ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

13-14 June 2019

ABSTRACT BOOKLET
4 projects - 4 themes – 22 abstracts
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

CONTENT
ABOUT THE CONFERENCE 1

MUFFIN 2

INXCES 3

SUrF 4

The Weather Radar project - Väderradarteknik inom VA-området test av metodik 5

THEME 1: END-USER ENGAGEMENT, TAILORED TOOLS & PUBLIC PERCEPTION 6

Knowledge exchange on Climate Adaptation Best Management Practices for

Sustainable water management in Resilient Cities 6

Are house owners’ willing to invest in self-protection against flooding? 8

Building flood resilience – together 10

A review of sustainability of urban flood management from the aspects of hydrology,

economy and the perceived urban design quality 12

Responding to end-user requirements before, during and after the flood 13

APPETIZER 14

RAINVIS: a real-time high-resolution high-intensity rainfall visualization prototype for

Sweden 14

The FLOODVIEW portal – a web-based prototype for innovative modelling and

visualization of heavy rainfall and urban flood risk 16
THEME 2: RAINFALL & ENVIRONMENTAL OBSERVATION & FORECASTING 18

Can we trust radar? High-intensity rainfall in operational radar 18

Accuracy of satellite-observed rainfall over Sweden 22

Use of weather radar in the water sector in Denmark; what can we learn? 23

Experiences of X-band weather radar in Skåne 25

Subsidence areas trends in bucharest city, based on psinsar analyses 26

APPETIZER 28

Subsidence in urban areas measured by InSAR (Sentinel1) related to flooding 28

THEME 3: URBAN FLOOD MODELLING & FORECASTING 30

High resolution modelling of urban flood modelling and its origin in the large-scale
catchment 30

Hydrodynamic vs. hydrological modelling: a comparative study in Aalborg and Helsinki

32
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

The relationship between rainfall and pluvial flooding in Rotterdam based on citizen
reports 34

A strategy to sustain the functionality of the existing urban drainage network 38

Which types of rainfall cause urban flooding 40

Urban Rainfall-Runoff Nowcasting with Open Data and Open Tools 42

APPETIZER 44

Flood and heatstress models and the need for (sub-) surface INnovations for eXtreme
Climatic EventS (INXCES) 44

A web-based visualization prototype of urban flood forecasts from a multi-scale

hydrologic-hydrodynamic flood forecasting system in Aalborg, Denmark 46
THEME 4: URBAN WATER MANAGEMENT; NATURE-BASED SOLUTIONS & CLIMATE ADAPTION 48

Multi-purpose urban water management in a complex urban basin 48

Seasonal Hydraulic efficiency of infiltration based SUDs (INXCES) 50

The governing of flood risk mitigation in complex and dynamic society 54

Children’s places in stormwater spaces 55

Characterisation of dissolved metal fractions in urban runoff 58

Vadose zone hydraulic assessment in urban areas with small scale variability 61

ClimateCafe for interdisciplinary active Knowledge exchange on Climate Adaptation:

25th edition Malmo Climatecafe 63

POSTERS 65

In situ mapping of pollutants in Sustainable Urban Drainage Systems, a new

methodology approach and preliminary results from the Netherlands 65

In-situ hyperspectral and fluorescence methods compared with remote sensing

Sentinel-2 satellite data for mapping chlorophyll-a/cyanobacteria concentrations 67
SPONSORS FOR THE CITIES, RAIN AND RISK CONFERENCE 69
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

ABOUT THE CONFERENCE

This is a conference open for an exchange of thoughts, highlighting results, tools, and mindset when it comes
to tackle the challenges in managing urban flooding. The conference present results from four innovative
research projects, from the perspective of the following themes:
- End-user engagement, tailored tools and public perceptions
- Rainfall and environmental observation and forecasting
- Urban flood modelling and forecasting
- Water management, nature-based solutions and climate adaptation
We welcome practitioners and experts dealing with urban flooding hazards in some way, as well as
academics and scientists in the field. The conference will consist of both short lectures and hands-on demos
of web-based tools.
The content of this booklet has not been peer-review and the information may have been published
elsewhere. The authors are responsible for content and copy rights.
The four projects:

The INXCES project The MUFFIN project

https://inxces.eu http://www.muffin-project.eu/about-muffin
Partners: Partners:
Norwegian University of Science and Technology Swedish Meteorological and Hydrological Institute
(NTNU) (SMHI),
University of Applied Sciences Groningen (HUAS) Aalborg University (AAU)
Luleå University of Technology (LTU) Delft University of Technology (TUD)
Technical University of Civil Engineering of Aalto University (AU)
Bucharest (UTCB-CCIAS) Swedish Geotechnical Institute (SGI)
Geological Survey of Norway (NGU) Sponsors: The Swedish Research Council Formas
Sponsors: The Research Council of Norway (NO), (SE), Netherlands Organisation for Scientific
The Executive Agency for Higher Education, Research (NL), Innovationsfonden (DK) and Maa- ja
Research, Development and Innovation Funding vesitekniikan tuki ry (FI) though the Water
(RO), Formas (SE) and Netherlands Organisation Challenges for a Changing World Joint Programme
for Scientific Research (NL) though Water Initiative (Water JPI)
Challenges for a Changing World Joint
Programme Initiative (Water JPI)

The Weather Radar project The SUrF project

http://www.svensktvatten.se/forskning/svenskt- http://www.surf.lu.se
vatten-utveckling/pagaende-svu- Partners:
projekt/vaderradarteknik-inom-va-omradet-- Lund University
test-av-metodik Sweden Water Research
Partners: Region Skåne
Lunds Tekniska Högskola Höje å Vattenråd
Lunds Universitet Länsförsäkringar Skåne
Sweden Water Research Göteborgs stad, kretslopp och vatten
SMHI Sponsor: The Swedish Research Council Formas
Sponsor: Svenskt Vatten Utveckling (SE)

1
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

MUFFIN has worked to bridge the gap between the urban and large-scale hydrological modelling
communities, providing mutual benefits and an arena for new thinking. The project has developed
innovative systems and solutions that diminish the adverse effects of urban flooding. The basic approach
was to analyze, develop, join, compare and evaluate observational and forecasting systems operating at
different scales (local, regional/national, continental). A flooding event may be generally divided into the
following three temporal stages, during which different users require different types of information.
• Before the flood: The main components in this stage are the rainfall and flood forecasts, signal-
ling when and where problems may occur.
• During the flood: Frequent real-time observations are required to follow how the event develops
and maximize situation awareness.
• After the flood: Properly quality controlled, stored and documented observations, forecasts and
other relevant information are needed to facilitate post-event analyses.
Furthermore, flood-related information (observations, forecasts) is available from systems operating over
different spatial domains.
• The city: A city may operate its own flood forecasting system or be included in a small regional
system.
• The region/country: Several regions and countries operate forecasting systems.
• The continent (Europe): A range of hydrological modelling systems have been set up for flood
forecasting at the Pan-European level.
Generally, local systems and information provide the highest value for the end-users in all flood phases.
However, local systems are lacking in many cities which are thus dependent on purely hydrological
information from national or even continental level. In MUFFIN, we have aimed at increasing the end-user
value of information related to urban floods by research and development at all scales as well as adaptation
and promotion of existing material. The research and development have been performed in three cities:
Aalborg (DK), Rotterdam (NL) and Helsinki (FI). During the first phase of the project, local forecasting systems
were developed and optimized for selected sub-basins in these cities. In parallel, the hydrological model
HYPE was developed for high-resolution modelling and set up for the same sub-basins. In the second phase
of the project, coordinated forecasting experiments were carried out in order to explore the benefits and
limitations of each type of model system as well as the prospect of combining them.
Partners:
Swedish Meteorological and Hydrological Institute, Norrköping in Sweden – Coordinator
Jonas Olsson ([email protected])
Aalborg University, Aalborg in Denmark
Søren Liedtke Thorndahl ([email protected])
Delft University of Technology, Delft in the Netherlands
Marc Schleiss ([email protected])
Aalto University, Helsinki in Finland
Teemu Kokkonen ([email protected]

2
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

INXCES is funded by the Research Council of Norway (NO), The Executive Agency for Higher Education,
Research, Development and Innovation Funding (RO), Formas (SE) and Netherlands Organisation for
Scientific Research (NL) though Water Challenges for a Changing World Joint Programme Initiative (Water
JPI). During the project period new innovative technological methods for risk assessment and mitigation of
extreme hydroclimatic events has been developed and including optimization of urban water-dependent
ecosystem services at the catchment level, for a spectrum of rainfall events. It is widely acknowledged that
extreme events, such as floods and droughts are an increasing challenge, particularly in urban areas. The
frequency and intensity of floods and droughts pose challenges for economic and social development,
negatively affecting the quality of life of urban populations. Prevention and mitigation of the consequences
of hydroclimatic extreme events are dependent on the time scale. Floods are typically a consequence of
intense rainfall events with short duration. In relation to prolonged droughts however, a much slower
timescale needs to be considered, connected to groundwater level reductions, desiccation and negative
consequences for growing conditions and potential ground – and building stability.
INXCES has taken a holistic spatial and temporal approach to the urban water balance at a catchment scale.
Perform technical-scientific research to assess, mitigate and build resilience in cities against extreme
hydroclimatic events with nature-based solutions. INXCES used and enhance innovative 3D terrain analysis
and visualization technology coupled with state-of-the-art satellite remote sensing to develop cost-effective
risk assessment tools for urban flooding, aquifer recharge, ground stability and subsidence. INXCES used
quick scan tools that will help decision makers and other actors to improve the understanding of urban and
peri-urban terrains and identify options for cost effective implementation of water management solutions,
which will reduce the negative impacts of extreme events, maximize beneficial uses of rainwater and
stormwater for small to intermediate events, and provide long-term resilience in light of future climate
changes. The INXCES approach optimizes the multiple benefits of urban ecosystems, thereby stimulating
widespread implementation of nature-based solutions on the urban catchment scale. https://www.inxces.eu
Partners:
Norwegian University of Science and Technology (NTNU), Trondheim in Norway – Coordinator
Tone M. Muthanna ([email protected])
Hanze University of Applies Sciences (HUAS), Groningen in the Netherlands
Floris C. Boogaard ([email protected])
Geological Survey of Norway (NGU), Trondheim in Norway
Guri Venvik ([email protected] )
Technical University of Civil Engineering of Bucharest (UTCB-CCIAS), Romania
Constantin Radu Gogu ([email protected])
Luleå University of Technology (LTU), Sweden
Maria Viklander ([email protected] )

3
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

SUrF stands for Sustainable Urban Flood Management and is a vision as well as a project. We, the project
partners, have reached a common view of the topic which we understand as a system of systems, illustrated
by the figure below.
This concept corresponds well with the title of the
conference. Rain is the input to the Hydrological system,
which may consist of both natural and man-made
components. In extreme cases, the Impact system
consisting of both physical elements and non-tangibles in
the City, may be negatively affected. This is a Risk which
has to be handled via a Flood Management system. Due to
the complexity of the problem the management system
requires understanding, and input from a broad spectrum
of individuals and organisations, private as well as public. In order to make useful research for this
multifaceted system we have established a multidisciplinary team of researchers with support from an
engaged group of funders which have also been partners in the project.
Our strategy has been to work with all the three systems in the figure above. Two special topics, which have
been especially important in SUrF are hydrological modelling and blue-green solutions, both of which can be
seen as dealing with the hydrological system and the management system. Issues related to responsibility,
legislation and organisational structure have been shown to be obstacles for efficient urban flood
management in Sweden (and elsewhere). Therefore, a lot of SUrF research activities have been focused on
these areas. The project team consists of researchers from six departments, representing three faculties, at
Lund University plus researchers from Malmö University and the Swedish University of Agricultural Sciences.
Partners:
Lund University– Coordinators
Rolf Larsson ([email protected] and Ronny Berndtsson ([email protected]
Sweden Water Research
Hans Bertil Wittgren ([email protected])
Göteborgs stad, Kretslopp och vatten
Lena Blom ([email protected])
Region Skåne
Jerry Nilsson ([email protected]) and
Höje å Water Council
Anna Helgeson ([email protected])
Länsförsäkringar Skåne
David Lamppu ([email protected])

4
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

The Weather Radar project - Väderradarteknik inom VA-området test av metodik

VA SYD together with the Faculty of Engineering (LTH), Lund University, Sweden, Water Research, and SMHI
have collaborated in the Weather Radar Project. The X-band weather radar was installed from July-September
2018 to test the facility on Dalby water tower, Lund municipality. The purpose with the test period was to
obtain an in-depth conclusion of how an X-band weather radar facility could be implemented in VA SYD´s
sewage utilities including weather radar data quality control and validation.
The purpose with the test period was to obtain an in-depth conclusion on how an X-band weather radar
facility could be implemented at VA SYD, including data quality control and validation. The following parts
constitute the report:
• A literature study and an evaluation of the critical parameters for a weather radar facility
• An evaluation with proposed sewerage system and wastewater treatment plant applications of the
weather radar data
• Validation of the radar data
VA SYD has collaborated with the Faculty of Engineering at Lund University (LTH), Sweden Water Research
(SWR) and the Swedish Meteorological and Hydrological Institute (SMHI) in this project. VA SYD and SMHI
made an analysis of the technology and localized a test site with ideal conditions for the first X-band weather
radar facility in Sweden to be on the top of the Dalby water tower outside the city of Lund. The X-band radar
offers higher resolution when compared to SMHI’s C-band radar. The weather radar was in operation
between 2018-07-03 and 2018-09-12. In addition to precipitation data from VA SYD, the Trelleborg
municipality has contributed their precipitation data.
The study included an analysis of processed radar data from Informatics (IT-supplier) and LTH, and from VA
SYD´s rain-, flow-, and sewerage overflow gauges. Four precipitations events from August 2018 have been
studied in which different data sets, including rain gauge data in Trelleborg municipality, have been
analysed. LTH processed and validated the radar data with a Matlab code while Informatics used a Python
code.
http://www.svensktvatten.se/forskning/svenskt-vatten-utveckling/pagaende-svu-projekt/vaderradarteknik-
inom-va-omradet--test-av-metodik

Partners:
VA Syd – Coordinators
Nicholas South ([email protected]) and Henrik Aspegren ([email protected])

Lund University
Hossein Hashemi ([email protected]) and Andreas Persson ([email protected])
SMHI
Jonas Olsson ([email protected])

Sponsor: Svenskt Vatten Utveckling

5
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

THEME 1: END-USER ENGAGEMENT, TAILORED TOOLS

& PUBLIC PERCEPTION
Knowledge exchange on Climate Adaptation Best Management Practices for Sus-
tainable water management in Resilient Cities
F. Boogaard 1+2, Guri Venvik3
1
Hanze University of Applied Science, Groningen, The Netherlands, Email: [email protected]
2
Global Center on Adaptation, Energy Academy Europe, Groningen, The Netherlands
3
Geological Survey of Norway (NGU), Trondheim, Email: [email protected]

The INXCES project

Cities are becoming increasingly vulnerable to climate change, and there is an urgent need to make them
more resilient. The Climatescan adaptation tool www.climatescan.nl is applied as an interactive tool for
knowledge exchange and raising awareness on Nature-Bases Solutions (NBS) targeting young professionals
in ClimateCafes. Climatescan is a citizen science tool created through ‘learning by doing’, which is interactive,
open source, and provide more detailed information on Best Management Practices (BMPs) as: exact
location, website links, free photo and film material. BMPs related to Innovations for Climatic Events
(INXCES) as stormwater infiltration by swales, raingardens, water squares, green roofs subsurface infiltration
are mapped and published on social media.
Climatescan is in continuous development as more data is uploaded by over 250 people around the world,
and improvements are made to respond to feedback from users. In an early stage of the international
knowledge exchange tool Climatescan, the tool was evaluated by semi-structured interviews in the
Climatescan community with the following result: stakeholders demand tools that are interactive, open
source, and provide more detailed information (location, free photo and film material).
In 2016 Climatescan (first stage of INXCES) was turned into an APP and within two years the tool had over
10,000 users and more than 3,000 international projects. More than 60% of the users are younger than 34
and 51% of users are female, resulting in engagement with an important target group: young professionals.
The tool is applied in Climatecafe.nl around the world (The Netherlands, Sweden, Philippines, Indonesia,
South Africa) where in a short period of time stakeholders in triple helix context (academia, public and
private sector) work on climate related challenges and exchange their knowledge in a café setting.
Climatescan has also been used in other water challenges with young professionals such as the Hanseatic
Water City Challenge and Wetskills.
During the INXCES project over 1000 BMPs related to Innovations for Climatic Events (INXCES) are mapped in
all partner countries (figure 1). The points of interest vary from just a location with a short description to a
full uploaded project with location, description and summary, photos and videos, presentations, links to
websites with more information and scientific papers and books (as Bryggen in Norway:
https://www.climatescan.nl/projects/16/detail ).

6
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

Fig. 1 Norway (73), Sweden (25), The Netherlands (>1000) and Bucharest, Romania (19).

In conclusion, there is a clear demand for a collaborative knowledge-sharing tool on BMPs, where first
impressions of different urban resilience projects can be quickly gained, and examples of climate adaptation
is easily accessible. Further work in linking events to the UN Sustainable Development Goals will further
empower the usability of this web-tool www.ClimateScan.nl. This tool helps policy makers and practitioners
to gather valuable data for decision-makers in a rapid appraisal at neighbourhood and city level. The results
provide insights, create awareness, and builds capacity with bringing together stakeholders in the
Climatescan community.

Acknowledgements: This study would not have been possible without the registered users from the public
and private sectors who have mapped their BMPs. This study would not have been possible without funding
from STOWA and collaboration within the JPI Water funded project INXCES and INTERREG IVb project
WaterCoG.

7
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

Are house owners’ willing to invest in self-protection against flooding?

Jonas Nordström
Lund University

The SUrF- project

Compared to insurances that reduces the severity of flooding for the individual, self-protection may reduce
both the probability and the severity of flooding. Increased self-protection may thus lower the socio-
economic costs of flooding and can be seen as an important part of the societies’ adaptation to climate
change. In this survey we study house owners’ self-protecting behaviour.
The results suggest that a relatively large fraction of the house owners is willing to undertake self-protecting
measures to reduce the probability and severity of flooding. The median expenditures that are spent on self-
protecting means were SEK 5000. There is, however, a large difference in the median and mean
expenditures, and about one fourth of the house owners report no (direct) costs for the self-protection
measures. For house owners that have been affected by flooding, the median expenditures on self-
protection measures were SEK 3000, while the median loss due to the latest flooding were estimated to SEK
10000.
The most common means of self-protection were “not to store or have valuable items in the basement”
followed by “draining”, “remove vegetation that can damage pipes” and “using building material that is less
sensitive to water”.
After the decision to carry our various self-protection measures the house owners stated the following
subjective probability of flooding for their property.

35
Percent of the house owners

30

25

20

15

10

0
0% 0,1% 0,5% 1% 2% 5% 10% 14% 20%
Subjective risk of flooding per year

no self-protection self-protection affected by flooding and self-protection

Figure 1: House owners' subjective probability of flooding per year

8
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

For half of the house owners that have not undertaken any self-protection measures, the subjective
probability of flooding is 0.5 percent or less. About 90 percent of the house owners that have not carried out
any measures, estimate that the probability of flooding is 2 percent or less per year, i.e. relatively low.
House owners that have carried out self-protection measures sated a higher subjective probability of
flooding compared to house owners that have not undertaken any measures. One third of these house
owners say that the probability of flooding is 0.5 percent or less per year, while about half of them stated
that the probability of flooding is between 1 and 5 percent. Half of the house owners that carry out self-
protecting measures thus have or have undertaken measures that reduce the subjective risk of flooding to
between 1 and 5 percent per year. The results also suggest that many of the house owners that have been
affected by flooding and that have carried out self-protecting measures, expect that their property will be
flooded again (during the coming 20 years). For these subjects, self-protection measures such as using
building material that is less sensitive to water and not store or have valuable items in the basement, can
have an important impact on the severity of the expected flooding.

9
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

Building flood resilience – together

Kristina Hall & Nina Steiner, VA SYD
VA SYD, Box 191, 201 21 Malmö, Sweden, E-mail: [email protected] ; [email protected]

The SuRF Project

Cloudbursts in urban areas can lead to massive flooding, in public areas as well as in private houses and
buildings. The costs of this are enormous, both for society and for individual citizens. The problem is not
new, so how come it hasn’t been solved?
Building flood resilience – together is exploring new ways to address the issue in the city of Malmö. The
project was initiated in 2017 by VA SYD, a regional water and wastewater services provider in south-western
Skåne, and operates in close collaboration with the city of Malmö. Four employees at VA SYD, three
engineers and one communications officer, work in the project.
Pre-requisites
There is a gap in the responsibility matrix regarding urban cloudbursts in Sweden (table below). Concerning
in particular ‘area with existing buildings’, there is nobody to hold responsible.

Normal rain Cloudburst

New urban development The wastewater The municipality

services provider

Area with existing The wastewater Nobody

buildings services provider

Another dilemma is the fact that 70% of Malmö consists of privately-owned land. It is therefore important to
make the private property owners take action and contribute to building flood resilience. Since there are no
legal demands on private property owners, actions are all voluntary. Making them want to act is one of the
main tasks for the project Building flood resilience – together.
Objectives and target groups
The project has identified two levels of objectives – a collective change of mind set in society and an increase
of individual actions.
The chosen target groups for the individual action part may be roughly divided into four categories:
1. Colleagues within VA SYD and the city of Malmö
2. Private homeowners (detached houses)
3. Private housing cooperatives (apartment houses)
4. Industries and businesses occupying large areas.

10
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

Strategic approach
“Hard” problems do not always have “hard” solutions. It is often more efficient to use “soft” techniques. We
can’t rebuild the city to make it fit todays conditions. But by involving everyone living and working in the city,
we can achieve change. Communication is therefore essential in order to reach the set objectives, and the
communicative strategy has from day one been to use positive enforcement in all communication. Emphasis
is on showing the benefits of doing things right, instead of showing the negative consequences of doing
things wrong.
In August 2017, a survey was made within target groups 1 and 2 to measure the knowledge about flood
resilience possibilities and the willingness to take individual action. Based on the answers, strategic and
communicative activities were decided on:
• Employing a variation of nudging techniques, ranging from tone of voice to delivering rain barrels to
peoples’ homes. They have all been evaluated and improved on a regular basis.
• Financial compensation of 2 500 SEK for each disconnected downpipe in Malmö has been promoted
strongly.
• Offering services of an engineer to do on site counselling to housing cooperatives.
• Seeking collaboration with housing cooperatives, real estate companies, private property owners,
builders, architects, landscape architects and others, in order to spread awareness and knowledge.
• Taking part in several workshops and giving lectures in order to put focus on cloudburst issues in
every part of the city planning process, locally as well as nationally.
• Not waiting to be found, but to find people through participation at garden shows and fairs
• Offensive PR work with focus on local media.
• Using the web site vasyd.se/platsforvattnet as the centre of all communicative activities. A lot of fo-
cus has been on improving the site and getting people to visit it.
Results
It is clear that there is no ‘one size fits all’-solution. All actions have to have an individual approach. This is
time consuming but effective. Through differentiated communication, offers and actions to each target
group, we have been able to reach more people and achieve more.
The awareness of the importance of building flood resilience has increased among colleagues and city
planners in Malmö.
An increasing amount of private house owners in Malmö are disconnecting their downpipes.
It is clear that there is a need of other solutions than offered today for big warehouses, industries and other
businesses.
A final conclusion is that building flood resilience together with the citizens, entrepreneurs, builders, city
planners and others takes time. But it is possible.

11
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

A review of sustainability of urban flood management from the aspects of

hydrology, economy and the perceived urban design quality
Misagh Mottaghi, Jonas Nordström, Catharina Sternudd, Salar Haghighatafshar, Karin Jönsson
Faculty of Engineering LTH, Lund University

The SurF-project

Flooding in urban areas is a problem of growing concern due to numerous reasons. It has been proved that
only relying on pipe system is not the solution. Urban areas are getting denser and large proportion of
impermeable surfaces makes built-up land more vulnerable to flooding than the surrounding environment.
Accordingly, applying surface solutions, called blue-green solutions and evolving the drainage systems are
essential steps for the reduction of flood impacts. Some municipalities are now well informed of the
necessity for applying green solutions as flood control methods. Many studies have shown that blue-green
solutions are efficient in controlling flooding and recycling the run-off water. However, we have limited
knowledge on the extent to which such solutions are sustainable and even less on how sustainable the
everyday life around them can be. Moreover, we have limited information on how individuals value blue-
green solutions in monetary terms. This information is important from both a welfare perspective and from
a socio-economic cost benefit analytic perspective.
The research is trying to assess the impact that blue-green solutions have on the quality of urban
environments. Here, we apply a multilevel approach to review socio-cultural values that these ecological
techniques offer to the environment. The research objective is the effect that such solutions have on
human’s everyday use and practice. In this study, an urban housing area called ‘Augustenborg Eco-city’ has
been chosen as the case study. The area is located in the city of Malmö in Sweden and includes almost 1970
households. The research team tries to gain knowledge from the user’s perspective to see how people
experience blue-green solutions in their everyday life as well as how they value them in relation to their
experience. The researchers have considered two criteria to come up with specific urban area as the case
study. First, the area needed to have different open stormwater techniques, which were deliberately
planned, designed, and implemented on site to deal with urban flooding. The second and very important
criteria was that the blue green techniques were implemented a while back in past, so the inhabitants have
had enough time to live, explore and experience the area. Augustenborg urban area has been retrofitted in
1998 with the goal of making the area flood resilient as well as more attractive. Augustenborg eco-city is a
well-known example being known as one of the best practices of sustainable urban project.
A questionnaire has been formulated and designed based on the affordance theory, which address questions
regarding urban design qualities. The questions are developed in the way to understand different
possibilities for action and perception that blue-green solutions may bring to the area and people’s everyday
life. Furthermore, a contingent valuation method was used to elicit the respondents’ valuation of the blue-
green solutions in monetary terms. The questionnaires were distributed to all households in the area in
November 2018 and collected by the end of December. The project is a collaboration between Lund
university of Sweden, VA SYD and MKB housing company. The research team is supposed to carry on the
project with a follow-up focus group as a new method for getting deeper knowledge on social and cultural
sustainability of blue-green solutions in Augustenborg Eco-city.

12
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

Responding to end-user requirements before, during and after the flood

Lisa Van Well
Swedish Geotechnical Institute

The MUFFIN project

The MUFFIN seeks to address the gap between the urban and large-scale hydrological modelling
communities and to contribute to improved integrated risk management solutions to urban floods. WP2 on
“End-user value”, for which the Swedish Geotechnical Institute (SGI) has been responsible, had the goal to
optimize the process and outputs of the project with respect to practical value for relevant end-user
categories. This is to facilitate that the flood forecasting meets the specific and concrete needs of the urban
users and can be integrated into their existing organizational structures and current use of forecasting.
To specify the end-user value, SGI used a three-prong or triangulation method to understand the needs and
requirements of the MUFFIN end-users. In the MUFFIN case these three methods consisted of 1) an
international workshop in February of 2017. 2) an end-user survey administered in December 2017 and 3) in-
depth telephone interviews with end-users in November 2017- February 2018.
End-users are had differing needs and conditions and thus MUFFIN cannot provide a one-size-fits all solution
to the problems of urban flooding. However, are some general conclusions about end-user specifications
that can be made from this three-pronged exercise:
• Local level forecasts and observations geared specifically to local specificities and conditions are most
important for end-users but there is also scoped to integrate data at larger scales to complement local level
data.
• End-users prefer data, forecasts and observations of rainfall/flooding at a local level which specific to their
conditions, but also request data on a larger scale such catchment areas. National stakeholders were
interested in extended geographic coverage of smaller watercourses and non-urban areas.
• The greatest need for more guidance and information tools on rainfall and flooding was requested at the
stage “before the flood”, followed by guidance and tools “after the flood”. But methods to integrate
observations and reports during the flood are also important.
• Accuracy and certainty of forecasts and observations appeared to take precedence over lead time or
timing of observations/analysis becoming available, although end-users and stakeholders were reluctant to
specify any trade-offs between accuracy and spatial resolution.
• Visualizations in GIS-formal and web-based visualizations at the different scales would be very useful for
end-users. Communication of observations through citizen observations can an important complement to
radar and rain gauge data to be further explored.
• MUFFIN is very ambitious, and the small steps achieved by the project can drive technology forward, even
if the project can’t solve all problems in each case study area.
• Better tools with higher accuracy within meteorological and hydrological forecasting and modelling are
needed, and if good tools exist for rainfall and flooding forecasts, end-users will find a way to use them.
The presentation will show how the MUFFIN case studies, joint experiments and the RainVis visualization
tool contribute to fulfilling these needs.

13
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

APPETIZER
RAINVIS: a real-time high-resolution high-intensity rainfall visualization prototype
for Sweden
Jonas Olsson1, Peter Berg1, Jim Hedfors2, Mats Öberg2
1
Research and Development (hydrology), Swedish Meterological and Hydrological Institute, 601 76
Norrköping, SWEDEN
2
Division of Climate Adaptation, Swedish Geotechnical Institute, Olaus Magnus väg 35, 581 93 Linköping,
SWEDEN. E-mail: jonas [email protected]

The MUFFIN Project

Real-time visualization of rainfall is today provided by many meteorological institutes and companies
worldwide. This includes both (recent) rainfall observations, often made by weather radar (Figure 1a), and
(short-term) rainfall forecasts by NWP or nowcasting (Figure 1b).

Figure 1. Examples of visualization of radar-observed (a) and forecasted (b) rainfall. A screenshot from the development of the RAIN-
VIS tool (c).

Although these available visualization products are clearly useful for assessing the risk of high-intensity
rainfall and subsequent pluvial flooding, from a hydrological perspective a number of limitations may be
identified:
- Although radar-based rainfall estimates are commonly adjusted towards gauges, to some degree,
systematic errors still exist that affect especially long-term accumulations but potentially also single
events.
- Rainfall is presented only at its highest resolution in time and space, without possibility to accumu-
late over relevant temporal and/or spatial domains.

14
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

- Rainfall intensity is given in arbitrary intervals with no connection to established levels or thresholds
representing e.g. frequency of occurrence.

In RAINVIS, we have attempted to overcome these limitations as follows:

- The radar-based rainfall estimates are more closely adjusted towards gauge observations, repre-
sented by daily gridded fields, which ensures accurate long-term accumulations (Berg et al., 2016).
- Concerning spatial resolution, the radar rainfall is averaged over hydrological basins; ~40 000 sub-
basins covering Sweden with a median size of ~7 km².
- Concerning temporal resolution, besides the highest available (1 hour), rainfall may be averaged
over durations of 2, 3, 6 or 12 hours.
- At each duration, the estimated rainfall depth is compared with national Depth-Duration-Frequency
statistics (Olsson et al., 2019) to provide an alert when 2-, 10- or 50-year return levels are observed
or forecasted.
- Furthermore, for multi-hour durations, observations from the recent hour(s) may be combined with
forecasts for the coming hour(s).

The ultimate aim of RAINVIS is to provide the user with the best possible information and decision support
both before a flood event (forecasts – for early warning), during the event (observation+forecast – for
situation awareness) and after the event (observations – for post-event analysis). Very welcome to have a
look at our demo.

References
Berg, P., Norin, L., and J. Olsson (2016) Creation of a high resolution precipitation data set by merging
gridded gauge data and radar observations for Sweden, J. Hydrol., 541, 6-13,
doi:10.1016/j.jhydrol.2015.11.031.
Olsson, J., Södling, J., Berg, P., Wern, L., and A. Eronn (2019) Short-duration rainfall extremes in Sweden: a
regional analysis, Hydrol. Res., nh2019073, doi: 10.2166/nh.2019.073.

15
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

The FLOODVIEW portal – a web-based prototype for innovative modelling and

visualization of heavy rainfall and urban flood risk
Jonas Olsson1, Heiner Körnich1, Linus Zhang2, Magnus Ihrsjö3, Mats Alexandersson3, Ramesh Saagi4, Kenneth
Persson2
1
Research and Development, Swedish Meteorological and Hydrological Institute, 601 76 Norrköping,
SWEDEN
2
Water Resources Engineering, Lund University, Box 118, 221 00 Lund, SWEDEN
3
4IT AB, IDEON Science Park, Scheelevägen 17, 223 70 Lund, SWEDEN
4
Industrial Electrical Engineering and Automation, Lund University, Box 118, 221 00 Lund, SWEDEN
e-mail: [email protected]

The MUFFIN Project

The FLOODVIEW project, utilizing concepts developed within the MUFFIN project, is aiming to provide a
web-based flood control decision support system for municipality managers/decision makers to identify
effective solutions to minimize urban flooding. It includes an early warning system through flood forecasting,
drainage system impact assessment, low impact development practices, insurance issues and management
framework. Three pilot sites at different municipalities from Canada and Sweden are chosen for the study.
The water sector benefits from deployment of solutions for flood control, decision support and quick market
access for water technologies.
The FLOODVIEW components with a direct link to the MUFFIN project include a prototype of real-time 1-
hour discharge simulations for Höje River in southern Sweden, which has a history of flooding problems.
Another envisaged component is a “cloudburst risk estimator”, which uses high-resolution meteorological
ensemble forecasts to estimate the risk of high-intensity rainfall in a certain location (e.g. a city), taking into
account the spatial forecast uncertainty.
SMHI/Jonas: description of the Höje Å 1-h forecast prototype and the concept of “rain risk estimation”
A functional prototype of the frontend application is currently under development. The frontend will be
completely web-based and fully responsive to be usable from both computers and mobile clients. From the
frontend users are able to search and browse results from various simulations and predictions. Depending
on the type of models and data used, the interface will present the results in different ways, but most of the
presentations will consist of a combination of maps and charts.
On the backend side the FLOODVIEW application will fetch data both ad-hoc and on schedule from various
data sources, like public API:s, files uploaded to FTP-servers specifically for the FLOODVIEW project etc.
There are also planned functionality for triggering simulations ad-hoc, storing the output in a database and
optionally making the results available for others to view.
A role-based permission system will be implemented where power users will have ability to trigger
simulations, upload data and share results with others, while other users will be limited to view results and
configure individual alerts etc. There will also be presentations that are publicly available without any login
required. Figure 1 show a screenshot of the FLOODVIEW portal
In addition to above prototype development, with the early warning and flood mitigation concern in mind,
and in order to understand how these changes will affect the existing infrastructure in urban areas, an
integrated urban water model is also needed to identify effective solutions to minimize urban flooding. In
this study, a newly developed urban water model software, the Tokyo Storm Runoff (TSR ) model was tested
and examined for the small urban area Augustenborg in Malmö, Sweden.
16
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

Figure 1. Example screenshot from the FLOODVIEW portal: real-time forecasting in Höje River.

17
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

THEME 2: RAINFALL & ENVIRONMENTAL

OBSERVATION & FORECASTING
Can we trust radar? High-intensity rainfall in operational radar
M. Schleiss,1 T. Niemi,2 T. Kokkonen,2 S. Thorndahl,3 R. Nielsen,3 and J. Olsson,4
1
Dept. of Geoscience and Remote Sensing, Delft University of Technology, Netherlands
2Dept. of Built Environment, Aalto University, Finland
3Dept. of Civil Engineering, Aalborg University, Denmark

4
Dept. of Hydrology, Swedish Meteorological and Hydrological Institute SMHI, Norrkoping, Sweden

The MUFFIN Project

Today, several high-resolution radar rainfall products for use in hydrology are readily available across the
globe. Com- pared with gauges, radar provides superior spatial coverage, leading to more insight into the
spatio-temporal characteristics of rain events and their link to hydrological response. But when it comes to
accurately measuring small-scale rain- fall extremes responsible for urban flooding, many challenges remain.
Large disagreements between radar and gauges associated and non-linear error propagation into
hydrological models strongly limit the reliability of radar for local flood predictions. The hope is that by
moving to higher resolution and making use of dual-polarization errors can be reduced. However, each
country seems to have developed its own strategy for this. Consequently, there is a strong need for an
objective, multinational and multiscale assessment of radar’s ability to capture heavy localized rain. This
study sheds more light on this issue by providing a detailed analysis of 4 different radar products in
Denmark, the Netherlands, Finland and Sweden. The top 50 events in the observational record for each
country are used to quantify the average disagreements between radar and gauges but also errors affecting
the peak rainfall intensities. By comparing different types of radar products (e.g., single vs dual pol,
composite vs single radar and bias corrected vs uncorrected) and analysing error propagation across scales,
important conclusions and recommendations can be drawn as to the use of radar in urban hydrology.

Data & Methods

Using automatic rain gauges, the top 50 events for each country were selected. For each event, the radar
pixels around the gauges were extracted and aggregated in time to match the sampling resolution of the
gauges. Table 1 provides an overview of the used radar products.
Denmark: A single radar product from a C-band radar 40 km south of Copenhagen. The radar scans at 9
different elevation angles are combined to generate a gridded product at 10 min and 500 m resolution.
Rainfall rate R is estimated based on a fixed Z-R relationship Z = 200R1.6. Rain rate values are corrected for
mean field bias using hourly data from a network of 67 rain gauges.
Netherlands: The used product is a 5 min, 1 km composite based on reflectivities from 2 C-band radars
operated by the Royal Netherlands Meteorological Institute (KNMI). Rainfall estimates are obtained by
applying a constant Z-R relationship given by Z = 200R1.6. It is adjusted for bias at hourly time scales using
a network of 35 rain gauges.

18
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

Finland: The Finnish radar product is an experimental product from the FMI OSAPOL-project that takes
advantage of dual-polarization and measurement geometry by combining data from 8 C-band dual-pol
Doppler radars. For heavy rain, rainfall is estimated using Kdp while for low to moderate intensities a
fixed Z-R relation given by Z = 223R1.53 is used. The OSAPOL is the only product that is not gauge-
adjusted.
Sweden: The Swedish radar product uses data from 12 single-polarization C-band Doppler radars. Rainfall
rates on the ground are estimated trough a constant Z-R relationship Z = 200R1.5. The product is adjusted
for range-dependent bias. Although several radars are available, the current system does not take
advantage of overlapping measurements and only uses the data from the nearest radar to estimate the
rainfall rate.

Table 1. Radar products used in the study

Country Radar type(s) Resolution Method

Denmark 1 single-pol Cband 500 m, 5 min Z-R

Finland Sweden 8 dual-pol Cband 1 km, 5 min Z-R Kdp

Netherlands
12 single-pol Cband 2 km, 15 min Z-R Z-R

2 single-pol Cband 1 km, 5 min

Results
Overall agreement between radar and gauges
Figure 1 shows the rainfall intensities of radar versus gauges at the highest temporal resolution for each
country. Looking at Figure 1, we see strong disagreements between radar and gauge estimates, which is a
common problem at high temporal resolutions. The relative root mean square errors are between
116.4% and 135.1% and the linear correlation coefficients between 0.72 and 0.83. The multiplicative bias
values of 1.59, 1.41, 1.56 and 1.77 show that radar systematically underestimates the rainfall rate by 41-
77% across countries. Overall, the Dutch product exhibits the best agreement, followed by Finland, Denmark
and Sweden. Differences between countries are hard to interpret due to the many confounding factors.
However, the fact that the Swedish product has the lowest performance is likely due to its lower resolution
of 2 km and 15 min. Interestingly, the high 500 m spatial resolution in the Danish product does not seem
to provide an obvious advantage over the 1 km resolution in the Netherlands and Finland. One possible
explanation for this could be that the Finnish and Dutch products combine data from multiple radars
whereas Den- mark only uses only one. The fact that the Dutch and Finish products have comparable
performance is interesting, since the Finish product has not been bias-adjusted but relied on polarimetry to
estimate rain rates in times of heavy rain.

19
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

Figure 1. Radar versus gauge intensities at the highest available temporal resolution for each country.

Agreement across scales

Figure 2 shows the relative root mean square error and correlation coefficient of radar versus gauges
estimates for aggregation time scales up to 2 h. We see that the radar products with higher spatial and
temporal resolution generally agree better with the gauges. Denmark (500 m and 5 min) comes out on
top, followed by the Netherlands, Fin- land (1 km, 5 min) and Sweden (2 km and 15 min). Note that even
at 2 h time scale, differences between radar and gauges remain substantial, ranging from 65.1% to 85.9%.

Figure 2. Relative root mean square error and correlation coefficients of radar versus rain gauge estimates at different aggregation
time scales.
20
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

Agreement in terms of peak intensity

Figure 3 shows the peak intensity ratios between gauges and radar as a function of aggregation time scale.
Compared to the average bias which is “only”41-77%, the underestimation of the peaks appears much larger,
in excess of 150-300%. Most of the bias is attributed to errors at mall time scales of 0-30 min. The mean
field bias correction does not seem to improve performance at these scales. The results for the Danish
product are particularly interesting, as it has the best overall agreement with gauges in terms of rmse but
also the worst performance in terms of peak intensities, showing that the ability to combine radar
measurements from multiple viewpoints appears to be more important for capturing the most intense
parts of a storm than spatial resolution.

Figure 3. Peak intensity ratios versus aggregation time scale for each country.

Conclusions
Overall, radar and gauges were in fair agreement with each other, with average biases of 40-80%. However,
much larger discrepancies of 150-300% were observed during the peaks. The spatial and temporal resolution
of the radar product seems to play an important role in determining the overall bias with respect to
gauges. But high resolution alone is not the key to success. The two best radar products (Netherlands and
Finland) all combined data from multi- ple radars to help mitigate attenuation in times of heavy rain,
performing better in terms of peak intensities than the Danish product which has higher resolution (i.e., 500
m) but only used data from a single radar. The rainfall estimation algorithm also plays an important
crucial role. While the Dutch used a fixed Z-R relationship and corrected for bias using gauges, the Finish
achieved similar performance by relying on polarimetric variables such as Kdp without bias adjustment.

Acknowledgments: This research was funded by ‘Water JPI Europe’, ERA-NET Co-fund
‘WaterWorks2014’ project MUFFIN (Multiscale Urban Flood Forecasting: From Local Tailored Systems
to a Pan-European Service). The Finish OSAPOL project was funded by the European Regional
Development Fund and Business Finland.

21
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

Accuracy of satellite-observed rainfall over Sweden

Hossein Hashemi, Faculty of Engineering LTH, Lund University

The SUrF Project

Precipitation is an important source of freshwater, and its accurate measurement is essential. Over the
years, several instruments, i.e., rain gauge, weather radar, and satellite, have been used to measure
precipitation. Ultimately, the choice of precipitation data depends on the particular application, the
catchment size, climate, and the time period of study. In poorly gauged regions the use of remotely sensed
precipitation products is an absolute necessity.
Rain gauge and weather radar are the most common ground-based precipitation measurement techniques.
However, both of these instruments have their own set of limitation, particularly concerning spatial
resolution and spatial coverage. The Global Precipitation Measurement (GPM) has been providing, since
February 2014, high-resolution precipitation data sets with larger spatial coverage relative to Tropical
Rainfall Measuring Mission (TRMM). GPM is expected to deliver the more accurate estimation of light
rainfall and snowfall, particularly, in the high latitudes such as Scandinavia. However, the evaluation of the
performance of the GPM mission is an essential step before the extensive use of its products in
hydrometeorological studies. Despite the unique spatial coverage, 60 N and 60 S globally, spatial
resolution, 0.1  0.1, and advanced retrieval algorithm, Integrated Multisatellite Retrievals for GPM
(IMERG), most recent studies have shown that the GPM products show some differences relative the
ground-based measurements. Since the GPM algorithm is based upon experiences from TRMM algorithm
(TMPA), therefore IMERG is experiencing new sources of error as the measurements have extended beyond
the subtropical region where the light rain and snowfall are more frequent.
Though GPM data product has a higher spatial and temporal resolution, it is quite often found to
overestimate or underestimate the true precipitation based on the region, topography, and type of
precipitation. In order to understand the performance of GPM and its limitations, this study aims to evaluate
the GPM products in daily and sub-daily scales against the gauge measurement in Sweden. The results show
that the GPM daily estimates perform better than hourly estimates. The satellite data agree well with gauge
data concerning the correlation coefficient. In the case of error analysis, such as MEA, RMSE, POD, and FAR,
the daily estimate outperforms hourly estimates. The results also show that the satellite sensor is
experiencing detection issues at the southwestern coastline of Sweden in both daily and sub-daily scales.
This might be related to the common precipitation type and temporal measurement issue with the GPM
core observatory (three-hourly) in this area. Overall, satellite performance relative to ground-based
measurement shows less accuracy in the coastal regions. These errors can be attributed to many factors like
the influence of wind, weather system, and the gap between satellite revisit of any given spot. Therefore,
further research is required to adjust the satellite data based on the investigated error.

22
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

Use of weather radar in the water sector in Denmark; what can we learn?

Søren Thorndahl1, Malthe Ahm2, Rasmus Nielsen3, Christoffer Bang Andersen4, Jesper Ellerbæk Nielsen5,
Michael R. Rasmussen6
1
Aalborg University, Department of Civil Engineering; [email protected] (MUFFIN)
2
Aarhus Water, [email protected]
3
Aalborg University, Department of Civil Engineering; [email protected] (MUFFIN)
4
Aalborg University, Department of Civil Engineering; [email protected]
5
Aalborg University, Department of Civil Engineering; [email protected]
6
Aalborg University, Department of Civil Engineering; [email protected]

The MUFFIN project

The interest in applying weather radar data for urban hydrological purposes has increased significantly in
recent years in Denmark. Until now radar data has only been applied in research or development projects,
but operations and services in urban hydrology are impending. Potential users of quantitate precipitation
estimates from weather radar counts water utility services, municipalities, consultants, authority, research
institutions and more. Potential usage of radar data covers both offline applications such as (Thorndahl et
al., 2017): analysis and statistics of rainfall, reanalysis of damaging extreme events, insurance claims,
management purposes, e.g. design or urban hydrological systems, as well as online applications such as
severe rainfall warning systems, flow/flood warning based on online hydrological models, real time control
of urban hydrological systems, etc.

In 2018 the largest water utility services in Denmark along with research institutions and private companies
to was funded the project : “VEVA” (in Danish: VEjrradar i VAndsektoren) with the overall purpose to
promote the application of weather radar data in the Danish water sector (www.veva.dk). The objectives of the
project are to:
• Create an open standard for use of radar data for hydrological and hydraulic purposes,
• Create a widely applicable quality controlled and adjusted data format,
• Develop and maintain data information models and data formats,
• Have expert knowledge and experience on use of radar data in the Danish water sector,
• Promote continuous research and development in use of radar data.

Danish radar data both cover X-band radars (primarily owned by water utility services) and C-band radars
(owned and operated by the Danish meteorological institute). The VEVA project includes both types of
radars which individually have different specifications and pros and cons (see e.g. Thorndahl et al. (2017).
Figure 1 shows an example of Danish radar data.

23
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

Figure 1: Example of data from the Danish Meteorological institute radar in Virring, Denmark on May, 10, 2018. The example shows a
narrow convective storm over the city of Vejle where almost no rain was recorded in the rain gauge network, but the radar recorded
up to 20 mm of rain during a few hours.

References
Thorndahl, S., Einfalt, T., Willems, P., Ellerbæk Nielsen, J., Ten Veldhuis, M.-C., Arnbjerg-Nielsen, K.,
Rasmussen, M.R., Molnar, P., 2017. Weather radar rainfall data in urban hydrology. Hydrology and Earth
System Sciences 21. doi:10.5194/hess-21-1359-2017

24
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

Experiences of X-band weather radar in Skåne

Nicholas South, VA SYD
VA SYD, Box 191, 201 21 Malmö, Sweden E-mail: [email protected]

The Radar Project

The purpose with the test period was to obtain an in-depth conclusion on how an X-band weather radar
facility could be implemented at VA SYD, including data quality control and validation. The following parts
constitute the report:
• A literature study and an evaluation of the critical parameters for a weather radar facility
• An evaluation with proposed sewerage system and wastewater treatment plant applications of the
weather radar data
• Validation of the radar data
VA SYD has collaborated with the Faculty of Engineering at Lund University (LTH), Sweden Water Research
(SWR) and the Swedish Meteorological and Hydrological Institute (SMHI) in this project. VA SYD and SMHI
made an analysis of the technology and localized a test site with ideal conditions for the first X-band weather
radar facility in Sweden to be on the top of the Dalby water tower outside the city of Lund. The X-band radar
offers higher resolution when compared to SMHI’s C-band radar. The weather radar was in operation
between 2018-07-03 and 2018-09-12. In addition to precipitation data from VA SYD, the Trelleborg
municipality has contributed their precipitation data.
The study included an analysis of processed radar data from Informatics (IT-supplier) and LTH, and from VA
SYD´s rain-, flow-, and sewerage overflow gauges. Four precipitations events from August 2018 have been
studied in which different data sets, including rain gauge data in Trelleborg municipality, have been
analysed. LTH processed and validated the radar data with a Matlab code while Informatics used a Python
code.
The weather radar facility recorded precipitation on several occasions in August 2018. In particular, the
event in Trelleborg on the 11th of August recorded a cloudburst event with local differences in accumulated
rain amounts (0-15 mm), which was confirmed and illustrated with the radar. However, some differences
were observed between the recorded amounts of precipitation by the rain gauges relative to radar
estimates. This shows the importance of accurate rain gauge measurements, as well as calibrating the radar,
as the first steps for the future facility.
The study indicated that the X-band radar could provide VA SYD with a future warning system with a margin,
at best, of 0,5-1 hour. A collaboration with the HOFOR/BIOFOS facility in Copenhagen could extend the
observational range of the radar further, based on the condition that most rain fronts come from the west.
The next step would be to initiate a project where the X-band radar will be integrated with other methods of
precipitation measurements, both observational and prognosis tools.
The weather radar in Dalby will measure precipitation in the southwestern part of Skåne from 2019, when a
permanent facility will be installed.
The weather radar technology will become a key component in a future warning system, which would
benefit the customers of Swedish water utilities.

25
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

Subsidence areas trends in Bucharest city, based on PS-InSAR analyses

Alina Radutu1,2*, John Dehls3, Guri Venvik3, Traian Ghibus1, Constantin Radu Gogu1
1
Technical University of Civil Engineering, Bucharest
2
Romanian Space Agency
3
Geological Survey of Norway
*Corresponding author: [email protected]

The INXCES Project

In Bucharest, a dynamic city, the consistently development of ground and underground works is associated
to groundwater dynamic changes. These impact the stability of the ground surface and can produce
subsidence. Using Sentinel 1 data for the period of October 2014-April 2018, the Norwegian Geological
Survey produced a map of vertical displacements for Bucharest by applying PSInSAR technique. Figure 1
presents the PSInSAR map obtained using Sentinel 1 data from ascending orbit 131. This map was used to
identify the city areas where subsidence is encountered. Hence, 13 areas with subsidence trends were
identified (see Fig.2). The comparison of these results to other vertical displacement maps for the historical
periods 1992-2000 (ERS data) and 2002-2009 (ENVISAT ASAR data) processed in the frame of INXCES project
as well as to the results of several previous projects studying the vertical ground dynamics of the same area,
revealed additional areas showing subsidence trends since the beginning of the 90s. Five of these areas have
been identified in the frame of PanGEO project (Vîjdea și Bindea 2013).

Fig. 1- PSInSAR ground deformation map of Bucharest Fig. 2- Identified subsidence areas in Bucharest

For 2014 – 2018 time period an area marked on the map as Cotroceni_04 has been identified as presenting
only subsidence. At the end of 2015 a collapse of the terrain occurred, as a Tunnel Boring Machine (TBM)
was crossing this area for the construction of a new subway line. The alluvial matrix, affected by the
groundwater barrier effect produced by the extensively channelized Dambovita River and another subway
line, collapsed. Poor studied geological and urban hydrogeological aspects, largely contributed to this
accident. As a consequence, many buildings of this area were damaged, even the ground surface was
registered as being stable during the previous monitoring periods.
26
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

Other three new observed areas affected by subsidence are newly built areas. For these, two hypotheses are
considered: the lack of subsidence trend shown by the previous periods (the subsidence trend could not be
observed due to the vegetation coverage) or the areas were previously stable, meanwhile the constructions
development process broke this stability. Beside these particular areas, a subsidence trend could be revealed
along Colentina River, one of the rivers crossing the city.

References
Vîjdea, Anca, and Gabriel Bindea. "D7.1.33 GeoHazard Description for Bucharest, Enabling Access to
Geological Information in Support of GMES (PANGEO)." 2013.

27
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

APPETIZER
Subsidence in urban areas measured by InSAR (Sentinel1) related to flooding
Guri Venvik1, Ane Bang-Kittilsen1, John Dehls1 & Floris C. Boogaard2
1
Geological Survey of Norway, Trondheim, Norway, [email protected]
2
Hanze University of Applied Science, Groningen, The Netherlands, [email protected]

The INXCES Project

Rapid changes in the urban environment due to growth puts the urban water cycle out of balance, hence,
affecting other surface and subsurface processes, such as subsidence and surface water management.
Subsidence of the ground is causing risk and hazard, as well as unexpected costs. This newly, November
2018, launched tool InSARNorge is Open Access and part of the Copernicus program.

Figure: Subsidence in three cities in Norway. Some areas clearly show subsidence, indicated by red points. The Bryggen (Wharf) in
Bergen has now stabilized due to mitigation, such as SuDS. www.insar.ngu.no

28
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

In a recent study (Venvik et al. submitted) datasets from InSAR satellites showing subsidence are combined
with data from flood modelling in two different analytical methods using ArcGIS tools to develop a risk
assessment map for areas most prone to the combination of both flooding and subsidence. Applying user-
centred principles, this work focuses on methods for risk assessment maps as a support tool to locate areas
where mitigation of subsidence and adaptation for surface water management will be most efficient and
measures can be implemented. The results of the methods for risk assessment maps show that one of the
methods give significant results compared to the other method. Such method will be a helpful tool for
decision-makers when prioritizing areas for measures such as Sustainable urban Drainage Systems (SuDS).
The study is related to the JPI Water funded project INXCES (www.inxces.eu).

Figure: Areas of subsidence and flooding in Bergen city centre, Western Norway. The darker colour of the dot the higher average ve-
locity (mm/year) for the subsidence. Blue color shows the flooded areas, with increasing water depth with darker colour (Venvik et al.
submitted).

This is an interactive tool that can showcase multiple real situations related to different problems such as,
areas prone to flooding, underground constructions, cultural heritage etc.
https://insar.ngu.no/

Aknowledgement: This research is supported through the JPI Water funded INXCES research project
“Innovation for eXtreme Climatic EventS”. Sentinel-1A/B images are provided by the European Space Agency
(ESA). We thank the Bergen Municipality, Tauw and Hogeschool van Amsterdam, University of Applied
Sciences, for support for this work.
Reference:
Venvik, G. Bang-Kittilsen, A. Boogaard, F.C. & Dehls, J. (Submitted) Risk assessment for areas prone to
flooding and subsidence - a concept model with case study from Bergen, Western Norway. Journal of
Hydrological Research Special Issue.

29
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

THEME 3: URBAN FLOOD MODELLING &

FORECASTING
High resolution modelling of urban flood modelling and its origin in the large-scale
catchment
Andreas Persson, Abdulghani Hasan, Petter Pilesjö and Hampus Nilsson
Lund University

The SuRF Project

Most of the hydrologic models currently used in flood forecasting are of the type catchment models, e.g.
EFAS, HYPE, SWAT. Catchment models are generally low-resolution models, intended for large scale
modelling, which use sub-catchments as the basic unit for the calculations (i.e. the size of sub-catchments
are large). As an example, the average resolution of the Swedish hydrologic model used for flood forecasting
(S-HYPE) [1] and two European operational floods forecasting systems (E-HYPE and EFAS) are around 13,
250, and 25 km2 respectively. Such low-resolution floods forecasting models are important to provide alerts
to when a flood may occur in the sub-catchments outlets and for a large model domain in reasonable
running time. The sub-catchment outlets coincide in general with watercourses, mainly rivers. The resolution
does not permit a detailed distributed flow or flood mapping as other high-resolution models do. However, a
high-resolution, grid-based, hydrological model (e.g. HEC-RAS, MIKE FLOOD, etc.) can provide detailed flood
mapping but have long processing time for the large areas of the sub-catchments. The models are usually
not public domain or open source models. The resolution is crucial when it comes to urban flood modelling.
Keeping track of flow paths and water depths site specifically is necessary to ensure an adequate model
result, showing the property, infrastructure and number of individuals potentially affected by the flood.
Tools or models that take both a larger contributing area and a smaller affected area into consideration in
the modelling process is sought for by authorities and municipalities as competence in specialized flood risk
modelling often is lacking. This is especially true for mid-sized and smaller municipalities and has been stressed
in a report from the Authority Network for Climate Adaptation (including Lantmäteriet, SMHI, Sjöfartsverket,
Länsstyrelserna). Municipalities specifically asked for methodologies in relation to flood mapping and risk anal-
ysis [2]. Municipalities prefer to use Geographical Information Systems (GIS) for data analysis, especially when
it comes to hydrology, since some competence in GIS is available.

To overcome the lack of high-resolution results for specific areas and provide a practical solution to deal with
the problem in a GIS-environment, a dynamic hydrologic model, TFM-DYN [3, 4], is further developed within
this project and coupled with the HYPE model. The TFM-DYN model is a grid-based model that move water
in a DEM and distribute it proportionally to lower lying cells. At the same time, it allows infiltration and
consider friction in order to calculate the water velocity of the overland flow. With the ability to choose a
time step, the water distribution and depth can be presented at different stages of a flood event.
The TFM-DYN do need an input in order to produce realistic results and therefore the output results from
the large domain of the HYPE model is interactively used as an input for the detailed high-resolution dynamic

30
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

flood modelling in TFM-DYN. With this coupling, the two models cover both the upstream large-scale general
water distribution and the high-resolution modelling required in urban areas.

Figure 1. Water depth calculations from TFM-DYN superimposed on an aerial image covering a few blocks in Lund municipality, Swe-
den. (Source: Nilsson, 2017)

References:
[1] Lindström, G., Pers, C., Rosberg, J., Strömqvist, J. and Arheimer, B. (2010). Development and testing of the HYPE (Hydrological
Predictions for the Environment) water quality model for different spatial scales. Hydrology Research, 41(3-4), pp.295-319
[2] Structor Miljöbyrån. Geodata För Klimatanpassning - Kommuners Och Länsstyrelsers Behov av Geodata och Geodatatjänster för
ett Förändrat Klimat. 2018. (LM M 1800088)
[3] Pilesjö, P., Hasan, A., 2014. A Triangular Form-based Multiple Flow Algorithm to Estimate Overland Flow Distribution and
Accumulation on a Digital Elevation Model. Transactions in GIS 18, 108–124. https://doi.org/10.1111/tgis.12015
[4] Nilsson, H., 2017. A dynamic and spatially distributed rainfall runoff model - Developing a model for overland flow in GIS, based
on a multiple flow direction algorithm (MSc). Lund University, Water Resources Engineering.

31
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

Hydrodynamic vs. hydrological modelling: a comparative study in Aalborg and

Helsinki
Rasmus Nielsen1, Tero Niemi2, Søren Thorndahl3, Yeshewatesfa Hundecha4, Teemu Kokkonen5, Jonas Olsson6,
Marc Schleiss7
1
Aalborg University, Dept. of Civil Engineering, [email protected]
2
Aalto University, Dept. of Built Environment, [email protected]
3
Aalborg University, Dept. of Civil Engineering, [email protected]
4
Swedish Meteorological and Hydrological Institute, Dept. of Hydrology, [email protected]
5
Aalto University, Dept. of Built Environment, [email protected]
6
Swedish Meteorological and Hydrological Institute, Dept. of Hydrology, [email protected]
7
Delft University of Technology, MUFFIN

The MUFFIN project

Flooding in urban areas is an environmental hazard and is a great cost for the communities affected. Urban
flooding is increasing due to the increasing urbanization of the city areas and climate changes. During rainfall
events, rainfall over urban catchments travels through anthropogenic surface pathways and is collected by
the stormwater sewer systems and routed to the recipient. Urban flooding is most often caused by the
sewer capacity being exceeded, which may happen during the extreme rainfall events, where water from the
sewer system may be surcharged to the surface. Accurate and fast flood simulation and prediction are
needed to perform diagnostics of historical flood events and decision-support tools to mitigate future
events.
Hydrodynamic models are the de facto standard for simulating flood risk in urban catchments, due to the
accurate model description of anthropogenic restructure of the natural surface and underground flow
pathways. The downside of the hydrodynamic models is long simulation times, which can be of critical
importance for real-time prediction of an incoming urban flood. In both reality and in the hydrodynamic
models, urban flood is often observed in the same general areas, where the sewer system lacks the capacity
to route incoming stormwater away from the area during extreme rainfall events. Knowing the flood prone
areas beforehand, obtained from either experience or using a hydrodynamic model as a diagnostic-tool,
limits the need for accurate spatial accuracy of predicting incoming flood, i.e. prediction of the exact location
on each surcharged manhole is no longer of great importance and simulation speed becomes the most
important factor.
Large-scale hydrological models have been used to predict flood in rural areas, where natural surface and
underground flow paths are well described by simple lumped basins. The use of large-scale hydrological
models to predict urban flooding have so far been limited to urban catchments, where the flood is mainly
controlled by river systems where the lead time between rainfall and is in the order of 12 hours to days. This
is mainly due to the limited ability to describe the water cycle in fast-reacting urban catchments, where high-
resolution models are needed to accurately capture flood. Recent developments of high-resolution
hydrological models with detailed land-cover and surface flow path descriptions may improve the simulation
accuracy of the urban water cycle while still being several magnitudes faster than hydrodynamic models.
In this abstract, a comparative study between two hydrodynamic models and a newly developed high-
resolution hydrological is presented. Two catchments in Helsinki, Finland and Aalborg, Denmark are used as

32
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

study cases to calibrate and test the high-resolution hydrological model’s performance against two
hydrodynamic models, see an example of comparison of runoff volume in Figure 1.

Figure 1: Total runoff volume per event for the two study cases. The fit between the hydrodynamic and high-resolution hydrological
model is considerately better for Pakila, Helsinki.

The high-resolution model’s ability to accurately predict urban flood is improved by linking the flood risk as
estimated by the hydrodynamic models to a variable in the high-resolution hydrological model. Flood risk is
defined in the hydrodynamic flood threshold based on the number of flooded manholes, see Figure 2. This
enables the high-resolution hydrological model to relate the simulated variables to a flood risk that is also
accurate in urban catchments, where anthropogenic changes the natural water cycle.

Figure 2: Flood severity indexes based on thresholds for the number of flooded nodes separating each index. The thresholds are cho-
sen when a steep gradient is observed in the number of flooded nodes.

33
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

The relationship between rainfall and pluvial flooding in Rotterdam based on

citizen reports
T. Xian,1 M.-C- ten Veldhuis,1 M. Schleiss,2 C. Bouwens,3 and N. van de Giesen1
1
Dept. of Water Management, Delft University of Technology, Netherlands
2
Dept. of Geoscience and Remote Sensing, Delft University of Technology, Netherlands
3
Beheer Water en Riolen, City of Rotterdam, Netherlands

The MUFFIN Project

Urban pluvial flooding is one of the most costly natural hazards worldwide. The complexity of the
involved processes and the lack of long-term field observations means that many crucial aspects related to
urban flood risks still remain poorly understood. In this paper, the possibility to gain new insight into
urban pluvial flooding using citizen flood observations is explored. Using a ten-year dataset of radar rainfall
maps and 70,000 citizen flood reports for the city of Rotterdam, we derive critical thresholds beyond
which urban pluvial flooding is likely to occur. Two binary decision trees are trained for predicting flood
occurrences based on peak rainfall intensities across different temporal scales. Results show that he
decision trees correctly predict 37%-52% of all flood occurrences and 95%-97% of all non-flood occurrences,
which is a fair performance given the uncertainties associated with citizen data. More importantly, all
models agree on which rainfall features are the most important for predicting flooding, reaching optimal
performance whenever short- and long-duration rainfall peak intensities are combined together to make a
prediction. The encouraging results suggest that citizen observatories, although prone to larger errors and
uncertainties, constitute a valuable alternative source of information for gaining in- sight into urban
pluvial flooding.

Materials and methods

The study focuses on the city of Rotterdam, which is located in a low-lying area in the delta of the Rhine
and Meuse rivers in the Netherlands. The city counts more than 640,000 inhabitants spread over a heavily
urbanized area with more than half of the total area being paved or semi- paved. Most areas of the city
are drained by a combined sewer system (1,800 km) whereas the remaining areas are drained by a
separate sewer system (500 km). Historically, stormwater systems in Rotterdam have been designed to be
able to cope with rainfall intensities of approximately 20 mm/h over short time windows of 10-15 minutes
with an estimated in-sewer storage capacity of 10-12 mm. However, the original infrastructure has aged
considerably, and some parts had to be renewed which means that the exact capacities are largely
unknown.

Citizen observations of urban pluvial floods

In 2004, the municipality of Rotterdam established a citizen observatory database with the objective to
collect flood reports by means of phone, email, mobile app, and web- page. Over a period of 10 years
between 2008 and 2017, about 70,000 citizen reports were collected this way. The total number of
reports after quality control and removal of weekends (when call centres are closed) was approximately
38,300. It is important to clarify that not all flood reports in the dataset are directly attributable to
rainfall. Even on days without precipitation, citizens will call in to re- port flooding in their house or
neighbourhood associated with other causes not directly related to rain (e.g., high ground- water levels,
34
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

bursts of water supply pipes etc). However, evidence shows that the average number of reports received
during dry days tends to be much lower than on rainy days. To better distinguish between the two, a dry-
weather base- line of 20 reports/day was established. Using this baseline, 443 weekdays with urban pluvial
flooding and 2,166 days with non-pluvial flooding were identified.

Radar rainfall observations

Information about rainfall intensity and spatial distribution was obtained from the national radar
network in the Netherlands. For the purpose of this study, a ten-year time series of radar rainfall maps
(from 2008 to 2017) at temporal and spatial resolutions of 5 min and 1 km, has been retrieved over the
city. To investigate the effect of temporal resolution, the radar data were aggregated from 5 min to 8
coarser temporal scales (10, 15, 30, 60, 120, 360, 720 and 1440 min).

The models
To model the relationship between peak rainfall amounts and the occurrence of pluvial flooding, we used a
binary decision tree in which ‘yes” means pluvial flooding (i.e., more than 20 reports per day) and “no”
means non-pluvial flooding (i.e., fewer than 20 reports per day). The features used to make the decisions
along the different branches of the tree are given by the peak rainfall amounts Rdx (in mm) (expressed in
mm) for each day, at nine different temporal aggregation scales. Two different models are considered:
• Model 1: uses all 9 available rainfall features at temporal aggregation scales from 5 min up to 24
hours. During training, the algorithm automatically chooses the most relevant features and
thresholds for classification.
• • Model 2: is trained using only 2 features: Rd5 and Rd1440 representative of the small-scale and
large-scale peak intensities. This model is expected to perform slightly worse than the full model.
However, since it only relies on 2 features for making its classification, it is operationally more
straightforward to implement and easier to interpret.

Results
The first trained model is shown in Fig. 1-(a). Al- though all 9 rainfall features were available for training,
only three of them (5 min peak, 60 min peak, and daily total) were used to make the classification. The
first and most important threshold appears to be the total daily rainfall amount Rd1440 . Days with Rd1440
< 12.7 mm were labelled as non-flood days while the days with accumulations above 21.19 mm were labelled
as flood days. The tree also con- siders two additional bifurcations depending on whether the hourly maxima Rd60 is
larger than 6 mm and the 5-min rainfall amount Rd5 larger than 0.35 mm (4.2 mm/h). The 6 mm threshold at
the hourly scale is particularly helpful for separating long lasting events with low intensities from those
with higher peak intensity but relatively modest ac- cumulations (e.g., thunderstorms). Interestingly, the
found thresholds are remarkably close to the design guidelines fol- lowed by the city of Rotterdam in the
1950-1980s when the stormwater and drainage systems were built. At that time, the system was designed
for approximately 5 mm of rain over 15 minutes and 10-12 mm of in-system storage capac- ity associated
with longer time durations. The threshold of 12.7 mm daily accumulation is likely associated with the 10-
12 mm storage. The short-duration threshold of 3.6 mm/h in 5 minutes is much lower than the design
value, which could be a sign that the local effective transport capacity is much lower than expected (e.g.,
due to blockages of sewer pipes or inlets).

35
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

Figure 1. Binary decision trees trained by using (a) rainfall intensities at nine selected temporal resolutions; (b) Maximum 5-mi-
nute (Rd5 ) and 24 hour (Rd1440 ) rain- fall depths.

Despite these encouraging results, it should be pointed out that the overall performance of model 1
remains low. Among the 443 flood events in the database, only 163 were successfully predicted, resulting in
a high miss rate of 63%. On the other hand, the non-flood days tend to be predicted very (97%). The

36
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

predictive rates for both cases are also relatively good: among the 219 predicted flood events, 163 were
actual flood days (positive predictive rate of 74%) and among the 2,380 predicted non-flood days, 2,110 (88%) were
correct. The overall error rate for both flood and non-flood days is only 13%. The second tree based on Rd5 and Rd1440
shown in Fig. 1-(b) performs very similarly, suggesting that for operational purposes, there is little added value in including
more than 2 temporal aggregation scales when making the flood predictions.

False positives and negatives

The absence of information about the origin of the flooding in citizen data clearly appears to be a
limiting factor influencing false positives and negatives. Plausible explanations for the large number of
false positives are (i) under- reporting when flooding happened in an area of lower population density, (ii)
under-reporting due to public holidays, lost records or lack of time, and (iii) the possibility that flood
reports may have been filed later, on the days following the rain event. This is hard to verify as citizen
flood reports do not come with a precise time stamp and make no distinction between the time of the
incident itself and the day on which it was reported. The most likely explanation for the large number of
false negatives is blockage in the stormwater system not caused by excess rainfall. Alternative explanations
are (i) an underestimation of rainfall by radar, and (ii) large amounts of antecedent rainfall on the days
prior to the flooding. The antecedent rainfall scenario is difficult to verify as the lack of precise time stamp
makes it impossible to determine the exact amount of rain that fell on the days leading up to the flood
event. Another differently important fact to keep in mind that the reports were made by non-experts,
which often evaluate situations than trained professionals. Lastly, the fact that models were trained on the
full datasets might also play a role, as it does not account for the growth in population and changes in
the way citizen report incidents. Indeed, a growing number of reports are now submitted through the
mobile app, resulting in approximately 1,000 additional reports per year since 2014. Reports submitted
via the app generally contain more de- tails, including the possibility to upload photos. This is an
encouraging development that hopefully will help increase the overall quality of the data in the future.

Conclusion
A ten-year dataset of 70,000 citizen flood reports for the city of Rotterdam and radar rainfall maps at 1
km, 5 minutes resolution were used to derive critical thresholds beyond which urban pluvial flooding is likely
to occur. Results show that our models correctly predict 37%-52% of all flood occurrences and 95%-97% of all
non-flood occurrences. Best performance is reached when short- and long-duration rainfall peak intensities
are combined. The study also highlighted a number of issues related to citizen flood observations which
need to be improved. The first concerns the lack of information about the type and origin of the
flooding, making it difficult to distinguish between events triggered by rain and other failure mechanisms
such as sewer blockages. The sec- ond concerns the timing between rainfall and flood reports; a more precise
time stamp would enable a better identification of rainfall conditions associated with the flood generation,
such as antecedent rain. Despite all these issues, results are highly encouraging, clearly highlighting the
possibility to infer physically plausible critical rainfall thresholds responsible for urban pluvial flooding.

Acknowledgments. This research was funded by JPI Urban Europe, ERA-NET Co-fund Smart Urban
Futures and NWO/VerDuS Smart Urban Regions of the Future (SURF). The third author acknowledges
funding through ‘Water JPI Europe’, ERA-NET Co-fund WaterWorks2014 project MUFFIN (Multi- scale
Urban Flood Forecasting: From Local Tailored Systems to a Pan-European Service).

37
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

A strategy to sustain the functionality of the existing urban drainage network

Salar Haghighatafshara, Mikael Yamanee-Nolinb, Maria Roldinc, Lars-Göran Gustafssonc, Anders Klintingd,
Henrik Aspegrena,e, Karin Jönssona
a Water and Environmental Engineering, Department of Chemical Engineering, Lund University, Lund,
Sweden
b Department of Chemical Engineering, Lund University, Lund, Sweden
c DHI Sweden, Malmö, Sweden
d DHI Denmark (Head office), Hørsholm, Denmark
e VA SYD, Malmö, Sweden

The SUrF – Project

Sustainable drainage systems (SuDS) might be placed in urban spaces based on different criteria. Space
availability, willingness of the landowner, morphologic properties of urban spaces are all among factors that
dictate the implementation point for the SuDS (Bach et al., 2013; Romnée et al., 2015). However, through a
piped-drainage perspective, the hydraulic interaction between SuDS retrofits and conventional sewer
systems is a key parameter in occurrence of pluvial floods. The effect of SuDS on the pressure relief of the
network is assumed to be a function of both the size and the location of the SuDS (Haghighatafshar et al.,
2018). This study aims to develop a methodology to efficiently size and locate SuDS at city-scale through
hydro-economic optimization.
A multi-domain hybrid hydrodynamic model consisting of hydrological, hydraulic and optimization modules
were constructed according to the logics presented in Figure 1. The model is fed with an initial catchment
characterization as well as the rainfall event of study in Figure 1-1 (Manual/automated feed). The provided
input data is employed by the hydrological module (Figure 1-2) which computes the total runoff (QTR). The
hydrological module consists of two components (domains); one for runoff simulation on the conventional
piped catchment using a standard runoff model based on runoff coefficients and time of concentration (D1)
and the other for the fast simulation of the SuDS based on the model introduced by Haghighatafshar et al.
(2019) (D2). The cumulated runoff from the hydrological module (QTR) is subsequently introduced to the
hydraulic module (Figure 1-3) which simulates the hydrodynamics of the pipe flow using the full Saint Venant
equations for dynamic flow (D3). These two modules (i.e. hydrological and hydraulic) were coupled in the
modelling software MIKE Operations. The hybrid model was facilitated with a derivative free numerical
optimization algorithm in the third module (Figure 1-4), namely constrained optimization by linear
approximation known as COBYLA (Conn et al., 1997; Powell, 1994) using a SciPy implementation.
The hybrid model was employed to optimize the required retention size and the location of the SuDS
retrofits for the cost-effective management of rainfall events. This was done by comparing the estimated
annual cost of the flooded manholes with the estimated annual cost of retrofits in various scenarios (Figure
1-4). Based on iterative simulations loops of the combined sewer network in Malmö, Sweden, conclusions
were drawn on where in the city the retrofits would lead to best effects regarding the hydraulic performance
of the combined sewer network.

38
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

The modelling methods and results from this study are included in a broader hydrological context in the
Future City Flow-project through which the long term effects on the urban water balance and the
inflow/infiltration of SuDS can be investigated and compared with other more traditional stormwater
solutions in terms of cost benefit efficiency.

Figure 1. The software architecture behind the hybrid model developed for citywide optimization of SuDS retrofits.

Acknowledgements
This study is realized through collaboration of DHI, VA SYD and Lund University and is financed by Sweden
Water Research AB, Richert Foundation at SWECO and Svenskt Vatten via VA-teknik Södra.

References
Bach, P.M., McCarthy, D.T., Urich, C., Sitzenfrei, R., Kleidorfer, M., Rauch, W., Deletic, A., 2013. A planning algorithm for quantifying
decentralised water management opportunities in urban environments. Water Sci. Technol. 68, 1857–1865.
doi:10.2166/wst.2013.437
Conn, A.R., Scheinberg, K., Toint, P., 1997. On the Convergence of Derivative-Free Methods for Unconstrained Optimization, in:
Iserles, A., Buhmann, M. (Eds.), Approximation Theory and Optimization: Tributes to M. J. D. Powell. Cambridge University Press,
Cambridge, UK, pp. 83–108.
Haghighatafshar, S., Nordlöf, B., Roldin, M., Gustafsson, L.-G., la Cour Jansen, J., Jönsson, K., 2018. Efficiency of blue-green
stormwater retrofits for flood mitigation – Conclusions drawn from a case study in Malmö, Sweden. J. Environ. Manage. 207, 60–69.
doi:10.1016/j.jenvman.2017.11.018
Haghighatafshar, S., Yamanee-Nolin, M., Larson, M., 2019. A physically based model for mesoscale SuDS – an alternative to large-
scale urban drainage simulations. J. Environ. Manage. 240, 527–536. doi:10.1016/j.jenvman.2019.03.037
Powell, M.J.D., 1994. A Direct Search Optimization Method That Models the Objective and Constraint Functions by Linear
Interpolation, in: Advances in Optimization and Numerical Analysis. Springer Netherlands, Dordrecht, pp. 51–67. doi:10.1007/978-
94-015-8330-5_4
Romnée, A., Evrard, A., Trachte, S., 2015. Methodology for a stormwater sensitive urban watershed design. J. Hydrol. 530.
doi:10.1016/j.jhydrol.2015.09.054

39
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

Which types of rainfall cause urban flooding

Johanna Sörensen
Water Resources Engineering, Lund University

The SuRF Project

Pluvial flooding is the most common type of flooding in Malmö. Only a very few flood claims have been
registered during high sea level caused by storms and there is only one minor watercourse in Malmö
(Riseberga Brook/Sege Brook), which seems not to be severely affected by riverine flooding. All of the eleven
biggest flood events during the 20-year period are caused by local or wide-spread rainfall events. These
eleven events account for about 80% of the flood claims reported to the water utility VA Syd and the
insurance company Länsförsäkringar Skåne. Three severe, pluvial flood events are presented in the study: 5
July 2007 with 150 and 169 flood claims to VA Syd and Länsförsäkringar Skåne respectively, 14 August 2010
with 210 and 148 flood claims, and 31 August 2014 with 2 109 and 2 649 flood claims. These flood events
were all caused by heavy rainfall distributed over the entire city. The 2010 and 2014 events were intense and
with a quick development, while the 2007 event was less intense, but with a long period of pre-event
rainfall. The 2014 event was heavier than a 100-year event for durations between 3 and 16 h (average for all
stations in Malmö).
There is a relation between large-scale topography and flooding in Malmö. Areas within 100 m from the
major system are more than twice as affected by flooding, compared to areas further away. During the
severe flood events in 2010 and 2014, areas close to the major system were even more affected by flooding
(3.0–4.2 times), compared to areas further away. During such downpours, runoff is quickly directed towards
low-lying areas, both through the pipe system and by overland flow. In Malmö, like probably in most other
places, the main sewers (minor system) are located under the main overland flow paths (major system), as
they follow the topography. The spatial distribution during these two, highly intensive rainfall events (2010
and 2014) were different than during other events, with more flood claims clustered around the main
sewers. For the other events, including the 2007 event, the flood claims were more evenly distributed within
the city (Figure 1).
The combined system is more exposed to flooding than the separate system. Even if only 31% of the urban
land in Malmö is connected to the combined system, 70% of the flood claims are reported from these areas.

40
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

Figure 1. Left: Flood claims reported by LF Skåne and VA Syd during two of the three severe flood events in Malmö (2010 and 2014).
Right: Flood claims reported by LF Skåne and VA Syd during all other events, except for two severe flood events (2010 and 2014).

During the 2010 and 2014 events, the combined system was 3.8–4.2 times more affected by flooding
compare to the separate system. Similar figures are found if all flood events are included. The 2007 event
shows a different pattern: the combined system areas were only slightly more than twice (2.3 times) as
severely affected by flooding during this event, compared to the areas with separate system. One reason
why the 2007 event differs from the other events might be the difference in flood causality, where
continuous rainfall during the preceding weeks saturated the ground with water. Flooding during this event
was therefore less related to type of drainage system. The dataset is biased as more people live in areas with
combined system. However, the difference in reported flood claims still exists when adjusted for this bias.
Locally, some flooding is caused by breakdown of the system, e.g. when a sewer pump stops pumping due to
system error. On the one hand, the phenomenon with local breakdowns could be seen as unique incidents
that are not likely to happen during future flood events. On the other hand, and in reality, it seems inevitable
that a few of these unique incidents happen during every flood event.

41
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

Urban Rainfall-Runoff Nowcasting with Open Data and Open Tools

Teemu Kokkonen1 and Tero Niemi2
1) Aalto University School of Engineering, Department of Built Environment, Espoo, Finland
2) Finnish Meteorological Institute, Space and Earth Observation Center, Helsinki, Finland

The MUFFIN – project

Cities and urban areas are prone to flooding due to the large extent of impervious surfaces hampering the
natural infiltration of rainfall. It is important to be prepared for the urban flood risk by developing methods
to forecast rainfall and runoff in immediate future. In this study, we explore the use of openly available data
and tools in nowcasting rainfall and runoff in urban setting.
Data and study site
The openly available weather radar data from the Finnish Meteorological Institute (FMI) was utilized for the
rainfall nowcasting. The radar images in 250 m x 250 m spatial and 5 min temporal resolution are provided
as a pre-processed radar composite for the entire country through the FMI open data portal for past 6 days
(https://en.ilmatieteenlaitos.fi/open-data). From these data, the Aug 12 2017 storm was selected for the
current study. The storm comprised a series of rapidly moving thunderstorms that bypassed the Helsinki
region in ca. four hours. The maximum hourly rainfall accumulation of 27.9 mm was measured in Kotka 115
km east of Helsinki. The storm generated close to 1 000 emergency service operations, mainly due to
damages caused by heavy winds.
Stormwater flow was assessed in an 85 ha Pakila catchment in Helsinki, which is known to be prone to
stormwater flooding. Model parameterizations utilize a 1x1 m2 digital elevation model openly available from
the city of Helsinki, land cover data freely available from the Helsinki Region Environmental Services
Authority HSY (https://kartta.hel.fi/), and drainage network data provided by HSY. The network data
represent in this study the only component that is not openly available.
Methods
Rainfall nowcasts were performed using an openly available pySTEPS software (Pulkkinen et al., 2019), which
aims to provide a community-driven effort to implement an open-source Python library for probabilistic
radar-based precipitation nowcasting (https://pysteps.github.io/). The theoretical framework of the model
is built around the Short-Term Ensemble Prediction System (STEPS) model (Bowler et al., 2006).
The rainfall nowcasts were used to feed the US EPA Storm Water Management Model (SWMM), which is a
widely used rainfall-runoff model for stormwater quantity and quality assessments (Rossman, 2015).
GisToSWMM5 tool (Niemi et al., 2019) was used to automatically divide the study area into subcatchments
and further to connect them to each other and to the outlets of the underlying drainage network.
Results
A 50 member ensemble of rainfall nowcasts was generated using pySTEPS. Figure 1 presents a nowcast 30
min ahead in time for a 250 x 250 km2 box around the capital area for one ensemble member. Table 1 lists
statistics for the Pakila catchment rainfall and stormwater flow at the catchment outlet across all 50
ensemble members. The large range in rainfall accumulations and in the resulting stormwater flows is mainly
controlled by the fact whether most intensive areas of rainfall in the advancing storm pass over the study
area.

42
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

Figure 1. A 30 minute rainfall nowcast for one ensemble member.

Table 1. Rainfall and stormwater flow statistics. Max refers to the ensemble member with highest rainfall accumulation and EnsMean
to the average value over all ensemble members.
Rainfall [mm] Flow [l/s]
Nowcast Event Accumulation Event Mean Event Max
Max 21.0 342 906
EnsMean 3.8 59 342

Summary and conclusions

Openly available precipitation and land cover data as well as openly available rainfall nowcasting and urban
hydrological models were utilized to produce rainfall nowcasts and corresponding flow nowcasts for an
urban catchment in Helsinki, Finland. Open data and tools allow easily approachable hydrological
assessments with realistic outputs without access to official forecast products produced e.g. by national
hydrometeorological services. Probabilistic nowcasts also allow for estimating the risk of stormwater
flooding in immediate future.
References
Bowler, N.E., Pierce, C.E., Seed, A.W., 2006. STEPS: A probabilistic precipitation forecasting scheme which merges an extrapolation
nowcast with downscaled NWP. Q. J. R. Meteorol. Soc. 132, 2127–2155. https://doi.org/10.1256/qj.04.100
Niemi, T.J., Kokkonen, T., Sillanpää, N., Setälä, H., Koivusalo, H., 2019. Automated Urban Rainfall–Runoff Model Generation with
Detailed Land Cover and Flow Routing. J. Hydrol. Eng. 24, 04019011. https://doi.org/10.1061/(ASCE)HE.1943-
5584.0001784
Pulkkinen, S., D. Nerini, A. Perez Hortal, C. Velasco-Forero, U. Germann, A. Seed, and L. Foresti, 2019: Pysteps: an open-source
Python library for probabilistic precipitation nowcasting (v1.0). Geosci. Model Dev. Discussion. Submitted.
Rossman, L.A., 2015. Storm Water Management Model, User’s Manual, Version 5.1 (No. EPA/600/R-05/040). U.S. Environmental
Protection Agency, Cincinnati, OH, USA.

43
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

APPETIZER
Flood and heatstress models and the need for (sub-) surface INnovations for
eXtreme Climatic EventS (INXCES)
F. Boogaard 1+2
1
Hanze University of Applied Science, Groningen, The Netherlands
2
Global Center on Adaptation, Energy Academy Europe, Groningen, The Netherlands
Email: [email protected]

The INXCES Project

Urban flooding and heatstress have become a key issue for many cities around the world. The project
'INnovations for eXtreme Climatic EventS' (INXCES) developed new innovative technological methods for risk
assessment and mitigation of extreme hydroclimatic events and optimization of urban water-dependent
ecosystem services at the catchment level. DEMs (digital elevation maps) have been used for more than a
decade now as quick scan models to indicate locations that are vulnerable to urban flooding. In the last years
the datasets are getting bigger and multidisciplinary stakeholders are becoming more demanding and
require faster and more visual results. In this paper, the development and practical use of DEMs is
exemplified by the case study of Bergen (Norway), where flood modelling using DEM is carried out in 2017
and in 2009. We can observe that the technology behind tools using DEMs is becoming more common and
improved, both with a higher accuracy and a higher resolution. Visualization tools are developed to raise
awareness and understanding among different stakeholders in Bergen and around the world. We can
conclude that the evolution of DEMS is successful in handling bigger datasets and better (3D) visualization of
results with a higher accuracy and a higher resolution. With flood maps the flow patterns of stormwater are
analysed and locations are selected to implement (sub-)surface measures as SuDS (Sustainable Urban
Drainage systems) that store and infiltrate stormwater. In the casestudy Bergen the following (sub-)surface
SuDS have been recently implemented with the insights of DEMS: settlement storage tank, rainwater
garden, swales, permeable pavement and I/T-drainage. The research results from the case study Bergen will
be shared by tools to stimulate international knowledge exchange. New improved DEMs and connected
(visualization) tools will continue to play an important role in (sub-)surface flood management and climate
resilient urban planning strategies around the world.
Next to flood modelling heatstress maps have been created within the INXCES project for Groningen and
Bucharest (figure 1c). The map illustrates points in the city that will have a high temperature and PET
(physical equivalent temperature) in the current climate and possible increase of temperature of 1 or 2 C
due to climate change. The calibration of the map and interaction with stakeholders was the subject of 25th
edition of the Wetskills Water Challenge in Romania where a Romanian-Dutch team of young water
professionals won this challenge creating the idea of a heat stressmap app for the city of Bucharest. The app
will show inhabitants and stakeholders where extreme heat can occur in cities on very hot days and where
you can take measures with a link to www.climatescan.nl. The team recorded the effects based on data from
local participants and interviews. The team designed a smart phone app that charts both problem locations
and possible solutions to stimulate climate adaptation in Bucharest. The jury assessed the winning team as
follows: “We chose a team with a market ready, feasible and flexible solution. Due to climate change,

44
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

extreme weather events are becoming more frequent. For Bucharest this has consequences such as heat
stress and flooding.

Fig. 1 Floodmodel Bergen, Norway in 2009 (left) and 2017 (right). Heatstess Bucharest Romania (right).

Acknowledgements: This study would not have been possible without funding from JPI Water funded
project INXCES.

References
Boogaard F., Kluck J, Schoof G, Bosscher M., The need for INnovations for eXtreme Climatic EventS (INXCES),
the progress of flood modeling case Bergen Norway, Procedia Engineering Volume 209, 2017, Pages 56–60,
https://doi.org/10.1016/j.proeng.2017.11.130, Elsevier 2017
https://www.dutchwatersector.com/news-events/news/25667-heat-stress-city-app-wins-romania-wetskills-
water-challenge.html

45
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

A web-based visualization prototype of urban flood forecasts from a multi-scale

hydrologic-hydrodynamic flood forecasting system in Aalborg, Denmark
Rasmus Nielsen1, Yeshewatesfa Hundecha2, Anna Johansson3, SørenThorndahl1, Jonas Olsson2
1
Department of Civil Engineering, Aalborg University, P.O. Box 159, DK-9100 Aalborg, DENMARK
2
Research and Development, Swedish Meterological and Hydrological Institute, 601 76 Norrköping, SWEDEN
3
Professional services, Swedish Meterological and Hydrological Institute, 601 76 Norrköping, SWEDEN
E-mail: [email protected]

The MUFFIN Project

Hydrological and hydrodynamic modelling are generally performed separately, which is largely
understandable; the purposes are different, the scales are different, the concepts and models are different.
However, we believe there are potential benefits in bridging this gap, exploring coupled solutions and
generally taking advantage of what the different types of modelling can offer. This issue has been one focus
in the MUFFIN project, where a case study in Aalborg, Denmark, has been conducted.

In Aalborg, an integrated system for hourly urban flood modelling/forecasting has been developed. A semi-
distributed hydrological model (HYPE), which estimates surface runoff and discharge from the upstream
rural catchment, has been coupled to a hydraulic 1D-2D urban drainage model (MIKE), including 1D sewer
river network and 2D overland flow, which describes flooding and inundation in central parts (Figure 1). The
system is designed for real-time flood forecasting, and this prospect has been evaluated by re-forecasting
several historical events characterized by a high rainfall intensity and/or high river discharge. The main
objective is to assess the added value of the coupled system, as compared with using only the MIKE model
with more generalized hydrological boundary conditions.

46
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

Figure 1. Schematic of the coupled system in Aalborg, with a meso-scale hydrological model (HYPE) providing
boundary conditions for the local-scale flood model (MIKE).

A web-based graphical interface has been developed to visually follow the evolution during the event, both
of rainfall and discharge and of any resulting inundated areas. At the demonstrations during CITIES, RAIN &
RISK, this interface will be mainly used for illustrating the performance of the coupled forecasting system,
but it also serves as a prototype for actual operational use.
Multi-basin hydrological models are already today being run operationally all over the world, and these
types of coupled forecasting systems may thus in principle be set up in any city. The temporal and spatial
resolutions of these existing operational models are, however, generally not sufficient for the type of
coupling explored in this case study, but development towards higher-resolution multi-basin hydrological
models covering national or continental domains will be performed in follow-up projects.

47
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

THEME 4: URBAN WATER MANAGEMENT; NATURE-

BASED SOLUTIONS & CLIMATE ADAPTION

Multi-purpose urban water management in a complex urban basin

Søren Thorndahl1, Rasmus Nielsen2
1
Aalborg University, Department of Civil Engineering; [email protected]
2
Aalborg University, Department of Civil Engineering; [email protected]

The MUFFIN Project

• The majority of the larger Danish cities are located close to the water - either by fjords/sea and/or by
rivers/streams/lakes. This means that many cities are in risk of flooding from both a high ground
water table, rivers and marine waters. Together with precipitation load and consequent drainage of
surface and wastewater, this makes the urban water management a challenge in the future climate.
The water current water management has multiple purposes, which is managed by different authorities,
e.g.: drainage of urban areas, treatment of wastewater, reduction of combined sewer overflow, certifying
high bathing water quality, minimizing flood risk, recreational use of water in the city, draining of agricultural
areas, certifying good quality drinking water.
This abstract presents a case study from the city of Aalborg, Denmark where all the above presented
challenges are present. Building upon the results of the MUFFIN project, a real time data and model based
monitoring system is being developed (Nielsen and Thorndahl, 2019; Thorndahl et al., 2019). This system will
consist of:
• Installation of multiple gauges for monitoring of ground water levels, stream water levels, drainage
system runoff, rainfall, etc. Gauges will both be professional and citizen gauges.
• Real-time models with assimilation of data of the urban water cycle
• A threshold warning system for different parts of the water cycle.

An online combined data and model-based surveillance system will help to improve water management
across borders of authority and reduce risk of failures such as flooding or pollution of surface waters.

An example of simulated return period of flooding in an urban catchment in Aalborg Denmark is shown in
figure 1 (Thorndahl et al., 2019).

48
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

Figure 1. Example of simulated flood return periods in an urban area of Aalborg. Flooding originates from multiple sources e.g.: point
1: surcharge from manholes, point 2: river flooding, point 3: flooding from pumping combined sewer overflows.

References
Nielsen, R.V., Thorndahl, S., 2019. Sensitivity Analysis of an Integrated Urban, Pluvial, Fluvial, and
Groundwater Flood Model. Journal of Environmental Management In Review.
Thorndahl, S., Murla-Tuyls, D., Nielsen, R.V., Schleiss, M., Olsson, J., 2019. Influence of Flood Water
Contribution from Multiple Sources in Extreme Event Statistics of Urban Flooding. Springer, pp. 397–401.
doi:10.1007/978-3-319-99867-1_67

49
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

Seasonal Hydraulic efficiency of infiltration-based SUDs (INXCES)

T.M. Muthanna
1
Norwegian University of Science and Technology, Trondheim, Norway
Email: [email protected]

The INXCES Project

Urban flooding and heatstress have become a key issue for many cities around the world. The project
'INnovations for eXtreme Climatic EventS' (INXCES) developed new innovative technological methods for risk
assessment and mitigation of extreme hydroclimatic events and optimization of urban water-dependent
ecosystem services at the catchment level. Urban flooding resulting from pluvial flooding is increasing with
climate change and driving the need for an increased focus on Sustainable Urban Drainage Systems (SUDS),
which can alleviate the already overloaded urban drainage systems. In cold and temperate climates seasonal
infiltration rates in infiltration-based SUDS is an important design consideration, together with rainfall
patterns. In addition, a changing climate changes the seasonal performance, making future performance
predictions challenging. In the INXCES project we have measured seasonal infiltration rates and models
impact for design in a future climate using case studies from Bergen and Trondheim.
In the Trondheim case two raingardens have been observed for seasonal infiltration over multiple seasons;
Åsveien, and Risvollan. At the Åsveien raingarden seasonal infiltration rates where measured regularly from
October through May in 2016/2017, based on a case study approach, combining technical (field
measurements), experimental (simulations) and qualitative approaches. For the second raingarden, Risvollan
infiltration measurements were carried out semi-regularly over the period 2016-2018. In addition, two
earlier measurement campaigns were added to the analysis, Paus et al., (2016), and Dalen (2012). This is a
research raingarden which has been in operation since 2010 and with continuous flow and meteorological
measurements. The Modified Phillip-Dunne Infiltrometers (MPD) (Ahmed et al. 2014) method was used for
measuring infiltration capacity, saturated hydraulic conductivity (Ksat) both in Bergen and Trondheim. The
MPDs were constructed of a plastic column of 50 cm length and 10 cm diameter. The columns were put 5 cm
down in the soil and filled with water. The water level and time was registered. The RECARGA model was
used for simulations (Dussaillant et al. 2005). The results show up to 70% reduction in winter infiltration
capacity, which should be accounted for in design, either by increased Ksat values in the soil media, or by a
large size raingarden.

50
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

Figure 2. Seasonal infiltration rates in the Risvollan raingarden in Trondheim measured by three different measurement campaigns
over a seven-year period from 2010-2017, sources: Paus et al., 2016; Balstad et al., 2018; and Dalen 2012).

The Bergen case study investigated the future performance of raingardens as source control measures for
retrofitting into dense urban areas. The study applied a combined spatial-temporal downscaling
methodology using the Statistical DownScaling Model-Decision Centric (SDSM-DC) and the Generalized
Extreme Value distribution to project future precipitation. This produced intensity-duration-frequency (IDF)
curves for future climate scenarios, which were used to investigate the performance of a raingarden using
the RECARGA raingarden modelling tool. The results showed that Ksat was the determining factor for peak-
flow reduction and that different values of Ksat had different benefits. Engineering flexible solutions by
combining measures holding different characteristics will induce robust adaptation. The implication for the
raingarden performance is that lag time is reduced while peak flow
reduction would be improved. Thus, a suitable combination of source control structures should compensate
for the reduced Ksat and lag time. Adding this flexibility ensures that the uncertainty in the scenarios are
accounted for and that the measure performs well under a range of scenarios, not a specific one. This,
ultimately, results in a robust adaptive measure.

51
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

Figure 3. Simulation of a raingarden performance for a set of future climate scenarios. Figure (2a) shows the performance of the
raingarden measured in % of the flow going into overflow for various Ksat values. Figure 2b shows a percentage comparison of peak
flow reduction with and without a rain garden, and figure 2c shows the effect on lag time before the overflow is activated for differ-
ent Ksat values and climate scenarios.

A second case from Trondheim investigated the placement of SUDS in the planning phase (Muthanna et al.,
2018). Again, the MPD method was deployed to investigate Ksat values for possible SUDS locations on the
NTNU campus in Trondheim. This case showcases the potential gain in considering SUDS placing early in the
planning process, which will ensure better performance of the system. The Ksat value varied greatly over an
area of 50 ha, which is typical for urban disturbed soil

Figure 4. Distribution of Ksat values over 20 sampling sites on the NTNU campus in Trondheim, Norway. The simple mean and the
geometric mean are reported together with the typical international recommended rates of 3 cm/hr and the Norwegian recommenda-
tion of 10 cm/hr, Source: Muthanna et al., 2018.

52
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

Acknowledgements: This work was funded by the Norwegian Research Council (contract #), through the JPI
Water funded project INXCES, from WaterWorks2014.

References
Ahmed, F., Nestingen, R., Nieber, J. L., Gulliver, J. S. & Hozalski, R. M. (2014). “A modified Philip-Dunne
infiltrometer for measuring the field-saturated hydraulic conductivity of surface soil”. Vadose Zone Journal 13
(10), 1–14.
Dalen, T. (2012). «Hydrologisk dimensjonering av regnbed i kaldt klima: Risvollan, Trondheim». Master thesis
NTNU (access: https://ntnuopen.ntnu.no/ntnu-xmlui/handle/11250/242219 )
Dussaillant, A. R., Cuevas, A. & Potter, K. W. (2005). “Raingardens for stormwater infiltration and focused
groundwater recharge – simulations for different world climates”. Water Science and Technology: Water
Supply 5 (3), 173–179.
Gregory, J.H., et al. (2006). Effect of urban soil compaction on infiltration rate”. Journal of Soil and Water 386
Conservation. 61(3): p. 117-124. 387
Kristvik, E.; Kleiven, G.H.; Lohne, J.; Muthanna, T.M. (2018). “Assessing the robustness of raingardens under
climate change using SDSM and temporal downscaling”. Water Sci. Technol. 2018, 77, 1640–1650.
Muthanna, T.M., Sivertsen, E., Kliewer, D., Jotta, L. (2018). “Coupling field observations and Geographical
Information System (GIS)-based analysis for improved Sustainable Urban Drainage Systems (SUDS)
performance”. Sustainability 10(12),4683
Paus, K. H., Muthanna, T. M. & Braskerud, B. C. (2016). “The hydrological performance of bioretention cells in
regions with cold climates: seasonal variations and implications for design”. Hydrology Research 47 (2), 291–
304.
Pitt, R., et al. (2008). “Compaction's impacts on urban storm-water infiltration”. Journal of Irrigation and Drainage 388
Engineering 134(5): p. 652-658.

53
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

The governing of flood risk mitigation in complex and dynamic society

Per Becker
Faculty of Engineering LTH, Lund University,

The SUrF Project

The purpose of this presentation is to share research findings of how the mitigation of flood risk is governed
in Sweden. Flood risk is a complex and transboundary challenge that cannot be managed by individuals or
organizations in isolation but requires collaborative networks of actors working together across
departments, administrative borders, levels of authority, and societal sectors. These networks are largely
based on individual choices made in particular relational and institutional settings. Previous work holds that
the emerging network-level patterns of interaction are fundamental for society’s capacity to mitigate risk,
but empirical research on the social organization of flood risk mitigation has lagged behind. Especially
concerning collaboration within a catchment area as a whole, which is the hydrologically significant unit for
analysing flood risk and the required level for addressing flood risk comprehensively by the EU Floods
Directive and Swedish legislation. Hence, the project collected unique empirical data comprising 217
interviews with all individual actors contributing actively to flood risk mitigation in a river catchment area in
Southern Sweden, complemented with documentary research of maps, legislation, policies, court rulings,
strategies, plans, and consultation minutes. The data were analysed structurally and interpretatively.
While the network of actors comprises impressive individual capacities in terms of knowledge and
experience, the findings suggest significant systemic issues that must be acknowledged and addressed for
the governing of flood risk mitigation to be able to meet the mounting challenges of the future. Most
pressingly, there is substantial fragmentation of the governing of flood risk mitigation in relation to how
water moves in the landscape. Water flows from upstream to downstream, crossing municipal borders,
while there are few links between the municipal administrations in the catchment area. The only institution
for coordination is a river council that is a voluntary association with no formal authority. There is a grave
need for improved coordination in the catchment area, which calls for increased attention of the County
Administrative Board regardless if Höje Å is currently prioritized or not. Moreover, the Swedish legal
framework governing flood risk mitigation comprises tightly coupled policies that demands coordination
between the actors implementing them within municipal administrations, but the findings suggest loose
couplings or even decoupling between particularly central actors. There is also fragmentation between the
management of urban drainage and flood risk mitigation, where the focus on the division of responsibilities
for rainfall less or more intense than the “10-year rain” is undermining pragmatic solutions. The
fragmentation of the governing of flood risk mitigation is also exacerbated by the planning process, where
the issue is analysed and addressed for each detailed planning area in isolation, regardless if these
boundaries rarely are hydrologically significant. This is further complicated by different actors generally
being involved in the planning process for adjacent planning areas, as well as different consultants procured
to support the work in different phases of the same planning process. Hence, undermining the possibility of
knowledge accretion. These systemic issues must be acknowledged and there are concrete measures to
address them.

54
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

Children’s places in stormwater spaces

Misagh Mottaghi, Sandra Kopljar, Maria Kylin, Catharina Sternudd
Faculty of Engineering LTH, Lund University

The SUrF-project

Due to the climate change, extreme rain events are expected to occur with higher frequency. At the same
time, the world urban population is increasing dramatically, and cities are getting denser which exposes
more people to the effect of urban flooding (UN, 2017). Cities focus on improving the resilience of urban
systems in order to handle natural disasters and their consequences on people and properties. Among other
approaches, above ground stormwater management has been introduced as a practice in order to improve
resilience and achieve urban sustainability. However, less focus has been put on the relationship between
urban resilience and urban sustainability. More specifically, cities have rarely reached sustainability of urban
environments, as regards to people’s health, wellbeing and liveability, through improved resilience (Pearson
et al., 2014). The gap is even more pronounced when it comes to children and their needs in urban
environments and a noticeable contemporary trend in Sweden is that places and areas traditionally
allocated for children’s play, such as schoolyards and neighbourhood parks are being questioned and used
for building purposes. Statistics Sweden (SCB, 2017) reports that schoolyards areas have shrunk with four
square meters per child in the latest three years. Cars and Strömberg (2005), point out that the driving
forces during the 1960’s and 1970’s in Swedish planning were politicians and professional planners with a
strong vision of a community based on equal social and family rights. This has slowly changed, and today’s
legislations partly leaves the responsibility for considering qualitative factors in spatial planning to the real
estate developers and constructing companies. Private interests with a focus on economic values are thus
given a bigger role, enabling them to become a driving force in community planning.
The process of landscape design is heavily affected by social and political interests. Concepts such as
“Resilience” and “Sustainable development” are constructions that are interpreted differently through
history and in different disciplines and not the least in different political contexts (Vardy & Smith 2017;
Seghezzo, 2009). The risk of “greenwashing” and that the concepts are used in different political obscure
agendas are shown in different studies (Tahvilzadeh et al, 2017). Sustainable development is often discussed
in terms of ecological, social and economic values and in many cases, it is stated that ecological and social
values can reinforce each other. The link between ecological and social values has been studied in different
ways, for instance in using the concept of ecosystem services (Chan et al., 2012; Lele et al., 2013). Egoz et al.
(2011) argue that the “right to landscape” can be used as a framework for negotiating physical rights of
individuals, communities, the powerless and the marginalized. If we identify the equitable landscape as a
design principle for landscape architecture, designing for children is of critical importance. Childhood can
partly be discussed in biological terms, but childhood is also a social construction affected by interpretations
made in a political, cultural and historical context (Holloway & Valentine, 2000). Children’s possibility to
literary “take space” in an urban context is significantly affected by political decisions in spatial planning and
processes. The right to “take space” in a landscape, urban or rural, has to do with power and social justice
and is part of an everyday experience (Setten & Brown, 2013).
The “compact city” is a predominant trend in current planning ideology and is often presented as a
sustainable way of increasing the urban population in urban areas. In this rhetoric, we often hear that the
smaller public green spaces can be optimized with good quality design so that ecological values (as in open
stormwater management) and social values (such as children’s everyday play/life) can coexist. The concept
55
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

of the compact city has however been questioned by several researchers (Jenks et al, 1996; Bramley et al
2009; Naess & Vogel, 2012; Tunström & Bradley, 2014).
The present paper explores the possibilities for open stormwater facilities to be used as children’s
playscapes and tries to gain knowledge by studying implemented facilities and their offered possibilities for
action. The study has two objectives. The first objective is to study the challenge of taking account for
various sustainability aspects at the same time. Through the example of stormwater management and
children´s play, we explore to what extent ecological and social values can reinforce each other. The second
objective is to use the case studies to initiate discussions on some of the challenges that contemporary
planning is facing. Making place for both local stormwater management and public green space in
dense/infill urbanization projects is a real challenge in the current planning paradigm.
In this project, we are examining a couple of urban housing areas through case studies. The criteria for
choosing the cases was that they had open stormwater features, planned and designed for the sites and
implemented in different scales - from a detailed scale like a pond to a neighbourhood park designed to
accommodate heavy rainfalls. These areas have been analysed with a reviewed method for assessing
environments based on affordances for children’s play (Lerstrup & Konijnendijk 2017) in order to investigate
if these areas have enhanced values for children´s everyday life and use. The study explores if specific design
and planning solutions for open stormwater management also enhance social values for children. Later the
paper discusses the results in terms of how the design could be reshaped or adjusted to better fit either
stormwater management or children’s everyday use and play. The aim is to highlight some points that can
raise awareness of the issues of diminishing green/outdoor open spaces in contemporary planning contexts
and explore how “sustainability” can be interpreted as a societal aspect of equality in an
architectural/planning context. A space that is efficient for stormwater management is not necessarily a
good space for children’s play. Later, a conclusion is drawn on how specific designs for open stormwater can
support or constrain children’s everyday use. One of the goals of the project is to encourage decision makers
to go for more effective policies and investments for stormwater green solutions in a way that benefits
children´s needs. Some discussion is also held on the challenges of using assessing methods of social values
in a technological and ecological context.

References
- Bramley, G., Dempsey, N., Power, S., Brown, C. & Watkins, D. (2009). Social sustainability and urban form:
evidence from five British cities, Environment and Planning A, vol. 41, no. 9, pp. 2125-2142.
-Cars, G., & Strömberg, K (2005). Att göra sin röst hörd. I Frank, G. (ed) Spelet om staden, Stockholm, Formas:
205-208.
-Chan, K,. Satterfield, T., Godstein., J (2012). Rethinking ecosystem services to better address and navigate
cultural values. Ecological Economics 74 (2012) 8-18.
-Egoz, S., & Makhzoumi, J., & and Pungetti, G., eds. (2011). The Right to Landscape: Contesting Landscape
and Human Rights. Farnham: Ashgate.
-Holloway, S., & Gill, V., eds. (2000). Children’s Geographies – Playing, Living, Learning. London and New
York: Routledge
-Jenks, M., Burton, E., & Williams, K. (Eds.), (1996). The Compact City: A Sustainable Urban Form? E & FN
Spon, London.
-Lele, S., Springate-Bainski, & O., Lakerveld, R., Deb, D., Dash, P., (2013) Ecosystem Services: Origins,
Contributions, Pitfalls, and Alternatives. Conservation and Society 11(4) 343-358
56
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

-Lerstrup, I,. Konijnendijk, C., (2017). Affordances of outdoor settings for children in preschool: revisiting
heft’s functional taxonomy. Landscape Research Vol. 42, No. 1, 47-62. Routledge
-Næss, P. & Vogel, N. (2012). Sustainable urban development and the multi-level transition perspective,
Environmental Innovation and Societal Transitions, vol. 4, pp. 36-50.
-Pearson, L. J., Newton, P. W., & Roberts, P. (2014). Resilient sustainable cities. [Elektronisk resurs] : a futur.
New York : Routledge, 2014. Retrieved from
http://ludwig.lub.lu.se/login?url=http://search.ebscohost.com/login.aspx?direct=true&db=cat01310a&AN=l
ovisa.005072642&site=eds-live&scope=site
-SCB (Statistiska centralbyrån). (2018). Grundskolor och friytor – Nationell kartläggning och uppföljning av
grundskoleelevers tillgång till friytor 2014-2017
https://www.boverket.se/contentassets/7f6b1586a5504fcfa11ea370f73b2b05/rapport-grundskolor-och-friytor.pdf

-Seghezzo, L., (2009). The five dimensions of sustainability. Environmental Politics, 18:4, 539-556.
https://doi.org/10.1080/09644010903063669

-Setten, G. & Myrvang Brown, K., (2013). Landscape and Social Justice. In P. Howard, I. Thompson, and E.
Waterton, (eds) Routledge Companion to Landscape Studies. New York: Routledge.
-Tahvilzadeh, N., Montin, S., & Cullberg, M. (2017). Functions of sustainability: exploring what urban
sustainability policy discourse “does” in the Gothenburg Metropolitan Area. The International Journal of
Justice and Sustainability. https://www.mistraurbanfutures.org/en/publication/functions-sustainability-exploring-what-urban-
sustainability-policy-discourse-does

-Thunström, M. & Bradley, K., (2014). Opposing the postpolitical Swedish urban discourse in Metzger, J.,
Allmendinger, P. & Oosterlynck, S. (eds) Planning against the political: Democratic deficits in European
territorial governance. New York: Routledge
-UN (United Nations) (2017). New Urban Agenda, Habitat III.
-Vardy, M., & Smith, S., (2017). Resilience. Environmental Humanities 9:1 May
2017https://read.dukeupress.edu/environmental-humanities/article-pdf/9/1/175/409168/175Vardy.pdf
-Vierikko, K., & Neimelä, J., (2016). Bottom-up thinking – Identifying socio-cultural values of
ecosystemservices in local blue-green infrastructure planning in Helsinki, Finland. Land Use Policy 50 (2016)
537-547 http://dx.doi.org/10.1016/j.landusepol.2015.09.031

57
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

Characterisation of dissolved metal fractions in urban runoff

S. Lindfors1, H. Österlund1, T. Meyn2, T.M. Muthanna2, L. Lundy1+3, M. Viklander1

1
Luleå University of Technology, Luleå, Sweden. Email: [email protected]
2
Norwegian University of Science and Technology, Trondheim, Norway
3
Middlesex University, the Burroughs, Hendon, London

The INXCES Project

Understanding the bio-physico-chemical behaviour of pollutants within the urban environment is key to
predicting their impact and identifying appropriate best management practices. The project 'INnovations for
eXtreme Climatic EventS' (INXCES) developed new knowledge on the behaviour of metals within the
dissolved fraction from different urban areas for use in risk assessment and their mitigation. Previous
research identified that several metals (e.g. Cu and Zn) occur in the dissolved fraction as free ions or in
association with light and organic colloids in urban runoff (Grout et al., 1999; McKenzie and Young, 2013).
However, the distribution of metals within these sub-dissolved fractions is not well understood, particularly
in terms of their labile behaviour. In contrast to the majority of EQS identified in the EQS Directive (2008),
the EQS for metals were established for dissolved concentrations, on the basis that only metals in the
dissolved fraction are available for active uptake. Following further research (e.g. Gandhi et al., 2010),
selected EQS were revised as part of the Priority Substances Directive (2013) which promotes the use of
bioavailability modelling to take into account the influence of dissolved organic carbon (DOC), pH and water
hardness to better estimate the fraction of the dissolved concentration which is bioavailable in practice.
Hence, there is an urgent need to increase knowledge on the occurrence and behaviour of metals in these
sub-dissolved fractions. Within the frame of the multi-disciplinary INXCES project, two sampling campaigns
have been conducted to evaluate the dissolved concentrations of selected metals (Cd, Cr, Cu, Ni, Pb, Zn) in
urban runoff in terms of identification of the bioavailable fraction, the colloidal content (and to what extent
metals in the colloidal fraction are labile) and to examine the relationship between these sub-dissolved
concentrations in terms of implications for urban runoff mitigation.
Runoff was sampled at two industrial parks (IA1: 15 ha / 75% impervious and IA2: 12 ha / 85% impervious)
and a car park (PL: 0.45 ha / 100% impervious) in Umeå, northern Sweden, and a section of highway (E18:
100% impervious) outside Enköping, southern Sweden. Highway runoff was collected from a 100-metre
stretch of the European Route 18, at “Testsite E18” managed by the Swedish Transport Administration. All
samples were analysed for the dissolved, colloidal and truly dissolved metal fractions using filtration (0.45µm
filter) and ultrafiltration (to determine the 3000 molecular weight cut off; 3kMWCO) together with basic
parameters (Ca, pH, DOC). In the first sampling campaign, samples from PL and IA1 were analysed using the
Bio-met bioavailability prediction tool to identify bioavailable metal concentrations. For the second sampling
campaign, samples from IA1, IA2 and E18 were compared with diffusive gradients in thin films (DGT)
measurements and the analysis was complemented with asymmetrical flow field-flow fractionation (AFFF)
measurements.

Results from fractionation using filtration and ultrafiltration show that the truly dissolved (<3 kMWCO)
concentrations of Cd, Cr and Ni are generally greater than those associated with the colloidal fractions (3
kMWCO-0.45µm) whereas truly dissolved Cu and Zn concentrations consistently dominates over colloidal Cu
and Zn concentrations. Truly dissolved concentrations of Pb is generally less than colloidal Pb
58
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

concentrations, Figure 1. Preliminary results from AFFF indicate that colloids occur in two dominant size
fractions: ranging from one to a few nm and around 100 to a few hundred nm (exact size determination was
not possible as the light scattering detector was unavailable). Low concentrations of DOC were reported in
both fractions, indicating colloids occurred in an inorganic form. Cd, Cu and Zn were not detected in
concentrations above the blank for these colloidal fractions. In contrast, dissolved Pb, and Cr predominantly
occurred in the larger colloidal fraction (around a few hundred nm) with Ni equally distributed between both
colloidal fractions at IA1 and IA2. Parallel measurements with passive samplers showed that labile
concentrations were in the same range as dissolved (<0.45 µm) concentrations, indicating that all sub-
dissolved fractions are labile for all metals at both sites.

Median concentrations of the truly dissolved and bioavailable fractions (as predicted using Bio-met) are
presented in Table 1. With the exception of Pb, the bioavailable fraction does not equate to the truly
dissolved fraction in any samples during either event type at either site. The percentage of the truly
dissolved fraction which consists of bioavailable metals varies on a metal by metal basis. For example, the
bioavailable fraction contributes >20% of the truly dissolved fraction of Zn in 89% of rainfall runoff and 90%
of snowmelt samples, of Ni in 75% of rainfall and 80% of snowmelt samples and of Cu in 16% of rainfall
samples (Cu was not reported in a bioavailable form in any snowmelt samples).

Figure 1. Metal concentrations in the analysed fractions total, dissolved (<0.45 µm) and truly dissolved (<3kMWCO) at site IA2 during
a rain event in October 2018. The <0.45 µm and <3kMWCO fractions of Cr and Pb are recorded in low concentrations relative total
concentrations, thus does not show in chart.

The remaining metals within the truly dissolved concentration are associated with small colloids in the <3
kMWCO fraction. This would indicate that a higher level of complexation occurs for Cu compared to Ni and
Zn. As noted earlier, the amount of Pb determined in the truly dissolved phase is considerably lower than for
the other metals (note: several samples Pb was below the limit of detection). Data presented in Table 1
indicates that any Pb that does occur in the truly dissolved phase is primarily bioavailable. However, it should
be noted that median bioavailable concentrations predicted are approximately two orders of magnitudes
below the Pb EQSbioavailable (1.2 µg/l).

59
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

Table 1. Cu, Zn, Ni and Pb site median (range) truly dissolved concentrations (<3 kMWCO) and bioavailable concentrations (Bio-met)
in runoff samples collected during rainfall and snowmelt events.

Industrial Car park

Rainfall Snowmelt Rainfall Snowmelt
Metal Fraction (µg/l) (n = 25) (n = 14) (n = 31) (n = 6)
Truly dissolved 9.1 (1.5-27.3) 14.3 (8.7-32.7) 3.9 (0.6-31.9) 4.9 (3.1-15.1)
Cu Bioavailable 0.8 (0.2-2.0) 0.7 (0.3-1.2) 0.4 (0.1-0.9) 0.2 (0.1-0.2)
Truly dissolved 1.9 (0.4-10.2) 2.1 (1.7-4.6) 0.6 (0.1-6.2) 2.0 (1.6-2.6)
Ni Bioavailable 0.5 (0.2-1.7) 0.6 (0.3-0.7) 0.3 (0.1-0.6) 0.6 (0.5-0.8)
Truly dissolved 75.3 (18.5-271) 84.3 (35.8-156) 13.6 (4.7-262) 2.6 (1.5-5.9)
Zn Bioavailable 38.2 (9.7-103) 28.5 (12.9-47.8) 8.1 (0.6-53.0) 1.0 (0.5-1.7)
Truly dissolved 0.03 (0.01-0.10) 0.04 (0.01-0.13) 0.02 (0.01-0.08) Not detected
Pb* Bioavailable 0.02 (0.00-0.06) 0.01 (0.01-0.01) 0.01 (0.01-0.02) Not detected
Key: *=n above detection limit for Pb: n=13 n=5 n=13

In conclusion, the data indicate that the truly dissolved concentrations dominate the sub-dissolved fractions
for Cu and Zn and contributes up to 30% and 50% of total Cu and Zn concentrations, respectively. Together
with the findings of two differently sized inorganic colloidal fractions (detected for Pb, Cr and Ni), this data
suggests that treatment systems should be designed to target removal of the sub-dissolved fractions,
especially due to the identified lability of these fractions suggesting runoff will have environmental impact.
Further, the truly dissolved concentrations are identified to be an overestimation of the bioavailable
concentrations and should therefore not be used as a direct surrogate.

This study was made possible with funding from the JPI Water funded project INXCES.

References
Gandhi, N., Diamond, M.L., Van de Meent, D., Huijbregts, M.A.J., Peijnenburg, W.J.G. and Guinée, J. (2010). New method for
calculating comparative toxicity potential of cationic metals in freshwater: Application to copper, nickel, and zinc. Env. Sci. & Tech.,
44, 5195-5201. Grout, H., Wiesner, M.R., Bottero, J.Y., 1999. Analysis of colloidal phases in urban stormwater
runoff. Environ. Sci. Technol. 33, 831–839. European commission (EC). (2008). Directive 2008/105/EC Environmental quality
standards in the field of water policy. OJ, L 348, 84-97. European Parliament. (2013). Directive 2013/39/EU priority substances in the
field of water policy. OJ, L 226, 1-17. McKenzie, E.R., Young, T.M., 2013. A novel fractionation approach for water
constituents-distribution of storm event metals. Environ. Sci. Process. Impacts 15, 1006–1016.

60
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

Vadose zone hydraulic assessment in urban areas with small scale variability
Traian Ghibus, Dragos Gaitanaru, Constantin Radu Gogu
Technical University of Civil Engineering, Bucharest
Bd. Lacul Tei, 124, 020369 – Romania
Phone: + 40 21 796 41 23
Email: [email protected]; [email protected]; [email protected]

The INXCES Project

In the urban areas, a correct determination of saturated hydraulic conductivity ks for the vadose zone has
multiple applications. Due to lithological heterogeneity of the urban soil, significant variations of this
parameter may occur in relatively close locations. To model the water infiltration from the terrain surface to
the shallow aquifer, the ks value it regularly used. In order to obtain solutions of the flow simulation model,
showing an accurate physical sense, it is necessary to take into account the complexity of this zone. As
consequence its heterogeneity and anisotropy conduct to mathematical models difficult to manage.
Using test devices with large infiltration area, the obtained ks values better characterizes the test zone due
to the variability of the lithological composition, cracks, vegetation, computing method and others. From the
theoretical point of view on natural and homogenous soils, the ratio between ks value and the infiltration
area A, have to decrease when A increases and indicates an inverse correlation between these parameters.
When this ratio increases, the conditional dispersion decreases and the residual dispersion increases. It is
also possible that other factors may disguise the influence of A on obtained ks values. In INXCES Project we
performed an analysis of obtained ks values on two study areas PL1 and PL2, by applying four methods for in
situ tests: Double ring infiltrometer DRI, Tube infiltrometer TI, Minidisk infiltrometer MDI and Inverse auger
method IA. The study areas are located on urban soils: PL1 on heterogeneous filling material (sandy clay,
silty clay with elements of gravel brick, glass, wood) with a light vegetation coverage and PL2 showing a
lower urban soil heterogeneity degree in comparison to PL1 (sandy silty clay with rare millimetre or smaller
elements of brick and concrete) but covered with a denser vegetation.

Fig. 1 – Study case PL1 Fig. 2 - Study case PL2

The analysis has been made by comparing the results obtained by each method, between several tests
carried out at the same location as well as at distinct locations within the same study area. The results
gathered from both study areas have been further compared. A statistical analysis indicates a direct and

61
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

good correlation between the hydraulic conductivity and the infiltration surface for PL1. Also, an inverse
correlation between the obtained ks values and ks/A ratio has been obtained. For PL2 it has been obtained a
moderate direct correlation between ks and A and no correlation between ks/A ratio and A. Also, a higher
value of 80% for the residual dispersion has been shown. A second analysis has been carried out for PL2
study area by applying the multiple regression method to understand to which degree the vegetation is
masking the direct connection between the two variables and to determine the proportion in which other
variables (that could not be considered by this analysis) influence the ks fluctuations.

62
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

ClimateCafe for interdisciplinary active Knowledge exchange on Climate

Adaptation: 25th edition Malmo Climatecafe
Floris Boogaard 1+2, Guri Venvik3, Allard Roest1, Antal Zuurman4, Ana C. Cassanti1, Rui Lima5
1
Hanze University of Applied Science, Groningen, The Netherlands
2
Global Center on Adaptation, Energy Academy Europe, Groningen, The Netherlands
3
Geological Survey of Norway (NGU), Trondheim
4
Municipality of Nijmegen, the Netherlands
5
INDYMO (INnovative Dynamic Monitoring), Email: [email protected]

Email: [email protected]

The INXCES Project

Our urban and rural environment (rivers, cities) are becoming increasingly vulnerable for climate change and
there is an urgent need to become more resilient. Climatecafe is developed by Groningen and Rotterdam
University of Applied Sciences to gather factual and objective data in a short period of time (1-2 weeks) by
young professionals and practitioners that enable them to assess the ‘level of resilience’ of a specific area.
The first step in climate adaptation strategy ‘Analyse, ambition, act’ is identification of bottlenecks in areas
such as flooding, drought, and heat. International exploratory city scans are set up in order to put the
necessary steps in place to go from ‘analyse’ to ‘action’. During a scan, an international ‘Quadruple Helix’
team (government, industry, academia, civil participants) work together towards tangible results. Bringing
these multidisciplinary viewpoints together in an environment that promotes team working, collaboration
and the sharing of ideas. By working together, this quadruple helix approach can create new shared value
that benefits all participants. Technology (such as interactive tools and quickscan modelling) plays a key role
in creating networks and connectivity. The ability to gather relevant information and measurement data in a
short time frame about the level of resilience at street or urban area level makes the climate scan unique.
This knowledge may help citizens and other stakeholders in urban areas or neighbourhoods to develop
support for, and implement, climate adaption measures. The participating parties work together to select a
variety of challenges.
Climatecafe aims to use low-cost and low-tech tools and instruments. Parameters that are assessed in rivers
cans and city scans: urban heat (temperature), urban floods (infiltration capacity), urban water quality
(several parameters as: nutrients, chlorophyll, oxygen), air quality (several parameters) and waste pollution
(plastic waste). Storytelling is recently added as a successful method to collect subjective data such as the
opinion and perspective of the community and integrate these data in the climate scan results. The
climatecafe method was tested in different areas around the globe (Asia, Europe, Africa) in groups of young
professionals and stakeholders in rapid urban appraisals.
Climatecafe Malmo is an interactive multidisciplinary fieldwork based event with practical examples of
climate adaptation. Climatecafe Malmo is an effort for bridging the gap between research, applicability and
societal effects. This 25th edition of Climatecafe will raise awareness and show how research relate to UN
SDGs Sustainable Development Goals, exchanging knowledge with: >50 young professionals, >15
nationalities, >15 disciplines. Climatecafe Malmo will measure, map, scan and assess different parameters
that give insight in the vulnerability and adaptation of a defined urban area of Malmo (Augustenborg). The
following actions are planned:

63
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

• measure, map and scan sustainable urban drainage systems (with the tool www.climatescan.nl)
• demonstration and participation in full-scale infiltration tests
• mapping of pollutants in SuDS
• measure water quality with your phone
• heat stress mapping
Climatecafe Malmo is organized before and during the end congress of 3 projects (14: INnovation for
eXtreme Climatic EventS (INXCES), MULTI-SCALE URBAN FLOOd FORECASTING (MUFFIN), SUrF (Sustainable
Urban Flood Management). Climatecafe Malmo will give you: International networking with 9 European
Universities and multiple disciplines.
In conclusion, there is a clear demand for an active collaborative knowledge-sharing activity as climatecafe.
This event helps raising awareness and capacity building on climate adaptation and will help policy makers
and practitioners to gather valuable data for decision-makers in a rapid appraisal at neighbourhood and city
level. The results of Malmo Climatecafe will be presented at 13 and 14th June 2019 by young professionals.

Fig. 1 Climatecafe Malmo (source: https://climatecafe.nl/2019/01/city-climatescan-malmo-will-be-held-10-14-june-2019)

Acknowledgements: This study would not have been possible without funding JPI Water funded project
INXCES and Interreg IVb proect WaterCoG (water-Co-Governance).

64
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

POSTERS

In situ mapping of pollutants in Sustainable Urban Drainage Systems, a new

methodology approach and preliminary results from the Netherlands
Guri Venvik1, Floris Boogaard2+3, Allard Roest2
1
Geological Survey of Norway (NGU), Trondheim, [email protected]
2
Hanze University of Applied Science, Groningen, The Netherlands
3
Global Center on Adaptation, Energy Academy Europe, Groningen, The Netherlands

The INXCES Project

Keywords: in situ mapping, XRF, pollutants, heavy metals, SuDS,

Stormwater runoff has severe negative and direct impact on the quality of surface waters and groundwater.
The impact can cause chemical and heavy-metal pollution. Applying well established methods to map
pollutants in urban areas and specifically in Nature-Based Solutions (NBS), such as Sustainable Urban
Drainage Systems (SuDS) is a step towards improving the water quality in the urban water cycle.
Traditional mapping of pollutants by the means of soil samples is costly, which is the main reason why the
environmental-technical functioning of rainwater facilities has not been investigated on a large scale and
systematically. X-ray fluorescence (XRF) is a known analysing method for finding metals and other
components, for laboratory analysis and portable instruments. In this work we propose a new approach of
mapping method for pollutants in-situ, such as heavy metals in soil in SuDS, with case studies from the
Netherlands where swales were implemented 20 years ago. In situ XRF measurements is a quick and cost-
efficient analysis for heavy meatal mapping in the respect to contaminated soil.
In situ XRF measures of various elements, including heavy metals is carried out in a quickscan and accurate
manner and measures both qualitatively and quantitatively. It makes the time-consuming and costly interim
analyses by laboratories superfluous. In this study, we suggest a new methodology approach for in situ
mapping of pollutants in various swales that were implemented from 20 to 5 years ago. The results differ
due to multiple factors (age, use of materials, storage volume, maintenance, run off quality, etc.). Several
locations reached unacceptable levels, above the national thresholds for pollutants. The spatial distribution
of pollutants in the over 30 swales mapped in the Netherlands show that the preferred water flow in the
SuDS controls the spreading of pollutants. The swales investigated are presented in an interactive way with
the open source tool www.climatescan.nl, containing more than 100 swales, part of which has been
investigated with in situ XRF measurements.
The research results are of great importance for all stakeholders in (inter)national cities that are involved in
climate adaptation. SuDS is the most widely used method for storing stormwater and infiltrating in the
Netherlands. However, there is still too little knowledge about the long-term functioning of the soil of these
facilities.

65
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

66
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

In-situ hyperspectral and fluorescence methods compared with remote sensing

Sentinel-2 satellite data for mapping chlorophyll-a/cyanobacteria concentrations
R. Lima1, F. Boogaard 2+3, N. Van der Vliet3, L.M. Dionisio Pires, A. Hommersom.

1
INDYMO (INnovative Dynamic Monitoring), Email: [email protected]
2
Hanze University of Applied Science, Groningen, The Netherlands
3
Tauw Group, The Netherlands
4
Delateres, The Netherlands

The INXCES Project

Climate change accentuates heat waves in The Netherlands. This, in combination with high nutrient
concentrations in Dutch surface waters, is intensifying the proliferation of toxic blue-green algae
(cyanobacteria), a threat for the health of the population and ecosystems. Lakes are resources for
recreational activities, such as bathing, sailing and fishing. Dutch water managers struggle with lack of high
spatial resolution information and updated/real-time alarm systems. Algal concentration data and its spatial
distribution is important to support and calibrate (early warning) forecasting models.
This study focuses on collecting and comparing simultaneous in-situ and remote sensing data of chlorophyll-a and
phycocyanin, and assessing the suitability of each method to characterize spatial (and with depth) variations of algae
concentrations. The in-situ measurements were performed using a portable water quality spectrometer (WISP-3) and a
hand-held fluorescence algae sensor (TriLux). Remote-sensing data was derived from Sentinel-2 multi-spectral earth
observation satellite using existing algorithms (WISP-3 measurements were used to force atmospheric corrections). The
data collection campaigns took place during overpasses of the satellite and when days were cloud-free, and between 10am
and 15pm.The spatial variations of algal concentrations were investigated using autonomous unmanned vehicles equipped
with the algae sensor to perform depth profiles and multiple transects along the canals and the lake. The study sites for this
research were Lake Paterswoldsemeer and its main tributary channels. This region is managed by the Water Authority
Noorderzijlvest that was also involved in this monitoring campaign.
The depth profiles measured with the algae sensor showed some variation, with the algal concentrations
increasing with depth to a maximum value a few centimeters from the surface, and then decreasing to a
stable concentration at higher water depths. Sentinel-2 data showed variations in the cross sections of the
channels that were not detected by the in-situ methods. The deviating values near the margins indicate that
surrounding land and trees/buildings influence pixel values. Finally, considering the different results
obtained with different methods, additional laboratory validation measurements should be conducted for
verification in future campaigns.
The collected data is now being incorporated in the calibration of algae forecasting models that are currently
being developed and applied in this region and up-scaled to many projects in the world. Output from these
models provides crucial information for decision-making on adequate measures and actions to address the
problem of harmful algal blooms.

Acknowledgements: Funding: European Union’s Horizon 2020 research and innovation programme
(EOMORES, n°730066), Noorderzijlvest, and Interreg WaterCoG (water-Co-Governance) and JPI Water
funded project INXCES

67
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

68
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

SPONSORS FOR THE CITIES, RAIN AND RISK

CONFERENCE

69
ABSTRACT BOOKLET for the conference CITIES, RAIN & RISK 13-14 June 2019

13th-14th June 2019 in Malmö, Sweden

70