...

Figure 3. Cluster analysis (CA) on the left side, PSCF in the middle
and weighted mean PM10 concentration with back-trajectories for 2019

air mass recirculating south (9.2%) and Cluster 5 (C5) is characterized by air mass
coming from the east (20.5%).
The 5 clusters obtained from the set of back-trajectories at 1500 m agl and
2000 m agl presented similar origins but different from the 500 m agl. Clusters
containing back-trajectories coming from the south western and north western (C2
and C4 respectively), are presented similar percentages of trajectories (~27%). C2
and C4 grouped around 60% of all trajectories for each data set. On the other hand,
short scale back-trajectories contained in C5 have represented less than 15% of all
the trajectories with origin in 1500 and 2000 m agl. C3 represented 17-19% of all
back-trajectories.
The PSCF and weighted mean PM10 concentration for three different highs
presented in Figure 3 give us information for source regions of the PM10 component
understudy or as preferred air mass pathways.
The lowest mean concentration values for all data sets were obtained for periods
with north, north-western air mass advection (C1 and C2) (Table 1).
175
Rozeta Neykova, Elena Hristova

Table 1. Mean PM10 concentration (µg.m-3) by trajectory cluster


at 3 different levels
Level 500 Level 1500 Level 2000
Cluster 1 36 38.3 35.6
Cluster 2 37.7 36.3 38
Cluster 3 64.4 40 40.4
Cluster 4 58.1 58.5 57.7
Cluster 5 44.4 52.5 55.1

Two regions are identified as the main source regions of contributing to


increasing PM10 levels in Plovdiv (C4 and C5). However, it should be considered
that with the CWT methodology and the back-trajectories obtained with the
HYSPLIT model it is not possible to correctly perform an identification of the
main local sources of PM10.

4. CONCLUSIONS
In this study different trajectory statistical methods were applied to a set
of 1 year of back trajectories in order to describe the air mass influence on
the atmospheric particulate matter (PM10) levels in Plovdiv. Potential Source
Contribution Function (PSCF) and Concentration Weighted Trajectory
(CWT) methods are applied for the first time on PM data from urban air
quality observations. Air mass back-trajectories were grouped into 5 clusters,
representing a typical meteorological scenario. The results obtained suggest that
the levels of atmospheric particulate matter in Plovdiv could be influenced by
cross-border natural and anthropogenic sources (south, south-wets). To prove
this more studies with input data with higher resolution is needed. The obtained
results are a step towards scientifically based studies on identifying different
source areas of different air pollutants.The development of new methods
enabling identification and assessment of the reasons behind this condition
and their further application for air quality management is necessary. Further
studies on combining observations on PM chemical composition and Trajectory
Statistical Methods are necessary.

ACKNOWLEDGMENTS
This study was performed with the financial support from the National Programme
“Young scientists and post-doctoral”. We acknowledge also to the NOAA Air
Resources Laboratory (ARL) for the provision of the HYSPLIT model and
READY website (http://www.ready.noaa.gov) and also to the R-project website
(https://www.r-project.org/) for Package ‘openair’.

176
Application of Backward Trajectories and Cluster...

REFERENCES
Brankov, E., Rao, S.T., Porter, P.S., (1998), A trajectory-clustering-correla-
tion methodology for examining the long-range transport of air pollut-
ants, Atmospheric Environment 32, 1525-1534.
Dorling, S.R., Davies, T.D., Pierce, C.E., (1992), Cluster analysis: a tech-
nique for estimating the synoptic meteorological controls on air and pre-
cipitation chemistry – method and applications, Atmospheric Environ-
ment 26A, 2575-2581.
EEA, (2018), Air Quality in Europe-2016 report, Report No 12, European
Environmental Agency, Copenhagen, ISSN 1977-8449
Fleming, Z.L., Monks, P.S. and Manning. A.J.,(2012), Review: Untangling
the influence of air-mass history in interpreting observed atmospheric
composition, Atmospheric Research,104-105:1–39, doi:10.1016/j.at-
mosres.2011.09.009.
López, V., Salvador, P., Artíñano, B. et al. (2019) Influence of the origin
of the air mass on the background levels of atmospheric particulate
matter and secondary inorganic compounds in the Madrid air basin.
Environ Sci Pollut Res 26, 30426–30443 https://doi.org/10.1007/
s11356-019-06205-8.
National State of the Environment (SoE) Report, (2019), MoEW, Sofia (in
bulg.)/МОСВ Доклад, 2019. Национален доклад за състоянието и
опазването на околната среда, 2019.
Pekney, N. J. , Davidson C. I., Zhou, L. and Hopke, P. K., (2006), Ap-
plication of PSCF and CPF to PMF-Modeled Sources of PM2.5 in
Pittsburgh.Aerosol Science and Technology, 40(10):952–961, doi:
10.1080/02786820500543324.
Salvador, P., Artíñano, B., Querol, X., Alastuey, A., (2008), A combined
analysis of backward trajectories and aerosol chemistry to characterise
long-range transport episodes of particulate matter: the Madrid air basin,
a case study, Science of the Total Environment 390, 495e506.
Seibert P, Kromp-Kolb H, Baltensperger U, et al (1994) Trajectory analysis
of aerosol measurements at high alpine sites. Transp Transform Pollut
Troposph 15:689–693
Seinfeld J.H. and Pandis N.S., (2006), Atmospheric Chemistry and Physics.
From Air Pollution to Climate Change, 2-nd Ed. John Wiley & Sons, Inc.
USA. 1225p.
Stein AF, Draxler RR, Rolph GD, et al (2015) NOAA’s HYSPLIT
atmospheric transport and dispersion modeling system. Bull Am
Meteorol Soc 96:2059–2077
Stephen S. Lim et al., (2012), A comparative risk assessment of burden of
disease and injury attributable to 67 risk factor clusters in 21 regions,

177
Rozeta Neykova, Elena Hristova

1990-2010: a systematic analysis for the Global Burden of Disease Study


2010. The Lancet, 380 (9859) 2224-2260.
Straif, K., Cohen, A., Samet, J. (Eds.), (2013), Air Pollution and Cancer,
International Agency for Research on Cancer (IARC), Lyon, France,
p. 487.

 Rozeta Neykova
National Institute of Meteorology and Hydrology
Sofia, Bulgaria
E-mail: [email protected]

 Elena Hristova
https://orcid.org/0000-0002-5681-4375
National Institute of Meteorology and Hydrology
Sofia, Bulgaria
E-mail: [email protected]

178
Proceeding of 1st International conference on Environmental protection and disaster RISKs

https://doi.org/10.48365/envr-2020.1.16 Air Pollution,


Climate and Health

RELATIONSHIP BETWEEN PARTICULATE


MATTER AND HEALTH INDICATORS FOR ACUTE
MORBIDITY IN SOFIA

Petar Zhivkov1, Aleksandar Simidchiev2


1
Institute of Information and Communication Technologies –
Bulgarian Academy of Sciences (IICT-BAS)
2
Medical Institute of Ministry of Interior

Abstract: The objective of this paper is to assess the relationship between fine
Particulate Matter (PM) concentrations and health indicators for acute morbidity.
The methods used are linear correlation and non-parametric correlation analysis
of a time series study conducted in Sofia from 1 January 2017 to 31 May 2019.
Data from 4 local health sources that cover hospital admissions and all Emergency
Medical Aid registrations from Sofia's population of 1.3 million. Results were
obtained: The mean daily PM10 concentration was 35 μg/m3 with a range from 5 to
336 μg/m3. The mean daily PM2.5 concentration was 11.7 μg/m3 with a range from
2 to 136 μg/m3. In 15.3% of the times, the daily PM2.5 or PM10 concentrations did
not meet the WHO Air Quality Guidelines target during the study period. At these
excesses, acute infections of the upper respiratory tract and pneumonia increase by
47% and 60%, respectively. COPD increases with 36%, there are also changes and
acute bronchitis. Conclusions: The findings provide a justification that exceeds the
concentrations of PM10 and PM2.5 were significantly associated with indicators of
acute morbidity in Sofia.
Keywords: particulate matter, acute morbidity, health indicators, PM10, PM2.5.

INTRODUCTION
The human health effects of exposure to outdoor air pollutants is considered a
global health concern (Thurston et al., 2017). The links between urban air pollution
and human health are consistently and clearly established by many researchers
(J. Samet & Krewski, 2007) (Li et al., 2013), especially for short-term effects
such as cardiovascular events (Vermylen, Nemmar, Nemery, & Hoylaerts, 2005),
neurovascular (Tallon, Manjourides, Pun, Salhi, & Suh, 2017; Wing et al., 2017)
and asthma (Ward & Ayres, 2004).
Significant literature of epidemiologic studies suggests a correlation between
acute morbidity and exposure to air pollution from particulate matter (Pope,
179
Petar Zhivkov, Aleksandar Simidchiev

Dockery, & Schwartz, 1995). Most of this data comes from time series analyses
(Bell, Samet, & Dominici, 2004) comparing the variations in hospitalization with
the average particulate matter variations (Dominici, Sheppard, & Clyde, 2003).
Also, multicity studies exist such as the European Air Pollution and Health: a
European Approach (APHEA) project (Katsouyanni et al., 1996) and the American
National Morbidity, Mortality, and Air Pollution Study (NMMAPS) (J. M. Samet
et al., 2000) that both provide a consistent evidence for association health and air
pollutants for multiple cities by covering a large geographic area.
The consequences of air pollution can be seen as an enhancement to a person's
risk of illness or injury or as an additional general well-being risk acquired by a
population (Gochfeld & Burger, 2011). The aim of air quality management is to
control or avoid adverse impacts on air pollution to public health. Therefore, it
is important to define such effects that are deemed "adverse," and to distinguish
them from those effects that are not deemed adverse, thus concentrating protection
efforts on the contaminants that cause the most extreme health impacts.

Air Quality Control - Difference in norms and legislation


There is difference between legislation and recommendations illustrated in Table
1. The WHO guidelines for air pollution can be looked more like recommendations,
they do not have a mandatory character. While the EU Air Quality Directive can be
looked as a legislation that every EU member has to follow.

Table 1. Difference between EU Air Quality Directive and WHO Guidelines


EU Air Quality
WHO Guidelines
Directive
Pollutant Period Concentration (limit value μg/m3) Concentration (limit value μg/m3)
24 Hours - 25
PM2.5
Annual 25 10
24 Hours 50 50
PM10
Annual 40 20

For the purpose of this research we will use the WHO guidelines as they are
advisable worldwide and having purely health aspects in the consideration, while
the EU Directive includes some also politico-economic aspects in recital.

Case study
Sofia is the only European capital situated in a valley and is characterized by
high quantity of anthropogenic emissions and by frequent occurrence of stagnant
meteorological conditions. The city has a population of 1.2 million people (NSI,
2019) and is situated in the Sofia valley. The area is recognized as a problematic
location where especially during winter there are numerous exceedances despite
180
Relationship Between Particulate Matter and Health...

the European legislation that aims at air pollution control. The combination of cold
winters and its situation predisposes to temperature inversions that last several days
to a week.
Many urban areas located in a valley without adequate air exchange encounter
significant air pollution problems that are linked with the local atmospheric pecu-
liarities (Rendón, Salazar, Palacio, Wirth, & Brötz, 2014). A strong inversion and
light precipitation and/or wind were the major causes for trapping pollutants in the
air mainly during winter time. The air quality in Sofia deteriorates significantly dur-
ing winter compared to summer. Hence it will be a good model to test our hypoth-
esis of a significant link between air pollution and health consequences.
The literature abounds with models that can be used to assess the pollution/
health hypothesis. A correlation between temperature and mortality around Sofia is
made by using models with linear and non-linear terms (Pattenden, Nikiforov, &
Armstrong, 2003). Other research that examines the genotoxicity of ambient air in
3 European cities, including Sofia shows that winter air pollution is six- to 10-fold
higher in comparison with summer air (Gábelová et al., 2004).
This will be the first such study with real data from official and civil sources
of information on particulate air pollution, comprehensive data from the activities
of the center for emergency medical care, data from two of the largest hospitals
related to the access to emergency care in Sofia. The results can be compared and
contrasted with other international studies with local data.

The aims of this paper are to:


(1) Compare data from air pollution in Sofia with health data originating from
hospitals and ambulance services.
(2) Highlight the key problem areas.
(3) Identify future prospects and summarize the key areas where further research
is needed to improve model performance.

DATA AND MODELLING


Health data
Due to the voluntary nature of data provision, there are different time
characteristics and formats from different sources:
• Summarized data on the activity of the Center for Emergency Medical Aid
Sofia by diagnoses of the center - from 01.01.2017 to 14.03.2019
• Information on the activity of diagnoses from DCC Tokuda - 02.01.2018 to
31.12.2018
• Information on the activity of diagnoses from Pirogov outpatient clinic -
01/01/2018 to 31/05/2019
• Information on the activity of hospitalized diagnoses from Pirogov - 01/01/2018
to 31/12/2018
181
Petar Zhivkov, Aleksandar Simidchiev

The research uses the International Classification of Diseases (ICD) and more
specifically ICD-10 for segmentation of the diseases and the identified morbidity.
ICD-10 is the 10th revision of the ICD, a medical classification list by the World
Health Organization (WHO) that is used by the time this study is conducted. It
contains codes for diseases, signs and symptoms, abnormal findings, complaints,
social circumstances, and external causes of injury or diseases (Organization,
1978, 2018).

Air pollution data


Air quality data in a form of hourly average concentrations of PM2.5 and
PM10 is obtained for 2019 as independent series for each element from official
sensors provided from (a) the Executive Environment Agency (EEA) and (b)
low-cost stations from the air.bg network (a contributors driven global sen-
sor network that creates Open Environmental Data). PM10 is particulate mat-
ter 10 micrometers or less in diameter, PM2.5 is particulate matter 2.5 microm-
eters or less in diameter . The data is collected from 5 official monitoring sites
(Druzhba, Nadezhda, Hipodruma, Pavlovo, and Mladost), and over 500 citizen
science monitors.
Beta attenuation method is used for PM10 and PM2.5 according to the Europe-
an Directive 2008/50/EC (Directive, 2008). Air quality monitoring networks are
situated in.
The use of air quality data from low-cost stations aims to increase the potential
benefits from traditional monitoring networks with additional geographic and tem-
poral measurement resolution (Castell et al., 2017). Furthermore, low cost stations
can contribute to data where limited air quality stations exist (Zimmerman et al.,
2017). In Sofia, at the moment of the research there are with additional geographic
and temporal measurement resolution.

Methods
For this research we use a time series analysis with correlation methods for
analyzing the air quality and health data. The statistical methods in the research fall
in the two categorizations: parametric and nonparametric. Parametric comparisons
are based on the premise that the variable is continuous and normally distributed.
Nonparametric approaches are used where data is continuous with non-normal
distribution or any other form of data other than continuous variables.
The parametric Pearson correlation test (1) is used for comparing the two sourc-
es of air quality data. It provides a measure of the linear association between the
two continuous variables (usually just referred to as correlation coefficient). To
conduct the test, correlation coefficients are calculated for each (x,y) pair, and the
values of x and y are subsequently replaced with their ranks. Application of the test
results in a correlation coefficient that ranges from -1 to 1.
182
Relationship Between Particulate Matter and Health...

(1)
Parametric methods are better ways to measure the difference between the
groups relative to their equivalent nonparametric methods, but due to certain strict
criteria, including data normality and sample size, we cannot use parametric tests in
any situation and instead use their alternate nonparametric methods:
• For correlation analysis between pairs of variables is also used the non-para-
metric method Spearman's rho (2).

(2)

• Intergroup comparison for two unpaired groups is used the non-parametric


Mann – Whitney U test between means.

Aggregation of air quality data


For PM10 all measurements above 1000 µg /m3, which is the limit of the sensor are
removed. A separate assessment based only on the days with below 70% humidity
is made. These conditions do not necessarily mean that the data is incorrect, but
under such conditions the manufacturer does not guarantee the predicted accuracy
of 10% (Jayaratne, Liu, Thai, Dunbabin, & Morawska, 2018).

RESULTS
The results for daily concentrations of PM10 and PM2.5 are summarized in Table
2. The mean daily PM10 concentration was 35 μg/m3 with a range from 5 to 336 μg/
m3. The mean daily PM2.5 concentration was 11.7 μg/m3 with a range from 2 to
136 μg/m3. 13.2% of the daily PM10 or PM2.5 concentrations did not meet WHO
Air Quality Guidelines target (50 μg/m3 for PM10 and 25 μg/m3 for PM2.5) during
the study period. Some of the sources of health effects were with missing data for
weekends (Saturday, Sunday, public holidays), so an analysis of gaps to exclude the
possibility of bias was undertaken.

Table 2. Daily concentration of PM10 and PM2.5 divided into working days
and weekends + holidays
Day of the
Source N Mean Median SD Min Max Percentiles Z P
week
working day 522 34,17 27,5 30,67 5 336 20 36
PM10 0,76 0,4
weekend 230 37,1 26 38,45 6 290 18 38
working day 522 11,15 9 12,61 2 104 4 14
PM2,5 0,09 0,9
weekend 230 12,93 8 17,13 2 113 3 16

183
Petar Zhivkov, Aleksandar Simidchiev

The conclusions are that there are no significant differences between weekends
and working days in PM levels, and the lack of medical data for the weekends
would not significantly affect the relevance of the results.
The correlation between the two sources of information for air quality (official
satiations from EEA and citizen stations from AirBG) is significant PM10 (n = 340)
PM2.5 (n = 296) as seen in Table 3. and corresponds to other sources comparing
these sources (National Institute of Meteorology and Hydrology and AirSofia.info).

Table 3. Pearson correlation coefficient (r) between the two sources


AirBG PM10 AirBG PM2.5
EEA PM10 0.967 0.937
EEA PM2.5 0.898 0.884

Correlation (Spearman's rho) of the measured values of PM10-AS and PM2.5-


AS and the number of EMA contacts by groups of diseases are shown in Table
4 and Table 5. Please keep in mind that the demand for EMA is usually initiated
by a patient, and diagnoses are usually preliminary. As exposure requires time to
develop effects outside the respiratory system, lag-shift analyzes of 1, 2 and 3 days
were also introduced. The red numbers are significantly correlated (p < 0.05). The
clearest correlation is observed to the total number of emergency contacts.

Table 4. Correlation (Spearman's rho) of the measured values of PM10-AS


and the number of EMA contacts by groups of diseases
Without LAG LAG period LAG period LAG period
Name of the disease / period 1 day 2 days 3 days
identified morbidity
R R R R
Individuals received EMA 0.211** 0.195** 0.226** 0.234**
Heart attacks 0,073* 0,058 0,013 0,085*
Strokes 0,069 0,108** 0,086* 0,002
Rhythmic diseases 0,058 0,052 0,063 0,111**
Acute heart failure 0,064 0,144** 0,121** 0,066

Table 5. Correlation (Spearman's rho) of the measured values of PM2.5


and the number of outpatient examinations in hospitals by groups of diseases
Without LAG LAG period LAG period LAG period
Name of the disease / period 1 day 2 days 3 days
identified morbidity
R R R R
Individuals received EMA 0.524** 0.516** 0.535** 0.561**
Heart attacks 0,148** 0,153** 0,120** 0,161**

184
Relationship Between Particulate Matter and Health...

Strokes 0,029 0,042 0,029 -0,001


Rhythmic diseases 0,166** 0,171** 0,185** 0,217**
Acute heart failure 0,161** 0,202** 0,199** 0,165**

The correlation is significant for cardiovascular and cerebrovascular conditions,


fully coinciding with the rapid increase in risk, albeit to a lesser extent, with a low
level fractions of PM2.5 contamination from published data.(Shin et al., 2019; Wang
et al., 2019; Zhang et al., 2018).
The results shown in Table 6 and Table 7 revealed an increase in EMA contacts
with 11% on days with exceedances of PM10 and 13.5% on days with exceedances
of PM2.5, for a period of at least 3 days. Mean increase (relative to background
levels) in neuroses on day 2 after PM10 exceedance by 1%, on heart attacks on day
3 after excess by 8%, on strokes immediately after excess by 9%, lasting up to day
3 after excess. The same applies to hypertensive conditions, which increase by 5%
immediately around the exceedance and until the third day following the event.

Table 6. Comparison of non-parametric data from EMA contacts


with WHO norms for average daily concentration of PM10
Without LAG LAG period LAG period LAG period
Name of the disease / period 1 day 2 days 3 days
identified morbidity
Z p Z P Z p Z p
Individuals received -6.49 <0.001 -5.95 <0.001 -7.11 <0.001 -7.81 <0.001
ЕМА
Neurosis -0,079 0,937 -1,102 0,271 -2,305 0,021 -0,412 0,680
Heart attacks -1,289 0,198 -0,882 0,378 -1,283 0,199 -2,555 0,011
Strokes -2,151 0,032 -2,153 0,031 -2,419 0,016 -1,500 0,134
Hip. diseases -2,101 0,036 -2,270 0,023 -2,182 0,029 -2,507 0,012
Ischemic heart disease -2,238 0,025 -2,225 0,026 -1,591 0,112 -1,779 0,075
Rit. diseases -0,762 0,446 -1,202 0,229 -2,121 0,034 -3,160 0,002
Acute heart failure -0,823 0,410 -2,413 0,016 -2,314 0,021 -1,670 0,095

Table 7. Comparison of non-parametric data from EMA contacts


with WHO norms for average daily concentration of PM2.5
Without LAG LAG period LAG period LAG period
Name of the disease / period 1 day 2 days 3 days
identified morbidity
Z P Z p Z p Z p
Individuals received ЕМА -6.34 <0.001 -6.12 <0.001 -7.08 <0.001 -7.34 <0.001
Neurosis -1,065 0,287 -0,738 0,461 -2,421 0,015 -1,058 0,290
Strokes -1,484 0,138 -2,140 0,032 -1,609 0,108 -1,424 0,154

185
Petar Zhivkov, Aleksandar Simidchiev

Hip. diseases -2,154 0,031 -2,365 0,018 -1,550 0,121 -2,476 0,013
Rit. diseases -1,536 0,125 -0,894 0,371 -1,490 0,136 -3,282 0,001
Acute heart failure -0,248 0,804 -3,369 0,001 -2,414 0,016 -2,028 0,043

In the data of hospitalized patients with more serious condition, we observe the
following correlations (Table 8 and Table 9). An increase (relative to background
levels) of respiratory and thoracic diseases from the 1st to the 3rd day after excess
of PM10 by 120%, as well as heart failure in the same time period by 18%. Excess
PM2.5 was associated with a 59% increase in pulmonary embolism rate on days 2
and 3, with a 19% increase in heart failure on days 1 to 3.

Table 8. Comparison of non-parametric data from hospitalized patients


in serious conditions with WHO norms for average daily concentration of PM10

Name of the Without LAG LAG period LAG period LAG period
disease / identified ICD-10 period 1 day 2 days 3 days
morbidity Z p Z p Z p Z p
Respiratory system C30-
-1,504 0,133 -2,211 0,027 -2,870 0,004 -2,862 0,004
and chest C39
Heart failure I50 -1,729 0,084 -3,656 <0,001 -3,475 0,001 -2,821 0,005

Table 9. Comparison of non-parametric data from hospitalized patients


in serious conditions with WHO norms for average daily concentration of PM2.5

Name of the Without LAG period LAG period LAG period


disease / identified ICD-10 LAG period 1 day 2 days 3 days
morbidity Z p Z p Z p Z p
Respiratory system
C30-C39 -0,361 0,718 -1,270 0,204 -1,265 0,206 -2,119 0,034
and chest
Pulmonary embolism I26 -1,601 0,109 -2,302 0,021 -2,894 0,004 -1,852 0,064
Heart failure I50 -1,586 0,113 -2,831 0,005 -3,195 0,001 -2,799 0,005

With regard to outpatient examinations in hospitals for mild conditions that do


not require hospitalization, we logically find an increased presence of respiratory
diagnoses (Table 10 and Table 11).
186
Relationship Between Particulate Matter and Health...

Table 10. Correlation (Spearman's rho) of the measured values of PM10-AS


and the number of outpatient examinations in hospitals by groups of diseases
Without LAG LAG period LAG period LAG period
Name of the disease / period 1 day 2 days 3 days
ICD-10
identified morbidity
R R R R
Respiratory system and chest C30-C39 0,025 0,046 0,097* 0,058
Cerebral infarction I63 0,109* 0,138** 0,100* 0,112*
Acute upper respiratory tract
J00-J06 0,159** 0,090* 0,102* 0,142**
infections
Bacterial pneumonia J13-J18 0,125** 0,145** 0,098* 0,131**
Chronic diseases of the lower
J40-J47 0,044 0,122** 0,072 0,079
respiratory tract
Acute bronchiolitis J21 0,053 0,097*
0,112*
0,127**
Asthma with a predominant
J45.0 0,053 0,040 0,056 0,123**
allergic component

Table 11. Comparison of non-parametric data from outpatient examinations


in hospitals with WHO norms for average daily concentration of PM10
Without LAG period LAG period LAG period
Name of the disease / LAG period 1 day 2 days 3 days
ICD-10
identified morbidity
Z p Z p Z p Z p
Respiratory system and chest C30-C39 -1,515 0,130 -2,950 0,003 -2,941 0,003 -2,423 0,015
Diseases of the ear and H65.0,
-0,754 0,451 -0,976 0,329 -1,580 0,114 -2,322 0,020
mastoid process H66.0, H81
Acute myocardial infarction I20-I25 -0,244 0,807 -1,393 0,164 -1,876 0,061 -2,104 0,035
Heart failure I50 -0,771 0,440 -2,840 0,005 -2,559 0,010 -2,972 0,003
Acute upper respiratory tract
J00-J06 -3,218 0,001 -3,507 0,000 -3,910 0,000 -4,605 0,000
infections
Bacterial pneumonia J13-J18 -3,171 0,002 -3,395 0,001 -2,556 0,011 -3,683 0,000
Acute bronchitis, Acute
J20-J21 -3,096 0,002 -2,848 0,004 -1,606 0,108 -1,588 0,112
bronchiolitis
Chronic diseases of the lower
J40-J47 -3,988 0,000 -4,049 0,000 -2,842 0,004 -3,481 0,001
respiratory tract
Skin and subcutaneous tissue
L00-L08 -2,552 0,011 -1,379 0,168 -0,267 0,789 -0,489 0,625
infections
Acute bronchitis caused by
J20.0 -2,330 0,020 -2,327 0,020 -0,430 0,667 -0,431 0,667
Mycoplasma pneumoniae
Acute bronchitis, unspecified J20.9 -2,496 0,013 -1,449 0,147 -0,039 0,969 -1,292 0,197
Asthma with a predominant
J45.0 -2,426 0,015 -2,808 0,005 -3,577 0,000 -3,643 0,000
allergic component
Other types of angina I20.8 -1,971 0,049 -0,763 0,445 -0,773 0,440 -0,674 0,500

187
Petar Zhivkov, Aleksandar Simidchiev

Acute infections of the upper respiratory tract and pneumonia 0-3 days increase
by 47% and 60%, respectively, Chronic Obstructive Pulmonary Disease (COPD)
after a day is 36% more, there are changes and acute bronchitis. Regarding asthma,
the allergic asthma records raise more in days with increased pollution in com-
parison to non-allergic asthma. In the other studied diseases shown in Table 10 and
Table 11, a smaller correlation is seen.

CONCLUSIVE COMMENTS
In conclusion, similar to data from published studies in cities in Europe, America
and Asia, increased levels of air pollution are associated with higher levels of
diagnosing health consequences. The process of exposure and impact is complex,
and it is very difficult to obtain accurate measurements except in an exposure
chamber, but with the methods of epidemiological studies and population statistics
it becomes clear that pollution leads to predictable increases in disease and hence
adverse economic consequences for society.
Low-cost stations have a significant error threshold (10%) and a problem
when humidity levels exceed 70%. Even if a better sensor of this type is used,
the fog still distorts the results as it is detected as dust (shows increased values).
But still they are highly valuable for explaining how air pollution affects health
as they show a lot more about what we actually breathe specifically near our own
home and vicinity compared to official sensors. This pollution data can include
the background pollution, as well as heating of neighboring buildings or homes,
burning garbage, barbecue nearby, cars under the window and even wind picking
up dust from the dirty facade. Official stations should meet the requirements for
distance, classification, etc. They are intentionally not placed directly next to
sources of pollution because their purpose is to account for background pollution.

Future research:
With few exceptions, short-term health effects are measured by using averaged
citywide air pollution concentrations for exposure indicators. This can lead to
exposure misclassification and thus to bias (Laurent et al., 2008). Measurements
of the relations between health effects and air pollution can be improved with more
geographically precise exposure measurements and a method to calculate personal
exposure during the day, and compare it to the health condition of each individual.
There are methods to reduce the error threshold (10%) in low-cost stations, but
there is a risk to have a lower consistency in the measurements and make the individual
station more expensive and thus make it not feasible for having such high quantity.
Lastly, economic effects of air pollution on a local level worth further
investigation. Existing software such as EPA’s BenMap can be adapted to local
aspects such as concentration-response relationships, population files, and health
and economic data to quantify these impacts.
188
Relationship Between Particulate Matter and Health...

ACKNOWLEDGEMENTS:
The authors would like to thank the anonymous reviewers for their valuable
comments to improve the paper, but also, to the hospitals Pirogov, Tokuda, the
Emergency Medical Aid Sofia, Faculty of Public Health at MU Sofia, Sofia
Municipality, Air for Health, Air Solutions, and air.bg the for kindly providing
the data for this research.

REFERENCES
Bell, M. L., Samet, J. M., & Dominici, F. (2004). Time series studies of
particulate matter. Annu. Rev. Public Health, 25, 247-280.
Castell, N., Dauge, F. R., Schneider, P., Vogt, M., Lerner, U., Fishbain, B., . . .
Bartonova, A. (2017). Can commercial low-cost sensor platforms contribute
to air quality monitoring and exposure estimates? Environment international,
99, 293-302.
Dominici, F., Sheppard, L., & Clyde, M. (2003). Health effects of air pollution:
a statistical review. International Statistical Review, 71(2), 243-276.
Gábelová, A., Valovičová, Z., Horváthová, E., Slameňová, D., Binková, B., Šrám,
R. J., & Farmer, P. B. (2004). Genotoxicity of environmental air pollution in three
European cities: Prague, Košice and Sofia. Mutation Research/Genetic Toxicology
and Environmental Mutagenesis, 563(1), 49-59.
Gochfeld, M., & Burger, J. (2011). Disproportionate exposures in environmental
justice and other populations: the importance of outliers. American Journal
of Public Health, 101(S1), S53-S63.
Jayaratne, R., Liu, X., Thai, P., Dunbabin, M., & Morawska, L. (2018). The
influence of humidity on the performance of a low-cost air particle mass
sensor and the effect of atmospheric fog. Atmospheric Measurement
Techniques, 11(8), 4883-4890.
Katsouyanni, K., Schwartz, J., Spix, C., Touloumi, G., Zmirou, D., Zanobetti, A.,
. . . Pönkä, A. (1996). Short term effects of air pollution on health: a European
approach using epidemiologic time series data: the APHEA protocol. Journal
of Epidemiology & Community Health, 50(Suppl 1), S12-S18.
Laurent, O., Pedrono, G., Segala, C., Filleul, L., Havard, S., Deguen, S., ... Bard, D.
(2008). Air pollution, asthma attacks, and socioeconomic deprivation: a small-
area case-crossover study. American journal of epidemiology, 168(1), 58-65.
Li, P., Xin, J., Wang, Y., Wang, S., Li, G., Pan, X., . . . Wang, L. (2013). The
acute effects of fine particles on respiratory mortality and morbidity in
Beijing, 2004–2009. Environmental Science and Pollution Research, 20(9),
6433-6444.
Organization, W. H. (1978). International classification of diseases:[9th]
ninth revision, basic tabulation list with alphabetic index: World Health
Organization.
189
Petar Zhivkov, Aleksandar Simidchiev

Pattenden, S., Nikiforov, B., & Armstrong, B. (2003). Mortality and temperature
in Sofia and London. Journal of Epidemiology & Community Health, 57(8),
628-633.
Pope, C. A., Dockery, D. W., & Schwartz, J. (1995). Review of epidemiological
evidence of health effects of particulate air pollution. Inhalation toxicology,
7(1), 1-18.
Rendón, A. M., Salazar, J. F., Palacio, C. A., Wirth, V., & Brötz, B. (2014).
Effects of urbanization on the temperature inversion breakup in a mountain
valley with implications for air quality. Journal of Applied Meteorology and
Climatology, 53(4), 840-858.
Samet, J., & Krewski, D. (2007). Health effects associated with exposure to
ambient air pollution. Journal of toxicology and environmental health, Part
A, 70(3-4), 227-242.
Samet, J. M., Zeger, S. L., Dominici, F., Curriero, F., Coursac, I., Dockery, D. W.,
Zanobetti, A. (2000). The national morbidity, mortality, and air pollution
study. Part II: morbidity and mortality from air pollution in the United States
Res Rep Health Eff Inst, 94(pt 2), 5-79.
Shin, S., Burnett, R. T., Kwong, J. C., Hystad, P., van Donkelaar, A., Brook, J. R.,
Chen, H. (2019). Ambient Air Pollution and the Risk of Atrial Fibrillation
and Stroke: A Population-Based Cohort Study. Environ Health Perspect,
127(8), 87009. doi:10.1289/EHP4883
Tallon, L. A., Manjourides, J., Pun, V. C., Salhi, C., & Suh, H. (2017).
Cognitive impacts of ambient air pollution in the National Social Health and
Aging Project (NSHAP) cohort. Environ Int, 104, 102-109. doi:10.1016/j.
envint.2017.03.019
Thurston, G. D., Kipen, H., Annesi-Maesano, I., Balmes, J., Brook, R. D.,
Cromar, K., Frampton, M. W. (2017). A joint ERS/ATS policy statement:
what constitutes an adverse health effect of air pollution? An analytical
framework. European Respiratory Journal, 49(1).
Vermylen, J., Nemmar, A., Nemery, B., & Hoylaerts, M. F. (2005). Ambient
air pollution and acute myocardial infarction. Journal of Thrombosis and
Haemostasis, 3(9), 1955-1961.
Wang, W., Liu, C., Ying, Z., Lei, X., Wang, C., Huo, J., Kan, H. (2019).
Particulate air pollution and ischemic stroke hospitalization: How the
associations vary by constituents in Shanghai, China. Sci Total Environ, 695,
133780. doi:10.1016/j.scitotenv.2019.133780
Ward, D., & Ayres, J. G. (2004). Particulate air pollution and panel studies in
children: a systematic review. Occupational and environmental medicine,
61(4), e13-e13.
Wing, J. J., Adar, S. D., Sanchez, B. N., Morgenstern, L. B., Smith, M. A., &
Lisabeth, L. D. (2017). Short-term exposures to ambient air pollution and

190
Relationship Between Particulate Matter and Health...

risk of recurrent ischemic stroke. Environ Res, 152, 304-307. doi:10.1016/j.


envres.2016.11.001
Zhang, R., Liu, G., Jiang, Y., Li, G., Pan, Y., Wang, Y., . . . Wang, Y. (2018). Acute
Effects of Particulate Air Pollution on Ischemic Stroke and Hemorrhagic
Stroke Mortality. Front Neurol, 9, 827. doi:10.3389/fneur.2018.00827
Zimmerman, N., Presto, A. A., Kumar, S. P., Gu, J., Hauryliuk, A., Robinson,
E. S., . . . Subramanian, R. (2017). Closing the gap on lower cost air quality
monitoring: Machine learning calibration models to improve low-cost sensor
performance. Atmos. Meas. Tech. Discuss, 1-36.
Bell, M. L., Samet, J. M., & Dominici, F. (2004). Time-series studies of particu-
late matter. Annu. Rev. Public Health, 25, 247-280.
Castell, N., Dauge, F. R., Schneider, P., Vogt, M., Lerner, U., Fishbain, B., . . .
Bartonova, A. (2017). Can commercial low-cost sensor platforms contribute
to air quality monitoring and exposure estimates? Environment internation-
al, 99, 293-302.
Dominici, F., Sheppard, L., & Clyde, M. (2003). Health effects of air pollution:
a statistical review. International Statistical Review, 71(2), 243-276.
Gábelová, A., Valovičová, Z., Horváthová, E., Slameňová, D., Binková, B.,
Šrám, R. J., & Farmer, P. B. (2004). Genotoxicity of environmental air pollu-
tion in three European cities: Prague, Košice and Sofia. Mutation Research/
Genetic Toxicology and Environmental Mutagenesis, 563(1), 49-59.
Gochfeld, M., & Burger, J. (2011). Disproportionate exposures in environmen-
tal justice and other populations: the importance of outliers. American Jour-
nal of Public Health, 101(S1), S53-S63.
Jayaratne, R., Liu, X., Thai, P., Dunbabin, M., & Morawska, L. (2018). The
influence of humidity on the performance of a low-cost air particle mass sen-
sor and the effect of atmospheric fog. Atmospheric Measurement Techniques,
11(8), 4883-4890.
Katsouyanni, K., Schwartz, J., Spix, C., Touloumi, G., Zmirou, D., Zanobetti,
A., . . . Pönkä, A. (1996). Short term effects of air pollution on health: a Eu-
ropean approach using epidemiologic time series data: the APHEA protocol.
Journal of Epidemiology & Community Health, 50(Suppl 1), S12-S18.
Laurent, O., Pedrono, G., Segala, C., Filleul, L., Havard, S., Deguen, S., . . . Bard, D.
(2008). Air pollution, asthma attacks, and socioeconomic deprivation: a small-area
case-crossover study. American journal of epidemiology, 168(1), 58-65.
Li, P., Xin, J., Wang, Y., Wang, S., Li, G., Pan, X., . . . Wang, L. (2013). The
acute effects of fine particles on respiratory mortality and morbidity in Bei-
jing, 2004–2009. Environmental Science and Pollution Research, 20(9),
6433-6444.
NSI. (2019). Population by cities. Retrieved from https://www.nsi.bg/bg/
content/2981/%D0%BD%D0%B0%D1%81%D0%B5%D0%BB%D0

191
Petar Zhivkov, Aleksandar Simidchiev

%B5%D0%BD%D0%B8%D0%B5-%D0%BF%D0%BE-%D0%B3%
D1%80%D0%B0%D0%B4%D0%BE%D0%B2%D0%B5-%D0%B8-
%D0%BF%D0%BE%D0%BB.
Organization, W. H. (1978). International classification of diseases:[9th] ninth
revision, basic tabulation list with alphabetic index: World Health Organiza-
tion.
Organization, W. H. (2018). International Classification of Diseases (ICD) In-
formation Sheet. 2018. In.
Pattenden, S., Nikiforov, B., & Armstrong, B. (2003). Mortality and tempera-
ture in Sofia and London. Journal of Epidemiology & Community Health,
57(8), 628-633.
Pope, C. A., Dockery, D. W., & Schwartz, J. (1995). Review of epidemiological
evidence of health effects of particulate air pollution. Inhalation toxicology,
7(1), 1-18.
Rendón, A. M., Salazar, J. F., Palacio, C. A., Wirth, V., & Brötz, B. (2014).
Effects of urbanization on the temperature inversion breakup in a mountain
valley with implications for air quality. Journal of Applied Meteorology and
Climatology, 53(4), 840-858.
Samet, J., & Krewski, D. (2007). Health effects associated with exposure to
ambient air pollution. Journal of toxicology and environmental health, Part
A, 70(3-4), 227-242.
Samet, J. M., Zeger, S. L., Dominici, F., Curriero, F., Coursac, I., Dockery, D.
W., . . . Zanobetti, A. (2000). The national morbidity, mortality, and air pol-
lution study. Part II: morbidity and mortality from air pollution in the United
States Res Rep Health Eff Inst, 94(pt 2), 5-79.
Shin, S., Burnett, R. T., Kwong, J. C., Hystad, P., van Donkelaar, A., Brook, J. R.,
Chen, H. (2019). Ambient Air Pollution and the Risk of Atrial Fibrillation
and Stroke: A Population-Based Cohort Study. Environ Health
Perspect, 127(8), 87009. Retrieved from https://www.ncbi.nlm.nih.gov/
pubmed/31449466
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6792368/pdf/ehp-127-
087009.pdf. doi:10.1289/EHP4883
Tallon, L. A., Manjourides, J., Pun, V. C., Salhi, C., & Suh, H. (2017). Cog-
nitive impacts of ambient air pollution in the National Social Health and
Aging Project (NSHAP) cohort. Environ Int, 104, 102-109. Retrieved
from https://www.ncbi.nlm.nih.gov/pubmed/28392067. doi:10.1016/j.en-
vint.2017.03.019
Thurston, G. D., Kipen, H., Annesi-Maesano, I., Balmes, J., Brook, R. D., Cro-
mar, K., Frampton, M. W. (2017). A joint ERS/ATS policy statement: what
constitutes an adverse health effect of air pollution? An analytical frame-
work. European Respiratory Journal, 49(1).

192
Relationship Between Particulate Matter and Health...

Vermylen, J., Nemmar, A., Nemery, B., & Hoylaerts, M. F. (2005). Ambient
air pollution and acute myocardial infarction. Journal of Thrombosis and
Haemostasis, 3(9), 1955-1961.
Wang, W., Liu, C., Ying, Z., Lei, X., Wang, C., Huo, J., . . . Kan, H. (2019).
Particulate air pollution and ischemic stroke hospitalization: How the as-
sociations vary by constituents in Shanghai, China. Sci Total Environ, 695,
133780. Retrieved from https://www.ncbi.nlm.nih.gov/pubmed/31416039.
doi:10.1016/j.scitotenv.2019.133780
Ward, D., & Ayres, J. G. (2004). Particulate air pollution and panel studies in
children: a systematic review. Occupational and environmental medicine,
61(4), e13-e13.
Wing, J. J., Adar, S. D., Sanchez, B. N., Morgenstern, L. B., Smith, M. A., &
Lisabeth, L. D. (2017). Short-term exposures to ambient air pollution and
risk of recurrent ischemic stroke. Environ Res, 152, 304-307. Retrieved
from https://www.ncbi.nlm.nih.gov/pubmed/27829205. doi:10.1016/j.en-
vres.2016.11.001
Zhang, R., Liu, G., Jiang, Y., Li, G., Pan, Y., Wang, Y., . . . Wang, Y. (2018).
Acute Effects of Particulate Air Pollution on Ischemic Stroke and Hemor-
rhagic Stroke Mortality. Front Neurol, 9, 827. Retrieved from https://www.
ncbi.nlm.nih.gov/pubmed/30333790. doi:10.3389/fneur.2018.00827
Zimmerman, N., Presto, A. A., Kumar, S. P., Gu, J., Hauryliuk, A., Robinson,
E. S., . . . Subramanian, R. (2017). Closing the gap on lower cost air quality
monitoring: Machine learning calibration models to improve low-cost sensor
performance. Atmos. Meas. Tech. Discuss, 2017, 1-36.

 Petar Zhivkov
https://orcid.org/0000-0001-5687-5277
Institute of Information and Communication Technologies
Bulgarian Academy of Sciences
Sofia, Bulgaria
E-mail: [email protected]

 Aleksandar Simidchiev
https://orcid.org/0000-0003-4643-9388
Medical Institute of Ministry of Interior
Sofia, Bulgaria
E-mail: [email protected]

193
Proceeding of 1st International conference on Environmental protection and disaster RISKs

https://doi.org/10.48365/envr-2020.1.17 Air Pollution,


Climate and Health

NOCTURNAL BOUNDARY-LAYER
CHARACTERISTIC IN A COSTAL SITE USING
LONG-TERM SODAR DATA

Damyan Barantiev, Ekaterina Batchvarova


Climate, Atmosphere and Water Research Institute –
Bulgarian Academy of Sciences

Abstract: The coastal Atmospheric Boundary Layer (ABL) structure is subject


of research in many countries as large number of cities and industrial activities are
located onshore seas, oceans or lakes. Mean characteristics of ABL at a Bulgarian
Black Sea coastal site are obtained based on unique long-term (August 2008 -
October 2016) acoustic remote sensing measurements. Marine and land nocturnal
air flow vertical profiles are studied. Exploring data of wind and turbulence
profiles with high spatial (10 m) and temporal (10 min) resolution from SCINTEC
MFAS sodar revealed typical parameters as the height of the Internal Boundary
Layer (IBL), nocturnal ABL height over sea and land, etc. The seasonal variability
of coastal ABL characteristics is explored. The reported results can be used for
verification of theoretical, mesoscale and air quality models. The study contributes
to the understanding of wind regime and turbulent structure of ABL in a region with
modest observation networks and data near the surface and no measurements and
data in vertical direction.
Keywords: SCINTEC MFAS sodar, remote sensing data, Black Sea region, wind
profiles, turbulent profiles, vertical structure, climatological studies, coastal area.

INTRODUCTION
The recent significant technological development of ground-based instru-
ments for remote sensing made them most reliable and indispensable to study of
the main meteorological parameters and turbulence in the ABL (Cimini, Marzano,
& Visconti, 2011; Coulter & Kallistratova, 2004; Dirk, 2009; Emeis, 2011; A.
Illingworth et al., 2013; Peña et al., 2016). Substantial interest in coastal ABL is
noted, due to the complexity of the transformation processes at the abrupt change
in the physical characteristics of the surface. The challenges in describing coastal
processes are related to IBL formation and its height growing with the distance
from the shore (Batchvarova, 2006; Hsu, 1986). Due to the formation of sublayers
and the presence of local circulation in coastal zones, the atmosphere is complexly
194
Nocturnal Boundary-layer Characteristic in a Costal...

stratified resulting in different diffusion characteristics, and hence intricate pro-


cesses of air pollution dispersion compared to regions with homogeneous surface
(Simpson, 1994).
A number of scientific experiments with doppler lidars, sodars, high meteoro-
logical masts, surface and aerological measurements aim to provide data to evalu-
ate the mesometeorological models performance in coastal areas and to ensure fur-
ther development of parameterizations.
Wilczak, Dabberdt, and Kropfli (1991) present the results of a complex experi-
ment to study varying scales of airflows in California, with a center in Santa Bar-
bara. The experiment was conducted on September 20, 1985. Doppler wind lidars,
sodars, multiple stations with surface observations, and radio sounding were used.
IBL study in a complex coastal area was based on airborne lidar measurements,
mesometeorological modeling with CSU-RAMS and use of analytical models for
the Pasific 93 experiment, Vancouver, Canada. Good agreement of modeled and
measured IBL height was noted (Batchvarova, Cai, Gryning, & Steyn, 1999).
Batchvarova and Gryning (1998) investigate the IBL development in Athens
during MEDCAPHOT-TRACE 1994 experiment which helped air quality manage-
ment for the 2004 Olympic Games. Using tethered balloons, IBL height of 400 m
was found at 4 km, and of 700 m at 13 km inland.
De Leo et al. (2008) analyze sodar data and mesometeorological modeling of
the breeze circulation in the area of Lamezia Terme, at the Tyrrhenian Sea in Cal-
abria in the summer of 2007.
Prabha, Venkatesan, Mursch-Radlgruber, Rengarajan, and Jayanthi (2002) de-
scribe observations with sodar and the formation of a Thermal Internal Boundary
Layer (TIBL) during sea breeze in India for 10 days in February 1998 at transition
from winter to summer.
Petenko et al. (2016) present turbulent characteristics and ABL height at Con-
cordia Station (Dome C) in Antarctica. A sodar with an exceptional resolution of 2
m and a range of about 200 m was used.
Sodar measurements are used in a number of coastal studies around the world,
but mainly for short periods. The analysis of long-term data proposed in this
paper brings new information on the coastal ABL not only for Bulgaria. Except
for theoretical research, the created database can be used for regime studies and
evaluation of model performance in coastal areas.

MEASURING SITE AND EXPERIMENT EQUIPMENT


Meteorological Observatory (MO) Ahtopol is located at the Bulgarian Black
Sea coast (Figure 1 - middle) at about 2 km southeast of the town of Ahtopol. Well
expressed breeze circulation is typical in the region during warm seasons. Local
circulation of lower frequency and smaller temporal and spatial scales is registered
also during the cold season (Barantiev, 2017). MO Ahtopol is situated on a primarily
195
Damyan Barantiev, Ekaterina Batchvarova

flat grassland at about 400 m inland and 30 m height above sea level. The coast line
is stretching out from NNW to SSE with a steep about 10m high cost (Figure 1 –
left and middle). For studying the wind profiles and turbulence in coastal ABL, an
acoustic mono-static Doppler remote sensing system - SCINTEC Flat Array Sodar
MFAS was used with a frequency range of 1650-2750 Hz, 9 emission / reception
angles (0 °, ± 9.3 °, ± 15.6 °, ± 22.1 °, ± 29 °), a vertical range from150 m to 1000
m and a vertical resolution of 10 m. The accuracy of wind speed measurement is
0.1 - 0.3 ms-1 and for the wind direction is 2 - 3 Deg. The sodar system is mounted
on the roof of the administrative building of MO Ahtopol (Figure 1 - right) at an
approximate height of 4.5 m. The data records were made every 10 minutes and
the average period is 20 minutes. The first measurement level is 30 m and the
maximum vertical range for the study reaches 700 m.

Figure 1. Location of MO Ahtopol in southeastern Bulgaria (middle)


with views of the terrene (left) and sodar system on the roof of the
administrative building of MO Ahtopol (right)

DATA AND ANALYSIS OVERVIEW


Data availability
The exploration of coastal ABL characteristics in this work covers a 3014-day
period from 1 August 2008 to 31 October 2016. The continuity of sodar operation
was disturbed by frequent cuts of power supply until 2011 and prohibited sounding
during night in 2008 and 2009 summers.
The data availability is presented in Table 1. The low data availability (yellow
markings) is in summer of 2008 and 2009, because of no sodar operation during
nights. Data with availability above 70% is given in green, between 40 and 70% in
yellow, and below 40% in red. The actual sodar range is also presented graphically
in Table 1 with one bar filled at a height of 510 m, 2 bars over 610 m, 3 bars over
196
Nocturnal Boundary-layer Characteristic in a Costal...

710 m and 4 bars over 810 m. In this study all profiles are up to 700 m due to low
availability of data above that height.

Applied analyzes
For the presented analysis, the direction from 0 to 120 Deg was assigned for
marine air masses and direction from 170 to 290 Deg for the air masses from land.
The entire 3014-day study period contains total of 2708 days with measurements,
representing nearly 90% availability. А summary of the analyzes of the marine air
masses and those from the land are presented in Table 2.
Firstly, all nocturnal land (brown color) and marine (blue color) air masses are
studied, and then separately for cold and warm parts of the year to reveal specific
coastal ABL characteristics. The cold part of the years is defined from November to
March and the warm part from May to September (Table 2- first column). Due to the
presence of breeze circulation in the studied area, the number of profiles involved
in the averaged characteristics of land air masses (10.5 %) are significantly higher
than marine air masses profiles (2.5 %). The conditions fulfilled in the various
analyzes (fourth column at Table 2) are as follows:
 1* - continuous profiles to fixed heights with simultaneous availability of 12
sodar output parameters (wind direction /WD/, wind speed and its dispersion /
WS, sigWS/, vertical wind speed and its dispersion /W, sigW /, horizontal wind
speed components and their dispersions /U, sigU, V, sigV/, eddy dissipation rate
/EDR/, turbulent intensity /TI/ and turbulent kinetic energy /TKE/);
 2 - continuous profiles with a minimum height of 110 m and simultaneous
availability of 12 sodar output parameters;
 3 - profiles consisting of a minimum of 3 points in height satisfying the wind
direction condition and permitting an interruption only for lack of data.
The complexity of different fulfilled conditions in the extracted profiles (1*, 2
or 3) determines the number of profiles involved in the various analyzes (Table 2 -
third column) and the maximum height to which the average profiles reach (Table 2
- second column). The lowest availability of profiles is observed under the strongest
condition 1* - continuous profiles of the 12 different parameters simultaneously
to fixed heights (150 m, 250 m, 270 m, 300 m, 320 m, 350 m, 450 m and 550 m).
The complexity of fulfilling this condition lies in the fact that there can be heights
with missing data or such rejected by the signal quality algorithms of the instru-
ment. Using conditions 2 and 3 results in averaged profiles based on different data
availability at different heights, while the condition 1* implies the same numbers
of values across all heights. In this paper we analyze data fulfilling conditions 2 and
3, in order to use all data up to 700 m. During nights the range of the sodar allows
to search the ABL height in the data as it can be stable ABL over land or stable/
convective ABL over sea, while during days the convective ABL height over land
is well above the range of the instrument.
197
Damyan Barantiev, Ekaterina Batchvarova

Table 1. Monthly data availability and maximum effective height reached


by the sodar (01.08.2008 - 10.31.2016)
I II III IV V VI VII VIII IX X XI
2008 - - - - - - 45,2% 40,3% 57,7% 59,5%
max range [m] - - - - - - 520 520 520 520
2009 99,9% 99,6% 99,2% 96,7% 98,3% 99,3% 94,6% 57,5% 59,9% 58,4% 96,7%
max range [m] 520 520 680 680 680 680 680 680 680 680 680
2010 97,5% 98,3% 89,0% 68,5% 96,8% 86,4% 99,9% 98,3% 99,9% 92,6% 99,6%
max range [m] 680 680 680 680 680 680 560 510 510 510 510
2011 96,8% 94,8% 99,8% 96,7% 96,8% 96,2% 92,2% 99,9% 38,8% 78,8% 99,2%
max range [m] 510 460 510 510 510 510 510 510 560 620 620
2012 75,3% 95,8% 99,7% 96,7% 99,3% 93,3% 99,1% 31,9% 96,6% 30,6% 23,3%
max range [m] 620 620 620 620 620 620 700 620 720 670 640
2013 96,6% 96,4% 50,6% 100,0% 98,3% 54,7% 95,3% 96,3% 96,7% 58,0% 74,7%
max range [m] 720 720 720 620 680 670 680 680 720 720 720
2014 100,0% 99,9% 99,7% 100,0% 97,1% 99,8% 90,0% 59,4% 51,7% 98,5% 97,5%
max range [m] 720 720 720 720 720 720 720 720 750 750 750
2015 99,9% 96,3% 94,9% 97,8% 69,5% 73,3% 57,4% 97,9% 99,5% 99,9% 65,9%
max range [m] 750 750 750 750 750 750 590 750 750 750 750
2016 80,1% 70,0% 76,1% 49,9% 35,6% 93,2% 99,9% 93,0% 75,0% 45,4% -
max range [m] 750 750 750 750 750 1000 1000 1000 1000 750 -

AVERAGED VERTICAL PROFILES OF NOCTURNAL


COASTAL BOUNDARY LAYER
Estimating the ABL height through the turbulent profiles characteristics
(sigW, TKE, BP, etc.) has been suggested in recent years in studies based on
Table 2. Period of analysis from 08/2008 to 10/2016 (total - 3014 days). Days with observations - 2708 days

Long-term Analysis of land air masses


Height [m] Available profiles Conditions
analyzes of /170 ÷ 290 Deg/
Nocturnal air masses /21 ÷ 05h/ 700/ 630/ 550*/ 450*/ 45 549/ 8 671/ 39*/ 244*/ 3/ 2/ 1*/ 1*/
night coastal (45 549 profiles) 10.5 % 350*/ 320*/ 300*/ 1 388*/ 2 027*/ 2 571*/ 1*/ 1*/ 1*/
air masses by from the total period 270*/ 250*/ 150* 3 523*/ 4 278* / 8022* 1*/ 1*/ 1*
wind direction Nocturnal /November÷ March/ 700/ 630/ 550*/ 450*/ 18 616/ 4 458/ 21*/ 178*/ 3/ 2/ 1*/ 1*/

and different (18 616 profiles) 4.3 % 350*/ 320*/ 300*/ 799*/ 1 224*/ 1 596*/ 1*/ 1*/ 1*/

conditions. from the total period


Nocturnal /May ÷ September/
270*/ 250*/ 150*
730/ 610/ 550*/ 450*/
2 143*/ 2540* / 4262*
18 769/ 2 603/ 18*/ 37*/
1*/ 1*/ 1*
3/ 2/ 1*/ 1*/
(18 769 profiles) 4.3 % 350*/ 320*/ 300*/ 379*/ 504*/ 602*/ 1*/ 1*/ 1*/
from the total period 270*/ 250*/ 150* 811*/ 1 021* / 2 311* 1*/ 1*/ 1*
Analysis of marine air masses
Height [m] Available profiles Conditions
/0 ÷ 120 Deg/
Nocturnal air masses /21 ÷ 05h/ 700/ 590/ 550*/ 450*/ 10 919/ 3 731/ 3*/ 9* 3/ 2/ 1*/ 1*
(10 919 profiles) 2.5 % 350*/ 320*/ 300*/ 25*/ 55*/ 89*/ 1*/ 1*/ 1*
from the total period 270*/ 250*/ 150* 196*/ 400*/ 3 120* 1*/ 1*/ 1*
Nocturnal /November÷ March/ 700/ 340/ 3 939/ 1 141/ 3/ 2/
(3 937 profiles) 0.9 % 320*/ 300*/ 22*/ 36*/ 1*/ 1*/ 1*
from the total period 270*/ 250*/ 150* 91*/ 159* / 854* 1*/ 1*/ 1*
Nocturnal /May ÷ September/ 700/ 590/ 550*/ 450*/ 4 666/ 1 863/ 3*/ 9* 3/ 2/ 1*/ 1*
(4 666 profiles) 1.1 % 350*/ 320*/ 300*/ 21*/ 26*/ 35*/ 1*/ 1*/ 1*
from the total period 270*/ 250*/ 150* 54*/ 151* / 1 654* 1*/ 1*/ 1*

198
Nocturnal Boundary-layer Characteristic in a Costal...

data from remote sensing measurements (A. J. Illingworth et al., 2015; A.


Illingworth et al., 2013). In the absence of temperature and humidity profiles,
these analyzes allow to retrieve more information from wind data. While with
the lidars the calculations of sigW are based on a backscatter signals from aero-
sols, the measurement of turbulent characteristics with the sodars are direct. In
this paper, all profiles of the turbulent parameters are presented and explored
first as information from long-term measurements with high spatial and tempo-
ral resolution that are unique for Bulgaria and secondly, as the basis for retriev-
ing of ABL characteristics, including its height.

Nocturnal land air masses


■ All nocturnal land air masses and condition 3
Characteristics of all nocturnal land air masses are shown in Figure 2
(averaged 45549 individual profiles representing 10.5% of the total period
- Table 2) through 12 averaged profiles (color lines) and their dispersions
(green area). A color indication expresses the availability of the data at
each height. The lowest profiles availability was observed in 4 parameters
(sigWS, sigU, sigV and TI) due to the fact that they are calculated as a second
statistical moment. Linear increase of wind speed is observed up to 540 m
(1318 individual profiles available at that height). Slightly expressed peaks
in the shape of the sigW, EDR and TKE profiles are observed at 430 m with
respectively 3625, 1756 and 3011 individual profiles involved in the averaging.
This suggests nocturnal stable ABL height of about 430 m. The surface layer
(SL) height is detected between 50 and 100 m with characteristic changes in
sigW, EDR and TKE profiles.

■ Nocturnal land air masses and condition 2 during the warm part of the year
Averaged nocturnal land air masses (May to September) are presented in
Figure 3. The analysis with condition 2 is chosen to ensure availability of
data on all levels up to 110 m and all 12 sodar parameters simultaneously.
This set allows to identify peculiarity of signal at specific heights with higher
confidence. These individual profiles constitute about 30% of all registered
nocturnal land air masses (Table 2). Well expressed changes near 400 meters
are observed in all graphics. The observed peaks in the sigW, EDR and TKE
mean profiles are with maximum at 410 m (98 individual profiles). Almost a
linear decrease in the averaged TI profile is observed after the main peak at
140 m (2412 individual pro files) and it is interrupted by a second smaller peak
at 430 m (59 individual profiles). This result is the reason for the peak in the
mean sigW, EDR and TKE profiles in Figure 2 at 430 m. Close to the ground, at
a height between 50 and 80 m, changes in the sigW, EDR, TI and TKE profiles
suggest the SL height.
199
Damyan Barantiev, Ekaterina Batchvarova

Figure 2. Averaged nocturnal land air masses characteristics with performed


condition 3. Mean profiles and their dispersions from left to right and from top
to bottom: WD, WS, sigWS, W, sigW, U, sigU, V, sigV, EDR, TI, TKE

Figure 3. Averaged nocturnal land air masses characteristics during warm part
of the year with performed condition 2. Details as in Fig. 2
200
Nocturnal Boundary-layer Characteristic in a Costal...

Nocturnal marine air masses


■ Nocturnal marine air masses during the cold part of the year and condition 3
Up to 3937 individual profiles of the period November to March are used for
the averaged profiles (fulfilling condition 3) and their dispersion in Figure 4. Peak
at 300 m exists in sigW and TKE profiles (406 and 289 individual profiles). Such
a pronounced peak is also observed in EDR profile, but at 340 m (103 individual
profiles). The ABL height of the slightly unstable or neutral marine air mass during
the cold months is 300 - 340 m. Positive peak at 40 - 50 m in TI profile, supported
by the almost constant values of sigW up to 50 m and EDR up to 40 m, as well as
the observed weak peaks of WD, WS and W at 40 m can be associated with height
of IBL with a dominant factor the roughness surface change.

Figure 4. Averaged nocturnal marine air masses characteristics during cold


part of the year with performed condition 3. Details as in Fig. 2

CONCLUSIONS
The presented results for all nocturnal land air masses with applied condition 3
show a super position of all seasonal profiles with imposed minimum restriction in
their selection. Reducing the number of profiles only during the warm part of the
year with applied condition 2 confirmed the results and increased the confidence in
them. Choosing of the warm period suggests neutral and slightly stable air masses
above land, which allows to search ABL height within the range of the sodar. The
third analysis is performed at minimum restrictions (condition 3) and for the cold
201
Damyan Barantiev, Ekaterina Batchvarova

period when the sea is relatively warmer than the air and neutral or slightly unstable
stratification of the ABL is expected. During warm periods, due to the land breeze,
the number of marine nocturnal flows is very small.
The analysis is based on 82% of the data on nocturnal land air masses. The
nocturnal air masses from the land were characterized by prevailing conditions
between neutral and slightly stable stratification. Indications of a nocturnal stable
ABL height at 410-430 m and corresponding SL height of 50-80 m were identified.
Prevailing conditions of neutral stratified and slightly unstable stratification
were revealed in the results for nocturnal marine air masses. Indication of nocturnal
IBL height of about 40 – 50 m and a nocturnal marine ABL of about 300 m were
identified.
Such analyzes allow not only climatological studies for a number of ABL
parameters, but also assessing the fraction of time when theoretical profiles can be
used in coastal areas.

ACKNOWLEDGMENTS
This work was supported by the Bulgarian Ministry of Education and Science
under the National Research Programme “Young scientists and postdoctoral
students” approved by DCM # 577 / 17.08.2018

REFERENCES
Barantiev, D., Batchvarova, E. and Novitsky, M. (2017). Breeze circulation
classification in the coastal zone of the town of Ahtopol based on data
from ground based acoustic sounding and ultrasonic anemometer.
Bulgarian Journal of Meteorology and Hydrology (BJMH), 22(5), 24.
Batchvarova, E. (2006). Theoretical and experimental studies of the
atmospheric boundary layerover different surface types. (Dissertation
for obtaining the scientific degree Doctor of Physical Sciences (Doctor
Habil.). Scientific specialty: Meteorology 01.04.11), National Institute
of Meteorology and Hydrology (NIMH) and the Bulgarian Academy of
Sciences (BAS), Sofia.
Batchvarova, E., Cai, X., Gryning, S.-E., & Steyn, D. (1999).
Modelling internal boundary-layer development in a region with
a complex coastline. Boundary-Layer Meteorology, 90(1), 1-20.
doi:10.1023/A:1001751219627
Batchvarova, E., & Gryning, S.-E. (1998). Wind climatology, atmospheric
turbulence and internal boundary-layer development in Athens during the
MEDCAPHOT-TRACE experiment. Atmospheric Environment, 32(12),
2055–2069. doi:https://doi.org/10.1016/S1352-2310(97)00422-6
202
Nocturnal Boundary-layer Characteristic in a Costal...

Cimini, Domenico, Marzano, Frank S., & Visconti, Guido. (2011).


Integrated Ground-Based Observing Systems (Domenico Cimini, Frank
S. Marzano, & Guido Visconti Eds.): Springer-Verlag Berlin Heidelberg.
Coulter, R. L. , & Kallistratova, M. A. . (2004). Two decades of progress
in SODAR techniques: a review of 11 ISARS proceedings. Meteorology
and Atmospheric Physics, 85, 3-19. doi:https://doi.org/10.1007/s00703-
003-0030-2
De Leo, L., Federico, S., Sempreviva, A.M., Pasqualoni, L., Avolio, E. ,
& Bellecci, C. (2008). Study of the development of the sea breeze and
its micro-scale structure at a coastal site using a Multi -Tone Sodar
system. Paper presented at the 14th International Symposium for the
Advancement of Boundary Layer Remote Sensing, 23-25 June 2008,
Copenhagen, Denmark.
Dirk, A.M. Engelbart et al. (Eds). (2009). COST 720 - Final Report:
Integrated ground-based remote-sensing stations for atmospheric
profiling : . Luxembourg Office for Official Publication of the European
Communities.
Emeis, Stefan. (2011). Surface-based remote sensing of the atmospheric
boundary layer (1st ed.): Springer Netherlands.
Hsu, S. A. (1986). A note on estimating the height of the convective internal
boundary layer near shore. Boundary-Layer Meteorology, 35(4), pp 311-
316. doi:https://doi.org/10.1007/BF00118561
Illingworth, A. J., Cimini, D., Gaffard, C., Haeffelin, M., Lehmann, V.,
Löhnert, U., . . . Ruffieux, D. (2015). Exploiting Existing Ground-
Based Remote Sensing Networks to Improve High-Resolution Weather
Forecasts. Bulletin of the American Meteorological Society, 96(12),
2107-2125. doi:http://dx.doi.org/10.1175/BAMS-D-13-00283.1
Illingworth, A., Ruffieux, D., Cimini, D., Lohnert, U., Haeffelin, M., &
Lehmann, V. (2013). COST Action ES0702 Final Report: European
Ground-Based Observations of Essential Variables for Climate and
Operational Meteorology. Retrieved from COST Office, PUB1062: http://
cfa.aquila.infn.it/wiki.eg-climet.org/Final_Report_EG_CLIMET.pdf
Peña, Alfredo, Floors, R. R., Sathe, A., Gryning, S-E., Wagner, R., Courtney,
M., . . . Hasager, C.B. (2016). Ten Years of Boundary-Layer and Wind-
Power Meteorology at Høvsøre, Denmark. Boundary-Layer Meteorology,
158(1), 1-26. doi:https://doi.org/10.1007/s10546-015-0079-8
Petenko, Igor , Argentini, Stefania , Casasanta, Giampietro , Kallistratova,
Margarita , Sozzi, Roberto , & Viola, Angelo (2016). Wavelike Structures
in the Turbulent Layer During the Morning Development of Convection
at Dome C, Antarctica. Boundary-Layer Meteorology, 1–19. doi: https://
doi.org/10.1007/s10546-016-0173-6

203
Damyan Barantiev, Ekaterina Batchvarova

Prabha, T. V. , Venkatesan, R., Mursch-Radlgruber, E., Rengarajan, G. , &


Jayanthi, N. . (2002). Thermal Internal Boundary Layer characteristics
at a tropical coastal site as observed by a mini-SODAR under varyng
synoptic conditoins. Journal of Earth System Science, 111(1), 63-77.
doi:https://doi.org/10.1007/BF02702223
Simpson, J. E. . (1994). Sea breeze and local winds. Cambridge, UK:
Cambridge University Press.
Wilczak, J. M., Dabberdt, W. F., & Kropfli, R. A. (1991). Observations and
Numerical Model Simulations of the Atmospheric Boundary Layer in the
Santa Barbara Coastal Region. Journal of Applied Meteorology, 30(5),
652–673. doi:https://doi.org/10.1175/1520-0450(1991)030<0652:OAN
MSO>2.0.CO;2

 Damyan Barantiev
https://orcid.org/0000-0001-9908-9014
Climate, Atmosphere and Water Research Institute
Bulgarian Academy of Sciences
Sofia, Bulgaria
E-mail: [email protected]

 Ekaterina Batchvarova
https://orcid.org/0000-0002-1293-9440
Climate, Atmosphere and Water Research Institute
Bulgarian Academy of Sciences
Sofia, Bulgaria
E-mail: [email protected]

204
Proceeding of 1st International conference on Environmental protection and disaster RISKs

https://doi.org/10.48365/envr-2020.1.18 Air Pollution,


Climate and Health

ENVIRONMENTAL IMPACT OF QUARRY


ACTIVITIES: THE CASE STUDY OF A GNEISS
DEPOSIT IN THE REGION OF BLAGOEVGRAD

Radostina A. Angelova1, Rositsa Velichkova1,


Sylvia Alexandrova2
1
Centre for Research and Design in Human Comfort,
Energy and Environment (CERDECEN),
Technical University of Sofia
2
Department of Hydroaerodynamics and Hydraulic Machines,
Technical University of Sofia

Abstract. The extraction of rock materials from the landscape is accompanied


by severe influence on the nature. The research presents the assessment of the
impact of the quarry activities in a gneiss deposit in the region of Blagoevgrad on
the environment. The effect of the blasting operations and other quarry works are
discussed in terms of the geomorphic impact, air and water pollution, noise, and
influence on biota. Some measures for improvement of the quarry works and further
quarry reclamation are suggested.
Keywords: quarry activities; blasting operations; environmental impact

INTRODUCTION
The nowadays knowledge on quarry activities, based on modern technologies
and research, has significantly reduced the effect of quarries on the environment.
However, the extraction of rock materials from the landscape cannot be performed
without causing environmental consequences.
The engineering processes, related to the extraction and treatment of the rock
materials, leads to a change in the use of the land, scenery, and geomorphology.
Responsible operational practices can be used to control and keep at tolerable lev-
els the engineering impacts (Luttig, 1994; Langer & Arbogast, 2002; Jin, Chen &
Soboyejo, 2015; Edokpayi, Odiyo & Shikwambana, 2016). However, the environ-
mental effects could have a cascade form: the engineering activities provoke a reac-
tion of the natural system, which, in turns, leads to another environmental reaction,
and on and on. Cascading effects could influence the environment beyond the limits
of the quarry activities (Langer, 2001).

205
Radostina Angelova, Rositsa Velichkova, Sylvia Alexandrova

Geomorphic impact
The principal geomorphic impact of quarrying is associated with rocks and
stones removal. It results in a dramatic effect on the landscape, destroys the natural
habitat of relicts, affects natural sinkholes and caves (Dávid, 2010). The size of the
geomorphic impact depends on the location of the quarry and its size. The disad-
vantage is that all quarries increase in size over time. Stanton (1996) researched
the advantages of deep quarrying, as it was found that the geomorphic impact of
numerous small quarries is worse than the effect of a large size quarry.
Blasting strongly preconditions the geomorphic impact of quarrying. Due to
poor control or design of the blasts, stones could be projected far beyond the blast
site that is considered to be a serious hazard (Langer, 2001). Besides the rock crush
from the quarry face, the blast energy provokes vibrations in both the ground and
air. The earth vibrations could lead to ground shaking and additional geomorphic
impact: fracture of the walls of the quarry, collapse of caves, change in the sub wa-
ter trajectory. Blasting could initiate even flooding events (Tan et al., 2020).

Air pollution
Air pollution in the form of dust particles of different size occurs due to blasting and
both digging and crushing the rocks. Dust is one of the most visible and potentially irri-
tating impacts of the quarry on workers, nearby settlements, and the ecosystem (Sayara,
2016). The amount of air pollution depends on the type of rocks, humidity, scale of the
excavation works, air currents and prevailing winds, as well as the presence of other
nearby sources of dust. As the quarry terrain recedes, the effects of dust air pollution,
dust deposition and hazardous impacts rapidly diminish (Gunn & Gagen, 1987).

Noise
The main source of quarry noise is blasting, drilling, the processing machines,
and stones-moving machines. Truck traffic can also be an important source of the
noise. The impact of noise strongly depends on the sound source, terrain, climatic
conditions. The noise is more noticeable in cold and humid days than in hot sum-
mer days (Angotzi et al., 2005) In an urban or industrial environment, background
noise can mask the noise of quarry operation, while the same noise level in a rural
or quiet residential suburb may be more noticeable to humans. The noise could
adversely affect biota (Christian, 2003).

Water impact
Usually, the first impact of the quarry on the landscape is the removal of superfi-
cial vegetation and soil. In temperate areas, this reduces evaporation and increases
effective rainfall. Excavation works can direct surface water to the groundwater
system. If runoff and sediment control measures are not taken, groundwater dete-
rioration is possible (Gunn & Hobbs, 1999).
206
Environmental Impact of Quarry Activities: The Case Study...

The quarry can cause the collapse of the sinkhole, which can lead to the capture of
surface water. At the same time, blasting operations can close existing groundwater pas-
sages or open a new passageway, leading to a change in the direction of groundwater
and risk of contamination. Blasting itself does not necessarily affect groundwater qual-
ity: Moore and Hughes (1979) found that there was no relationship between blasting
and water quality in springs. Large amounts of slime and other quarry products (fuel,
waste, oil) can pollute both rivers and groundwater far beyond the quarry field.

Biota
Any species that inhabit the rocks, destroyed by the quarry, are at risk. Some
species are limited to special habitats (e.g. single cave systems) and are little known.
White et al. (1995) reported the discovery of 47 species of aquatic and terrestrial
invertebrates in the Movile Cave and its surrounding springs, 30 of which have
been completely unknown and are considered endemic.
The quarry destroys the rocky habitats or the passages to them. Only animals that are
mobile and can find new homes survive. The other species die. The quarry can disturb
active groundwater courses or lead to their blockage with adverse consequences for the
species. The risk to groundwater and the diversion of surface water may lead to the dry-
ing up of surface and groundwater hydrological systems with the relevant biota.
It has been found that noise and air concussions can disrupt habitats at a dis-
tance of 1500 m from the quarry (Vermeulen & Whitten, 1999). Noise interferes with
communication signals between animals, masks the sounds of predators and prey,
and can even cause temporary or permanent hearing loss (Fletcher & Busnel, 1978).
In dry weather, dust, if uncontrolled, can spread through the air, penetrate the
soil, and create harmful conditions for flora and fauna (Fletcher & Busnel, 1978).
If it suffocates the leaf surface, the vegetation can be damaged due to inhibited gas
exchange and reduced photosynthesis (Howard & Cameron, 1998).
The present study was performed in a real quarry for extraction of gneiss:
the “Buchaka” quarry, near the city of Blagoevgrad. The aim of the study was to
assess the impact of the quarry on the environment and to suggest activities for
improvement of the present situation.

GENERAL DESCRIPTION OF THE DEPOSIT


Location and weather conditions
“Buchaka” quarry for making crushed stone for road pavements is located about
2 km northeast of the village of Buchino, Blagoevgrad municipality. According to
data for inhabitants at a permanent and current address, 61 persons live in the village,
which is situated at approx. 10 km from the city of Blagoevgrad – Fig. 1 (Google
Maps, 2020). The quarry occupies part of the eastern side of the hill of the same name.
The quarry has final area dimensions of 300 x 200 m. The terrain is extremely
mountainous, rocky and sparsely forested - only in the northern slope, there are
207
Radostina Angelova, Rositsa Velichkova, Sylvia Alexandrova

single low-stemmed trees and shrubs. The slope of the terrain is about 30 - 40 ° in
the southeast direction. The displacement between the lowest and highest parts of
the deposit is 50 m on average. Its altitude varies between 350 - 410 m. The lowest
exploitation level of the deposit is above the level of the local erosion base (Struma
River that flows in the vicinity, east of the deposit). The dominated cracked and
broken rocks facilitate the drain of the fallen atmospheric precipitation in depth.

Figure 1. Location of the village of Buchino (Google Maps, 2020)

Quality of the gneisses


The geological structure of the deposit involves biotite and two-mica gneisses
and gneiss-schists of Precambrian age, overlapped in places with Quaternary clays
mixed with rock fragments. The two-mica striped gneisses with quartz veinlets
(10-15 m thickness) predominate. According to the assessment of the Ministry of
Environment and Water (Report NB-12, 2007), the quantity of the inert materials
in the “Buchaka” deposit is as mentioned in Table 1.

Table 1. Amount of inert materials


Field number Opencast Deposit
(m3) (m3)
1 7292 148565
2 36921 1279945
Total 44213 1428510

The gneisses (Fig. 2) can be used for making road groundworks (not treated
with binders) for very light, light, and medium category of traffic. They could also
208
Environmental Impact of Quarry Activities: The Case Study...

be involved in asphalt mixtures, as the material corresponds to the requirements of


the BDS EN 13043 (2017) standard. The extracted from the deposit crushed gneiss
sand is not suitable for the production of concrete and mortars, as it does not cor-
respond to the requirements of the BDS EN 12620 (2017) standard.

Engineering characteristics
The characteristics of the excavation works in the quarry are as follows:
• Height of the working step:
○ up to 17 m for level 390 m;
○ 20 m for level 370 m;
○ up to 14 m for level 356 m.
• Angle of inclination of the working step: 75°;
• Angle of inclination of the final non-working board: 65°;
• General angle of inclination on a non-working board: 51°;
• Minimum width of the work site: 40 m;
• Width of the safety berms: 5 m;
• Slope for gravity drainage of the levels: 1% to the south-east.

Figure 2. The processed material from the quarry

ASSESSMENT OF THE IMPACT OF THE QUARRY ON


THE ENVIRONMENT AND SUGGESTIONS FOR IMPROVEMENT
Blasting operations
The blasting works in the “Buchaka” quarry are carried out by inclined drilling
from the upper horizon. They are divided into:
209
Radostina Angelova, Rositsa Velichkova, Sylvia Alexandrova

• primary works – with the help of which the mineral is diverted from the
massif;
• secondary – for additional fragmentation of the oversized pieces obtained
during the primary blasting.
The primary blasting works are carried out with via blasting wells. Secondary
drilling and blasting works are performed by open charges or blasting holes
(depending on the size of the oversized piece) or are crushed with a hydraulic
excavator-hammer. The annual opencast in the “Buchaka” quarry is 9180 m3, which
corresponds to 75 m3 per working day.
Before the explosion, a signal is given with a siren. After the explosion, but not
earlier than 30 minutes, the detonation field is inspected for failures. People and
equipment are allowed working in the field no earlier than 14 hours after the blasting.
The nearest neighbouring objects are outside the range of the safe distance of
400.0 m. Due to the proximity of the gravel pit, the work there during the blasting
works stops and both people and machinery are evacuated outside the safe area.
The posts guarding the blasting works are located outside the protected zone with a
radius in all directions of the blasting field of not less than 400.0 m.
To prevent gassing after blasting, the safety procedures in the quarry state that
people should enter the quarry after a mandatory measurement with a gas analyzer.
In addition, it is recommended to carry out mass blasting once a quarter – on the
last day of the working week.
A systematic survey of the behaviour of the slopes is organized to establish movements
in the working sides of the quarry and to take effective measures to prevent landslides or
falling of individual rocks. To protect the neighbouring agricultural and forest lands from
disturbance and pollution, no movement of machines and disposal of materials and waste
is allowed on them. The produced crushed fractions and the technological waste from the
processing of rock materials are placed at designated places.
On the 26th of September 2017 (Tuesday) the blasting works in the quarry
however led to effect of a shock wave, which was felt in the city of Blagoevgrad
in the form of an earthquake (http://pirinsko.com). The check in the European-
Mediterranean Seismological Centre (EMSC) showed that there were no registered
earthquakes in the region that could be felt in city of Blagoevgrad. The further
questioning showed that the population of the Buchino village complained en
masse about the explosions in the quarry. People also commented their worry about
the possible change in the sub water trajectory, as all households in the village are
supplied with water from underground sources. The real water impact of the quarry
is commented below.

Air pollution
Blasting operations and the additional on-site fragmentation of the oversized
pieces (Fig. 3) provoke the appearance of in-air particles of different size. The
210
Environmental Impact of Quarry Activities: The Case Study...

access of dust particles into the human lungs depends on their physical size and
chemical composition (MacNee & Donaldson, 2003). Large particles may not be
able to penetrate the upper respiratory system. Most particles with an aerodynamic
diameter of over 10 μm are thought to be unlikely to enter the airways of the lungs
(Tranter, 2004).
As the air-pollution is strongly related to the humidity, irrigation of both work
sites and transport routes in the quarry are performed during each shift. The ban
on access 14 hours after the blasting operations leads to the deposition of dust
particles before the presence of workers in the quarry. Mining equipment and
vehicles that do not meet the requirements for the concentration of pollutants in
the exhaust gases are not allowed to operate. All the equipment passes annual
technical inspections.

Figure 3. On-site machine operations

Noise
The quarry and its management are responsible for assuring that the emitted
noise is within the levels, set by regulations. The quarry site is located away from
populated areas. However, the relief could hardly be used as a sound-deadening
structure. То decrease the noise of transportation, conveyors could be used for in-
pit movement of materials.
Noisy operations could be limited to certain times of the day. Similar to blasting,
the noisy operations could be scheduled. A problem is the very close location of the
in-pit cite to the village (Fig. 4), which could be seen from Fig. Personal protective
211
Radostina Angelova, Rositsa Velichkova, Sylvia Alexandrova

equipment is foreseen for all workers in the quarry, including the drivers of the
heavy machinery and trucks.

Figure 4. In-pit site: the neighbour houses are visible

Water impact
The quarry exploits the high parts of the sloping hill, which is intensely cracked,
cavernous in places. Аt the establishment of the quarry, the risk assessment for the
water resources were assessed as low. During the exploration works, no outflows
of water or water levels in the deposit were recorded. Fallen precipitation quickly
drains to a depth below the lowest operating level on the horizon + 390 m. The
hydrogeological conditions could be defined as simple.

Biota
A preliminary assessment of the effect of the quarry on the biota had not been
performed. In any case, the quarry works have changed the habitat of some animals,
due to the human presence, blasting and noise. The effect on plants is smaller, as the
low-strength soil layer is 5-10 cm thick and disappears on several places. However,
all species, inhabiting the rocks, are at risk.
The start of reclamation is considered to be the beginning of the end of a career's
environmental impact (Langer, 2001). Just as the natural resource of the quarry is
used to improve the quality of human life, so the reclamation of the landscape im-
proves the quality of human life. A design plan is needed for the good formation of
the land from the field. There are examples of the use of quarries for recreational,
tourist and even residential purposes [26]. But for the biota, it is much more impor-
tant to recultivate the massif. There are examples of organized reclamation of rock
212
Environmental Impact of Quarry Activities: The Case Study...

quarries to almost natural conditions. There is also evidence of natural ecosystem


restoration in abandoned quarries after a period of 20 to 100 years.

CONCLUSIONS
The performed analysis highlighted the impact of “Buchaka” quarry works
on the environment. The assessment showed that the quarry has a considerable
impact on the natural landscape, natural resources, and biota. However, the right
management and planning could minimize the negative effects and put them into
control. The analysis showed that most of the prescriptions of the legislation are
followed in “Buchaka” quarry. However, blasting operations and noise from drilling,
machinery works, and trucks are the main sources of risks for the population in the
vicinity. The impact on biota is still not highlighted enough. Plans for the further
re-cultivation of the area should be elaborated.

ACKNOWLEDGMENTS
This work has been carried out in the framework of the National Science Program
“Environmental Protection and Reduction of Risks of Adverse Events and
Natural Disasters”, approved by the Resolution of the Council of Ministers №
577/17.08.2018 and supported by the Ministry of Education and Science (MES)
of Bulgaria (Agreement № ДО-230/06-12-2018).

REFERENCES
Angotzi, G., Bramanti, L., Tavarini, D., Gragnani, M., Cassiodoro, L.,
Moriconi, L., ... & Bovenzi, M. (2005). World at work: Marble quarrying
in Tuscany. Occupational and environmental medicine, 62(6), 417 – 421.
Arbogast, B. F., Knepper, D. H., & Langer, W. H. (2000). The human factor
in mining reclamation (Vol. 1191). US Department of the Interior, US
Geological Survey.
BDS EN 12620:2002+A1:2008 (2017). Aggregates for concrete – National
Annex (NA)
BDS EN 13043:2005+AC:2005 (2017). Aggregates for bituminous
mixtures and surface treatments for roads, airfields and other trafficked
areas – National Annex.
Christian, J. (2003). White’s Cove quarry blasting: potential impacts on
American lobster. LGL Limited, St. John’s, NL.
Dávid, L. (2010). Quarrying and other minerals. In Anthropogenic
Geomorphology, 113 – 130, Springer, Dordrecht.
Edokpayi, J. N., Odiyo, J. O., & Shikwambana, P. P. (2016). Seasonal
variation of the impact of mining activities on Ga-Selati River in
Limpopo Province, South Africa. International Journal of Biological
and Ecological Engineering, 10(2), 156 – 161.
213
Radostina Angelova, Rositsa Velichkova, Sylvia Alexandrova

Fletcher, J. L., & Busnel, R. G. (Eds.). (1978). Effects of noise on wildlife.


Academic Press.
Google Maps, 2020. Buchino. Imagery @2020 CNES / Airbus, Maxar
Technologies, Map data @2020 [Accessed 19 August 2020].
Gunn, J., & Gagen, P. (1987). Limestone quarrying and sinkhole development
in the English Peak District. In Multidisciplinary conference on sinkholes
and the environmental impacts of karst. 2, 121 – 125.
Gunn, J., & Hobbs, S. L. (1999). Limestone quarrying: hydrogeological
impacts, consequences, implications. Karst hydrology and human
activities–Impacts, Consequences and Implications Handbook,
192 – 201.
Howard, B., & Cameron, I. (1998). Dust control: Best Practice Environmental
Management in Mining. Environment Australia.
http://pirinsko.com/moshten-grym-ot-kariera-na-agromah-raztrese-blago-
evgrad-36299.html
Jin, R., Chen, Q., & Soboyejo, A. (2015). Survey of the current status of
sustainable concrete production in the US. Resources, Conservation and
Recycling, 105, 148 – 159.
Langer, W. H. (2001). Potential environmental impacts of quarrying
stone in karst: a literature review. US Department of the Interior, US
Geological Survey.
Langer, W. H., & Arbogast, B. F. (2002). Environmental impacts of mining
natural aggregate. In Deposit and geoenvironmental models for resource
exploitation and environmental security, 151 – 169, Springer, Dordrecht.
Luttig, G. W. (1994). Rational management of the geo-environment. In
Aggregates-Raw Malerials' Giant Report on the 2n" International
Aggregate Symposium (pp. 1 – 34).
MacNee, W., & Donaldson, K. (2003). Mechanism of lung injury caused by
PM10 and ultrafine particles with special reference to COPD. European
Respiratory Journal, 21(40 suppl), 47 – 51.
Moore, J. D., & Hughes, T. H. (1979). Effect of quarry blasting on ground-
water quality in a limestone terrane in Calhoun County. Alabama: Water
Resources Research Institute Bulletin, 38.
Report NB-12 (2007). Report from 07.03.2007, Ministry of Environment
and Water, Republic of Bulgaria.
Sayara, T. (2016). Environmental impact assessment of quarries and
stone cutting industries in Palestine: case study of Jammain. Journal of
Environment Protection and Sustainable Development, 2(4), 32 – 38.
Stanton, W. (1966). The impact of quarrying on the Mendip Hills.
Somerset: Proceedings of the University of Bristol Speleological
Society, 11, 54 – 62.

214
Environmental Impact of Quarry Activities: The Case Study...

Tan, Y., Zhang, S., Guo, W., Liu, X., & Shen, B. (2020). Mine Stability and
Water Inrush in Coal Mine. In Modelling Rock Fracturing Processes,
361 – 403. Springer, Cham.
Tranter, M. (2004). Occupational hygiene and risk management. Allen &
Unwin.
Vermeulen, J., & Whitten, T. (1999). Biodiversity and cultural property
in the management of limestone resources: lessons from East Asia. The
World Bank.
White, W. B., Culver, D. C., Herman, J. S., Kane, T. C., & Mylroie, J. E.
(1995). Karst lands. American scientist, 83(5), 450 – 459.

 Radostina A. Angelova
https://orcid.org/0000-0002-6943-7695
Centre for Research and Design in Human Comfort, Energy and Environment
Technical University of Sofia
Sofia, Bulgaria
E-mail: [email protected]

 Rositsa Velichkova
http://orcid.org/0000-0003-3757-8685
Centre for Research and Design in Human Comfort, Energy and Environment
Technical University of Sofia
Sofia, Bulgaria
E-mail: [email protected]

 Sylvia Alexandrova
Department of Hydroaerodynamics and Hydraulic Machines
Technical University of Sofia
Sofia, Bulgaria
E-mail: [email protected]

215
Proceeding of 1st International conference on Environmental protection and disaster RISKs

https://doi.org/10.48365/envr-2020.1.19 Biodiversity

PENIOPHORA PROXIMA (PENIOPHORACEAE,


BASIDIOMYCOTA): NEWLY RECORDED
FOR THE BULGARIAN MYCOTA

Aneta Lambevska-Hristova, Svetlana Bancheva


Department of Plant and Fungal Diversity and Resources,
Institute of Biodiversity and Ecosystem Research – Bulgarian Academy of Sciences

Abstract: Peniophora proxima Bres. is corticioid fungus belonging to family


Peniophoraceae (Basidiomycota). It is easily distinguished by the numerous
encrusted cystidia, gloeocystidia and ellipsoid spores. The species is considered
a common white rot saprobe restricted only to the Buxus representatives.
One of the most popular species of the Old world is Buxus sempervirens L.
(European boxwood), native to western and southern Europe, northwest Africa,
and southwest Asia. Although the species does not belong to the native flora of
Bulgaria, for many years, the European boxwoods have been one of the most
popular and widely used landscaping plants, in the form of ornamental shrubs
or dense, evergreen hedges in shady places. P. proxima has been reported from
different parts of Europe, but it has not been previously reported for Bulgaria.
This article reports this new find from three localities in Bulgaria, while providing
information about the morphological and ecological features of the species and
commenting on the pattern of its distribution.
Keywords: first record, Bulgarian mycota, Buxus sempervirens, lignicolous
fungi.

INTRODUCTION
Buxus L. (with common names box or boxwood) is a genus of about 100 species
in the family Buxaceae, native to western and southern Europe, southwest, southern
and eastern Asia, Africa, Madagascar, northernmost South America, Central
America, Mexico and the Caribbean (http://www.theplantlist.org1). One of the most
popular species of the Old world is Buxus sempervirens L. (European box), native
to western and southern Europe, northwest Africa, and southwest Asia. It represents
a small tree or shrub with small evergreen leaves and can reach 800 years of age.
Although the species does not belong to the native flora of Bulgaria, for many
years, boxwoods have been one of the most popular and widely used landscaping
plants, in the form of ornamental shrubs or dense, evergreen hedges in shady places
216
Peniophora Proxima (Peniophoraceae, Basidiomycota)...

(Jordanov, 1979). It is found in parks, gardens and yards throughout the country.
In the past, boxwood has been used also to treat persistent and recurrent fever
due to malaria, gout, rheumatism, urinary tract infections, hemorrhoids (Tashev
& Tsavkov, 2008) and in some in folkloric traditions and customs (Boycheva &
Marinova, 2017).
B. sempervirens can be a host to other organisms that use it as a substrate on
which they grow (likes the fungus Peniophora proxima Bres.) or feed on parts of
it and cause damage by destroying its tissues and organs, the lepidoptera Cydalima
perspectalis (Walker, 1859).
Mordecai Cubitt Cooke (1879) described Peniophora as the first genus of
corticioid fungi based on microscopic features (incrusted cystidia). According to
Andreasen & Hallenberg (2009) it is a member of the section Gloeopeniophora
Höhn. & Litsch. (with P. incarnata-group) which possess ellipsoid to narrow
ellipsoid spores.
P. proxima belongs to the corticoid fungi with restricted range that follows
the spread of B. sempervirens. The species is known only from Buxus, which
makes it easy for determination (Bernicchia & Gorjón, 2010). In addition, it
can be easily recognized by the numerous encrusted cystidia, gloeocystidia and
ellipsoid spores (Lambevska, Rusevska & Karadelev, 2013). P. proxima has
been reported from different parts of Europe, but it has not been previously
reported for Bulgaria.
The box tree moth [Cydalima perspectalis (Walker, 1859) (Lep.: Crambidae)]
is an invasive species of pest, native to Japan, China and Korea, but from 2006 was
introduced accidentally in Europe where it spreads extremely rapidly on ornamen-
tal Buxus communities (Leraut, 2012). The larvae feed on the leaves and shoots and
causes very serious damage. It was registered as new for Europe by Billen (2007)
in Germany, where the species was probably introduced in 2005, while his first
finding in Bulgaria was in 2014 by Beshkov, Abadjiev & Dimitrov (2015) in three
remote locations.

Material and methods


For identification of lignicolous fungi standard methods have been applied,
implying microscoping and application of reagents (Melzer’s reagent and 5%
KOH). Measurements and photographs were examined at magnification up to
1000×, with a LW scientific microscope and MiniVID camera.
The specimens were identified by the following sources: Еriksson & Ryvarden
(1973); Breitenbach & Kränzlin (1986); Andreasen & Hallenberg (2009);
Bernicchia & Gorjón (2010), Yurchenko (2010). The nomenclature follows
Index Fungorum2 and Mycobank3. The studied specimens were deposited in the
Mycological Collection of the Institute of Biodiversity and Ecosystem Research,
Bulgarian Academy of Sciences (SOMF).
217
Aneta Lambevska-Hristova, Svetlana Bancheva

Results and discussion


During field research in the natural habitats of the genus Juniperus L. and in
some parks in Bulgaria, in connection with a project, in connection with a project
under the National Research Programme “Young scientists and postdoctoral stu-
dents” of the Bulgarian Ministry of Education and Science, a species of corticoid
fungi, Peniophora proxima Bres., was accidentally found on the Buxus sempervi-
rens, which later turned out to be new to the Bulgarian mycota.

Peniophora proxima Bres., Bull. Soc. Mycol. Fr. 28(4): 402 (1912).
Morphological description. Basidiome resupinate, effused, hymenophore
smooth to tuberculate, reddish-pink, margin loosening from the substrate, hyphal
system monomitic, hyphae with clamps, hyaline, thin-walled, somewhat aggluti-
nated, gloeocystidia cylindrical to fusiform, thin-walled, lamprocystidia numerous,
fusiform, 15-40 x 5-7 μm, basidia up to 50 x 10 µm, with 4-sterigmata, and with a
basal clamp, basidiospores ellipsoid, thin-walled, smooth, 8-14 x 6-7 µm, hyaline
(Fig. 1, A-B).
Ecological remarks and host. It is known to be exclusively found on B.
sempervirens.
General distribution. P. proxima has been reported from different parts of Europe,
Russia, France, Poland, Hungary, Belgium, United Kingdom, Slovakia, Portugal,
Turkey, Sweden, Denmark, Norway, Finland, Ukraine, Macedonia, Turkey and the
Caucasus (Bernicchia & Gorjón, 2010; Lambevska, Rusevska & Karadelev, 2013).
Distribution in Bulgaria. Sofia region and Central Balkan Range.
Material examined. Sofia Valley, Sofia city, “King Boris’Garden” park, ur-
ban vegetation, old tree on B. sempervirens, N 42.68861° E 23.33953°, 545 m,

Fig. 1. A) Macroscopic view of basidiome of P. proxima; B) Microscopical


features of P. proxima: section of basidiome, a) presence of numerous incrusted
cystidia (lamprocystidia), b) gloeocystidia with glanular content, c) basidiospores.
218
Peniophora Proxima (Peniophoraceae, Basidiomycota)...

28.VI.2020, leg. A. Lambevska-Hristova & G. Hristov, det. A. Lambevska-Hristova


(SOMF 30 274); Park-Museum “Vrana”, urban vegetation, old tree on B. sempervi-
rens, N 42.639763° E 23.433025°, 572 m, 04.IV.2018, leg. A. Lambevska-Hristova
& G. Hristov, det. A. Lambevska-Hristova (SOMF 30 275); Stara Planina Mt.,
Bozhentsi village, urban vegetation, old tree on B. sempervirens, N 42.872408° E
25.424716°, 531 m, 01.III.2020, leg. A. Lambevska-Hristova & G. Hristov, det. A.
Lambevska-Hristova (SOMF 30 276).
P. proxima is microscopically easily diagnosed by its typical ellipsoid spores,
gloeocystidia and the numerous encrusted cystidia. It is considered a common
white rot saprobe that is apparently restricted to Buxus sp. (Lambevska, Rusevska
& Karadelev, 2013). Bondartseva & Seman (1978) mentioned it as destructor of
processed wood materials.
Up to now, P. proxima is known only from threе localities in Bulgaria: two in
Sofia region, “Vrana” and “King Boris’Garden” parks and one in Central Balkan
Range in the Bozhentsi village.

ACKNOWLEDGEMENTS
Financial support by the Bulgarian Ministry of Education and Science under
the National Research Programme “Young scientists and postdoctoral students”
approved by DCM # 577 /17.08.2018 is gratefully acknowledged.

NOTE:
1. http://www.theplantlist.org/
2. http://www.indexfungorum.org/Names/Names.asp
3. http://www.mycobank.org/

REFERENCES
Andreasen, M. & Hallenberg, N. (2009). A taxonomic survey of the
Peniophoraceae. Synopsis Fungorum, 26, 56-119.
Bernicchia, A. & Gorjón, S.P. (eds.) (2010). Corticiaceae s.l. Fungi Euro-
paei. Edizioni Candusso, Italia.
Beshkov, S., Abadjiev, S. & Dimitrov, D. (2015). Cydalima perspectalis
(Walker, 1859) (Lepidoptera: Crambidae: Spilomelinae) – new invasive
pest moth in Bulgaria. The Еntomologist's Record and Journal of Varia-
tion, 127, 18-22.
Billen, W. (2007). Diaphania perspectalis (Lepidoptera: Pyralidae) – a new
moth in Europe. Mitteilungen der Entomologischen Gesellschaft Basel,
57(2/4), 135-137.
219
Aneta Lambevska-Hristova, Svetlana Bancheva

Bondartseva, M. A. & Seman, E. O. (1978). Fungi ex excavationibus artifi-


cialibus subterraneis RSS Armeniae. Nov. Syst. Pl. non Vasc., 15, 76-81.
Boycheva, P. & Marinova, V. (2017). An Ethnobotanical study of the
medicinal plants in the Beloslav area, Northern Black Sea Coast region
(Bulgaria). International Journal of Advanced Research, 5(12), 205-218.
Breitenbach, J. & Kränzlin, F. (1986). Fungi of Switzerland. Vol. 2: A
Contribution to the Knowledge of the Fungal Flora of Switzerland. Non-
Gilled Fungi. Heterobasidiomycetes (jelly fungi), Aphyllophorales (non-
gilled fungi), Gasteromycetes (puffballs), Verlag Mykologia, Lucerne.
Eriksson, J. & Ryvarden, L. (1973). The Corticiaceae of North Europe. Vol.
2: Aleurodiscus-Confertobasidium. Oslo: Fungiflora.
Jordanov, D. (ed.). (1979). Flora Reipublicae Popularis Bulgaricae. Vol. 7.
Aedibus Acad. Sci. Bulgaricae, Serdicae (in Bulgarian).
Lambevska, A., Rusevska, K. & Karadelev, M. (2013). New data on the
taxonomy, distribution and ecology of the genus Peniophora Cooke
(Basidiomycota, Fungi) in the Republic of Macedonia. Macedonian
Journal of Ecology and Environment, 15(2), 69-79.
Leraut, P. (2012). Moths of Europe. Vol. 3: Zygaenids, Pyralids 1 and
Brachodids. N.A.P. Editions, Verrières-le Buisson, 600.
Tashev, A. & Tsavkov, E. (2008). Medicinal plants of the Bulgarian
dendroflora. Phytologia balcanica, 14(2), 269-278.
Yurchenko, E. (2010). The genus Peniophora (Basidiomycota) of Eastern
Europe: morphology, taxonomy, ecology, distribution. National Academy
of Sciences of Belarus, VF Kuprevich Insitute of Experimental Botany.

 Aneta Lambevska-Hristova
https://orcid.org/0000-0003-0061-1015
Department of Plant and Fungal Diversity and Resources
Institute of Biodiversity and Ecosystem Research
Bulgarian Academy of Sciences
Sofia, Bulgaria
E-mail: [email protected]

 Svetlana Bancheva
https://orcid.org/0000-0001-7365-9971
Department of Plant and Fungal Diversity and Resources
Institute of Biodiversity and Ecosystem Research
Bulgarian Academy of Sciences
Sofia, Bulgaria
E-mail: [email protected]

220
Proceeding of 1st International conference on Environmental protection and disaster RISKs

https://doi.org/10.48365/envr-2020.1.20 Biodiversity

APPLICABILITY OF MTDNA CONTROL REGION


FOR ASSESSMENT OF TURBOT POPULATIONS
ALONG THE BULGARIAN BLACK SEA COAST

Petya Ivanova1, Nina Dzhembekova1, Ivan Atanassov2,


Krasimir Rusanov2, Violin Raykov1, Ivelina Zlateva1 , Maria Yankova1
1
Institute of Oceanology – Bulgarian Academy of Sciences
2
AgroBioInstitute – Agricultural Academy (ABI)

Abstract: Turbot is a valuable commercial fish species classified as


endangered. Knowledge of the level and distribution of genetic diversity in turbot
is important for designing conservation strategies for their sustainable survival and
to preserve their evolutionary potential. Mitochondrial control region sequences
were investigated to evaluate variability in population genetic structure of one
population along the northern Black Sea coast. 19 CR haplotypes were found
in the Shabla population, which were shared with turbot haplotypes, previously
described for the Black Sea and the Mediterranean. A pattern with a high level of
haplotype diversity (h = 0.954 ± 0.022) and a very low level of nucleotide diversity
(π = 0.00563 ± 0.00063) indicated a high number of closely related haplotypes and
suggested that this population may have undergone a recent expansion. Tajima’s
D test and Fu’s FS test both suggest recent population growth. The haplotypes
found in the Shablа population were typical for the north Black Sea populations
and could be used for distinguishing the population along the Bulgarian Black
Sea coast.
Keywords: control region (mt DNA), turbot, population structure, Bulgarian
Black Sea coast

INTRODUCTION
The turbot (Scophtalmus maximus L.) is a fish species widely distributed
throughout the European waters, from Northeast Atlantic to the Arctic Circle.
It occurs in the Baltic and the Mediterranean, as well as in the Black Sea. It is
one of the most valuable commercial species in the Black Sea and represents an
endangered fishery resource. Turbot is represented by several local populations
that could be considered independent units of the stock (Daskalov & Ratz, 2011),
and state-of-the-art knowledge of local Black Sea turbot populations is needed for
accurate assessment of general turbot stock.
221
P. Ivanova, N. Dzhembekova, I. Atanassov, K. Rusanov, V. Raykov, I. Zlateva, M. Yankova

Understanding the evolutionary dynamics of population structure and genetic


diversity of marine fish species is of great importance for fishery management and
resource conservation (Xu et al., 2018). In recent years, different molecular mark-
ers (nuclear and mitochondrial DNA) provided new information concerning the
genetic variability of the Black Sea turbot populations (Nikolov et al., 2015, Turan
et al., 2019a, Bessonova & Nebesikhina, 2019, Turan et al., 2019b) permitting the
delineation of management units and allowing assessment of conservation priorities.
Mitochondrial DNA has been widely used in the studies of genetic structure, evo-
lution, and population biology. Due to its simple structure, the lack of recombination,
multi-copy status in a cell, maternal inheritance, and high evolutionary rate, the mtDNA
has been extensively used for population genetic study and phylogenetic analysis (Shi
et al., 2011). Therefore, mtDNA variations can be extremely useful for identifying and
managing stocks of fish species (Billington et al., 1992, Martins et al. 2003).
The control region (CR) has been shown to evolve faster than the coding genes
present in fish mitogenomes (Mate´ et al., 2004), and has been utilized to trace
population differentiation, phylogeographic distribution and phylogenetic recon-
struction (Faber & Stepien, 1998; Martinez et al., 2006; Sivasundar et al., 2001,
according to Jondeung & Karinthanyakit, 2015).
The analysis of the control region haplotype sequence data of turbot along the
west Black Sea coast did not provide clear indications on the phylogeographic dif-
ferentiation (Atanassov et al., 2011) and support the Suzuki et al. (2004) opinion
proposed that the Mediterranean turbot populations belong to two main genetically
distinct lineages, ‘western Mediterranean’ and ‘eastern secluded Mediterranean ba-
sins’.
The genetical structure of the turbot population along the Bulgarian Black Sea coast
is unclear and more population and markers have to be investigated to specify it.
The aim of this study is to assess the population structure of one turbot
population along the northern Bulgarian Black Sea coast based on the application
of one mitochondrial DNA marker (CR).

MATERIAL AND METHODS


Sample collection and DNA extraction
DNA analysis was performed on 30 turbot samples caught by a local fishing
vessel in March 2019 in Shabla region (43° 30' 49.802''N 28° 44' 25.962''E), the
north part of Bulgarian Black Sea coast. All tissue samples were taken from dorsal
fin and stored in 96% ethanol at 40°C until the analyses. Genomic DNA was
extracted using DNeasy Blood & Tissue Kit (QIAGEN).

PCR and Sequence analysis of mitochondrial DNA (CR)


The polymerase chain reaction (PCR) using mitochondrial primers
(CR) was carried out in a reaction volume of 50 µl containing 2 µl of each
222
Applicability of mtDNA Control Region for Assessment...

primer, 25 µl of mastermix (MyTaqTM HS Mix) and 2 µl of target DNA. The


mitochondrial control region was PCR amplified using universal primers
(L15924: 5’- AGCTCAGCGCCAGAGCGCCGGTCTTGTAAA and H16498-
5’-CCTGAAGTAGGAACCAGATG, according to Atanassov et al., 2011). The
conditions of PCR amplification included the following parameters: 35 cycles (940C
for 5min, 950C for 45 sec, 500C for 1min, 720C for 1 min), and 720C for 5 min. PCR
product quality control was performed by electrophoresis on 2% agarose gel. The
sequencing was performed by Macrogen Europe B.V. The newly described CR
sequences were deposited in the GenBank under the accession numbers MN556856
– MN556885.
The new sequences and sequences extracted from the GenBank database cov-
ering all available complete CR sequences of different populations of S.maximus
were compiled and analyzed.

Statistical analyses
MEGA7 (Kumar et al., 2018) was used to analyze mtDNA sequence data. After
sequence alignment, the number of haplotypes, haplotype connectivity, and TCS
Networks using Popart (Clement et al., 2002) was also constructed. The statistical
robustness in the nodes of the resulting tree was determined by 1000 bootstrap
replicates. Analysis performed by Tajima’s D test (Tajima, 1989) tested the
conformity of DNA sequence evolution to neutrality.

RESULTS AND DISCUSSION


The mitochondrial genome has been used as an effective tool for phylogenetic
and population genetic analyses in vertebrates (Saitoh et al., 2016). Mitochondrial
DNA polymorphism is widely used to determine population structure, species
differences, and evolutionary relationships. In this study, S. maximus individuals
from one population in the northern Black Sea were sampled to examine genetic
diversity and population structure by investigating sequences of the mitochondrial
control region.
The DNA sequencing of the CR PCR products produced 431 - 432 bp
sequence for each turbot samples. Among the 30 individual turbot sequences, 19
CR haplotypes were found (Table 1). Nine of the haplotypes observed in Shabla
were new and 10 haplotypes were identical to some haplotypes previously
reported (Atanassov et al., 2011, Suzuki et al., 2004). Among the haplotypes
observed in Shabla Hap7, Hap9, Hap13, and Hap16 were previously found in
the north Bulgarian Black Sea coast and Hap19 was detected in the Romanian
coast (Atanassov et al., 2011). The haplotypes Hap2 and Hap5 (Shabla, Black
Sea) correspond to Sma15-16, both reported in the Black Sea and the second
one found also in the Atlantic Ocean. The haplotypes Hap3, Hap6, and Hap14
(Sma05-08) were shared between the Shabla population (Black Sea) and
223
P. Ivanova, N. Dzhembekova, I. Atanassov, K. Rusanov, V. Raykov, I. Zlateva, M. Yankova

the Azov Sea population, with Hap6 found also in the Mediterranean. Two
haplotypes Hap2 and Hap17 were shared with the Sea of Marmara and the
former one found also in the Aegean (Suzuki et al., 2004). The data received
did not support the Suzuki et al. (2004) opinion about the lack of common
haplotypes between “western” and “eastern” lineages and the existence of two
distinct turbot mitochondrial lineages. At the same time, there are common
haplotypes between north Black Sea populations according to our data and data
provided by Atanassov et al. (2011). The haplotypes of the south Black Sea
populations (Atanassov et al., 2011) were not found in the Shabla population,
which proposed that these markers are enough appropriate for distinguishing
north and south Bulgarian population but could not be used for differentiation
at a global scale as western and eastern lineages.

Table 1. Distribution of CR haplotypes of S. maximus populations in Shabla


(SHA) (current study) and reference data from north Bulgarian coast (BS/N)
(Atanassov et al., 2011), the Black Sea (BS), the Azov Sea (AS), the Aegean
Sea (AGS), the Marmara Sea (MS), the Mediterranean Sea (MDS) and the
Atlantic Ocean (AO) (Suzuki et al., 2004).
Haplo SHA BS/N BS AS AGS MS MDS AO
type
Hap 1 1
Hap 2 2 4 2 1
Hap 3 2 1
Hap 4 1
Hap 5 5 4 1
Hap 6 4 1 1
Hap 7 1 1
Hap 8 1
Hap9 1 1
Hap10 1
Hap11 1
Hap12 1
Hap12 1
Hap13 1 2
Hap14 2 2
Hap15 1
Hap16 1 1
Hap17 2 1
Hap18 1
Hap19 1 1 (Ro)

224
Applicability of mtDNA Control Region for Assessment...

The CR datasets with haplotype distribution were star-like (Fig. 1), with a central
predominant haplotype. The star-like topologies are often viewed as evidence of
recent population expansion (Slatkin and Hudson, 1991).

Figure 1. Haplotype network obtained from the TCS analysis.


The size of the circle represents the frequency of each haplotype.
Small lines represent substitutions between haplotypes.

The Shabla population of turbot showed high levels of haplotype diversity


(0,954) but very low nucleotide diversity (0,00563) in the mtDNA control region
225
P. Ivanova, N. Dzhembekova, I. Atanassov, K. Rusanov, V. Raykov, I. Zlateva, M. Yankova

(Table 2). A high level of haplotype diversity in the same region was also reported
for turbot along the Bulgarian Black Sea coast (Atanassov et al., 2011). The reason
for that is that the majority of the identified haplotypes originate from a small set
of dominant haplotypes (Hap 5, 2and 14) following single nucleotide substitution.

Table 2. The genetic information of the turbot S. maximus populations in Shabla


(SHA) (current study) and reference data from north and south Bulgarian coast
(BS/N and BS/S) (Atanassov et al., 2011), the Black Sea (BS), the Azov Sea (AS),
the Aegean Sea (AGS), (Suzuki et al., 2004).
Population Hap pHap Hd π k D Fs
SHA 19 9 0,954±0,022 0,00563±0,00063 2,42529 -1,71288 -15,924
BS/N 26 14 0,946±0,00035 0,00722±0,0036 3,11224 -1,19452 -19,796
BS/S 9 4 0,978±0,054 0,00700±0,00085 9,0222 -1,299864 -5,201
BS 9 2 0,901±0,039 0,00689±0,00074 2,97076 -0,49028 -1,951
AS 12 7 0,844±0,054 0,00513±0,00078 2,20899 -1,45864 -5,201

Hap: the number of haplotypes; pHap: the number of private haplotypes;


Hd: haplotype diversity; π: nucleotide diversity; k: the mean number of pairwise
nucleotide differences; D: Tajima’s D value; Fs: Fu’s Fs value.
High haplotype diversity suggests a large, stable, effective population size over
time in the continental shelf fishes (Stepien, 1999), and in concurrence, with low
nucleotide diversity, it has been linked to population growth after a period of low
effective population size (Grant & Bowen, 1998). Low nucleotide-diversity values
indicated that haplotypes were closely related; most haplotypes differed by only
1 or 2 nucleotides. Common haplotypes among populations and haplotypes from
several populations indicate high genetic connectivity among the populations.
High levels of haplotypic diversity and low nucleotide diversity is indicative of a
population bottleneck followed by rapid population growth and accumulation of
mutation (Avise, 2000). Concerning the demographic history of the population, the
negative value of Tajima's D indicates a bias towards rare alleles, the latter being a
signature of recent population expansion. Fu's FS test is based on the distribution
of alleles or haplotypes, and negative values can indicate recent population growth
(Jong et al., 2011). Tajima'sD and Fu's Fs were all non-significant and negative,
as expected for a recent population expansion or selection (Alcaraz & Gholami,
2018).

CONCLUSION
The mt DNA haplotypes (CR region) found in the Shabla population could be
used for distinguishing of north and south Bulgarian Black Sea turbot populations.
The obtained data can be useful in selecting fish stocks that preserve a better genetic
diversity of S. maximus in conservation and/or hatchery programs.
226
Applicability of mtDNA Control Region for Assessment...

The assessment of the genetic structure detected provides new perspectives on


population structuring of turbot along the Bulgarian Black Sea coast, which can en-
hance the scientific foundations for the sound management of this fishery species.
Effective management and conservation programs require knowledge of species'
genetic diversity, population structure, and mediating factors. The data for Shabla
population genetic diversity could be an essential background for the scientific
management of the turbot populations along the Bulgarian Black Sea coast. To
specify the genetic structure of the Black Sea turbot populations further studies are
needed for increasing the number of populations analyzed and investigating more
informative genetic markers.

ACKNOWLEDGMENTS
This work has been carried out in the framework of the National Science Program
"Environmental Protection and Reduction of Risks of Adverse Events and
Natural Disasters", approved by the Resolution of the Council of Ministers №
577/17.08.2018 and supported by the Ministry of Education and Science (MES)
of Bulgaria (Agreement № D01-230/06.12.2018).

REFERENCES:
Alcaraz, C. & Gholami, Z. (2020). Diversity and structure of fragmented
populations of a threatened endemic cyprinodontid (Aphanius sophiae)
inferred from genetics and otolith morphology: Implications for
conservation and management. J Zool Syst Evol Res.; 58, 341–355.
https:// doi.org/10.1111/jzs.12333
Atanassov, I., Ivanova, P., Panayotova, M., Tsekov, A., and Rusanov, K.
(2011). Mitochondrial control region DNA variation in turbot popula-
tions from the Bulgarian and Romanian Black Sea coasts. Biotechnol.
Biotechnol. Equip. 25(4), 2627-2633.
doi:10.5504/BBEQ.2011.0068.
Avise J. (1998). The history and preview of phylogeography: A personal
reflection. Mol Ecol 7, 371–379.
Avise J. (2000). Phylogeography: The history and formation of species.
Cambridge, MA: Harvard University Press.
Bessonova, N.A. & Nebesikhina, N.A. (2019) Investigation of the
genetic structure of the Black Sea turbot Scophthalmus maeoticus
(Scophtalmidae) population in the Azov and Black Seas, Aquatic
Bioresources & Environment, vol.2, 4, 60-67.
Billington N, Barrette RJ & Hebert PDN (1992) Management implications
of mitochondrial DNA variation in walleye stocks. North Am J Fish
Manag ,12,276-284.
227
P. Ivanova, N. Dzhembekova, I. Atanassov, K. Rusanov, V. Raykov, I. Zlateva, M. Yankova

Clement M, Snell Q, Walke P, Posada D, Crandall, K (2002). TCS:


estimating gene genealogies. Proc 16th Int Parallel Distrib Process Symp
2:184.
Daskalov, G., & H. J. Rätz. (2011). Scientific, Technical and economic
Committee for Fisheries (STECF), Opinion by written procedure:
Assessment of Black Sea Stocks (STECF-OWP-11-06). EUR 25020 EN
– 2011. Publications Office of the European Union, Luxembourg. [online]
URL: http://stecf.jrc.ec.europa.eu/documents/43805/218009/2011-11_
OWP+11-06++Black+Sea+Stock+Assessments_JRC67414.pdf
Grant W.S. & Bowen B.W. (1998) Shallow population histories in deep
evolutionary lineages of marine fishes: insights from sardines and
anchovies and lessons for conservation. Genetics, 89, 415–426.
Jondeung, A. & Karinthanyakit, W. (2015). Mitochondrial DNA control
region of three mackerels, genus Rastrelliger: structure, molecular diver-
sity and phylogenetic relationship, Mitochondrial DNA, Early Online:
1–6, 2015 Informa UK Ltd. DOI: 10.3109/19401736.2015.1028047ISS
N: 1940-1736 (print), 1940-1744 (electronic)
Jong MA, Wahlberg N, van Eijk M, Brakefield PM, Zwaan BJ (2011)
Mitochondrial DNA Signature for Range-Wide Populations of Bicyclus
anynana Suggests a Rapid Expansion from Recent Refugia. PLoS ONE
6(6): e21385. doi:10.1371/journal.pone.0021385
Kumar S, Stecher G, Li M, Knyaz C. & Tamura K. (2018). MEGA X:
Molecular Evolutionary Genetics Analysis across computing platforms,
Molecular Biology and Evolution 35:1547-1549.
Martins,C. Wasko,A.P., Oliveira, C. & Foresti, F. (2003). Mitochondrial
DNA variation in wild populations of Leporinus elongatus from the
Paraná River basin. Genetics and Molecular Biology, 26, 1, 33-38.
Li, C., Bessert, M. L., Macrander, J., & Orti, G. (2009). Low variation but
strong population structure in mitochondrial control region ofthe plains
topminnow, Fundulus sciadicus. Journal of Fish Biology, 74,
1037–1048.
Mate´ ML, Di-Rocco F, Zambelli A, Vidal-Rioja L. (2004). Mitochondrial
DNA structure and organization of the control region of South American
camelids. Mol Ecol Notes, 4 ,765–767.
Nikolov, V., Ivanova, P., Dzhembekova, N. Panayotova, M., Raykov,
V., and Dobrovolov I. (2015). Application of allozyme markers for
screening of turbot populations along the Western Black Sea coast.
ZooNotes ,79, 1-15.
Satoh, T.P., Miya, M., Mabuchi, K. & Nishida1, M. (2016). Structure and
variation of the mitochondrial genome of fishes. BMC Genomics,17,
719, https://doi.org/10.1186/s12864-016-3054-y

228
Applicability of mtDNA Control Region for Assessment...

Shi W, Kong XY, Wang ZM, Jiang JX (2011) Utility of tRNA genes from
the complete mitochondrial genome of Psetta maxima for implying a
possible sister-group relationship to the Pleuronectiformes. Zool Stud,
50(5):665 681.
Slatkin M, Hudson RR. 1991. Pairwise comparisons of mitochondrial DNA
sequences in stable and exponentially growing populations. Genetics.
129:555–562.
Stepien C. (1999). Phylogeographical structure of the Dover sole
Microstomus pacificus: The larval retention hypothesis and genetic
divergence along the deep continental slope of the northeastern Pacific
Ocean. Mol Ecol 8, 923–939.
Suzuki, N., Nishida, M., Yoseda, K., Ustundag, C., Sahin, T., & Amaoka,
K., (2004). Phylogeographic relationships within the Mediterranean
turbot inferred by mitochondrial DNA haplotype variation. Journal of
Fish Biology, 65(2), 580- 585.
Tajima F. (1989). Statistical method for testing the neutral mutation
hypothesis by DNA polymorphism. Genetics. 123:585–595.
Turan, C., Ivanova, P., Gürlek, M., Yağlioğlu, D., Ergüden, D., Karan, S.,
Doğdu, S. A., Ali Uyan, A., Öztürk, B., Nikolov V., Raykov, V., Do-
brovolov, I., Khanaychenko, A. (2019 a). Phylogenetic Relationships of
Turbot Species (Scophthalmidae) Inferred from the Mitochondrial COIII
Gene and Morphological Characters. NESciences, 2019, 4(1): 28-41doi:
10.28978/nesciences.522593.
Turan, C., Ivanova, P., Raykov, V., Gürlek, M., Ergüden, D., Yaglioglu, D.,
Karan, S., Doğdu, S., Uyan, A., Oztürk, B., Nikolov, V., Dobrovolov, I.,
Khanaychenko, A., Giragosov, V. (2019 b). Genetics Structure Analysis
of Turbot (Scophthalmus maximus, Linnaeus, 1758) in the Black and
Mediterranean Seas for Application of Innovative Management Strate-
gies. Frontiers in Marine Science. 6. 10.3389/fmars.2019.00740.
Wang, S-Y., Shi, W., Miao X-G., & Kong X-Y., (2014) Complete mitochon-
drial genome sequences of three rhombosoleid fishes and comparative
analyses with other flatfishes Pleuronectiformes), Zoological Studies,
53:80 http://www.zoologicalstudies.com/content/53/1/80
Xiao, Yo., Ren, G., Song, Na., Li, J., and Gao, T. (2014). Genetic diversity
and population structure of largehead hairtail, Trichiurus japonicus,
based on mtDNA control region. Mitochondrial DNA, 25(6): 473–481,
ISSN: 1940-1736 (print), 1940-1744 (electronic).
Xu, S., Wang, Zh. Zhang, Y., Gao, T., Zou, J.& Zhang, X. (2018). Evidence
for intra-mitochondrial variation in population genetic structure of
Platycephalus sp.1 In the Northwestern Pacific, Mitochondrial DNA Part
A, DOI: 10.1080/24701394.2018.1484119

229
P. Ivanova, N. Dzhembekova, I. Atanassov, K. Rusanov, V. Raykov, I. Zlateva, M. Yankova

 Petya Ivanova
https://orcid.org/0000-0002-7487-9033
Institute of Oceanology
Bulgarian Academy of Sciences
Varna, Bulgaria
E-mail: [email protected]

 Nina Dzhembekova
http://orcid.org/0000-0001-9620-6422
Institute of Oceanology
Bulgarian Academy of Sciences
Varna, Bulgaria
E-mail: [email protected]

 Ivan Atanassov
https://orcid.org/0000-0002-7771-3629
AgroBioInstitute
Agricultural Academy
Sofia, Bulgara
E-mail: [email protected]

 Krasimir Rusanov
https://orcid.org/0000-0001-8131-3364
AgroBioInstitute
Agricultural Academy
Sofia, Bulgara
E-mail: [email protected]

 Violin Raykov
http://orcid.org/ 0000-0003-4322-6352
Institute of Oceanology
Bulgarian Academy of Sciences
Varna, Bulgaria
E-mail: [email protected]

 Ivelina Zlateva
https://orcid.org/0000-0003-4133-5627
Institute of Oceanology
Bulgarian Academy of Sciences
Varna, Bulgaria
E-mail: [email protected]

 Maria Yankova
https://orcid.org/0000-0002-3333-7131
Institute of Oceanology
Bulgarian Academy of Sciences
Varna, Bulgaria
E-mail: [email protected]

230
Proceeding of 1st International conference on Environmental protection and disaster RISKs

Informatics, Remote Sensing, High Performance Computing


and Gis for Environmental Monitoring and Management
https://doi.org/10.48365/envr-2020.1.21

OPEN SOURCE GIS DECISION SUPPORT TOOLS


FOR WILDFIRE AND FLOOD NATURAL HAZARDS
WITH EMBEDED INTERACTIVE TRAINING
RESOURCES

Stefan Stefanov1), Nina Dobrinkova1) , Emilia Tosheva2)


1)
Institute of Information and Communication Technologies –
Bulgarian Academy of Sciences (IICT-BAS)
2)
South-West University “Neofit Rilski”, Technical Faculty

Abstract. Training of volunteers that participate in natural hazard suppression


measures is very important and time consuming. The volunteers do not have
unlimited amount of time to dedicate on educational activities. Thus, interactive
training platforms are the next step of the training processes. In our paper we will
describe an open source Geographic Information System (GIS) tool specifically
designed to support decision making for field operations in cases of natural disasters.
The training activities for the tool implementation were based on Google G-Suite
cloud application. The GIS application was developed to deliver field situation
alerts in operational room and G suite was used before the field work as preparatory
training for the teams working with it. The combination of the online tools described
in the paper is presented as potential first step of real time volunteer training system
that in future can be further elaborated and improved.
Keywords: G Suite; volunteer's training; open source GIS applications

INTRODUCTION
GIS tools in general are very costly licensed applications that can support
georeferenced datasets analyses. This is the reason why such applications are
very hard to get in everyday use for volunteer or operational groups activities in
most Bulgarian municipalities. In the last two years our work was focused on GIS
application development, that is based on free data and open source GIS solutions
supporting decision making in cases of wildland fires and flood events. The main
goal was improving accessibility, time response and communication between field
operation teams. The application had to be able to visualize Points of Interest (POIs)
which operational teams on the field need to know during suppression measures.
Such POIs could be logistic centers for water supplies and firefighting tools for
example. All of this information is very good source if the people having the tools
231
Stefan Stefanov, Nina Dobrinkova, Emilia Tosheva

know how to operate with georeferenced databases. This was not the case with
most of the volunteer group members in the municipalities we were working with.
Thus, we decided to use Google cloud-based platform called G Suite (G-Suite) for
developing of training application along with the main one. The second tool could
deliver in an easy way an interactive training interesting for the volunteers and
applicable in real time. G Suite is cloud-based and gives the opportunity to be used
on different devices such as smart phones, tablet devices, laptops and etc. Even
after the main training the resource can be used as follow up reference (Tosheva,
2019).

STUDY AREAS
The study areas used as base for development of our application have been
selected based on two different international projects operating in two countries
– Bulgaria and Armenia. The Bulgarian test area is situated on the territory of
Zlatograd forestry department and was focused on forest fires. The flood related
area was selected for Armenia in the catchment of river Voghji – Kapan region.

Bulgarian test area


The Zlatograd forestry department covers the territories of the municipalities
of Zlatograd, Madan and Nedelino. All three of the municipalities are located in
southern Bulgaria, at the end of the Eastern Rhodopes, to the last Rhodope hill, after
which the Aegean Plain begins. The relief of the municipalities is low and medium
mountainous and extends in the south-western part of the Eastern Rhodopes. The
area belongs to the continental – Mediterranean climate area, the South-Bulgarian
climatic sub-region and the low mountain climatic region of Eastern Rhodopes.
The temperature is continental and prevails in the low mountainous regions, with
a marked influence of the Aegean Sea, while in the mountainous areas there is a
decrease in temperature, an increase in the amount of precipitation and the wind
speed. The wildfire regime is occurring in the periods of the late spring (mid-March
by mid-May) and during the second half of the summer season. Most of the wildfires
have potential to cross the Bulgarian-Greek border zone and monitoring of the fire
prone areas is the main concern of the local municipal authorities.

Armenian test area


The test areas covering the flood events in Armenia used in our application
web-based tool are covering the territories of The Geghi reservoir and Geghanoush
Tailing Storage Facility (TSF).
The Geghi reservoir is located in Syunik, the southernmost province of Arme-
nia (Figure 1). The reservoir is situated on the Geghi River, the left-bank tributary
of the river Voghji. The maximum water level discharge occurs during the spring.
Due to the high-altitude nature of the area, snow melt increases gradually as does
232
Open Source GIS Decision Support Tools for Wildfire...

the level of the river and the reservoir. Snow melt typically lasts from March to
August (Armenian State Hydrometeorological and Monitoring Service officially
used sources).
The surface of the Geghi reservoir is 50 ha and the elevation above sea level
is nearly 1400 m. The height of the dam is 70 m and the length along the crest is
270 m. The total volume of reservoir is 15 million cubic meters, but the effective
volume is about 12 million cubic meters. Nearly 4,300 people would be affected by
in a case of a dam break of the reservoir.
The Geghanoush TSF is located in the gorge of the Geghanoush River, in the
southern part of Kapan (Figure 2). The difference of relative heights between the
tailing dam, on one hand, and city buildings and transport infrastructure, on the
other hand, is 75 meters. In case the reservoir dam is broken due to an earthquake,
the sliding mass could cover industrial and residential buildings, and as a result of
barrage, the polluted water could flood central quarters of the city. The existing
Geghanoush Tailings Repository was designed in early 1960’s and had been operated
between 1962 and 1983, when the Kajaran Tailings Repository at Artsvanik was
commissioned. The Geghanoush tailings repository was re-commissioned in 2006
after the completion of the diversion works and continues to be used today along
with an upstream extension currently under construction. The volume of the tailing
is 5.4 million m3 and the dam height is 21.5 m.

Figure 1. The location of Geghi reservoir. The inset shows its location
within Armenia. Background image: Sentinel-2, RGB composite
(source: American University of Armenia)
233
Stefan Stefanov, Nina Dobrinkova, Emilia Tosheva

Figure 2. The location of Geghanoush Tailing Dam. The inset shows


its location within Armenia. Background image: Sentinel-2, RGB
composite (source: American University of Armenia)

DESKTOP APPLICATION STRUCTURE


The wildland fire and flood events applications include features and tools which
can be used by volunteers or operational teams on the field for improved response
in cases of wildland fires or floods.

Wildland fires application


The developed Wildland Fire Application was implemented in the work of
volunteer groups of Zlatograd municipality. It presents fire zones and POI's in the
municipalities of Zlatograd, Madan and Nedelino as predefined layers and base map
layers via the OpenStreetMaps (OSM) platform (Figure 3). The used techniques are
based on (Singh & Gambhir, 2012), (Das, Prakash, Sandilya & Subhani, 2014),
(García, Perotto-Baldivieso & Hallett, 2010). This application is able to show both
active and previous fires on the tested area together with the logistic centers for
water supply and other firefighting tools. The Wildland Fire Application passed the
testing phase and now targeted users are working with it in order to produce better
feedback for its future development and improvements.
234
Open Source GIS Decision Support Tools for Wildfire...

Figure 3. Wildland fires application main screen

The Wildland Fire Application has different features that can be used by
firefighting or volunteer groups on the field. It is suitable for operational room
operators too. The module maintains the Geolocation feature which is helpful when
the first arrived team alert the rest of the groups where the problem is. Via the Draw
feature they are able to mark the locations of fires so that they become visualized
and possible to be seen in the operational room. As a result, firefighters can request
help or firefighting tools faster. The Draw feature is able not only to mark signs but
also to depict the fire area with polygons. The Measure and Distance features are
used to measure the distances between the fires and the nearest POI's and also can
determine the areas of the fire zone. The module has the ability of switching the
predefined layers and the base map layers. The base map layers include Street maps,
Satellite maps, Shaded relief maps and NatGeo maps. These can be used mainly
for analysis in the operational rooms. The Pop-Up feature visualizes information
about the number and tools available in the POI’s together with their locations.
The Export feature is able to save the maps with the new data in map-printable
format files. These outcomes can be applied for further analysis. The application
provides detailed information about the current or future weather conditions via
the Openweather platform. In addition, it is connected to the EFFIS emergency
management services, which offer extra satellite data if the fire is big enough to be
captured.

Flood events application


The Flood Event Application is able to analyze conditions that can cause flood
events for the area of Armenian region Kapan located in the south east part of the
country (Figure 4). The tool was used by the local civil protection teams. The software
235
Stefan Stefanov, Nina Dobrinkova, Emilia Tosheva

is based on the algorithms for open source GIS build up tools discussed in details
in (Singh & Gambhir, 2012), (Das, Prakash, Sandilya & Subhani, 2014), (García,
Perotto-Baldivieso & Hallett, 2010). The tool is using inputs from DG ECHO funded
project called “Alliance for Disaster Risk Reduction (ALTER)”. ALTER is targeting
the establishment of public-private partnerships for prevention of flood risks that may
stem from water and mining dam failures. The project was focused on three pilot
areas where dams and other activities present risks to local communities: the Akhtala
and Teghut areas of Lori Marz along the Shamlugh river, the Vorotan Cascade and its
associated dams in the Syunik region, and the Kapan and Voghji river basin of Syunik
region. The developed GIS application covered only the case of Voghji river basin.
Its capabilities were built upon flood events, caused by different probabilistic waves
in the catchment area of Voghji river. The Flood Event Application has been tested in
Kapan region with a real field test exercise and it was used for the table top education
of the local civil protection teams.

Figure 4. Flood events application main screen

The application has the ability to switch the predefined layers and also the base
map layers. The predefined layers include a variety of information. By turning them
On or Off users can easily analyze the risks in the current situation. The layers can
be downloaded as GeoJSON file. They can easily be located and their opacity can
be changed. The base map layers include Street map, Satellite map, Shaded relief
map and NatGeo map, which can help in operational room analysis.
The Geolocation feature could be used for localization of users via the GPS of
their device or via the network location. The Draw feature is able to mark the zone
236
Open Source GIS Decision Support Tools for Wildfire...

of interest by polygon or by line in order to be visualized and to be seen in the


operational rooms. As a result, civil protection teams could request more resources
or specialized equipment. The Pop-Up feature visualizes information about the
most vulnerable buildings such as schools, kindergartens and others. The Export
feature can supplement the maps with new data, such as picture format files. These
options can be used in analysis for future actions of the teams. The application is
connected to the OpenWeather platform, which provides detailed information about
the current or future weather conditions. It is connected to the EFAS emergency
management service that provides extra satellite data about the situation through
the European Emergency Services.

TRAINING
The presented open source tools that have been developed for the wildland fires
and the flood events include different features, which are basic GIS skills, but if not
trained, hard to be utilized. Simple and understandable training is needed in order
to facilitate the work with them. To support the learning process, we used some of
the Google G Suite options (G-Suite is open source platform to develop an inter-
active courses). G Suite is a set of cloud computing, productivity and collaboration
tools, software and products. It comprises of Gmail, Hangouts, Calendar for com-
munication, Currents, Drive for storage, Docs, Sheets, Slides, Forms, and Sites for
productivity and collaboration. To develop our teaching platform, we selected to
use Docs, Slides, Drive and Sites in order to set up the training platform (Tosheva,
2019). Video materials, presentations and documents were included in the web re-
sources of both applications. The training for the Wildland Fire Tool was developed
in Bulgarian language (Figure 5).

Figure 5. Wildland fires application training web resource main screen


237
Stefan Stefanov, Nina Dobrinkova, Emilia Tosheva

It included lessons for different tools that are available in the application.
General orientation how the application works, navigation instruction through the
main menu, getting familiar with layers and how to add a new information were the
main topics in the educational materials. The training set of lessons for the flood
event application were developed in English language (Figure 6).

Figure 6. Flood events application training web resource main screen

The flood educational platform included lessons for different capabilities


related to the flood phenomena. The lectures were based on main skills on how the
application works, how navigation can be done among all features, work analyses
with layers and usage of pop-up options.

CONCLUSION
The presented article gave an overview on the way of building open source
web based GIS tools in support of volunteer or civil protection operational teams
from the field. The main ideas how the architecture of the application can be
established is described in details in (Singh & Gambhir, 2012), (Das, Prakash,
Sandilya & Subhani, 2014), (García, Perotto-Baldivieso & Hallett, 2010). The
future use of the platforms can be mainly from volunteer groups, firefighting or
civil protection teams in their decision making on the field or in the operational
rooms (Traneva, Atanassova & Tranev, 2019), (Traneva & Tranev, 2019), (Traneva,
Tranev & Atanasova, 2019). This kind of ICT solutions are essential due to the
growing requirements in efficient actions in the context of wildland fires and flood
monitoring activities. Our tools provide opportunity for delivering critical data and
knowledge in a timely manner to all participants in hazard response operations.
The main goal of the Wildland Fire Application was to deliver information about
fire locations and the available equipment and resources in the nearest Points
of Interest (POIs) – logistic centers for water supplies and firefighting tools. In
this way the Information and Communication Technologies (ICT) were able to
238
Open Source GIS Decision Support Tools for Wildfire...

meliorate the orientation of the volunteer’s groups on field and the reaction of the
first respondents in operational rooms. The Flood Event Application main goal was
to visualize the most vulnerable areas and buildings in cases of floods. It presented
different scenarios that can be analyzed in operational room in order to support and
improve the reaction and management of the current or future situations of flood
events. The tools have been tested during both projects’ lifetime. They proved to
be suitable for access from everywhere where Internet connection is in place by
different devices like desktop computers, smart phones, tablets and laptops.

ACKNOWLEDGMENT
This paper has been supported partially by the Bulgarian National Scientific Fund
project number DFNI DN12/5 “Efficient Stochastic Methods and Algorithms
for Large-Scale Problems” and the National Scientific Program “Environmental
Protection and Risk Reduction Under Adverse Events and Natural Disasters” –
NNP-OS contract number Д01-322/18.12.2019.

REFERENCES
E. Tosheva (2019), Technological model for training students in the cloud
environment, Conference on Information Technology and Development
of Education – ITRO 2019 Zrenjanin, Republic of Serbia, ISBN 978-86-
7672-322-5 (Tosheva, 2019).
H.Singh, D. Gambhir, “An Open Source Approach to Build a Web GIS
Application”, International Journal of Computer Science and Technology
(IJCST), Volume 3, ISSN: 2229-4333, p110 – p112, 2012 (Singh &
Gambhir, 2012).
A. K. Das, P. Prakash, C. Sandilya, S. Subhani, “Development of Web-Based
Application for Generating and Publishing Groundwater Quality Maps
Using RS/GIS Technology and P. Mapper in Sattenapalle, Mandal, Guntur
District, Andhra Pradesh”, In ICT and Critical Infrastructure: Proceedings
of the 48th Annual Convention of Computer Society of India, Vol. 2, 2014,
pp. 679 – 686. (Das, Prakash, Sandilya & Subhani, 2014).
V. G. García, H. L. Perotto-Baldivieso, S. Hallett,“A prototype design for
a Web-GIS disaster support system: The Bolivia Amazon case study”,
2010. ( García, Perotto-Baldivieso & Hallett, 2010).
[Online] Available: https://gsuite.google.com (G-Suite)
Traneva V., Atanassova V., Tranev S. Index matrices as a decision-
making tool for job appointment, Springer Nature Switzerland AG, G.
Nikolov et al. (Eds.): NMA 2018, LNCS 11189, 1 – 9, 2019. https://
doi.org/10.1007/978-3-030-10692-8_19 (Traneva, Atanassova &
Tranev, 2019).
239
Stefan Stefanov, Nina Dobrinkova, Emilia Tosheva

Traneva V., S. Tranev. An interval-valued intuitionistic fuzzy approach to


the assignment problem. Advances in Intelligent Systems and Computing,
Intelligent and Fuzzy Techniques in Big Data Analytics and Decision
Making. In: C. Kahraman et al. (Eds.). Proceedings of the INFUS
Conference, July 23 – 25, 2019, Istanbul, Turkey, AISC, Vol. 1003, 1-9,
2020, ISBN 978-3-030-23755-4
DOI: 10.1007/978-3-030-23756-1_72 (Traneva & Tranev, 2019).
Traneva V., Tranev S., Atanassova V. Index Matrices as a Cost Optimization
Tool of Resource Provisioning in Uncertain Cloud Computing
Environment. In: S. Fidanova (ed.), Recent advances in computational
optimization, Studies in computational intelligence, vol 838. Springer,
Cham, 2020, 1 – 20. https://doi.org/10.1007/978-3-030-10692-8_19
(Traneva, Tranev & Atanasova, 2019).

 Stefan Stefanov
https://orcid.org/0000-0001-6012-533X
Institute of Information and Communication Technologies
Bulgarian Academy of Sciences
Sofia, Bulgaria
E-mail: [email protected]

 Nina Dobrinkova
https://orcid.org/0000-0002-6506-748X
Institute of Information and Communication Technologies
Bulgarian Academy of Sciences
Sofia, Bulgaria
E-mail: [email protected]

 Emilia Tosheva
https://orcid.org/0000-0002-3409-125X
Technical Faculty
South-West University “Neofit Rilski”
Blagoevgrad, Bulgaria
E-mail: [email protected]

240
Proceeding of 1st International conference on Environmental protection and disaster RISKs

Informatics, Remote Sensing, High Performance Computing


and Gis for Environmental Monitoring and Management
https://doi.org/10.48365/envr-2020.1.22

METHODOLOGY FOR CALCULATING


THE PARAMETERS OF RADIO PATHS
WITH IONOSPHERIC REFLECTION

Rumiana Bojilova, Plamen Mukhtarov


National Institute of Geophysics, Geodesy and Geography –
Bulgarian Academy of Sciences (NIGGG-BAS)

Abstract: A methodology for calculating the parameters of radio paths operating


with ionospheric reflection that depends on the state of the ionosphere and at a given
distance of radio communication is presented. Based on the critical frequencies of
the ionospheric regions E (foE) and F (foF2) and the maximum usable frequency
at a distance of 3000 km (MUF3000), the electron density profile is reconstructed
and the geometry of the radio wave reflection from the ionosphere is determined.
As a result of the modeling, the frequency range of the radio waves at which radio
communication is possible is determined; the latter is limited by the lowest usable
frequency (LUF) and the highest usable frequency (MUF) illustrated by an example.
Keywords: radio waves, radio paths, ionosphere, critical frequencies.

INTRODUCTION
Long-distance radio communication dates back to the early twentieth century
(1901 or 1902) with Marconi’s experiments on radio communication across the
Atlantic Ocean. At the same time, Heaviside and Kennelly suggested the existence
of an ionized layer in the atmosphere that allowed radio waves to reflect off it. In
the first half of the twentieth century, the work of Appleton and Ginzburg built the
theory of propagation and reflection of radio waves from plasma (Appleton, 1925;
Ginzburg, 1964). This theory is based on Maxwell’s system of equations. The value
of complex dielectric constant is introduced specifically for the electromagnetic
field in plasma (Gadzhev et al., 2011; 2012). This makes it possible to use the laws
of geometric optics, to derive values of the refractive index and to determine the
trajectory of radio waves in the ionosphere (Rawer, 2013).

METHOD
The task for the calculation of radio paths contains the determination of the fre-
quency range at which a radio communication at a given distance between the two
241
Rumiana Bojilova, Plamen Mukhtarov

radio communication points can take place. The frequency range is limited by the
minimum and maximum usable frequencies (LUF and MUF respectively), which
depend on the distance of the radio communication and the state of the ionosphere
at the respective time.
To solve this problem, the altitude profile of the electron concentration up to the
ionospheric maximum in the reflection area of radio waves (it is the middle of the
radio path) has to be known. The calculation methodology is based on the theory of
radio wave propagation without taking into account the influence of Earth’s mag-
netic field. We note that this simplification does not reduce the accuracy of the cal-
culation. Upon entering the ionosphere, radio waves suffer the so-called magnetic-
ion splitting, i.e. they break up into two separate waves that propagate at different
speeds. Traditionally, the data from the ionograms take into account ionospheric
characteristics (e.g. critical frequencies) of the reflections of the ordinary wave.
The propagation of an ordinary wave is not affected by the presence of the Earth’s
magnetic field.
The following theorems of the vertical propagation of an ordinary wave in an
ionized layer prove to be sufficient for the calculation of LUF and MUF.
The formula (1) represents the dependence of the virtual height of reflection on
the height profile of the electron concentration N(z) in 1/cm2, where z is the height
above the Earth in km, and the frequency of the radio signal f in MHz.
z0
dz
 f 
hvert  N z 
0
1  80.8
f2 (1)
N  z0 
80.8 1
f2

The reflection height z0 is determined by the condition for equality of the


operating and plasma frequency. The integral is improper (indefinite) because at the
height of reflection the denominator of the sub-integral function is canceled. The
ratio allows the ionogram for the vertical radio-wave propagation to be calculated,
if the electron concentration profile is known.
Figure 1 shows the geometry of the oblique propagation of radio waves in the
ionosphere. The secant and equivalence theorems apply to it is:

h'  f ob   AC  hvert
'
 f vert sec0
f ob  f vert sec0
(2)

242
Methodology for Calculating the Parameters of Radio...

Figure 1. Basic geometrical ratios at oblique reflection of radio signals


from the ionosphere

The secant theorem shown in the formula (2) reflects the following regularity.
An oblique ray with a frequency fob incident on the ionosphere at an angle j0 is
reflected by the ionosphere at the same height as a vertical ray with a frequency fvert.
The shown in Figure1 geometry contains a triangle with vertices at the two
endpoints of the radio path (points A and B) and the upper part of the virtual
reflection height h’vert (point C) located in the middle of the path of the oblique
frequency fob. The frequency of the oblique ray is equal to the frequency of the
vertical ray multiplied by the secant of the angle j0, which is determined by the
following formula (3), following the geometry shown in Figure1 (right plot).
a sin   (3)
tg0  '
h
vert  a1  cos 

Where a is the radius of the Earth (about 6370 km), the angle q is half of the
central angle between the two endpoints of the radio path. If the length of the radio
path on the Earth’s surface is denoted by r, as in the Figure 1 (right plot), then its
angle q in radians can be calculated by formula (4):
r
 rad   (4)
2a
The procedure for calculating the radio path begins with the synthesis of a model
electron density profile based on the predicted critical frequencies foF2, MUF3000
and foE. The used model, called by the names of its creators, is the profile of Di
Giovanni - Radicella (Di Giovanni & Radicella, 1990).
243
Rumiana Bojilova, Plamen Mukhtarov

For daily conditions the profile is constructed according to the following


formula (5):

h  hmE ;
h h
N h   N mE sec h 2  mE 
 2 BEb 
h  hmE ;
  h  h   h  hmE  (5)
N h    N mE  N mF sec h 2  mF 2 mE   sec h 2   
  2 B F    B Et 
    
 N mF 2  N mE sec h 2  hmF 2  hmE   sec h 2  h  hmF 2 
  2B   B 
  Et   F 

For night conditions the respective representation is formula (6):

f oE  0;
h h (6)
N h   N mF 2 sec h 2  mF 2 
 B 
 f 

This model is based on the representation of the electron density profile with
hyperbolic secant functions (Stankov et al., 2003). The values of the three critical
frequencies are sufficient to calculate the model electron density profile for heights
up to the height of the F-region maximum; the latter is calculated by the given
below formulas (7). The maximum electron concentration of the E and F- layers are
determined directly by the critical frequencies.
The height of the E-layer maximum is assumed to be fixed 120 km. The model
uses the empirically determined by Dudeney & Kressman (1986) dependence
of the height of the F2-layer maximum hmF2 on the parameter, the ratio foF2/
foE and the correction value Dm according to the given below formulas (8).
The parameter Bf, related to the half-thickness of the ionospheric F- layer, is
expressed by an empirical dependence of the derivative of the electron profile
dN/dh, which in turn is expressed by the values of foF2 and M3000F2. The
half-thicknesses of the E- layer (BEb-below and BEt-above the maximum of the
E-region) take fixed values.

244
Methodology for Calculating the Parameters of Radio...

 
N mE cm2  1.24.10 4 foE2 MHz
N mF 2 cm  2
  1.24.10 4
foF22 MHz

hmE  120 km; BEb  7.5km; BEt  7km


Mf
hmF 2  1470  176
M 3000 F 2  Dm

MUF 3000
M 3000 F 2  (7)
foF 2

foF 2 2
B f  0.04774 .1011
dN
dh

0.0196 M 3000 F 22  1
M f  M 3000 F 2
1.2967 M 3000 F 22  1
dN
dh
  
 1.10 9 exp  3.467  0.857 lg foF22  2.02 lg M 3000 F 2 
foE  0; Dm  0;
foF 2 foF 2
 1.807 ; Dm  0.725  1.725
foE foE
(8)
0.253
Dm   0.012
foF 2
 1.215
foE

After calculating the electron profile with a vertical resolution of about 1


km, the model ionogram is calculated by numerical integration. Figure 2 shows
an exemplary electron density model profile (left plot) and its respective model
ionogram (right plot).
The values of the ionospheric characteristics are accepted on the basis of the data
from the virtual sounding of the ionosphere with radio waves in the ionospheric
station “Plana” at the National Institute of Geophysics, Geodesy and Geography -
Bulgarian Academy of Sciences.
245
Rumiana Bojilova, Plamen Mukhtarov

After calculating the model ionogram, which represents the function h’vert(fvert),
i.e. the dependence of the virtual height on the frequency of the vertically
propagating radio waves, then the calculation of the model oblique ionograms at a
given distance of the radio communication (or the coordinates of the two endpoints
of the radio path) according to the formulas above can be obtained.

400 500

400
300

300
Height [km]

Height [km]

200

200

100
100

0 0
0 2 4 6 8 10 12 14 16 0 2 4 6 8 10 12 14 16
Plasma frequency [MHz] Frequency [MHz]
Figure 2. An example model of the electron density profile in daily conditions
(left panel) and the relevant model ionogram (right panel).

The results for distances respectively 0-100 km, 250 km и 500 km are shown
in Figure 3.
It can be assumed that for radio distances of up to 100 km, with sufficient
accuracy for practice, the reflection of the radio waves is close to the vertical
propagation. At longer distances, it is possible certain frequencies to reflect from
the E and F regions. Traditionally, stable radio communications take place at
reflections in F- region. Reflections from E region are accompanied by large
absorption of radio wave energy due to reflection from a region with a high
neutral air concentration.
The presented in Figure 3 oblique ionograms for given distances define the
limits of the frequency range for radio waves reflecting from the F region at the
respective distances.
246
Methodology for Calculating the Parameters of Radio...

500

Oblique propagation
Distance 0-100 km
250 km
500 km
400
Virtual height [km]

300

200

100

0 2 4 6 8
Frequency [MHz]

Figure 3 Oblique ionograms based on the model profile of Figure 2


at radio communication distances 0-100 km (points),
250 km (crosses) и 500 km (diamonds)

CONCLUSION
The methodology proposed in the present paper is intended for servicing users
who carry out long-distance radio communications, as radio amateurs and gov-
ernment organizations. The methodology is suitable for organizing automatic data
processing. In order to make a forecast for the propagation of radio waves at a
given hour of the day, only three ionospheric characteristics are needed: foE, foF2
and MUF3000, which values can be obtained from both direct measurements and
forecasting models (Pancheva & Mukhtarov, 1996, 1998).

ACKNOWLEDGMENTS
The present work is supported by the Bulgarian Ministry of Education and Science
under the National Research Programme “Young scientists and postdoctoral
students” approved by DCM № 577/ 17.08.2018. The presentation of the
results is financed by Contract No D01-282/17.12.2019 - Project “National
Geoinformation Center (NGIC)” funded by the National Roadmap for Scientific
Infrastructure 2017-2023 of Bulgaria.

247
Rumiana Bojilova, Plamen Mukhtarov

REFERENCES
Appleton, E. V., (1925), Geophysical influences on the transmission of wireless
waves, Proc. Phys. Soc., 37, 16D.
Gadzhev G., Syrakov, D., Ganev, K., Brandiyska, A., Miloshev, N., Georgiev, G.,
Prodanova, M., (2011), Atmospheric composition of the Balkan region and
Bulgaria. Study of the contribution of biogenic emissions, AIP Conference
Proceedings, 1404, pp. 200-209. DOI: 10.1063/1.3659921
Gadzhev G., Ganev, K., Syrakov, D., Miloshev, N., Prodanova, M. (2012)
Contribution of biogenic emissions to the atmospheric composition of the
Balkan Region and Bulgaria. International Journal of Environment and
Pollution, 50 (1-4), pp. 130-139. DOI: 10.1504/IJEP.2012.051187
Ginzburg, V. L., (1964), Propagation of Electromagnetic Waves in Plasmas,
Pergamon Press, Oxford.
Di Giovanni, G., & Radicella, S. M. (1990). An analytical model of the electron
density profile in the ionosphere. Advances in Space Research, 10(11), 27-30.
Dudeney, J. R., & Kressman, R. I. (1986). Empirical models of the electron
concentration of the ionosphere and their value for radio communications
purposes. Radio Science, 21(3), 319-330.
Pancheva, D. V., & Mukhtarov, P. Y. (1996). A single-station spectral model of
the monthly median F-region critical frequency. Annals of Geophysics, 39(4).
Pancheva, D., & Mukhtarov, P. (1998). A single-station spectral model of the
monthly median foF2 and M (3000) F2. Studia Geophysica et Geodaetica,
42(2), 183-196.
Rawer, K. (2013). Wave propagation in the ionosphere (Vol. 5). Springer Sci-
ence & Business Media.
Stankov, S. M., Jakowski, N., Heise, S., Muhtarov, P., Kutiev, I., & Warnant,
R. (2003). A new method for reconstruction of the vertical electron density
distribution in the upper ionosphere and plasmasphere. Journal of Geophysi-
cal Research: Space Physics, 108(A5).

 Rumiana Bojilova
National Institute of Geophysics, Geodesy and Geography
Bulgarian Academy of Sciences
Sofia, Bulgaria
E-mail: [email protected]

 Plamen Mukhtarov
National Institute of Geophysics, Geodesy and Geography
Bulgarian Academy of Sciences
Sofia, Bulgaria
E-mail: [email protected]

248
Proceeding of 1st International conference on Environmental protection and disaster RISKs

Informatics, Remote Sensing, High Performance Computing


and Gis for Environmental Monitoring and Management
https://doi.org/10.48365/envr-2020.1.23

GIS IN CRISES MANAGEMENT USE OF SMART


PHONE GIS APPLICATION IN THE EVENT
OF NATURAL DISASTER

Aleksandar Petrovski, Nenad Taneski, Jugoslav Ackoski


University “Goce Delchev” Shtip
Military Academy “General Mihailo Apostolski” – Skopje, North Macedonia

Abstract. Today it has become both the government and citizens concern
to care human beings against ever-increasing crisis around the globe. There
are so many natural disasters like: forest fire incidents, earthquakes, floods,
eruptions, tsunami and etc. despite all this kinds of natural and man- made
disasters from a social perspective, it their potentiality to seriously and
irreversible damage regions. In addition the extent and severity of such
incidents may significantly affect the population. For that reason and the
fact that Geographic information system is recognized as a useful system
for alleviation the risk during the crisis management we develop the idea
to promote a new model to have a practical and applicable method of GIS
adoption within the crisis management in order to save as many lives as it can
be done. This study aims to come out with a model that will not allow drop
links. Each smart phone will have application in the event of natural disaster
and it switched due to the change of the vital function of man in trouble. This
will help departments respond as soon as they can and save lives as many as
they can and to reduce the percentage of people who could not call for help.
This work highlights the design challenges and required technical innovations
towards the goal of making GIS much more useful.
Keywords: GIS, disaster, crisis management, rescue, emergency response.

INTRODUCTION
The risks posed by climate change are real and its impacts are already taking
place. The UN estimates that all but one of its emergency appeals for humanitarian
aid in 2007 were climate related. Climate change is best viewed as a threat
multiplier which exacerbates existing trends, tensions and instability. The core
challenge is that climate change threatens to overburden states and regions which
are already fragile and conflict prone. The effects of climate change are being felt
now: temperatures are rising, icecaps and glaciers are melting and extreme weather
249
Aleksandar Petrovski, Nenad Taneski, Jugoslav Ackoski

events are becoming more frequent and more intense1). The following section
outlines some of the forms of conflicts driven by climate change which may occur
in different regions of the world. Some impacts from increasing temperatures are
already happening: Ice is melting worldwide, especially at the Earth’s pole. Sea
levels are rising faster over the last century. Precipitation (rain and snowfall)
has increased across the globe, on average. Other effects could happen later
this century, if warming continues. Sea levels are expected to rise between 7
and 23 inches (18 and 59 centimeters) by the end of the century, and continued
melting at the poles could add between 4 and 8 inches (10 to 20 centimeters).
Hurricanes and other storms are likely to become stronger. Floods and droughts
will become more common. Some diseases will spread such as malaria carried
by mosquito. Forest fires will became everyday situation, because of the rising
temperatures caused by climate changes and will be real danger in every place
on the earth, which will increase people movement in northern countries. This is
a huge problem for all of us, and if we don’t use all our technological resources
and knowledge to small the effects of climate changes and to put all investments
in prevention and reduction on it.

GIS IN CRISES MANAGEMENT


Climate change is a geographic problem, and we believe solving it takes a
geographic solution. GIS has a long history of driving environmental understanding
and decision making. Policymakers, planners, scientists, and many others
worldwide rely on GIS for data management and scientific analysis. GIS users
represent a vast reservoir of knowledge, expertise, and best practices in applying
this cornerstone technology to climate science, carbon management, renewable
energy, sustainability, and disaster management. When natural disasters alter the
world around us, high-quality data must be collected and examined to accurately
assess the damage and plan restoration efforts. Rapid response and attention to detail
are both of the utmost importance. Through the use of mobile GIS, these often-
opposing objectives can be achieved simultaneously. Operating PDAs equipped
with Arc Pad and GPS, trained field personnel can take advantage of custom forms
and other time-saving data entry features that render obsolete the use of pen, pad,
and paper map. Upon return to the office, field data is integrated into the project
geo-database with no additional data entry needed, allowing more time to analyze
the data and develop maps and reports.

SMART PHONE GIS APPLICATION IN THE EVENT


OF NATURAL DISASTER
For that porpoise it should be made a complex system of software for managing
crisis situations and disasters. The system it will be able to help in all aspects of
crisis management.
250
GIS in Crises Management Use of Smart Phone...

The system must comply with the objectives, functions and powers of the
territorial bodies of executive authorities and departments of the countries, in the
field of prevention, combating and elimination of criminal crises, terrorism, natural
and manmade, anti-terrorist and anti-criminal protection, safety and security.
Creating a system should be comprehensive interdepartmental aimed at improving
information technology and information and communication infrastructure of
public administration in general. To create a comprehensive system is to ensure the
safety of protection by reducing the probability of threats of natural, man-made,
criminal, terrorist and other situations due to:
– Effective monitoring of the current situation and providing information for the
authorities;
– Providing information about the current state of security facilities, protection
services and operational headquarters;
– Providing analysis and management of the threats, natural, man-made,
criminal, terrorist and other character crisis;
With an integrated system must be provided:
– Development of technical regulations (conditions) to equip the objects of
protection (and their components) technical means of security and control, as well
as the functioning of the monitoring tools and equipment (systems).
– The ability to monitor the status of security protection objects (their elements),
as well as the movement of persons and transport facilities for protection;
– Automation of the process of collecting and transmitting information to the
monitoring objects of protection (their elements);
– The possibility of collecting and transmitting information from existing and
emerging security products to protect and control objects;
– The possibility of collecting and transmitting information to the services of
territorial bodies;
– The possibility of sharing information between services (including situational
centers) with the task of ensuring human safety both on a bilateral and multilateral basis;
Must be implemented automated control mechanisms, information analysis
monitoring, forecasting situations. Integrated system it will be based on
geographically distributed principle and will be consisting of a single point of
management, transport network, as well as places of gathering, processing and
analysis of information.
The whole system it will consists of four phases.
• Phase one: Software that helps in locating, guiding and directing the people;
• Phase two: Flying drones that search and detect disasters and people;
• Phase three: Specialized machines for helping people;
• Phase four: human resources or specialized people working on field.
1st phase: Is to create a GIS software that will detect the crisis situation or disaster
and to classify it. (What sort of disaster (crisis situation) and the proportions of it).
251
Aleksandar Petrovski, Nenad Taneski, Jugoslav Ackoski

Then it will be created an appropriate application for all types of smart phones,
for all types of operating systems (Android, IOS, Microsoft ....). The application it
will have access to the location of the smart phone, calls, messages, emails and the
phone contacts. This will allow locating phones in crisis areas and access to vital
content on the phone (name of the user, numbers of the closest or most used phone
calls), for those phones that are turned on (operational)2).
The application will consist of three main functions / tools:
• Prevention or procedures and actions what to do before the disaster,
• Instructions what to be done during the crisis and
• Instructions after the crisis, how to take un appropriate measures and procedures.
1. The first function of the application will inform the users about the possible
natural disaster, what to do and how to do it. Directions on how to protect provide
or avoid possible situation (flood, forest fire, tsunamis, NHB pollution ...). It will
inform the consumers where to head, with the appropriate address, phone numbers
and how to reach the place with a particular map.
2. The second function it will be based on the morale and stability during the
disaster. How the users to remain calm and not to panic, but to think realistically
and rationally. Of course, the application will have all the previous functions,
where, how and what to do. The most useful part in this phase is that every person
can send real time SMS to crises centre with useful geo-data information that will
help to people on the ground about them and their condition.
3. The third tool will consist of basic survival manuals with pictures. And
measures where people to go and how to reach the most basic needs.
For all non-smart phones and cell phones that don’t have GPS, this will also be
possible. It can still be tracked. This is because a cell phone is basically an advanced
two way radio in which communication are made via cell towers. These cell towers
are within a network of cell towers which its main function is send and receive the
radio signals emitting from your phone. And cell phone is basically transmitting
radio signal to the nearest cell tower. The closer your phone to a cell tower, the
stronger the signal that is emitted. So, by measuring the signal strength and also
through triangulation method with other cell tower; that is by measuring how
strong or weak signal emitted by your phone, your phone location can be pinpointed
almost accurately. However, there are some disadvantages to this technique as big
buildings or trees can affect the signal and therefore affect the signal strength. All
users of these phones will receive un SMS and MMS with instructions.
2ndphase: The second phase will be consisting of a series of little flying drones
which it will be equipped with a various sensors and cameras (video live stream,
thermal scanning, structural laser scanning, Wi-Fi hotspots servers, signal scanner
and receiver). Sensors for identifying the crisis and sensors for tracking and
founding people. All flying drones will be communicating with one HQ that will
control the drones.
252
GIS in Crises Management Use of Smart Phone...

Figure 1. Parrot beob drone

3rdphase will be for those who cannot reach the emergency centers and basic
needs. On them will be sent help from ground and air with adequate machines
(robots and drones) in the form of food, water, first aid and medicinе. The machines
will choose where to go alone and where to deliver the packages.
4thphase will consist of specialized teams for different kind of disasters and
teams dedicated for managing crisis situations. Every team will be equipped with
equipment and tools for dealing with every situation on field and in every HQ. This
phase will have an option of including the Army of the country that has a crisis
situation or every available people.

Figure 2. Structure of Geo HQ organization


253
Aleksandar Petrovski, Nenad Taneski, Jugoslav Ackoski

NAMED DISTANCE MODEL


The second model named distance model is more complex and its without UAV,
the position of the soldier is send via the VHF military radio from the built-in GPS.
The distance and the angle to the target is determinate by the two team members
using Laser Rangefinder and send to the TOC using voice communication3). There
will be addition to the algorithm that will calculate the coordinates of the target
using the coordinated of the rescue member, and the known distance to the target.
This is done using simple geometry to determine third point when two points and all
sides (distance) are known (Charles & Grinstead) and is calculating the coordinated
of the victim and send them as an input to the algorithm with the received GPS data
from the radios. The following steps are same as the UAV model.
System architecture of this model is represented in the following figure. Two
team members use standard radio and laser rangefinder, and then the data is send
via land mobile radio, to the team leader, and forwarded via Harris RF7800M and
VHF Radio network to the TOC where the algorithm calculates the coordinate of
the victim and makes the decision in accordance to the created algorithm.

Figure 3. System architecture for the distance model

ALGORITHM
Steps of the algorithm:
First: for all our team members we are checking the visibility to the goal by our
GIS connected by UAV orvoice communication.
Second for team member that have visibility we measure the distance to the victim.
Third we get the information for weather conditionsfrom nearby weather stations
or our meteoritical data.
Fourth from the appropriate tables we get the appropriate probabilities to save
the victim.
254
GIS in Crises Management Use of Smart Phone...

Fifth (and finally) the team member with the biggest probability to save the
victim gets an order to save the man in trouble.

EXPLAINING ONE REAL SITUATION ON THE GROUND


We are giving real description of our algorithm. We have five team members
and for them we know the number of saves to the goal from 100 attempts. Also we
have number of saves in different weather condition: sun weather, haze weather,
poor rain, and straight rain. Another important characteristic that we must to now is
distance from the team member to the man in trouble.

Table 1. Sun weather


SUN WEATHER
Soldier 1 100 attempts 600 meters 0,93 goal
Soldier 2 100 attempts 800 meters 0,749 goal
Soldier 3 100 attempts 800 meters 0,89 goal
Soldier 4 100 attempts 800 meters 0,91 goal
Soldier 5 100 attempts 500 meters 0,98 goal

Table 2. Haze weather


HAZE WEATHER
Soldier 1 100 attempts 600 meters 0,793 goal
Soldier 2 100 attempts 800 meters 0,65 goal
Soldier 3 100 attempts 800 meters 0,79 goal
Soldier 4 100 attempts 800 meters 0,79 goal
Soldier 5 100 attempts 500 meters 0,86 goal

Table 3. Poor rain weather


POOR RAIN WEATHER
Soldier 1 100 attempts 600 meters 0,80 goal
Soldier 2 100 attempts 800 meters 0,82 goal
Soldier 3 100 attempts 800 meters 0,89 goal
Soldier 4 100 attempts 800 meters 0,85 goal
Soldier 5 100 attempts 500 meters 0,73 goal

Table 4. Straight rain weather


STRAIGHT RAIN WEATHER
Soldier 1 100 attempts 600 meters 0,39 goal
Soldier 2 100 attempts 800 meters 0,41 goal
Soldier 3 100 attempts 800 meters 0,31 goal
Soldier 4 100 attempts 800 meters 0,55 goal
Soldier 5 100 attempts 500 meters 0,54 goal

255
Aleksandar Petrovski, Nenad Taneski, Jugoslav Ackoski

On the this picture we have example of one possible situation on the field. Our
three (blue circle) and one man in trouble (red circle) team members are at the
positions.

Picture 1. Real position on the field

ALGORITHM WORKFLOW
First: From the ground situation team member with number 1 and 3 have
visibility to the goal.
Second: The distance from team member with number 1 to the victim is 600 m
and the distance from team member 3 to the victim is 800m.
Third: The team member from the field reports us that the weather is sunny, and
we are checking weather condition on android application.
Fourth: From the tables, we can see that probability for team member with
number 1 is 0,93 and 0,89 for team member with number 3. This calculating will be
automatically done by the application in our TOC center.
Fifth: Algorithm gives us final result that team member with number 1 is most
suited to finish the task.
Six: The commanding centar gives the executive order.

CONCLUSION
This integrated system should allow:
256
GIS in Crises Management Use of Smart Phone...

– Ensure compatibility of existing and emerging information and technical


systems to improve the efficiency of management in the field of safety of life of the
population;
– Increase the effectiveness of control objects with mass stay of people and
critical facilities;
– Improve the efficiency of solving the territorial problems;
– Increase the effectiveness of measures to eliminate the consequences of natural
and man-made disasters;
– To reduce the loss of life in emergencies, fires, earthquakes.
Every day, we are confronted with disasters of varying degrees. Those that
have adequately developed, maintained, and exercised their contingency plans
will survive. Yet many people continue to take the uninhibited operations of
their lives for granted. They remain complacent, assuming that the power will
always be available, the telephone system will not fail, there will be no fire or
earthquake--everything will always be normal. Very few people plan for their own,
if we want to survive, organizational “strategic” and “tactical” battle planning is
essential. However, it is only as good as the foundation upon which it was built. The
foundation is, of course, the concept.
This document is the means by which a particular mission, program, or policy
directive is translated into a fundamental organizational and operational methodology.
Once the system is developed, and is sanctioned by both management and the operating
elements, construction of the contingency plan may commence. A fundamental
premise of successful contingency planning is that plans are developed by those who
must actually carry them out in the event of an actual disaster. Disaster planning is
truly a vital part of the overall business plan. With every year rising disasters the
people are more and more threatened. We are hoping that with this software will
help them all. The software is expected to decrees the casualties all around the world,
in the crisis areas. And also organize the management in these situations. The low
cost of the software is a big advantage in the development of the system (Petrovski,
Bogatinov & Boshkovski, 2017).

NOTES
1. http://environment.nationalgeographic.com/environment/global-
warming;
2. http://myphonelocater.com/category/triangulation;
3. Disaster Recovery Plan Strategies and Processes February 2002 pp 11-12.

REFERENCES
Charles M. Grinstead, J. Laurie Snell. Introduction to Probability. 520.
257
Aleksandar Petrovski, Nenad Taneski, Jugoslav Ackoski

Tortonesi M, Stefanelli C, Benvegnu E, Ford K, Suri N, Linderman M.


Multiple-UAV coordination and communications in tactical edge
networks. IEEE Communications Magazine. 2012 Oct 1;50(10):48-55.
Petrovski A., Bogatinov D., Boshkovski D., GIS IN ARMY: Usage of
GIS in shooting with weapon in purpose to minimize resources and
maximize the results, DOI: http://dx.doi.org/10.18509/GBP.2017.15,
UDC: 528:004]:623.6.09, proceedings Geobalcanica 2017, Skopje

 Dr. Aleksandar Petrovski, Assoc. Prof.


ORCID iD: 0000-0002-5265-5813
University “Goce Delchev” Shtip
Military Academy “General Mihailo Apostolski”
Vasko Karangjeleski bb
Skopje, North Macedonia
E-mail: [email protected]

 Nenad Taneski
https://orcid.org/0000-0002-6471-0189
University “Goce Delchev” Shtip
Military Academy “General Mihailo Apostolski”
Vasko Karangjeleski bb
Skopje, North Macedonia
E-mail: [email protected]

 Jugoslav Ackoski
https://orcid.org/0000-0003-2782-3739
University “Goce Delchev” Shtip
Military Academy “General Mihailo Apostolski”
Vasko Karangjeleski bb
Skopje, North Macedonia
E-mail: [email protected]

258
Proceeding of 1st International conference on Environmental protection and disaster RISKs

Informatics, Remote Sensing, High Performance Computing


and Gis for Environmental Monitoring and Management
https://doi.org/10.48365/envr-2020.1.24

FRAMEWORK FOR WILDFIRE DANGER


PREDICTION SYSTEM

Nina Dobrinkova1, Maria Asenova2


1
Institute of Information and Communication Technologies – Bulgarian Academy of Sciences
2
University of Forestry, Faculty of Forestry, Department of Forest Management

Abstract: Wildland fires are natural hazard that occur every year in different
parts of the world. European Mediterranean countries often are suffering from this
disaster and in some cases human casualties plus infrastructure damages are in
place. Bulgaria as part of the north east Mediterranean basin is affected by wildland
fires in the last three decades more often than before. In our article we will present
a framework for wildfire danger prediction system by usage of wireless sensor
networks collecting real data and elaborating it into alerts. The test area is located
on the territory of South-West state-owned forestry district. The described approach
is based on the national fire danger rating algorithm, but in our case downscaled to
daily prognosis. The system is under construction at the time of the paper writing
thus the framework of the system main functionalities is the main scope of the
article.
Keywords: wildfire, fire danger, sensor network.

INTRODUCTION
Wildland fires are becoming a global threat in the last decades worldwide. They
are damaging forests and every economy sphere connected to the forestry products.
As a result, after wildland fire vast amounts of valuable timber is lost. Other direct
negative impacts are disruption and degradation of forest plantations, deforestation
of big areas, deterioration of the protective functions of forests, reduction of the
forest ecosystem water protection role, increasement of soil erosion, especially in
mountain areas with steep slopes. Agricultural conditions become less productive
in zones after fires. Animals, people, industrial buildings and housing, critical
infrastructure, other important facilities are endangered if they are affected directly
or indirectly by firestorms.
Since the middle of the last century, the problem of wildland fires gets a critical
dimension due more often and more dangerous scale of the fire behaviour. This
problem mostly affected large forest areas in the USA, Canada, Australia, Russia
and Southern Europe countries. Substantial human, financial, material and scientific
259
Nina Dobrinkova, Maria Asenova

resources are allocated for finding a solution of the above problems associated with
the adverse effects of forest fires and finding a solution for them at national, regional
and international level. With this regard significant funds are allocated also in EU
programs for wildland fire protection of the Union’s forests. Main objectives of the
resource provision programs are measures for fire protection with reduction of the
number and size of the burnt areas.

Wildland Fire Occurrence in Bulgarian Forests


In 2016 (Lyubenov, 2016) has done an analysis of fire activity in Bulgarian
forests based on the official published data from the European Fire Database (EFD).
The analysis and the follow-up assessments are based on data for the period 1990 –
2013 (24 years). The total wildland fires in numbers for the period is 12 905 fires or
538 average fires per year. Total burnt areas for the same period are 234 858 ha that
represents 6.02% of the forest land of Bulgaria, or 0.25% thereof is burnt average
per year. 0.14 fires occur average per year on area of 1000 ha (10 km2), and the area
burnt thereby is 2.51 ha per 1000 ha forest land. Average area of a medium fire is
18.2 ha.
The dynamics of the number of forest fires occurred for the 1990 – 2013 period
shows that the number of fires per year is explicitly dynamic figure characterising
with significant unevenness in different years for the different Bulgarian regions.
Other specification of this fire activity indicator is that it is characterised with
relative permanence and clearly demonstrate cyclical peaks over the years, as for
the period 1990 – 2007 these peaks occur exactly over seven years — 1993, 2000

Figure 1. Dynamics of the number of forest fires occurred in Bulgaria


for the 1990 – 2013 period (based on (Lyubenov, 2016)).
260
Framework for Wildfire Danger Prediction System

and 2007. The last peak for the period occurs after 5 years — in 2012.The values of
peaks (maximum levels) for 1993, 2000, 2007 and 2012 exceed the average value
for the period for the number of fires occurred respectively 2,21; 3,18; 2,75 and
1,63 times more (Figure 1).
Figure 2 shows the dynamics of magnitude of burnt areas for the same period
(1990 – 2013). The chart clearly shows the large unevenness and direct relation of
this magnitude with the data about the number of fires occurred (Fig. 1). For the
peak years of the fire ocurence — 1993, 2000, 2007 and 2012 the size of burnt area
exceeds the average for the period respectively 1,86; 5,87; 4,39 and 1,3 times more.

Figure 2. Dynamics of the burnt areas in Bulgaria


for the period 1990 – 2013 (based on (Lyubenov, 2016)).

The analysis of fire activity and forest fire risk in the Bulgarian forest areas
evaluated by (Lyubenov, 2016) was proved to be constant, with dynamic character
and peaks on every 5 to 7 years.

METHODOLOGY DETERMINING FOREST FIRE RISK IN BULGARIA


The first attempt for determining the “fire risk” in Bulgaria was made in 2000
(Lyubenov-Panov, 2000). When determining the so called “fire risk” as main
indicator it was used the average annual burnt area for the period 1991 – 2000
by Regional Forest Departments (RFD). The existing forest fire databases were
elaborated and summarised for the first fire risk estimation results for Bulgaria.
A scale for determining the “risk level” by RFD and the relevant ranking was
261
Nina Dobrinkova, Maria Asenova

developed at the same time. Table 1 shows the results for “fire risk level” by RFD
based on (Lyubenov-Panov, 2000) for 9 years of the analysed period.

Table 1. Fire risk level by RFD (1991 – 2000)


(based on (Lyubenov-Panov, 2000))
Fire risk level Regional forest departments
Ruse, Varna, Blagoevgrad, Shumen, Veliko Tarnovo,
low
Smolyan
medium Sofia, Plovdiv, Burgas, Pazardzhik
high Berkovitsa, Lovech, Kyustendil
very high Kardzhali, Stara Zagora, Sliven

Ranking according to that indicator was concluded that gives significant


disparities between different RFD. For example: RFD-Kardzhali reported values
of burnt average annual area exceeded the values in RFD-Smolyan and RFD-
Blagoevgrad respectively with 38.5 and 12.3.
On the other hand, the number of fires occurred average per year in RFD-
Kardzhali exceeds these in RFD-Smolyan and RFD-Blagoevgrad respectively
with 1.5 and 2.2 times. If calculating the average area of a contingent fire for
the example RFDs used only the average fire area than the Kardzhali case was
exceeding the values for Smolyan and Blagoevgrad respectively with 27.5 and 5.9
times. This approach was not the best solution for future fire risk estimations. Thus,
the conclusion for the first fire risk attempt for Bulgaria was that the forest fire
risk established only by average annual burnt area indicator is not sufficient for
Bulgarian regions.
In 2008, immediately after the second major peak of fire activity in Bulgarian
forests, a new methodology for determining the forest fire danger risk was proposed
by (Lyubenov-Konstantinov, 2008). The essence of the method consisted from
a common (inclusive) criterion for determining the “fire danger (forest fire risk
level)” in a forest area. The key indicators for characterizing fire activity in a forest
area that were used for the determination of this criteria were: average number
of fires occurred in a given territory and the average area annually burnt for the
same zone. The outcome of combining these two indicators was the integrated fire
danger assessment criteria called after for short: “forest fire risk in the area”. The
cases could be refined where the number of fires was high and burnt area relatively
small or the other way round — a small number of fires and large burnt area. Such
approach indirectly took into consideration the average burnt area per fire. The
summarised results for the second attempt for fire risk estimation was based on
forest fire data sets per RFD for a period of 18 years (1991 – 2008) by (Lyubenov-
Konstantinov, 2008). In Table 2 is given a summary of the main outcomes of the
second forest fire risk assessment for Bulgaria:
262
Framework for Wildfire Danger Prediction System

Table 2. Fire danger level by RFD (1991 – 2008)


(based on (Lyubenov-Konstantinov, 2008))
Fire danger level Degree level Regional forest departments
I low Shumen, Blagoevgrad, Veliko Tarnovo, Ruse
II medium Smolyan, Burgas, Varna, Pazardzhik
III high Sofia, Plovdiv, Kyustendil, Sliven
IV very high Lovech, Stara Zagora, Berkovitsa, Kardzhali

Table 3 represents the summarized fire danger data in forest areas of the
country of Bulgaria according to the second proposed methodology by (Lyubenov-
Konstantinov, 2008) and its following relevant estimates for the 18 years period.

Table 3. Summarized data on forest fire danger


(based on (Lyubenov-Konstantinov, 2008))
Fire danger extent Number of fires Average burnt area per year,
Burnt area, %
and level occurred, % % of the total for Bulgaria
I – low 13.88 5.23 0.065
II – medium 22.40 13.18 0.138
III – high 29.60 26.57 0.288
IV – very high 34.12 55.02 0.605
For the country 100.0 100.0 0.270

After these two attempts for establishment of common methodology for fire
danger level in Bulgaria general conclusion was made that the dynamic values in
the estimations are very dependent on the period that is determined for evaluation.
In 2016 has been done a third version of the Methodology determining forest
fire risk in Bulgaria by (Lyubenov, 2016). The basic inputs that he used for his
estimations were:

1. Historical period of assessment of only 10 years (2006-2015),


2. Number of fires reported regardless of the size of the area burnt,
3. Division of the country of Bulgaria by administrative units – districts under
NUTS3 code (28 for the territory of Bulgaria)
4. The areas (wooded or non-wooded) were considered burnt if signs of fire
were visible on combustible matter in the living or dead forest,
5. The total classes for fire danger extend level was limited to only 3 classes
(low, medium and high).

The risk of forest fires is determined by an integrating indicator (Rf.risk), which


is a product of the numerical values for the density of forest fires (Rdens.) and the
263
Nina Dobrinkova, Maria Asenova

actual flammability (Rf.comb.) of the territories per districts in Bulgaria (demonstrated


in eq. (1)):

R f. risk = R dens. × R f. comb. (1)

The classification according to the degree of forest fire risk is performed


according to the scale shown in Table. 4, according to the indicator (Rf.risk).

Table 4. Determining the degree of forest fire risk (based on (Lyubenov, 2016))
Values of the integrated indicator Rf.risk Degree of forest fire risk
Rf.risk ≤ 0,1 Low (green)
Rf.risk > 0,1 и ≤ 0,3 Medium (yellow)
Rf.risk > 0,3 High (red)

Based on these criteria the third methodology developed by (Lyubenov, 2016),


estimated that the average values for the country which are: average forest cover -
34.52%, average density of fires - 0.13, average flammability of forest areas - 2.11
and integrated indicator for the forest fire risk for the state is - 0.273.

Figure 3. Ranking of the districts in Bulgaria according to an integrated


indicator of forest fire risk based on the third methodology for
the estimation period (2006-2015) (based on (Lyubenov, 2016)).
264
Framework for Wildfire Danger Prediction System

OTHER FOREST FIRE RISK APPROACHES


The current Canadian Forest Fire Danger Rating System (CFFDRS) has been
developed for more than 10 years combining four separate modules:

1. Fire danger assessment according to the weather conditions (FWI);


2. Forest fire behaviour prediction (FBP);
3. Assessment of flammable materials moisture content (AFMS), and
4. Forest fires occurrence predicting system.

The Canadian system attempts to unite the three individual components of the
overall forest fire danger in one indicator (constant fire danger, dynamic fire danger
and fire risk). This system works only in Canada and is recommended for use in
the Russian boreal forests (Sofonov, 2005). The FWI module determining only
dynamic fire danger is used by the Joint Research Centre (JRC) of EC to predict the
dynamic fire danger (according to weather conditions) for all countries using the
European Forest Fire Information System (EFFIS).
In USA the national Fire Danger Rating System (NFDRS) is established for the needs
of the USA Forest Department. According to this system forest areas of USA from Alaska
to Florida are divided in four groups by types of flammable materials: grasslands, bushes,
wooded and lands occupied by clearing residues. Additionally, the system requires
division of combustible materials in 20 separate types. Separate combustible model is
developed for each type serving as basis for defining the fire danger level. This system is
in limited use in New Zealand and South African Republics (Sofonov, 2005).
In France based on (Lyubenov, 2016) for the most fire dangerous French
region — Mediterranean, has been adopted methodology for determining the risk
measure named “average annual risk” (AR, %) based on statistical data for certain
period of years. This methodology also attempts to combine both indicators for
fire activity assessment in the given region — percentage of burnt area for the
period and likelihood of fire occurrence in the same area expressed in years. Such
approach and evaluation for determining the “forest fire risk level” is not applicable
for Bulgaria. Such values in percentage of burnt forest area annually for a region
can not be reached in other countries and regions. Furthermore, this methodology
does not take into account important measurable indicator as the number of fires
occurred averagely per year or for the reported period.
In Russia has been consistently worked on the issue of classification of forest
areas by “fire danger level”. A scale for determining the so-called “real combustibility
of forests” (Lomov, 2007) was developed and validated for Russia. The Russian
approach complies the general requirements for classification of forest areas by
“forest fire risk”. However, its direct application in Bulgaria is not possible, given
the scales of regions and field work done for the Russian Federation forestry zones.
A newer methodology and scale including integrated indicator associated with both
265
Nina Dobrinkova, Maria Asenova

the number of fires and burnt average annual area for a given forest area (Sofronov,
2005) is proposed in Russia to the current date, which will increase the accuracy
of the assessment of “real combustibility of forests”. Although more precise this
scale with the integrated indicator “aggregate perimeter” is likely to be used for
plain terrains where in equal other conditions the shape of burnt area is relatively
correct, approaching circular one. For oblique, steep and rugged terrain the shape is
extremely complex and this classification is difficult to be applied.
In Germany the fire activity assessment is determined in “fire danger classes”
by evaluating different forestry units (Misbach, 1989). This approach is applied in
the former German Democratic Republic (east German provinces) where forest fire
activity recently is increasing.
This methodology divides forestry in four fire danger classes:

1. Fire danger class A1: forests where three or more fires have occurred per
year, with sizes of burnt area larger than 100 ha and with more than 50 ha burnt
area per 100 000 ha forest land.
2. Fire danger class A: forests with over 50 ha completely burnt area per 100
000 ha per year.
3. Fire danger class B: forests with completely burnt area of 5 to 50 ha per 100
000 ha per year.
4. Fire danger class C: forests with less than 5 ha completely burnt area per 100
000 ha per year.

The first attempt for classification of forest fire risk areas in EU is for the Southern
European countries (Portugal, Spain, Italy, France and Greece). That study has been
launched with a pilot project for creation of an forest fire information system using
the database for the period 1989 – 1993 described in details in (Regulation (EES)
№ 2158/92). For areas with high fire risk are considered those where 4,000 or more
fires occur averagely per year. The number of fires may also be smaller, but if the
average size per fire exceeds 30 hectares the risk remains high again. As high-risk
under these conditions are classified only areas in Portugal, Spain and Greece plus
7 regions in France and 11 in Italy defined in (Regulation (EEC) No 2158/92). In
1997, the European Commission established a working group for development and
proposal of methods for forest fire risk assessment and evaluation of burnt areas
in the European Union discussed in details in (Forest Fires in Europe: Report №
11/2010). This group is part of the Institute for Environment and Sustainability of
the European Commission and is part of the Joint Research Centre (JRC) in Ispra.
The work of this group ends with the development of the European Forest Fire
Information System (EFFIS) launched in 2000. This system was approved by the
European Council and European Parliament with Regulation (EC) No 2152/2003
(Forest Focus) as a tool for monitoring of forests and environment.
266
Framework for Wildfire Danger Prediction System

On 23.02.2005 was held the third meeting of the Working Group on forest fires,
which initiated a discussion called: “Key proposals to EU 25 for the classification
of forests by fire risk”. The concept of “fire risk” and components included therein
was introduced for the first time at that meeting and since than every EU member
state is obligate to provide information for its forest fire risk zones. However there
is still lack within the European Union of a unified methodology for determining
the classes (degree) of fire risk or forest fire risk. Because of that reason is required
each EU Member State to develop its own methodology to classify their forest areas
by forest fire risk and monitor them with state specific systems and technologies.

FRAMEWORK FOR WILDFIRE DANGER


PREDICTION SYSTEM
The chronological overview of the fire r isk a ssessment m ethodologies in
Bulgaria and the most famous approaches used worldwide gave to the team of IICT-
BAS a good start to work for a pilot study case as part of the National Scientific
Program for Environmental protection. Our test area is located in the south-west
state forestry district. The zone is classified with medium range of forest fire risk
based on the methodology of (Lyubenov, 2016). That part of Bulgaria has large
number of touristic zones in the mountains of Pirin and Rila. This is the reason why
most of the wildfires in south-west district are caused by human mistakes working
with fire in open spaces.
After careful evaluation we have estimated that a combination between Wireless
Sensor Network and the Canadian Fire Danger Rating System approaches will be
the best for our test area. The sensor technology offers many benefits and a lot of
features in detecting fire in the forest which cannot be achieved through traditional
physical observations. Our approach includes module-based system that consist in
each node of specially designed sensor stations collecting information for: rain, soil
temperature, soil moisture for shallow depth, wind speed and wind direction, air
temperature, humidity and atmospheric pressure, data logger, solar panel, battery
and SIM card for transmission of the collected data.
The features of Wireless Sensor Technology that have sensors deployed in the
field provide high coverage for the analyzed area, low costs, scalability, network
access, greater resiliency, hypervisor protection against network attacks and less
maintenance in energy compared to the other technologies (Lee et. all, 1999). Our
sensor stations will have solar panels with battery that will feed the sensors with the
needed power consumption.
The proposed solution is based on three core components Wireless Sensor
Network, Processing software and Web Application with embedded alerting
functionalities. The components are coupled with each other so that the early
detection of the forest fires can be achieved and alerts to predefined groups of
responsible actors can be sent out.
267
Nina Dobrinkova, Maria Asenova

The System deployment consists of three phases – preparatory, instalment with


tests and implementation into every day work.
The preparatory phase took us a year and covered the following activities –
selection of wildland test territory, field observations and sensor network map
elaboration. The map on (figure 4) is the final outcome of this first phase of
evaluations and selections. The test zone is located in the south-west forestry district
nearby Kresna Gorge. It has been first evaluated about its geography – flat or hilly
and what kind of vegetation is covering that area. As second step has been done
sensor network estimation for high coverage with fewer nodes. The final phase of
the preliminary evaluations ended with creation of the node map and achieving
special permission by the south-west forestry district director for instalment of the
sensor network stations.

Figure 4. The test case sensor network in the zone of south-west forestry district
of Bulgaria located nearby Kresna Gorge.

The second phase where is planned to be the instalments with tests is still in its
beginning. It will cover the system architecture development in practice (figure 5).
The system main three components are Fire Weather Observation, Fuel Moisture
and Fire Behaviour. This modules are based on the Canadian Fire Danger Rating
System and are discussed in details in (Bagheri 2007) and (Groot, 1998). Our
main goal is to build up the physical network and develop the software collecting
the sensor data, elaborating it into alerts and presenting it into daily maps that in
cases of potential fire occurrence will transmit automatically alerts to predefined
stakeholders.
268
Framework for Wildfire Danger Prediction System

Figure 5. The test case sensor network in the zone of south-west


forestry district of Bulgaria located nearby Kresna Gorge.

The test case sensor network is supposed to be physically available to transmit


its data by March 2021, with the idea to work on a small area of south-west forestry
giving demo case working in practice for vulnerable zone. In case this test case is
giving sufficient results, this architecture will be further expanded and elaborated.

CONCLUSIONS
The proposed framework for a sensor network with the capacity of fire danger
rating and early identification of starting forest fire is something that has not been
done so far in Bulgaria. In the last seven years Firewatch towers have been placed
in vulnerable areas, however this technology is very expensive and after the end
of the project that funded its building the hardware maintenance is gone. This led
to the situation that facilities with average price 300 000 euro each start to fail
doing its job and eventually stop working. Our approach is aiming to provide low
cost, but effective tool for everyday use and alerting system that can be reliable to
capture fire danger and the initial source of fire start.

ACKNOWLEDGMENTS
This work has been supported by the NNP-OS contract number Д01-322/18.12.2019
and the Bulgarian National Scientific Fund project number DFNI DN12/5
“Efficient Stochastic Methods and Algorithms for Large-Scale Problems”.

269
Nina Dobrinkova, Maria Asenova

REFERENCES
Lyubenov, K. 2016. “Methodology for determining the risk of forest fires
on the territory of the country of Bulgaria” for the needs of measure 8
"Investments in forest areas - development and improvement of forest
viability" of the funding period (2014 - 2020). (in Bulgarian)
Lyubenov, K., P. Panov. Forest fires in Bulgaria (1991 – 2000). Forester
Ideas journal, No. 2/2000.
Lyubenov, K., Vl. Konstantinov. Analysis and evaluation of fires and forest
fire danger. Forester Ideas journal, No. 1-2/2008
Sofonov, M.A., I.G. Holldammer et al. Fire danger in natural conditions.
Krasnoyarsk, 2005 (book).
Lomov, V.D., S.N. Volkov Easy Pyrology Moscow, 2007 (book).
Karl Misbach. Forest fires. Sofia, 1989 (book)
Forest Fire in the South of the European Union (Regulation (EES) №
2158/92). Brussels, 1996
Council (Regulation (EEC) No 2158/92) on protection of the Community’s
forests against fire OJ No. L 217, 31.07.1992.
Lee, S.-J., Gerla, M., & Toh, C.-K. (1999). A simulation study of table-
driven and on-demand routing protocols for mobile ad hoc networks.
Network, IEEE, 13 (4), 48-54.
Bagheri, M. (2007). Efficient K-Coverage Algorithms for Wireless Sensor
Networks and Their Applications to Early Detection of Forest Fires.
Computing Science, SIMON FRASER UNIVERSITY. MSc: 75.
Groot., W. J. (1998). Interpreting the Canadian Forest Fire Weather Index
(FWI) System. In Proc. of the Fourth Central Region Fire Weather
Committee Scientific and Technical Seminar, Edmonton, Canada.

 Nina Dobrinkova
https://orcid.org/0000-0002-6506-748X
Institute of Information and Communication Technologies
Bulgarian Academy of Sciences
Sofia, Bulgaria
E-mil: [email protected]

 Maria Asenova
https://orcid.org/0000-0002-9660-0275
Department of Forest Management
Faculty of Forestry
University of Forestry
Sofia, Bulgaria
E-mail: [email protected]

270
Proceeding of 1st International conference on Environmental protection and disaster RISKs

Informatics, Remote Sensing, High Performance Computing


and Gis for Environmental Monitoring and Management
https://doi.org/10.48365/envr-2020.1.25

EVALUATION OF SENTINEL-3A OLCI OCEAN


COLOR PRODUCTS IN THE WESTERN BLACK SEA

Violeta Slabakova, Snejana Moncheva,


Natalyia Slabakova, Nina Dzhembekova
Institute of Oceanology – Bulgarian Academy of Sciences (IO-BAS)

Abstract: Ocean color remote sensing has entered a new era with the successful
launch of the Ocean and Land Colour Instrument (OLCI) on board Copernicus
Sentinel-3 satellites. The representativeness and accuracy of the OLCI geophysical
products are poorly quantified in the Black Sea which is featured by high bio-
optical complexity. This paper addresses the evaluation of the OLCI/3A sensor's
performance for retrieving ocean color products of the western Black Sea waters,
by carrying out qualitative and quantitative match-up analysis between in situ
and OLCI data. Results indicate a systematic underestimate of the water-leaving
radiances (more pronounced at 412 and 665 nm) in contrary to the OLCI standard
Chlorophyll (CHL) products which significantly overestimate the in situ records.
These findings confirm the need of consistent reference in situ bio-optical data with
improved spatial and temporal coverage for further satellite ocean color missions
calibration and validation activities in the Black Sea.
Keywords: OLCI, ocean color, Black Sea, validation

INTRODUCTION
The Ocean and Land Color Instrument (OLCI) on board the Sentinel-3
satellites as a joint European Space Agency (ESA) and EC initiative for Earth
observations – Copernicus has been designed to provide continuous global and
regional observations of the land and ocean surface at high accuracy allowing
operational monitoring of environmental parameters and climate studies. Of partic-
ular interest are the ocean colour measurements related to coastal zone monitoring
in order to support environmental assessment of water quality and risk phenomena
(such as harmful algal blooms) and to improve management of coastal habitats
(EUMETSAT, 2018а). The OLCI sensor is a follow up of Medium Resolution
Imaging Spectrometer (MERIS) on board the ESA/ENVISAT platform (Don-
lon et al.,2012). It incorporates several significant technological improvements
comparing to MERIS including a) increased number of spectral bands from 15 to
21, b) reduction of sun glint effects, c) better global coverage, d) improved spatial
271
Violeta Slabakova, Snejana Moncheva, Natalyia Slabakova, Nina Dzhembekova

resolution of 300 m and signal-to-noise ratio. Calibration and validation of ocean


color products in open-sea and coastal waters are critical element of each satellite
ocean color mission (Hooker et al., 2000). The most fundamental among the pa-
rameters determined from space are the water-leaving radiance LWN and the remote
sensing reflectance Rrs. The uncertainties in quantification of these primary ocean
color products may affect the subsequent estimation of optical (e.g. absorption and
back-scattering coefficients) and biogeochemical products (e.g. chlorophyll (CHL)
and Total Suspended Matter (TSM) concentrations), which further may challenge
the assessment results.
The Black Sea is an extraordinarily complex water body for ocean colour
remote sensing, as it belong to Case 2 waters (Kopelevich et al., 2004), which are
characterized by relatively high absorption by Colored Dissolved Organic Matter
(CDOM) while the concentration of non-pigmented particulate matter does not co-
vary in a predictable manner with chlorophyll a (Babin et al., 2003; Churilova et
al., 2017). The optical complexity of the Black Sea is the reason why the standard
bio-optical algorithms developed for Case 1 waters, are the source of large uncer-
tainties (of the order of hundreds of percent) of chlorophyll a concentration in the
coastal areas (Sancak et al., 2005, Oguz, & Ediger, 2006, Slabakova et al., 2011).
Nonetheless, results demonstrate reasonable performance of ocean color remote
sensing, at least at local scales and/or in the open Black Sea. Different remote sens-
ing algorithms for determination of optically active constituencies have been devel-
oped or adopted for the Black Sea (Zibordi et al, 2015; Suslin& Churilova, 2016;
Kopelevich et al, 2013). Some of the studies rely on MERIS data and the results
can be implemented on Sentinel-3 OLCI data as well. The aim of this research is to
evaluate the performance of Sentinel-3A OLCI primary ocean color products and
standard Level 2 algorithms for CHL retrieval based on comprehensive bio-optical
in situ reference data set collected in western Black Sea.

DATA AND METHODS


In situ data
In order to increase the number of match-ups in the assessment of Sentinel-
3A OLCI ocean colour products, we used in situ data from three different sources
(Fig.1).
AOP of seawater were determined from multispectral free-fall system (Satlantic
Micro profiler) which simultaneously measures the upwelling radiance Lu (z, λ),
the downward irradiance Ed (z, λ) and the upward irradiance Eu (z, λ) as a function
of depth z and wavelength λ, in addition to the above–water downward irradiance
Ed (0+, λ). Data products from the free-fall optical profiler include spectral values
of: irradiance reflectance, remote sensing reflectance, normalized water–leaving
radiance, diffuse attenuation coefficient and the so called Q-factor. The processing
and quality assurance of data products are made using the Optical Processor
272
Evaluation of Sentinel-3A OLCI Ocean Color Products...

Figure 1. Location of measurement stations: 1) star, LWN obtained


from the Galata AERONET-OC; 2) Triangles, Apparent optical properties (AOP)
measurements and subsurface water samples for CHL analysis collected
during Bio-Opt 2016 and 2019 campaigns and 3) Circles, subsurface water
samples for CHL analysis collected during seven national monitoring
campaigns onboard R/V Akademik” in the period 2016-2018.

developed at the JRC (D'Alimonte et al., 2002). The processing steps, defined in
agreement with consolidated protocols are documented in Zibordi et al., 2011.
CHL was determined by filtering the water samples through Whatman GF/F glass
microfibre filters (pore size 0.7μm, diameter 47 mm), extracting the pigments with
acetone (90%) and spectrophotometrically measuring absorption at wavelengths of
480 630,645, 663, 664, 665nm and 750 nm (Edler, 1799). The values of CHL were
calculated with the trichromatic equations (Jeffrey & G. Humphrey, 1975).
LWN from the Galata AERONET–OC site for the period April 2016 - August 2019
were obtained from AERONET Ocean Color web site. The match-up analyses were
based on fully quality controlled AERONET-OC Level 2.0 radiometric products
determined at different center wavelengths λ in the 412–1020 nm spectral region
(Zibordi et al., 2009).

Satellite data
Match-ups for the in situ data were constructed using Sentinel-3 OLCI Level-2
(L2) data with 300 m spatial resolution, mode “Non Time Critical” (NTC),
Processing Baseline 2.23 (EUMETSAT, 2018a, 2018b) for the period 26 April 2016
273
Violeta Slabakova, Snejana Moncheva, Natalyia Slabakova, Nina Dzhembekova

– 29 November 2017 over the western Black Sea obtained from Copernicus Online
Data Access Reprocessed archive (CODAREP). The operational OLCI L2, Full
Resolution (FR) mode “NTC” data from 30 November 2017 to August 2019 for
the same geographic domain were downloaded from the EUMETSAT Data Centre.
The OLCI spectral reflectance data ρw were converted to normalized water-
leaving radiance LWN according to:

(1)

where E0 is mean extraterrestrial solar irradiance (Thuiller et al., 2003).


Sentinel-3 Toolbox Kit Module (S3TBX) in Sentinel Application Platform
(SNAP) version 7.0.3 on Windows 10 was used to process the OLCI images.

Match-Up Construction
The match-up constructions were obtained from arithmetic average of valid
OLCI full-resolution data calculated over the box of 3 × 3 – image elements
centered on the measurements location. These OLCI averages were discarded
when data: 1) were affected by the main OLCI standard flags (EUMETSAT, 2018a)
with exception of ANNOT flags (Zibordi et al., 2018); 2) sensor and sun zenith
angles were greater than 60° and < 70°, respectively; 3) the time window between
in situ measurements and satellite overpass were higher than ± 3 h; and 4) the
coefficients of variation of all 9 pixels within the box were higher than 0.2. Finally,
the percentage of valid pixels in each box was checked, and when this percentage
was not less than 50%, the mean values of the valid pixels in the box was calculated
and compared to the in situ data.

Statistical Method
OLCI L2 ocean color products were evaluated by standard statistics - coefficient
of determination R2, root-mean-square error (RMSE), Mean Percentage Difference
(MPD), Mean Absolute Percentage Difference (MAPD). The log_RMSE was used
for the assessment of OLCI pigment concentrations because of lognormal natural
distribution of CHL (Campbell, 1995).

RESULT AND DISSCUSSION


Assessments of the OLCI/SENTINEL 3A normalized water – living
radiance data
The analysis of Galata and satellite LWN spectra was restricted to the bands for
which in situ data exist (i.e. those identified by the center wavelengths at 412, 443,
490, 560, and 665 nm). The comparison between the two data sets benefited of 115
matchups. The quantitative analysis of match-up spectra and their averages and
standard deviations indicated agreement between the satellite and AERONET-OC
274
Evaluation of Sentinel-3A OLCI Ocean Color Products...

water-leaving radiances, with exception of OLCI LWN data in the blue spectral band
(412 nm), where negative values frequently occurred (Fig. 2).

Figure 2. (Left) in situ LWN and (Right) OLCI LWN spectra for the Galata
match-ups. The red continuous lines indicate the spectral averages
and the red dashed lines indicate ± one standard deviation.

The spectra of Galata and OLCI radiance data showed extremely high LWN
(λ) values for both data sets in June 2017 (Fig. 2). The phytoplankton samples
collected on 23.05.2017 in the area around Galata platform confirmed the presence
of coccolithophore Emiliana huxleyi with abundance of 2 265 120 cells/l during the
initial phase of the bloom. The full development of the bloom during late spring of
2017 in the Black Sea is clearly visible in the OLCI RGB images (Fig. 3).

23. 05.2017 12.06.2017 13.07.2017

Figure 3. Sentinel-3 OLCI RGB images for 23.05.2017, 12.06.2017


and 13.07.2017.

In general there was a good correlation between OLCI LWN and Galata records for
all the wavelengths (Fig 4. Tab.1). The coefficient of determination R2 varies from
0.84 at 665 nm to 0.98 at 490 and 560 nm, indicating that most of the OLCI data
275
Violeta Slabakova, Snejana Moncheva, Natalyia Slabakova, Nina Dzhembekova

products appear to agree with the in situ data. Still, a significant number of OLCI
negative LWN values were observed at 412 and 665 nm. This indicates that OLCI
LWN data are likely underestimated. The highest RMSE of 0.37 mW cm−2μm−1sr−1 is
determined for LWN at 412 nm that considerably decreases to 0.07 for LWN at 665 nm,
in agreement with the lower LWN values. The lower MPD of -10.63% and -11.20%
were estimated at 490 and 560 nm, respectively.

Figure 4. Scatter plots of OLCI and Galata AERONET OC LWN match-up


data at 412, 443, 490, 560, and 665 nm.

Table 1. OLCI LWN(λ) vs Galata LWN (λ) statistics (N=115).


λ, nm MPD, MAPD, RMSE, R2
% % mW
cm−2μm−1sr−1
412 -47.11 72.7 0.37 0.86
443 -24.06 39.22 0.27 0.95
490 -10.63 17.21 0.19 0.98
560 -11.20 14.80 0.29 0.98
665 -42.89 52.10 0.07 0.84

The comparison of in-water radiometric LWN (λ) and satellite LWN spectra was
performed for the bands for which in situ data exist (i.e. those identified by the
276
Evaluation of Sentinel-3A OLCI Ocean Color Products...

center wavelengths at 412, 443, 490, 510, 560, and 665 nm). The total number of
matchup pairs was 13. The OLCI estimates displayed systematic negative errors
(Tab.2), thus, an underestimation of LWN (λ) showed up at 412, 443, 490, 560 and
665 nm compared to sea-truth data. The positive MPD was found only for 510 nm.
The highest underestimate of about -32% was measured at 665 nm while the lowest
(bellow 5%) occurred at 443, 490 and 560 nm. Relatively low correlation of OLCI
LWN and in situ records was observed in the blue spectral band - 412 nm (Tab. 2).

Table 2. OLCI LWN(λ) vs in-water radiometric LWN (λ) statistics (N=13)


λ, MPD, MAPD, RMSE, R2
nm % % mW cm−2μm−1sr−1
412 -14.19 21.51 0.24 0.36
443 -3.35 13.78 0.22 0.60
490 -2.91 11.53 0.16 0.72
510 8.2 13.27 0.24 0.60
560 -2. 34 12.18 0.17 0.77
665 -31.85 34.10 0.07 0.52

Generally, the highest uncertainty of OLCI normalized water-leaving radiances


with respect to the in situ data were observed in the blue (412nm) and red (665
nm) spectral bands which could be linked to CDOM dominated Black Sea waters
with significantly varying concentrations of non-pigmented particles, displaying
maxima at either 490 nm or 555 nm and minima at 412 nm and 665 nm (Churilova
et al., 2017) as well as issues in the atmospheric correction and its calibrations
coefficients in shorter wavelengths bands (Toming et al.,2017). The OLCI LWN (λ)
bands with the best performances were 490, 510 and 560 nm, which agree with the
results presented in Zibordi et al., (2018).

Assessments of the OLCI/SENTINEL 3A CHL products


The OLCI L2 standard CHL products were evaluated in respect to the 26 in
situ measurements. The CHL concentrations were derived from the two standard
bio-optical algorithms: i) "OC4Me" Maximum Band Ratio (MBR) semi-
analytical algorithm, developed by Morel et al., (2007) following the approach of
O'Reilly et al., (1998) and ii) Neural network (NN) inverse radiative transfer bio-
optical model, originally developed for MERIS by Doerffer &Schiller, (2007),
and updated to the Case 2 Regional / Coast Color (C2RCC) processor. The CHL
values used in the analysis ranged between 0.15 and 2.3 mg m-3, with average
concentration >0.8 mg m-3.
The CHL_OC4ME product shown an overestimate of in situ chlorophyll by
about 160% (MAPD), while CHL_NN was about two times lower (81.29%) – Ta-
ble 3. A moderately good correlation (R2=0.65) was observed between satellite and
277
Violeta Slabakova, Snejana Moncheva, Natalyia Slabakova, Nina Dzhembekova

in situ data for CHL_OC4ME, whereas the CHL_NN products exhibit a determina-
tion coefficient of 0.09. The estimated log_RMSE values for both satellite products
are 0.42 and 0.43 mg m-3 for CHL_OC4ME and CHL_NN, respectively. Despite
the significantlly high MPD error estimated for OLCI OC4ME algorithm it was
almost 2 times lower comparing to the standard MODIS (OC3M) and SeaWiFS
(OC4) algorithms, while the results of CHL_NN agreed to those obtained for re-
gional algorithm (Suslin et al, 2018).

Figure 5. Scatter of OLCI derived pigments concentration


[i.e., (Left) OC4ME and (Right) CHL_NN] versus in situ CHL data.

Table 3. OLCI L2 CHL products vs in situ data (N=26)


Product MPD, MAPD, RMSE, R2
% % mg m -3
CHL_OC4ME 160.47 160.47 0.42 0.65
CHL_NN 2.77 81.29 0.43 0.09

CONCLUSIONS
The comparison carried out over almost a three year period of Sentinel 3A/
OLCI Level 2 FR data based on in situ measurements obtained from three different
data sources in western Black Sea, indicated:
 systematic underestimation of satellite normalized water-leaving radiances
LWN (λ) with effects more pronounced in the blue and red spectral regions;
 significant overestimation of CHL_OC4ME (160% MPD) and CHL_NN
(about 80% MPD) derived pigments concentration.
The results of the study showed that there are strong needs of consistent
reference in situ bio-optical data set with improved spatial and temporal (including
bloom events in the coastal area) coverage for calibration and validation of present
and future satellite ocean color missions in the Black Sea.
278
Evaluation of Sentinel-3A OLCI Ocean Color Products...

ACKNOWLEDGMENTS
The authors would like to thank Dr. Zibordi from Joint Research Center for providing
the quality controlled Level 2 data from Galata AERONET-OC site and for his sup-
port with the AOP data processing and quality determination. Thanks also go to the
EUMETSAT for the distribution of OLCI data products. This research was supported
by contract №4000123951/18/NL/SC “Bio-optics for ocean color remote sensing of
the Black Sea under the ESA Programme for European Cooperating States for Bul-
garia and contract DO1-161/28.08.2018 “NGIC – National Geoinformation Center
for monitoring, assessment and prediction natural and anthropogenic risks and
disasters” under the Program “National Roadmap for Scientific Infrastructure 2017-
2023” financed by Ministry of Education and Science of the Republic of Bulgaria.

REFERENCES
Babin, M., Stramski, D., Ferrari, G. M., Claustre, H., Bricaud, A., Obo-
lensky, G., and Hoepffner, N. (2003). Variations in the light absorption
coefficients of phytoplankton, nonalgal particles, and dissolved organic
matter in coastal waters around Europe. J. Geophys. Res., 108(C7), 3211.
Churilova, T., Moiseeva N., Efimova, T., Suslin, V., Krivenko, O., & E.
Zemlianskaia (2017). Annual variability in light absorption by particles
and colored dissolved organic matter in coastal waters of Crimea (the
Black Sea). Proc. SPIE 10466, 23rd International Symposium on Atmo-
spheric and Ocean Optics: Atmospheric Physics, 104664B; https://doi.
org/10.1117/12.2288339
Campbell, J. (1995). The Lognormal Distribution as a Model for Bio-Opti-
cal Variability in the Sea. Journal of Geophysical Research: Oceans 100
(C7), 13237–54.
D'Alimonte, D., Zibordi, G. &Berthon, J.-F. (2002).The JRC processing
system. In In results of the second SeaWiFS data analysis roundrobin,
NASA Tech. Memo. 206892, 15, 15.
Donlon, C. Berruti, B., Buongiorno, A., Ferreira, M., Féménias, P., Frerick,
J., Goryl, P., Klein, U., Laur, H., Mavrocordatos, C., et al. (2012). The
global monitoring for environment and security (GMES) Sentinel-3 mis-
sion. Remote Sensing of Environment,120, 37–57.
Doerffer, R. & Schiller, H. (2007) The MERIS Case 2 Water Algorithm.
International Journal of Remote Sensing. 28, 517-535.
Edler, L. (1979). Recommendations on methods for marine biological
studies in the Baltic Sea. Phytoplankton and Chlorophyll. Baltic Marine
Biologists, Publication No. 5, 1- 38.
EUMETSAT. (2018а). Sentinel-3 OLCI Marine User Handbook. Doc. No.
EUM/OPS-SEN3/MAN/17/907205.v.1H. 2018.
EUMETSAT. (2018b). Sentinel-3A Product Notice – OLCI Level-2 Ocean
Colour Operational Products and Full-Mission Reprocessed Time Series.
279
Violeta Slabakova, Snejana Moncheva, Natalyia Slabakova, Nina Dzhembekova

Jeffrey, S., & Humphrey, G. (1975). New spectrophotometric equations for


determining chlorophylls a and c2 in higher plants, algae and natural
phytoplankton. Biochem. Physiol. 167, 194 –204.
Kopelevich O., Burenkov V., Ershova S., Sheberstov S., & Evdoshenko M.
(2004) Application of SeaWiFS data studding variability of bio optical
characteristics in the Barents, Black and Caspian Seas. Deep-Sea Research.
Part II: Tropical studies in Oceanography. V.51, N 10-11, 1063-1091.
Kopelevich, O., Sheberstov, S., Burenkov, V., Vazyulya, S., & I. V. Sahling.
(2013). Bio-Optical Characteristics of the Russian Seas from Satellite Ocean
Color Data of 1998–2012. In Current Problems in Optics of Natural Waters:
Proceeding of VII International Conference (ONW’2013), 168–171.
Morel, A., Gentili, B., Claustre, H., Babin, M., Bricaud, A., Ras, J., et al.
(2007). Optical properties of the “clearest" natural waters. Limnology
and Oceanography, 52, 217-229.
Oguz, T. & Ediger, D. (2006). Comparison of in situ and satellite-derived
chlorophyll pigment concentrations, and impact of phytoplankton bloom
on the suboxic layer structure in the western Black Sea during May–June
2001. Deep Sea Res.II, 53, 1923-1933.
Sancak, T., Besiktepe, A., Lee, M., & Frouin, R. (2005) Evaluation of Sea-
WiFS chlorophyll-a in the Black and Mediterranean Sea. International
Journal of Remote Sensing, 26, 2045-2060.
Slabakova, V., Slabakova, N., Hristova, O., & Dzhurova, B. (2011). Assess-
ment of MERIS ocean color products using in situ data collected in the
Northwestern Black Sea. Proceedings of 5th International Conference
on Recent Advances in Space Technologies - RAST2011, 132-136.
Suslin, V., & Churilova, T. (2016). The Black Sea regional algorithm of
separation of light absorption by phytoplankton and colored detrital mat-
ter using ocean color scanner’s bands from 480-560 nm. Int. J. Remote
Sens. 37, 4380–4400.
Suslin, V., Churilova, T., Lee M., Moncheva S. &. Finenko, Z. (2018).
Comparison of the Black Sea chlorophyll-a algorithms for SeaWiFS and
MODIS instruments. Fundamentalnaya i Prikladnaya Gidrofizika. 11,
3, 64–7.
Thuillier, G., Hersé, M., Labs D., Foujols, T., Peetermans, W., Gillotay, P.,
Simon, C., and Mandel, H. (2003). The solar spectral irradiance from
200 to 2400 nm as measured by the SOLSPEC spectrometer from the
Atlas and Eureca missions. Solar Physics 214, 1–22.
Toming, K., Kutser, T., Uiboupin, R., Arikas, A., Vahter, K. & Paavel, B.
(2017). Mapping Water Quality Parameters with Sentinel-3 Ocean and
Land Colour Instrument Imagery in the Baltic Sea, Remote Sens., vol.
9, 1070.

280
Evaluation of Sentinel-3A OLCI Ocean Color Products...

Zibordi, G., Mélin, F., Berthon, J. F., Holben, B., Slutsker, I., Giles, D., et
al. (2009). AERONET-OC:A network for the validation of ocean color
primary products. Journal of Atmospheric and Oceanic Technology, 26,
1634–1651.
Zibordi, G., Berthon, J.-F., Mélin, F. & D. D’Alimonte. (2011). Cross-site
consistent in situ measurements for satellite ocean color applications:
the BiOMaP radiometric dataset. Remote Sensing of Environment,115,
2104–2115.
Zibordi, G., Mélin, F., Berthon, J.-F., and Talone, M. (2015). In situ
autonomous optical radiometry measurements for Satellite Ocean color
validation in the Western Black Sea, Ocean Sci. 11, 275–286.
Zibordi, G., Mélin, F., Berthon, J.-F. (2018). A regional assessment of
OLCI data products. IEEE Geoscience and Remote Sensing Letters,
15, 1490–1494.

 Violeta Slabakova
https://orcid.org/0000-0002-3089-0126
Institute of Oceanology
Bulgarian Academy of Sciences
Varna, Bulgaria
E-mail: [email protected]

 Snejana Moncheva
http://www.orcid.org/0000-0002-2679-5062
Institute of Oceanology
Bulgarian Academy of Sciences
Varna, Bulgaria
E-mail: [email protected]

 Natalyia Slabakova
https://orcid.org/0000-0002-3294-5346
Institute of Oceanology
Bulgarian Academy of Sciences
Varna, Bulgaria
E-mail: [email protected]

 Nina Dzhembekova
https://orcid.org/0000-0001-9620-6422
Institute of Oceanology
Bulgarian Academy of Sciences
Varna, Bulgaria
E-mail: [email protected]

281
Proceeding of 1st International conference on Environmental protection and disaster RISKs

Informatics, Remote Sensing, High Performance Computing


and Gis for Environmental Monitoring and Management
https://doi.org/10.48365/envr-2020.1.26

A MODELLING SYSTEM FOR NUMERICAL


WEATHER PREDICTION IN THE BULGARIAN
ANTARCTIC BASE AREA

Boriana Chtirkova, Elisaveta Peneva, Gergana Georgieva


“St. Kliment Ohridski” University of Sofia, Faculty of Physics,
Department of Meteorology and Geophysics

Abstract: The weather forecast of good quality is essential for the humans living
and operating in the Bulgarian Antarctic Base (BAB), located on the Livingston
Island coast at 62.64⁰ S and 60.36⁰ W. The numerical weather prediction models
in southern high latitude regions still need improvement as the user community
is limited, little test cases are documented and validation data are scarce. The
modelling system is based on the WRF model, configured in three nested domains
down to 1 km horizontal resolution, centered over BAB. The main objective of the
study is to configure and validate the WRF model and to recommend certain set-up
parameters like model area coverage, topography, land use, vertical levels, physics
parametrization schemes, preparation of initial and boundary conditions and run
configuration on HPC machines. Three test periods of “bad” weather events are
considered during the summer seasons in 2016, 2017 and 2020. The ability of the 3
days forecast to capture the basic meteorological events in this period is discussed.
Sensitivity experiments to the land use type are conducted and while the MODIS
land data is found to be the most accurate for the region, a slight increase in the soil
thermal capacity results in better model performance.
Keywords: numerical weather prediction, Antarctica, horizontal resolution, land
use type

INTRODUCTION
The Bulgarian Antarctic base St. Kliment Ohridski (BAB) is a national scientific
facility, located on Livingston Island, South Shetland Islands. A specialised
weather forecast of good quality is essential for the people working and living
on this base. The purpose of this work is to set up а limited BAB area forecasting
system based on the Weather Research and Forecasting (WRF) model and validate
its performance. The elevation of the base is 12 to 15 m above sea level, an average
of 25 people work there during the austral summer, usually from late November
until early March. The weather in the region is mostly influenced by extratropical
282
A modelling System for Numerical Weather Prediction...

cyclones, which form and propagate as Rossby planetary waves. The cyclones
move west to east and tend to diffuse over land, due to friction. Thus, the weather
is highly variable with intense storms and strong wind events occurring regularly.
Several test cases with shifting weather patterns are used to evaluate the model
performance.

MODEL CONFIGURATION
The modelling system is based on the Weather Research and Forecasting model,
version 4.0, developed by the National Center for Atmospheric Research (NCAR)
and the National Centers for Environmental Prediction (NCEP). The numerical
model uses a staggered Arakawa C-grid (Skamarock, 2019) and the nesting is
performed in a ration 9:3:1 km. The domain configuration is shown in Figure 1: it
is centered over BAB with coordinates 62.64° S and 60.36° W and consists of three
nested domains d01, d02 and d03. The outermost domain – d01, with a resolution of
9 km, has horizontal dimensions of 999 km in both directions and covers the northern
part of the Antarctic Peninsula. It is a parent domain to the first nested domain – d02,
with a horizontal resolution of 3 km and dimensions of 342 km in both directions. The
finest domain – d03, is nested in d02, and covers the area of Livingston Island and its
neighbouring small islands; the horizontal dimensions of d03 are 129 km in west-east
direction and 111 km in south-north direction. The domain configuration is made so
that there are no high mountains or complex relief near the domain borders.

Figure 1. Three domains coverage- d01, d02 and d03. Background image
from NASA Visible Earth - Blue Marble.

The land surface boundary condition of the modelling area is assembled via
the WRF Preprocessing system program – geogrid, which interpolates topography
283
Boriana Chtirkova, Elisaveta Peneva, Gergana Georgieva

and land use type data into the model grid. The topography data is taken from the
GMTED2010 dataset, developed by the United States Geological Survey (USGS)
and the National Geospatial-Intelligence Agency (NGA) and has a horizontal
resolution of 225 m. The fitted in the finest domain topography is shown of Figure
2-a. A comparison with regional maps such as the map from Ivanov (2017) show
that this dataset does not represent accurately Livingston Island. The mountain
range Tangra mountains, reaching heights up to 1700 m, is represented as a flat
surface with elevation of ~50 m. Figure 2-a represents the topography grid in the
finest domain, which has a number of points in the x-direction imax=129 and in the
y-direction – jmax=111, each grid point covers an area of 1 km2.
The WRF model can perform mainly with two land use datasets, which cover
the entire globe – USGS and MODIS. The USGS data is based on satellite advanced
very-high-resolution radiometer (AVHRR) data, collected in the period April 1992
– March 1993. They contain 24 land use types and have a resolution of 1 km.
The MODIS data (Moderate resolution imaging spectroradiometer) are gathered
by NASA satellite missions in the period 2001-2005. They are made up of 20
land use type categories and their resolution reaches 500 m. A visual comparison
between the MODIS and USGS land use type data is given in Figures 2-b and 2-c.
According to both datasets, the entire domain area is described with only 2 land use
types – snow/ice and water. The land use type is presented in numerical modelling
through the following parameters: albedo α [%], soil moisture availability M [%],
surface emissivity, ϵ [%], roughness length z0 [m], thermal inertia λT [Jm−2K−1s−1/2]
and surface heat capacity C [Jm−3K−1]. The values of these parameters differ
throughout the seasons and are used to describe the energy, momentum, water and
heat fluxes. They slightly differ between the two datasets, mainly in the parameters
z0 and λT,which are slightly higher within the MODIS data, but the difference
should not result in large computational differences. Comparing Figures 2-b and
2-c, one can conclude that the coastal line, formed by the USGS data is not as

Figure 2. Topography (a) and land use type (b) of the finest domain –
d03 with a resolution of 1 km, imax=129, jmax=111. The coastlines on the image
are taken from the Natural Earth Database (https://www.naturalearthdata.com/)
and have a horizontal resolution of 10 m.
284
A modelling System for Numerical Weather Prediction...

continuous as the MODIS one. There is a slight displacement of the grid between
the two datasets but this problem is eliminated with a manual choice of grid point
to represent BAB. Having taken this into account, and the fact the MODIS data is
more recently collected, the authors conclude it is better suited for the modelling
system. However, the description of the whole Livingston island as covered with
snow and ice may not still be accurate in recent years. This problem is addressed
further in the text.
For a vertical coordinate in WRF, version 4.0, one can choose a terrain following
(TF) coordinate or a hybrid vertical coordinate (HVC). In the present study, hybrid
η-levels are used, unevenly distributed from the surface up to isobaric level 50 hPa.
In order to determine the optimal number a vertical levels, a sample model run of
a 72 hour forecast with a different number of levels has been performed. All WRF
experiments in the present study are run on the Sofia University Parallel Computer
Center cluster PHYSON. Three different vertical level configurations have been
tested – with 35 vertical levels (close to the WRF minimum number), 50 and 70
levels. The η-levels from the three configurations are distributed with equal density
near the surface but their distribution difference manifests after certain height.
Hence, the description of the atmospheric state near the surface is similar, but the
configuration with the least densely distributed levels in height fails to represent
adequately high altitude phenomena, such as the polar jet stream. All simulations
are performed using the same number of computer cores, 88 in this case, on the
PHYSON cluster. Although being a heterogenic HPC, the PHYSON cores operate
on similar frequencies and the time difference that arises from the fact that different
simulations are run on different CPU types is negligible. The computer time for
the simplest simulation with 35 vertical levels is 9 h 22 min, the forecast with 50
vertical levels is computed for 13 h 18 min and the one with 70 levels – 18 h 33
min. For an optimal configuration, performed for a minimal time, the 50 η-levels
configuration is chosen.
Even though the high horizontal and vertical resolution of recent numerical
models allows us to describe smaller scale phenomena, other physical processes on
scales smaller than the model grid still need to be parametrized. The parametrization
schemes and their combinations has a profound effect on numerical forecasting,
especially in larger time scales and therefore the schemes should be chosen
carefully. The subgrid processes that are or may be parametrized in WRF comprise
of microphysics, convection, turbulence in the planetary boundary layer (PBL),
interactions between the atmosphere and the surface layer and the longwave and
shortwave radiation. The choice of parametrization schemes for the present study has
been made through a literature review. The Antarctic Mesoscale Prediction System1
(AMPS) produces numerical forecasts for the Antarctic region, made through a
modified PolarWRF (Hines & Bromwich, 2008). The physics parametrization in
the present configuration are chosen to be coherent and done in accordance with
285
Boriana Chtirkova, Elisaveta Peneva, Gergana Georgieva

AMPS. The following schemes are used:


• Boundary layer: Mellor-Yamada-Janjic (Eta) TKE scheme
• Surface layer: Monin-Obukhov (Janjic Eta) scheme
• Land-surface interactions: Unified Noah Land Surface Model
• Microphysics: WSM 5-class scheme
• Long-wave radiation: RRTMG longwave radiation scheme
• Short-wave radiation: Goddard shortwave radiation scheme
• Convection: Kain-Fritsch (new Eta)
For a more detailed description of each scheme, the reader is referred to
Skamarock et. al. (2019) or Chtirkova (2020). The Kain-Fritsch convection
parametrization scheme is not used in the finest domain, because its horizontal
resolution of 1 km can resolve convective processes.
The regional models need suitable atmospheric initial and lateral boundary
conditions. They are taken from the GFS 0.25 Degree Historical Archive (NCEP,
NWS, NOAA, U.S. 2015) and the lateral boundary conditions are updated every 3
hours of the simulation. The sea surface temperature is also taken as a time varying
surface boundary condition and is updated every 3 hours into the simulation. The
GFS model analysis in 0 UTC is taken as an initial condition for each of the three
domains, while the lateral boundary conditions are given only to the outermost
domain. Two-way nesting is performed, which means that the forecast in the parent
domains is affected by the solutions in the finer domains.

MODEL VALIDATION
At latitudes around 60°, two of the general circulation cells are closed – namely
the Ferrel and polar cells, enabling the formation of a convective zone with westerly
dominating winds. The Rossby waves form and propagate over this low-pressure
system, with extratropical cyclones influencing the weather by producing storms
and high-speed winds. Three test cases with rapidly changing weather are chosen
among the recent years weather records from the GFS model analysis and in-situ
observations at BAB: 16-19 December 2016, 26 February-1 March 2020 and 25-28
January 2020. The model configuration has been validated against measurements
from an automatic meteorological station at BAB, synoptic measurements in
the nearby meteorological stations2 and ERA-5 climatic hourly reanalysis data
(Copernicus Climate Change Service, 2017). The in-situ measurements come from
an automatic meteorological station Davis Vantage Vue for the 2016 and 2017 test
cases, and an automatic station assembled by MeteoRocks3 for the 2020 test case.
The validated meteorological variables are temperature at 2 m, surface pressure,
wind speed and wind direction. The total number of synoptic stations in the largest
domain is 18, two of which lie in the finest domain. The synoptic observations
from “Base Arturo Prat” with WMO index 890057 and coordinates 62.3° N,
-59.41°E, are used in the comparison as indicative of the weather pattern in the
286
A modelling System for Numerical Weather Prediction...

region. They are in agreement with the observations from BAB in all test cases. The
validation process is divided into 3 parts – the first one consists of a comparison of
model simulations with different horizontal resolution, the second one is a general
validation for all test cases and lastly a sensitivity experiment to the surface type is
performed.

Comparison of simulations with different horizontal resolution


The “nesting” of domains allows us to significantly decrease the horizontal
resolution of WRF, as the finest resolution of the current configuration is down
to 1 km. This section aims to compare model results with a different resolution
– 9 km, 3 km and 1 km, and the role of the sea is also taken into account.
The comparison is conducted for the grid point, where BAB is located, and
is manually chosen, as mentioned above. The choice criteria consist of an
approximate location in the South Bay of Livingston island, a land surface type
and sea boarders in two sides. In the outermost domain, the grid coordinates of
the point are xd01=55, yd01=55; in the first nested domain BAB falls into xd02=55,
yd02=55; and in the finest domain it is taken as xd03=61, yd03=53. The following
comparison is conducted only for the test case in 2020 with the meteorological
variables, which have been measured at BAB. The comparison is visually
presented on Figure 3. The black points mark the values, measured at BAB;
brown rectangles mark the ERA-5 reanalysis data; dark red triangles are from
the synoptic station “Base Arturo Prat” and are shown as an instance of the
weather pattern in the region. The red, orange and yellow curves correspond to
model simulations with different resolutions in the BAB point. The blue curve
shows the values of the respective meteorological variable in point xd03=60,
yd03=53, which has a surface type of water and represents the sea in front of
BAB. The green curve is calculated as the mean of the BAB land point and its
neighbouring sea point (i.e. the yellow and blue curves) – this calculation is
done, because BAB is located close the edge of the corresponding land point,
and the effects of the sea should not be neglected.
The comparison of the 2 m temperature time series on Figure 3-a shows
the WRF forecasts lower values than the observed ones. It is possible that this
difference is due to the lack of proper calibration of the temperature sensor, but
the measurements from “Base Arturo Prat” also suggest warmer air temperature
and larger temperature amplitudes. The graph clearly shows that decreasing the
horizontal resolution leads to closer to measurement simulated values. The rapid
cooling in 10 UTC on 27 January 2020 is also described quite well. The simulations
with higher resolution generally show higher temperatures, but they nicely describe
the minimum at 0.5°C. The higher temperatures of the sea point next to BAB are
explained not only with the thermal flux differences between the two points, but
also with the higher sea surface temperature – at this point the cooling is less
287
Boriana Chtirkova, Elisaveta Peneva, Gergana Georgieva

Figure 3. Time series of the meteorological variables 2 m temperature (a), sea


level pressure (b), wind speed (c) and wind direction (d) with different horizontal
resolution and different methods of accounting for the influence of the sea,
compared with observational data ant the ERA-5 reanalysis.

pronounced. The combination of the BAB land and its neighbouring sea point,
which is comprised by the average value in every time instant, simulate the rapid
cooling and the higher temperatures relatively well. In order to give a quantitative
measure of the forecast skill, the BIAS, Root mean square error (RMSE) and Mean
absolute error (MAE), are calculated as:

where n is the number of discrete values, Mi is the model simulated value and Oi – the
corresponding (closest in time) observed value. The WRF values are output every
hour, therefore for a 3 days forecast n=72, in the case of no missing observations.
These statistical measures possess the dimensionality of the measured variable and
are calculated for each of the model points described above relative to the BAB
measurements and to the ERA-5 reanalysis separately. The results are shown in
Table 1.
288
A modelling System for Numerical Weather Prediction...

Table 1. Forecast statistics calculated against measurement data and ERA-5


data for simulations with different horizontal resolution and different means
of accounting for the influence of the sea.

BAB observations ERA-5 reanalysis


BIAS RMSE MAE BIAS RMSE MAE
Temperature [°C]
BAB in d01 (9 km) -2.31 2.45 2.31 -0.74 0.92 0.75
BAB in d02 (3 km) -2.17 2.32 2.17 -0.6 0.82 0.63
BAB in d03 (1 km) – land -2.07 2.23 2.07 -0.5 0.75 0.57
BAB in d03 (1 km) – sea -1.73 1.93 1.73 -0.16 0.56 0.44
BAB in d03 (1 km) – land and sea -1.9 2.08 1.9 -0.33 0.64 0.48
Sea level pressure [hPa]
BAB in d01 (9 km) -0.54 0.86 0.76 -1.65 1.69 1.65
BAB in d02 (3 km) -0.44 0.79 0.69 -1.54 1.59 1.54
BAB in d03 (1 km) – land -0.27 0.72 0.6 -1.37 1.42 1.37
BAB in d03 (1 km) – sea -0.21 0.71 0.58 -1.31 1.36 1.31
BAB in d03 (1 km) – land and sea -0.27 0.72 0.6 -1.37 1.42 1.37
Wind speed [ms-1]
BAB in d01 (9 km) -0.12 3.68 2.84 -0.38 2.73 2.23
BAB in d02 (3 km) -0.2 3.79 2.91 -0.45 2.86 2.33
BAB in d03 (1 km) – land 0.12 3.95 3.13 -0.14 3.04 2.49
BAB in d03 (1 km) – sea -0.09 3.75 2.93 -0.35 2.81 2.33
BAB in d03 (1 km) – land and sea 0.12 3.95 3.13 -0.14 3.04 2.49
Wind direection [deg]
BAB in d01 (9 km) -54.71 59.56 55.11 -17.44 27.55 21.27
BAB in d02 (3 km) -52.29 57.74 53.14 -15.02 26.46 20.23
BAB in d03 (1 km) – land -52.9 58.27 53.68 -15.63 26.76 21.16
BAB in d03 (1 km) – sea -54.34 -59.71 -54.94 -17.07 28.36 22.83
BAB in d03 (1 km) – land and sea -52.9 58.27 53.68 -15.63 26.76 21.16

The BIAS is a measure for the direction of the deviation. All its values for
the temperature at 2 m are negative, which again indicates the tendency for lower
temperatures, calculated by the model. The numerical comparison of the simulated
2 m temperature with BAB observations and ERA-5 data shows best results in the
sea point next to BAB.
The pressure time series, shown on Figure 3-b, suggest that this field is predicted
extremely well by the numerical forecast. The negative BIAS shows that the
forecasted values are slightly under the measured ones, the smallest deviations of
289
Boriana Chtirkova, Elisaveta Peneva, Gergana Georgieva

RMSE=0.71 hPa against BAB observations and 1.36 hPa against ERA-5 reanalysis
are as well observed at the sea point in the finest domain.
The wind speed is simulated quite well, which is seen by the amplitudes of the
model curves on Figure 3-c and the close to 0 BIAS. The forecast statistics are similar
in the simulations with different resolution, as the largest deviations is RMSE=3.95
ms-1 in the sea point. The wind speed maximum around 21 UTC on 27 January 2020
is described fairly well despite the higher peak in the model with ~ 5ms-1.
In order to calculate the statistics for the wind direction, Δd needs to be introduced
as the difference in degrees between the simulated and the observed wind direction,
it varies in the interval Δd‪[-180°, 180°]. This quantity is calculated differently,
depending on the difference between Mi and Oi (Jiménez & Dudhia, 2013):

In this case Mi and Oi have the dimensionality of degrees, indicating a direction, in


which 0° corresponds to North. The statistical characteristics are calculated as follows:

The results of the comparison between the forecasted direction and the measured
one give RMSE values reaching up to 60°. The comparison of the simulated wind
direction with ERA-5 data shows more promising results with RMSE values less
than 30°. This difference may be due to various factors, one of them being the
inadequate topography data, used by the WRF and ERA-5 simulations. Another
possible error may have arisen from an imperfect placement of the mast resulting
in a Venturi effect or turbulent eddies at heights below 10 m. According to the
forecast statistics and Figure 3-d, the horizontal resolution does not have a profound
effect on the wind direction. The wind speed is slightly higher in the sea point,
due to the smaller roughness length z0 over the sea surface. The wind direction
is approximately the same in all grid points in the comparison. There is a scarce
distinction between the WRF and ERA-5 values in the period from 9 UTC on 26
January until 3 UTC on 27 January 2020.

Validation for all test cases


The results from the numerical forecast validation suggest that the horizontal
resolution mainly affects the 2 m temperature field, which is improved by ~0.5°C in
290
A modelling System for Numerical Weather Prediction...

the finest domain. The pressure values are also affected. The sea grid point produces
the slightest deviations with observed data, which suggests that the location of BAB
is strongly influenced by the sea. The forecast statistics for all covered test cases
and their average value are presented in Table 2. The test cases in 2016 and 2020 are
described relatively well by the model, whereas the 2017 test case proved to be a
numerical challenge and the deviations between simulated and observed values are
quite high. The WRF shows an average temperature RMSE of 2.59°C against BAB
measurements, the pressure field is forecasted very well with an average RMSE of
just 1.34 hPa for the three cases. The wind speed RMSE ranges from 2 to 9 ms-1 for
the different test cases, the average value being 5 ms-1. The wind direction is highly
biased with RMSE values up to 70°, which could be due to the poor topography
representation or flawed in-situ measurements.

Table 2. Forecast statistics calculated against measurement data and ERA-5 data
for the three test cases and their average value for each meteorological variable.

BAB observations ERA-5 reanalysis


BIAS RMSE MAE BIAS RMSE MAE
Temperature [°C]
Test case 2016 -1.44 1.69 1.44 -0.17 0.98 0.62
Test case 2017 -1.21 3.86 3.31 -1.67 3.95 2.83
Test case 2020 -2.07 2.23 2.07 -0.50 0.75 0.57
Average for all test cases -1.57 2.59 2.27 -0.78 1.89 1.34
Sea level pressure [hPa]
Test case 2016 0.57 1.10 0.81 -0.82 1.27 1.14
Test case 2017 0.76 2.21 1.86 -1.16 2.52 1.71
Test case 2020 -0.27 0.72 0.60 -1.37 1.42 1.37
Average for all test cases 0.35 1.34 1.09 -1.12 1.74 1.41
Wind speed [ms-1]
Test case 2016 -0.96 2.67 2.18 0.61 1.39 1.10
Test case 2017 -4.70 8.36 6.51 -1.36 4.61 3.36
Test case 2020 0.12 3.95 3.13 -0.14 3.04 2.49
Average for all test cases -1.85 4.99 3.94 -0.30 3.01 2.32
Wind direction [deg]
Test case 2016 -6.15 35.14 27.48 -3.60 35.61 26.76
Test case 2017 -32.18 119.22 109.55 -53.07 99.88 83.46
Test case 2020 -52.90 58.27 53.68 -15.63 26.76 21.16
Average for all test cases -30.41 70.88 63.57 -24.10 54.08 43.79

291
Boriana Chtirkova, Elisaveta Peneva, Gergana Georgieva

Land cover type sensitivity


The land use type in the BAB grid point, as well as the whole Livingston island,
is depicted as snow and ice. The available land use data – MODIS, is assembled
in the period 2001-2005, but due to the climate changes and the dynamics of land
and sea ice, these data may prove unsuitable for 2020 simulations. Recent Landsat
images, as well as some summer photos in the region, indicate the presence of
rock fields and even grasslands. The turbulent fluxes over bare rock areas are quite
different from those over snow and ice. In order to quantify the effect of the surface
parameters only in one grid point, three additional experiments of altering 1 km2
of snow area with another land type are performed. Technically, this narrows to
altering the WRF variables LU_INDEX and LANDUSEF in the grid point xd03=60,
yd03=53 of the finest domain. The following land use types are used: Grasslands
(LU index=10), Barren or Sparcely Vegetated (LU index=16), Barren Tundra (LU
index=20). The parameters of these land use types for the summer season are shown
in Table 3. It can be seen that the thermal capacity C of the different surfaces differs
by a factor of 1020. The differences in the roughness length z0 are around 6-7 cm.

Table 3. MODIS land use type parameters for the summer,


as in LANDUSE.TBL of WRF, v4.0.
λT [4,184.102
No. α [%] M [%] ϵ [%] z0 [10-2 m] C [105 Jm-3K-1] Name
Jm-2K-1s−1/2]
10 19 15 92 7.5 2.37 20,8.105 "Grasslands"
15 55 95 98 1 5 9.1025 "Snow and Ice"
"Barren or Sparsely
16 25 2 85 6.5 0.81 12.105
Vegetated"
17 8 100 98 0.01 6 9.1025 "Water"
20 15 2 90 6 1.6 12.105 "Barren Tundra"

Figure 4. Time series of the meteorological variables 2 m temperature


(a) and wind speed (b) corresponding to the different land use type experiments,
compared with observational data ant the ERA-5 reanalysis.
292
A modelling System for Numerical Weather Prediction...

The simulation results for the 2020 test case are shown on Figure 4-a for the 2
m temperature and on Figure 4-b for the 10 m wind speed. The BAB measurements
and the ERA-5 reanalysis are given as real data representatives. The dark green
curve depicts the “Grasslands” experiment, the light blue one represents the “Barren
or Sparsely Vegetated” experiment, and the violet one is for the “Barren Tundra”
experiment. The yellow curve corresponds to the original simulation with a land
use type of “Snow and Ice” and it is the same as in the previous figures.

Table 2. Forecast statistics calculated against measurement data and ERA-5


data and the unmodified model run for the different land use types.
BAB observations ERA-5 reanalysis WRF – snow/ice
BIAS RMSE MAE BIAS RMSE MAE BIAS RMSE MAE
Temperature [°C]
Snow/ice -2.07 2.23 2.07 -0.50 0.75 0.57 - - -
Grasslands -1.07 1.30 1.10 0.50 1.01 0.79 1.01 1.18 1.02
Barren/Sparsely Vegetated -1.20 1.39 1.22 0.37 0.92 0.75 0.87 1.07 0.93
Barren Tundra -1.10 1.34 1.15 0.47 1.01 0.79 0.97 1.15 0.99
Sea level pressure [hPa]
Snow/ice -0.27 0.72 0.60 -1.37 1.42 1.37 - - -
Grasslands -0.27 0.73 0.61 -1.37 1.42 1.37 0.00 0.02 0.01
Barren/Sparsely Vegetated -0.27 0.72 0.61 -1.37 1.42 1.37 0.00 0.02 0.01
Barren Tundra -0.27 0.72 0.61 -1.37 1.42 1.37 0.00 0.02 0.01
Wind speed [ms-1]
Snow/ice 0.12 3.95 3.13 -0.14 3.04 2.49 - - -
Grasslands 0.25 3.79 3.04 0.00 2.88 2.42 0.13 0.41 0.31
Barren/Sparsely Vegetated 0.49 3.85 3.14 0.23 2.96 2.47 0.37 0.53 0.43
Barren Tundra 0.29 3.78 3.02 0.03 2.87 2.43 0.17 0.47 0.35
Wind direction [deg]
Snow/ice -52.90 58.27 53.68 -15.63 26.76 21.16 - - -
Grasslands -53.93 59.00 54.41 -16.66 27.05 21.19 -1.03 2.97 1.72
Barren/Sparsely Vegetated -53.41 58.82 54.01 -16.13 27.05 21.43 -0.50 2.23 1.34
Barren Tundra -53.91 58.93 54.41 -16.64 27.02 21.40 -1.01 3.18 1.89

The 2 m temperature is highly sensitive to the land use type. On Figure 4-a, one can
see that the diurnal cycle is more pronounced, when the surface is different form snow
and ice, thus the thermal capacity is higher. Table 4 presents the forecast statistics for
these simulations calculated against BAB measurements, ERA-5 reanalysis and the
unmodified model run – the last group is given in order to evaluate the model response
to this change. The numerical comparison shows that the 2 m temperature is the highest
293
Boriana Chtirkova, Elisaveta Peneva, Gergana Georgieva

with grasslands, slightly lower with barren tundra, even lower with barren or sparsely
vegetated soil and the lowest with snow and ice, which corresponds precisely to the
respective thermal capacities in Table 3. Accordingly, with higher temperatures, the
forecast BIAS approaches 0. The smallest values of RMSE and MAE against BAB
measurements are obtained with grasslands, because the temperature in this experiment
is the highest. The best results against ERA-5 data are obtained with the original
simulation with snow and ice, which suggests that the WRF and ERA-5 models
are configured with similar land type data. The atmospheric pressure field is almost
unaffected by the change in the land type. The wind speed at 10 m is only slightly
altered with values of ~0.5 ms-1 – the results with the least RMSE=3.78 against BAB
measurements are obtained with the barren tundra surface type. Figure 4-b shows that
the 10 m wind speed timeseries curve is similar in all experiments. The average wind
speed is the highest with the barren or sparsely vegetated soil, although its roughness
length z0 is relatively high. Overall, the change in land use type affects all surface
meteorological variables, but the largest differences are observed in the 2 m temperature
field, which is relatively increased by ~1°C with grasslands, compared to snow and ice.
Although the “Grasslands” experiment produces the best results, it does not ideally
describe the situation. The experiments are conducted with a change of only 1 km2 of
the land surface, while the Landsat images and summer photos in the region suggest
a higher area without snow. The most adequate description of the land surface would
be the use of more recent satellite data to alter a larger area in the finest domain, as
well as parts of the larger model domains. The barren or sparsely vegetated soil also
produces fairly good results. The differences are explained by the values of the surface
parameters, indicative for each land type. The lower thermal capacity C, as well as the
lower thermal inertia λT of the barren soil, in combination with the low albedo α, lead
to the absorption of more short-wave radiation and a faster warming from the Sun.
The relatively low surface emissivity reduces radiative cooling during the night, which
explains the higher night temperatures of barren soil, compared to snow and ice land
type in WRF. As a conclusion, the snow and ice land use type does not adequately
represent Livingston island in the summer season.

CONCLUSION
A complete configuration of the WRF model for forecasting the weather in the
Livingston island area is set up. The GMTED2010 topography data is proven to
be flawed in the region, for this reason the authors advise the use of the Reference
Elevation Model of Antarctica (REMA) in further simulations, which is a high
resolution, time-stamped Digital Surface Model (DSM) of Antarctica at 8-meter
spatial resolution (Howat et al., 2019). The model grid is centred over BAB, the
length of each simulation is 72 hours and the results are compared with measurements
at BAB, nearby synoptic stations and the ERA-5 reanalysis. Various cases with
passing cyclones and shift of weather patterns are addressed in our study. The WRF
294
A modelling System for Numerical Weather Prediction...

is found to represent satisfactory the temperature at 2 m, sea level pressure and wind
at 10 m in the point of BAB during weather shift events. A comparison of results
with different horizontal resolution suggests that the forecast quality increases with
the model grid resolution. The BAB point is found to be strongly influenced by the
sea and its effect should be considered in the forecast representation. Further details
on the influence of sea surface temperature on the forecast skill can be found in
Chtirkova (2020) and Chtirkova & Peneva (2020).
A sensitivity experiment with altering the land use type in one grid point (1 km2) in
the finest domain is carried out. This grid point corresponds to the location of the base,
where the snow melts completely during summer and reveals the rocky soil underneath.
A change in the land use type, respectively the thermal capacity of the underlying
surface, significantly increases the forecast quality. This experiment shows that MODIS
data, collected with satellite missions during 2005, does not adequately represent the
current situation on places, affected by climate change, such as Livingston Island. For
accurate simulations in these regions, a revision of the land use type is advisory.

ACKNOWLEDGMENTS
The authors would like to express their gratitude to all people and projects, involved
in the operation and maintenance of the PHYSON cluster (http://physon.phys.
uni-sofia.bg/about-physon-bg).

NOTES
1. NCAR, UCAR – The Antarctic Mesoscale Prediction System (AMPS): https://
www2.mmm.ucar.edu/rt/amps/
2. Synoptic data from the weather information service OGIMET: https://www.
ogimet.com/
3. The MeteoRocks project: https://meteo.rocks/page/aboutus

REFERENCES
Skamarock, William C. et al. (2019). A Description of the Advanced
Research WRF Model Version 4, UCAR/NCAR. doi: 10.5065/1DFH-
6P97, url:https://opensky.ucar.edu/islandora/object/opensky:2898 .
L.L. Ivanov (2017). Antarctica: Livingston Island and Smith Island. Scale
1:100000 topographic map, Manfred Wörner Foundation, ISBN 978-
619-90008-3-0 .
Hines, K. M., and D. H. Bromwich (2008). Development and testing of
Polar WRF. Part I. Greenland ice sheet meteorology. Mon. Wea. Rev.,
136, 1971-1989, url: https://polarmet.osu.edu/PolarMet/PMGFulldocs/
hines_bromwich_mwr_2008.pdf.
295
Boriana Chtirkova, Elisaveta Peneva, Gergana Georgieva

Chtirkova, B. (2020). A study on the impact of sea surface temperature


on the weather forecast for the Bulgarian Antarctic Base on Livingston
Island (Master's Thesis). Sofia University "St. Kliment Ohridski", Sofia,
Bulgaria.
National Centers For Environmental Prediction/National Weather Service/
NOAA/U.S. Department Of Commerce (2015). NCEP GFS 0.25 Degree
Global Forecast Grids Historical Archive, doi: 10.5065/D65D8PWK,
url: https://rda.ucar.edu/datasets/ds084.1/ .
Copernicus Climate Change Service (C3S) (2017). ERA5: Fifth generation
of ECMWF atmospheric reanalyses of the global climate, url: https://cds.
climate.copernicus.eu/cdsapp#!/home .
Jiménez, Pedro A. and Jimy Dudhia (2013). On the Ability of the WRF
Model to Reproduce the Surface Wind Direction over Complex Terrain,
Journal of Applied Meteorology and Climatology 52.7, p. 1610—1617.
doi: 10.1175/jamc-d-12-0266.1, url: https://doi.org/10.1175/jamc-
d-12-0266.1 .
Howat, I. M., Porter, C., Smith, B. E., Noh, M.-J., and Morin, P. (2019). The
Reference Elevation Model of Antarctica, The Cryosphere, 13, 665-674,
https://doi.org/10.5194/tc-13-665-2019 .
Chtirkova, B. and Peneva, E (2020). The impact of SST on the weather
forecast quality in the Bulgarian Antarctic Base area on Livingston
Island, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-
9347, https://doi.org/10.5194/egusphere-egu2020-9347

 Boriana Chtirkova
https://orcid.org/0000-0001-5743-8511
Department of Meteorology and Geophysics
Faculty of Physics
“St. Kliment Ohridski” University of Sofia
E-mail: [email protected]

 Elisaveta Peneva
https://orcid.org/0000-0003-1325-685X
Department of Meteorology and Geophysics
Faculty of Physics
“St. Kliment Ohridski” University of Sofia
E-mail: [email protected]

 Gergana Georgieva
https://orcid.org/0000-0003-3515-9524
Department of Meteorology and Geophysics
Faculty of Physics
“St. Kliment Ohridski” University of Sofia
E-mail: [email protected]

296

You might also like