See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/223945414

SOLS: A lake database to monitor in the Near Real Time water level and
storage variations from remote sensing data

Article in Advances in Space Research · May 2011

DOI: 10.1016/j.asr.2011.01.004

CITATIONS READS

717 972

11 authors, including:

J. Cretaux Adalbert Arsen

Université Toulouse III - Paul Sabatier CalVal
212 PUBLICATIONS 10,795 CITATIONS 13 PUBLICATIONS 1,867 CITATIONS

SEE PROFILE SEE PROFILE

Stéphane Calmant Alexei Kouraev

Institute of Research for Development Université Toulouse III - Paul Sabatier
312 PUBLICATIONS 10,474 CITATIONS 102 PUBLICATIONS 3,460 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Rodrigo Abarca-del-Rio on 24 May 2022.

The user has requested enhancement of the downloaded file.

This article appeared in a journal published by Elsevier. The attached
copy is furnished to the author for internal non-commercial research
and education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information
regarding Elsevier’s archiving and manuscript policies are
encouraged to visit:
http://www.elsevier.com/copyright
 Author's personal copy

Available online at www.sciencedirect.com

Advances in Space Research 47 (2011) 1497–1507

www.elsevier.com/locate/asr

SOLS: A lake database to monitor in the Near Real Time water

level and storage variations from remote sensing data
J.-F. Crétaux a,⇑, W. Jelinski a, S. Calmant b, A. Kouraev c, V. Vuglinski d, M. Bergé-Nguyen a,
M.-C. Gennero a, F. Nino b, R. Abarca Del Rio e, A. Cazenave a, P. Maisongrande a
a
CNES/LEGOS, 14 av. Edouard Belin, 31400 Toulouse, France
b
IRD/LEGOS, 14 av. Edouard Belin, 31400 Toulouse, France
c
Universite de Toulouse, UPS (OMP-PCA), LEGOS, 14 av. Edouard Belin, F-31400 Toulouse, France
d
State Hydrological Institute of Roshydromet 23, Second Line, SU-199053 St. Petersburg, Russia
e
DEFAO, Universidad de Concepcion, Casilla 160-C, Conception, Chile

Received 13 July 2009; received in revised form 25 June 2010; accepted 2 January 2011
Available online 11 January 2011

Abstract

An accurate and continuous monitoring of lakes and inland seas is available since 1993 thanks to the satellite altimetry missions
(Topex–Poseidon, GFO, ERS-2, Jason-1, Jason-2 and Envisat). Global data processing of these satellites provides temporal and spatial
time series of lakes surface height with a decimetre precision on the whole Earth. The response of water level to regional hydrology is
particularly marked for lakes and inland seas in semi-arid regions. A lake data centre is under development at by LEGOS (Laboratoire
d’Etude en Géophysique et Océanographie Spatiale) in Toulouse, in coordination with the HYDROLARE project (Headed by SHI:
State Hydrological Institute of the Russian Academy of Science). It already provides level variations for about 150 lakes and reservoirs,
freely available on the web site (HYDROWEB: http://www.LEGOS.obs-mip.fr/soa/hydrologie/HYDROWEB), and surface-volume
variations of about 50 big lakes are also calculated through a combination of various satellite images (Modis, Asar, Landsat, Cbers)
and radar altimetry. The final objective is to achieve in 2011 a fully operating data centre based on remote sensing technique and con-
trolled by the in situ infrastructure for the Global Terrestrial Network for Lakes (GTN-L) under the supervision of WMO (World Mete-
orological Organization) and GCOS (Global Climate Observing System).
Ó 2011 COSPAR. Published by Elsevier Ltd. All rights reserved.

Keywords: Lake; Database; Radar altimetry; Satellite imagery; Essential climate variable

1. Introduction with no outflow. In any cases they are very sensitive to cli-
mate change.
The knowledge of the lake water storage variations In some regions highly ephemeral lakes provide infor-
in time over a long period is fundamental for understand- mation on the events like severe drought or inundation.
ing the impact of climate change and human activities on Closed basin lakes are sensitive to changes in regional
the terrestrial water resources. Variations in air tempera- water balance. Therefore a Near Real Time (NRT1)
ture and precipitation have impact on the water balance monitoring of this type of lakes is essential in the frame
of a lake, and in an extreme case a lake can entirely of such extreme events. For small lakes of this class, the
disappear. There are two main types of lakes: open lakes sensitivity to a small change in inputs is proportionally
with outflow draining rivers, and in opposite, closed lakes

1
The NRT mode corresponds to a delay of the products delivery, in
⇑ Corresponding author. SOLS, of not more than one week when the satellite altimetry measure-
E-mail address: [email protected] (J.-F. Crétaux). ments were collected.

0273-1177/$36.00 Ó 2011 COSPAR. Published by Elsevier Ltd. All rights reserved.

doi:10.1016/j.asr.2011.01.004
 Author's personal copy

1498 J.-F. Crétaux et al. / Advances in Space Research 47 (2011) 1497–1507

higher, principally in terms of their level variations. In a coastal areas, damage in transportation system, tourism,
given region covered with several lakes, if the records of recreational, fishery and hydroelectricity industries.
their level variations are long enough, they could reveal Finally, large open lakes can themselves affect local climate
recurrence of trends in a very reliable and accurate manner. through evaporation and albedo effects.
In this way lakes can be considered as an excellent proxy Hence it is essential to build up global database on lakes
for climate change. and reservoirs worldwide providing information about
For example the Andean chain in South America is cov- level, surface and volume variations, ice cover duration
ered with hundreds of lakes. They are located in region, and ice cover thickness, as well as surface water tempera-
which is under the influence of several climatic forcing: ture variations. The availability of these datasets in NRT
Southern Atlantic Oscillation, Pacific Decadal Oscillation, mode would also be beneficial worldwide, as it would
El Nino, Glacier melting, etc. (Garreaud and Battisti, increase the quantity and quality of information on water
1999; Garreaud and Aceituno, 2001; Zola and Bengtsson, management, regional network enhancement, model assim-
2006). From North to South, over a distance of more than ilation, climate change’s impact on the lake system and sur-
4000 km, lakes are distributed in the Semi Arid plateau rounding human activities, etc.
region (Peru-Bolivia) which receives extremely small rain- Yet, the existing gauge networks on a regional and a
fall and suffers from intensive evaporation, and in the bor- global scale are declining and observations are often non-
eal region (Patagonia) where inter-annual fluctuations of existent, particularly in developing countries. Long term
precipitation are high, and water release from glacier thaw- change and inter-annual variability of water storage are
ing has increased last century. Studying of these lakes in a also often lacking. When existing, data are neither harmo-
continental framework would be therefore very useful for a nised nor expressed in the same reference frame. Remote
better understanding of climate change impact on surface sensing hence could offer an opportunity to improve this
water resources, in particular for the big Andean cities situation.
fed in summer by the melt of water outpouring from the The main purpose of this article is to present the poten-
glaciers stocking the winter snowfalls and rainfalls. As tiality of remote sensing techniques, to fulfil the require-
illustration, for South America, the IPCC (International ment of the lake data centre development in addition to
Panel for Climate Change) scenarios predict an increase the compilation of the in situ historical and current data.
of temperature going from 1 to 6 °C, and an increase of For this an initiative was taken in 2009 to setup such a data
precipitation anomalies by about 20% before the end of centre under the GCOS auspices with two main compo-
the XXIth century (Bates et al., 2008). But the geographical nent, one, the HYDROLARE centre hosted by SHI, and
distribution of these anomalies is not homogeneous: Most two, a web portal under construction, SOLS (Service
GCM (Global Climate Model) projections indicate larger d’Observation des Lacs par Satellite: French acronym)
(positive or negative) rainfall anomalies for the tropical hosted by LEGOS. Section 2 presents the different tech-
region, and smaller for the extra-tropical part of South niques that allow producing the ECVs (Essential Climate
America. This will have an impact on the runoff of the riv- Variables) for lakes. Section 3 discusses the international
ers that feed the South American lakes. institutional context and the SOLS component of the
Another region with dense lakes coverage is the East HYDROLARE data centre. Section 4 shows the prototype
African rift. Understanding of the climate variability in of the future web portal of SOLS. Conclusions are given in
Africa, and therefore the tropical variability is one of the Section 5.
key goals for the climate research. Also African lakes them-
selves modify greatly local climate (Nicholson and Yin, 2. Remote sensing techniques and lakes survey
2002). Level records analysis of the Great East African
Lakes such as Victoria, Tanganyika, and Nyassa-Malawi, 2.1. Satellite imagery as a tool for lakes surface monitoring
and the regional climatic variability show synchronicity,
which can be explained by a large-scale linkage (Ropelewski Satellite imagery is basically used to classify the Earth
and Halpert, 1996). Several studies have investigated the surface for several purposes: land use, ecology, sustainable
effect of ENSO (Mistry and Conway, 2003) but also of the development, agriculture, water management, farming
Indian Ocean Dipole (Richard, 1994), which is partly development, urbanisation, fire monitoring, etc. Observa-
responsible for driving climate variability over eastern tions of the solar reflected energy at different frequencies,
Africa (Marchant et al., 2007). usually from visible to IR, allow discrimination between
Global warming impact on lakes has ecological conse- different types of land use. Several satellites carry instru-
quences that are not precisely predictable. For example, ments providing data that can be thus classified, with var-
warming of the lake Tanganyka has lead to a reduction in ious spatial resolutions: usually from very high resolution
photosynthesis productivity, following by drastic decrease of few decimetres for military applications, to kilometre
in fish population with obvious societal consequences. In resolution for global monitoring of the Earth.
another region, the general reduction of water level of Recent scientific literature is full of articles describing
the Great Lakes of North America has several societal methodologies using remote sensing (optical and radar)
and environmental consequences: pollution on the rise in for water extent monitoring, mainly with purpose of
 Author's personal copy

J.-F. Crétaux et al. / Advances in Space Research 47 (2011) 1497–1507 1499

studying floodplains and wetlands (Töyrä et al., 2001; Owing to the efforts of the scientific community, altimetry
Töyrä and Pietroniro, 2005; Frappart et al., 2005a,b; for continental water bodies has grown up and acquired a
Henry et al., 2003; Peng et al., 2005; Sakamoto et al., 2007). relative autonomy with respect to the altimetry for oceanog-
Recently NASA has put on the Internet a complete set of raphy. Indeed, altimetry for continental water bodies has
high resolution landsat imagery covering the Earth over the specific requirements, different from those for the oceanic
last 20 years: http://edcsns17.cr.usgs.gov/EarthExplorer. domains. These differences include range estimates by algo-
The MODIS full database is also freely available rithms tuned for echoes different from the ocean paradigm
on the web portal of NASA with regular up-to-date (Frappart et al., 2006), altitude-dependent dry tropospheric
releases: http://edcimswww.cd.usgs.gov/pub/imswelcome/. corrections wet tropospheric corrections from global meteo-
The Brasilian space Agency (INPE) has done the same rological models to replace the radiometer deficiency in the
for South America with images from Landsat and land environment (Cretaux et al., 2009), and use of high fre-
CBERS-2 satellites, with spatial resolution of few metres: quency (10 Hz, 18 Hz or 20 Hz depending on the mission)
http://www.inpe.br/ingles/index.php. As for the European sample values and coordinates instead of 1 Hz averages.
Space Agency it provides access to a database of the ASAR Over the past 5 years the LEGOS has developed a web
imagery from Envisat satellite: http://www.eopi.esa.int/ database (HYDROWEB: http://www.LEGOS.obs-mip.fr/
esa/esa. soa/hydrologie/HYDROWEB) containing time series over
In the range of decametres to hectometres these data can water levels of large rivers, lakes and wetlands on a global
be successfully used to detect water presence on the Earth, scale.2 Due to technical problems on the satellite, GEOSAT
and if series of imagery over a specific lake are collected data are not suitable to estimate water height over conti-
over a long period of time one can estimate the water sur- nental water, and for ERS-1 and 2, the GDRs (Geophysi-
face at different level of those lakes, from low to high level cal Data Records) provided by ESA are limited to ocean
stages. We hence use the different sources of satellite imag- surface. Moreover the performances of Jason-1 and GFO
ery to calculate surface variations of the GTN-L lakes. satellites over narrow inland water are also poor, which
It is worth noting that once pairs of both level and sur- has limited the use of these satellites to only big lakes where
face are available for a given lake ranging from the driest to it has provided accurate lake height variations (see Figs. 2a
wettest situations, rating curves can be established, and in and 2b). Therefore the lake levels are based on the merged
case one of the two parameters is missing, it can be deduced Topex/Poseidon, Jason-1 and 2, ENVISAT and GFO data.
from the rating curve in order to fill the gap in the series. Almost 150 lakes and reservoirs monthly level variations
All of these sources of data made available recently by deduced from multi-satellite altimetry measurements are
space agencies will therefore allow us to reach one of the freely provided. (Figs. 1a and 1b). Potentially the number
ECVs requested by GCOS, which are the lakes surface of lakes monitored could be significantly increased.
variations in time. Operating at dual frequencies 13.6 GHz (Ku Band:
Jason T/P) and 5.3 GHz (C Band) or 13.6 GHz and
2.2. Lakes water height inferred from radar altimetry 3.2 GHz (S Band: Envisat), each altimeter emits a series
of microwave pulses towards the nadir direction that are
Despite some limitations, altimetry is a technique that examined in the time domain. By noting the two-way time
has a proven potential for hydrology science since the data delay between pulse emission and echo reception, the sur-
are freely available worldwide, and it is the only source of face height is determined by the difference of the satellite
information for most lakes in remote areas. Water level orbit and the altimeter range measurement. Depending on
measurement by satellite altimetry has been developed and the size of the lake, it is necessary to correct for the slope
optimised for open oceans. Several satellite altimetry mis- of the geoid (or equivalently, the mean lake level) which is
sions were launched since the late 1980s: GEOSAT (1986– represented by a mean lake surface. The EGM2008 geoid
1988), ERS-1 (1991–1996), Topex/Poseidon (1992–2005), is used as a priori but does not fit correctly short wavelength
ERS-2 (1995–), GFO (2000–2008), Jason-1 (2001–) Jason- variations over a lake surface hence must be corrected. It is
2 (2008–) and ENVISAT (2002–). ERS-1, ERS-2 and done by averaging the altimetry derived height levels from
ENVISAT have a 35-day temporal resolution (duration of the whole cycles along each track. The water levels are fur-
the orbital cycle) and 70 km inter-track spacing at the equa- ther referred to this ‘mean lake level’. If different satellites
tor. Topex/Poseidon, Jason-1 and Jason-2 have a 10-day cover the same lake, the lake level is computed in a 3-step
orbital cycle and 350 km equatorial inter-track spacing.
GEOSAT and GFO have a 17-day orbital cycle and
170 km equatorial inter-track spacing. The combined global 2
The US Department of Agriculture has developed a similar database
altimetry data set has more than two-decade-long history for water level variations for almost 100 lakes monitored from the satellite
and is expected to be continuously updated in the coming missions T/P, Jason-1 and Jason-2. The data are available on the web site:
http://www.pecad.fas.usda.gov/cropexplorer/global_reservoir/. The Euro-
decade with missions such as Altika, Jason-3, Cryosat-11,
pean Space Agency, in cooperation with the De Montfort University in
Sentinel-3 and SWOT. The technique is now applied to UK has also developed a database which provides instantaneous water
obtain water levels of inland seas, lakes, rivers, floodplains level products for rivers and lakes in NRT mode at the following address:
and wetlands (Aladin et al., 2005; Crétaux et al., 2005). http://earth.esa.int/riverandlake/.
 Author's personal copy

1500 J.-F. Crétaux et al. / Advances in Space Research 47 (2011) 1497–1507

Fig. 1a. Map of the existing lakes and reservoirs of the HYDROWEB database.

Fig. 1b. Schema of use of HYDROWEB.

process. Each satellite data are processed independently. (Shum et al., 2003; Crétaux et al., 2010, Figs. 2c and 2d)
Potential radar instrument biases between different satellites For the Lake Victoria, comparison between Jason-1
and geoid differences are removed using Topex/Poseidon inferred water level variations and in situ Gauges records,
data as reference. Then lake levels from the different satel- from September 2004 to March 2006 has given an RMS
lites are merged on a monthly basis. We generally observe of differences of 2.6 cm and a correlation of both time
an increased precision of lake levels when multi-satellite series of 0.99 (Figs. 2a and 2b). For such a big lake, the
processing is applied. altimetry clearly provides very accurate level data in a con-
Hydrology is a growing field of application for satellite tinuous manner (Figs. 2a–2d, Table 1).
altimetry. However, very few studies have been dedicated Although the technique is less accurate for smaller lakes
up to now for checking the quality of the altimetry data the derived level variations are generally an order of
over the continental waters. A lake database will be a magnitude higher than the total error budget (Crétaux
unique opportunity for Calibration/Validation activities and Birkett, 2006). An ongoing research is examining the
over the continental waters. trade off between minimum target size observable and
Comparisons of altimetry results with gauge data for the acceptable height accuracy for the various hydrological
Great Lakes, for example, are accurate to 3–5 cm rms applications.
 Author's personal copy

J.-F. Crétaux et al. / Advances in Space Research 47 (2011) 1497–1507 1501

Fig. 2a. Comparison of water level of Lake Victoria from in situ gauges (green) and from radar altimetry (Red: Jason-1). (For interpretation of the
references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 2b. Scatter plot of in situ/Jason-1 level data over the Lake Victoria.

Fig. 2c. Comparison of water level of Lake Superior from in situ gauges (green) and from radar altimetry (Red: Jason-1). (For interpretation of the
references to colour in this figure legend, the reader is referred to the web version of this article.)

2.3. Lake ice monitoring from remote sensing data affects the energy exchange between water and atmosphere,
as well as various hydro physical and hydro biological pro-
Many boreal continental water bodies such as lakes or cesses (Kozhova and Izmest’eva, 1998). The presence of ice
enclosed seas have seasonal ice cover. Ice dramatically and its transparency affects the bloom of diatoms and
 Author's personal copy

1502 J.-F. Crétaux et al. / Advances in Space Research 47 (2011) 1497–1507

Fig. 2d. Scatter plot of in situ/Jason-1 level data over the Lake Superior.

Table 1
Some examples of comparison of lake level calculated from satellite altimetry data and measured by in situ gauges. Country and size of the lake are also
indicated. For the biggest lakes (Victoria, Superior and Erie) the accuracy is better than 10 cm, the intermediate (Kariba, Mar de Chiquita, and Titicaca)
the accuracy is at the decimetre level, while for some small water bodies, as Lake Powel it is closer to 1 m, particularly due to the fact that this lake is long
but very thin. The Lake Issykkul, which is intermediate lake in size, presents surprisingly very accurate results.
Lake name Country Size of the lake (km2) RMS of the differences in situ/altimetry level (cm)
Erie USA, Canada 25,821 5
Issykkul Kyrgyzstan 6236 4
Kariba Zambia, Zimbabwe 5400 24
Mar de Chiquita Argentina 6000 13
Powell USA 380 80
Superior USA, Canada 82,367 4
Titicaca Peru, Bolivia 8372 17
Victoria Tanzania, Uganda, Kenya 68,800 3

primary productivity (Mackay et al., 2003). In spring, as (Kouraev et al., 2008). Although not included initially by
soon as enough light can penetrate the ice, development GCOS in the ECVs for lakes, the SOLS/HYDROLARE
of phytoplankton, zooplankton and benthos starts under data centre will include lake ice parameters for boreal lakes
the ice cover. Parameters such as timing of ice formation inferred from remote sensing data (radar altimetry and
and break-up, ice cover extent and thickness, are good indi- SSM/I) but also from in situ measurements when available.
cators of regional and large-scale climate changes and are Fig. 4 shows the product that may be delivered through
widely used for various climatic studies, relating ice condi- SOLS/HYDROLARE on this topic.
tions with global climatic variability (Livingstone, 1999).
Ice dynamics influences transport and navigation, fisheries 3. SOLS/HYDROLARE data centre
and other industrial activities. Temporal and spatial vari-
ability of ice processes in seas and lakes is influenced not 3.1. Institutional context
only by meteorological conditions (mainly by thermal
regime), but also by wind and currents, bottom morphol- Over the last decade several countries joined in an inter-
ogy and other factors (Wüest et al., 2005). Studies and national treaty, UNFCC (United Framework Convention
monitoring of ice cover conditions are thus providing valu- on Climate Change) to investigate the issue of the global
able information for climate research, maritime safety and warming reduction. An Implementation Plan has been
sustainable environmental management (Mackay et al., developed by the GCOS endorsed by UNFCCC, to elabo-
2005). rate appropriate policies on climate change. To face these
The studies of ice and snow cover of continental water challenges, UNFCCC emphasises the crucial need to setup
bodies can benefit from the synergy of more than 15 years- a global system capable of acquiring the earth observations
long simultaneous active (radar altimeter) and passive required for these different components: terrestrial, atmo-
(radiometer) observations from radar altimetry satellites spheric, oceanic. The Global Terrestrial Observing System
complemented by SSM/I passive microwave data to improve (GTOS) and its various technical panels are in charge of
spatial and temporal coverage. An ice discrimination the development of the required observing strategy. One
approach based on a combined use of the data from the four of the main goals is to provide systematic, continuous, glo-
altimetry missions and SSM/I was developed and validated bal, harmonised and reliable observations. GCOS has
 Author's personal copy

J.-F. Crétaux et al. / Advances in Space Research 47 (2011) 1497–1507 1503

emphasised in its successive reports the difficulties encoun- basis, in the form of time series based on satellite radar
tered with respect to these requirements in the terrestrial altimetry and in situ gauges.
domain, which remains the least well-developed compo- Products T1.3: surface temperature of all the lakes on the
nent of the climate observing system. GTN-L list: Daily temperature with accuracy of 0.2° and
The Terrestrial Observing Panel for Climate (TOPC) is a a stability of 0.1° with 1 km spatial resolution.
joint panel of GTOS and GCOS. Its main objective is to
define and assess the availability of the ECVs in order to The HYDROLARE data centre seeks to provide the
improve the knowledge of the terrestrial component of product T1.2 and T1.3 from an in situ database for lakes
the climate system, the causes and impact of its changes. having permanent stations in terms of historical observa-
The GCOS implementation plan identified 13 terrestrial tion, and current records regularly updated.
ECVs including lake levels (T04) for which it has proposed Last year CNES developed a prototype for a data centre
actions towards the development and operation of an inter- for lakes and reservoirs exclusively based on Remote sens-
national lake and reservoirs data centre. ing data: satellite radar altimetry and satellite imagery:
Lake levels are currently measured by numerous organ- SOLS. CNES is involved in the Committee on Earth
isations worldwide. Despite the clear and long established Observing Satellite (CEOS) in charge of the international
requirement for such a centre, it does not exist yet and, civil space missions’ coordination in Earth observation.
at present, only two global databases exist for lakes (from As body representing users, GCOS meets CEOS to ensure
International Lake Environment Committee: ILEC) and “continuity of satellite measurements, systematic data gener-
reservoirs (from International Commission Of Large ation, safeguarding of the records, data access and interna-
Dams: ICOLD) that maintain a database with a limited tional coordination addressing future measurements needs”
set of information (lake name, morphometry, average level (Biennal report 2008 of the GTOS). Under this framework,
depth and surface, etc.). Nothing about water level one of the CNES contributions (under the responsibility of
dynamic and storage is provided. the LEGOS) will be the participation in HYDROLARE in
cooperation with SHI.
3.2. SOLS/HYDROLARE data centre This will be done through the extension of the HYDRO-
WEB database to the GTN-L lakes and by adding infor-
However in 2008 The WMO has signed an agreement mation from the satellite imagery to determine surface
with the SHI to establish the International Data Centre variations in time for each of the selected lakes.
on Hydrology of Lakes and Reservoirs (HYDROLARE), Satellite radar altimetry provides level variations in time
hosted by the SHI, in support of the GCOS implementa- for a given lake. Satellite images collected at different dates
tion plan. from low to high water stage will allow determining a sur-
The objective of the Centre is to establish, develop and face variation of each lake in the database (see Section 4.1).
regularly update an international database on hydrological It is also planned to apply the methodology developed
regime of lakes and reservoirs. One main requirement of by LEGOS for large seasonally-frozen lakes and enclosed
GTOS is to develop “basic processing and presentation tools seas and to derive from this approach time series of specific
for lakes and reservoirs data for distribution to stakehold- dates of ice events (the first appearance of ice, the forma-
ers”, to conduct “analysis and assessment of spatial and tion of stable ice cover, the first appearance of open water
temporal tendencies of hydrological elements of lakes and and the complete disappearance of ice) for each water body
reservoirs” and also to “establish agreed methodologies or its sub-region.
and standards” (Biennal report 2008 of the GTOS3). The Currently, HYDROWEB is updated every year for lake
objective is to bring together observations (mostly lacking products (water level variations). The SOLS project will
for lakes and reservoirs), prediction, and decision-support manage two complementary data sets:
systems and establish links to climate and other types of
data. A list of products for the T04 ECVs is proposed for  A reliable, long-term series based on the GDR data
large open lakes, highly ephemeral lakes and closed basin (Geophysical Data Record), computed using the best
lakes in the frame of the GCOS implementation plan: quality orbit, but available only 5–6 weeks after data
acquisition) currently used by HYDROWEB.
Products T1.1: maps of the lakes on the GTN-L: Grids of  An intermediate-quality but rapidly available time-series
georef maps of 250 m spatial resolution on monthly based on the IGDR dataset (the interim GDR, based on
basis for about 100 lakes with a requested accuracy of the preliminary computation); typically, data can be
5%. delivered in ca. 3 days after acquisition.
Products T1.2: level of all the lakes on the GTN-L list:
The requested accuracy is 10 cm on a weekly/monthly This two-mode database can then be used to tackle both
the long-term trend analysis and the NRT applications that
require a high reactivity of the decision makers. Further-
3
http://www.fao.org/gtos/doc/pub50.pdf and http://www.fao.org/gtos/ more, the automation of data collection and time-series
doc/ECVs/T04/T04.pdf. calculation procedures make the interactive download of
 Author's personal copy

1504 J.-F. Crétaux et al. / Advances in Space Research 47 (2011) 1497–1507

the products possible. As for HYDROWEB, the SOLS/ storage variability is purely natural and well documented.
HYDROLARE data policy along with the ancillary data It is a shallow lake with maximum depth of 10 m. Given
(e.g. lake images from remote-sensing images) bundled the flatness of the surrounding ground, it exhibits a highly
with the on-line data will be made available. Open – free variable extent with a level range from 5000 km2 for the
of charge access to ECVs for the GTN-L lakes will be given low level of the dry period to 6500 km2 at high stages dur-
on the SOLS/HYDROLARE. ing the wet season. The daily in situ levels are available and
Based on this principle, lake level and surface variations can be compared to the remote sensing data, making this
have already been calculated for almost 20 lakes and reser- lake a possible site for continuous – although delayed –
voirs. They will constitute the basic input for the prototype check of the altimetry products. The large variations of
of a lake data centre, which is described in this manuscript. the surface extent can be monitored with multi-spectral
In the following pages we describe an overview of the medium resolution sensors like the Modis/Terra instru-
future products for two selected lakes (Mar de Chiquita ment. It can also be compared with an extent determined
and Baikal) that will be delivered through the SOLS using High Resolution image (Landsat, CBERS: Fig. 3).
project. Finally, a precise hypsometric function can be calculated
and used to compute continuous variations of the level,
4. Examples of the SOLS/HYDROLARE ongoing web surface and hence volume for this lake. The web page of
pages the future SOLS/HYDROLARE database will include
lakes as shown in Fig. 3, three types of variables for each:
In the following two paragraphs we show what will be the monthly level variations from the in situ gauges and radar
future web pages of HYDROLARE data centre for ECVs altimetry data, the hypsometry curves parameters, and
inferred from the remote sensing data. The purpose of this monthly surface variations. The re-processed images at a
section is not only to demonstrate the capacity of the remote different epoch from low to high stage will also be available
sensing techniques to produce suitable information in the directly from the data centre.
field of hydrology, but also to introduce the data centre to
users. For details on the radar altimetry applications on lake 4.2. Lake Baikal
monitoring, see Birkett, 1995, 2000; Mercier et al., 2002; Coe
and Birkett, 2005; Crétaux et al., 2005; Crétaux and Birkett, There is an example of the products to be delivered on
2006; Kouraev et al., 2007, 2008; Medina et al., 2008; the SOLS/HYDROLARE for the lake ice phenology.
Swenson and Wahr, 2009; Becker et al., 2010. The Lake Baikal, located in Siberia, is completely frozen
every year (from late December to May) due to continental
4.1. Mar de Chiquita climate conditions with long and cold winter.
From the methodology described in Kouraev et al.
Mar de Chiquita is the largest closed saline lake in South (2008), three parameters can be derived from remote sens-
America, located in the central plain of Argentina. This ing data: time series of ice formation, break-up and dura-
lake is a remarkable target to study current and past envi- tion from year to year, starting from 1992 when the T/P
ronmental changes in a saline lacustrine basin as its water satellite was launched. Fig. 4 shows the time series of these

Fig. 3. Mar de Chiquita future web page. For some lakes the relation dh/dS (hypsometry) is not linear. This explains why the time series of level and
surface variations have different slopes.
 Author's personal copy

J.-F. Crétaux et al. / Advances in Space Research 47 (2011) 1497–1507 1505

Fig. 4. Time series of historical and satellite-derived observations of (a) ice formation (days since 31 December), (b) ice break-up (days since 31 December)
and (c) fast ice duration (days). Southern Baikal: thick grey line – NSIDC/ LIN RAS, data for Listvyanka, thick black line – satellite-derived estimates for
region near Listvyanka. Northern Baikal: thin black line – NSIDC data for Nizhneangarsk, satellite-derived estimates are for the whole Northern Baikal
(NB, thick line) and for sub-region near Nizhneangarsk (NA, thin line). Note that all vertical scales are identical. Straight lines indicate positive and
negative tendencies discussed in the text – they are not statistically significant and do not represent calculated trend lines, but merely serve as a visual aid.
The future users of SOLS/HYDROLARE will have access to both figures like these one and the associated historical and remote sensing data.

parameters for the c. This will be extended to all lakes in the addressed by UNFCCC. In the frame of the GCOS imple-
northern hemisphere that are under similar climatic condi- mentation plan, few ECVs were identified, among them,
tions (lakes of Siberia, Aral Sea and Caspian Sea, Ladoga, the water level and surface variations for a group of 100
Onega, and all lakes in Canada). We will update the data- lakes located throughout all continents. Some of them
base for these products (plots as in Fig. 4 and ASCII files are already operationally monitored by national agencies
with the corresponding data) every year after winter time. (Russia, United States or Canada). Others are not
As for many continental water bodies, the observations of equipped with the in situ gauges yet, or when they are,
ice regime for the Lake Baikal started in the end of the for different reasons the data are not available. Those
19th–to beginning of the 20th century at coastal stations. two ECVs when they measured allow estimating the water
Later on, mainly between 1940s and 1970s, field trips and storage variation which depends on many factors: climate
aerial surveys made it possible to significantly extend the change forcing and anthropogenic stress in the lake basin
research from a point of observation to lake wide spatial watershed (through irrigation, dam construction, water
scale. On the SOLS/HYROLARE, both the remote sensing consumption, etc.)
data (radar altimetry and SSM/I) and these in situ data will Over the last 10–15 years, new satellites techniques have
be supplied to users as shown on Fig. 4. demonstrated their capabilities to fulfil the GCOS require-
ments. Satellite radar altimetry can provide lakes level vari-
5. Conclusion ations for hundreds of lakes (they just need to be under the
track of the satellite) with an accuracy ranging from centi-
For the next few decades, there will be a need to a con- metres to decimetres depending on their size, and also in a
tinuous and automatic survey of lakes worldwide as lesser extent on their location (mountain lakes for examples
 Author's personal copy

1506 J.-F. Crétaux et al. / Advances in Space Research 47 (2011) 1497–1507

are more difficult to track). To be able to calculate a lake’s References

water surface variation, one needs complementary infor-
mation, which usually is quite impossible to derive from Aladin, N.V., Crétaux, J.-F., Plotnikov, I.S., et al. Modern hydro-
ground measurements Lately, national space agencies such biological state of the Small Aral Sea. Environmetric 16, 1–18,
doi:10.1002/env.709, 2005.
as NASA, ESA and INPE have setup free public web data Bates, B.C., Kundzewicz, Z.W., Wu S., Palutikof, J.P. Climate Change
bases that provide imagery of satellites instruments like and Water. Technical Paper of the Intergovernmental Panel on
Landsat or CBERS optical and IR sensors, Terra/MODIS, Climate Change. IPCC Secretariat, Geneva, 210 pp., 2008.
SRTM or Envisat/ASAR. They all have different spatial Becker, M., Llowel, W., Cazenave, A., Güntner, A., Crétaux, J.-F. Recent
and temporal resolution, but depending on the size and hydrological behaviour of the East African Great Lakes region
inferred from GRACE, satellite altimetry and rainfall observations.
on the magnitude of water surface variability of each tar- C. R. Geosci., doi:10.1016/j.crte.2009.12.010, 2010.
get, a large panel of instruments can now provide surface Birkett, S. Contribution of TOPEX/POSEIDON to the global monitoring
variations of hundreds of lakes worldwide. of climatically sensitive lakes. J. Geophys. Res. 100 (C12), 25179–
With current progress in information technology, the 25204, 1995.
Birkett, C.M. Synergistic remote sensing of lake Chad: variability of basin
processing power and storage capacities of research labora-
inundation. Remote Sens. Environ. 72, 218–236, 2000.
tories are sufficient to cope with data volumes, which were Coe, M.T., Birkett, C.M. Water resources in the Lake Chad basin:
manageable only by large data centre 10 years ago. Lever- prediction of river discharge and lake height from satellite radar
aging this processing power with knowledge sharing made altimetry. Water Resour. Res. 40 (10), doi:10.1029/2003WR002543,
possible with an open source software, a pre-operational 2005.
lake database can be created with a modest effort, and act Crétaux, J.-F., Kouraev, A.V., Papa, F., Bergé Nguyen, M., Cazenave, A.,
Aladin, N.V., Plotnikov, I.S. Water balance of the Big Aral Sea from
as a proof-of-concept to build up the dynamics necessary satellite remote sensing and in situ observations. J. Great Lakes Res.
to create a next-generation international fully operational 31 (4), 520–534, 2005.
database and making possible the rapid distribution of Crétaux, J.-F., Birkett, C. Lake studies from satellite altimetry. C. R.
the key ECVs and metadata with the NRT updating, cross Geosci., doi:10.1016/J.cre.2006.08.002, 2006.
calibrated with in situ measurements and dedicated field Cretaux, J.F., Calmant, S., Romanovski, V., Shibuyin, A., Lyard, F.,
Berge-Nguyen, M., Cazenave, A., Hernandez, F., Perosanz, F. An
campaigns. absolute calibration site for radar altimeters in the continental domain:
An agreement was signed in 2008 between the SHI and lake Issykkul in Central Asia. J. Geod., doi:10.1007/s00190-008-0289-
WMO to develop such a lake data centre, called HYDRO- 7, 2009.
LARE, which will collect and distribute in situ lake level Crétaux, J.-F., Calmant, S., Abarca Del Rio, R., Kouraev, A., Bergé-
Nguyen, M., Maisongrande, P. Lakes studies from satellite altimetry.
variations, surface temperatures and metadata. The remote
Handbook on Coastal altimetry, Springer-Verlag Berlin Heidelberg,
sensing techniques that are mature enough to complete the Chapter 19, doi:10.1007/978-3-642-12796-0_19, 2010.
requested information (lake level, surface and volume vari- Frappart, F., Seyler, F., Martinez, J.-M., et al. Floodplain water storage in
ations, lake ice regime) will be under the responsibility of the Negro River basin estimated from microwave remote sensing of
LEGOS, in coordination with SHI. In the last 6 years inundation area and water levels. Remote Sens. Environ. 99, 387–399,
LEGOS has developed a lake and river database, with 2005a.
Frappart, F., Dominh, K., Lhermitte, J., et al. Water volume change in the
water level variations for a group of 150 lakes and reser- lower MEKONG basin from satellite altimetry and other remote
voirs fully derived from current satellite altimetry. Algo- sensing data. Geophys. J. Int., 2005b.
rithms of radar and optical satellite imagery processing Frappart, F., Calmant, S., Cauhopé, M., Seyler, F., Cazenave, A.
were developed by LEGOS in order to complete this data- Preliminary results of Envisat RA-2-derived water levels validation
base with information on lake’s surface and volume varia- over the Amazon basin. Remote Sens. Environ. 100, 252–264, 2006.
Garreaud, R., Aceituno, P. Interannual rainfall variability over the South
tions that will be added to the database. It will result in an American Altiplano. J. Clim. 14, 2779–2789, 2001.
international lake data centre, in cooperation between SHI Garreaud, R.D., Battisti, D.S. Interannual (ENSO) and interdecadal
and LEGOS, under the auspices of WMO. (ENSO-like) variability in the Southern Hemisphere tropospheric
circulation. J. Clim. 12, 2113–2123, 1999.
Henry, J.-B., Chastanet, P., Fellah, K., et al. Envisat multi-polarised
ASAR data for flood mapping. 0-7803-7929-2/03/$17.00ÓIEEE, pp.
Acknowledgements 1136–1138, 2003.
Kozhova, O.M., Izmest’eva, L.R. (Eds.). Lake Baikal, Evolution and
This work has been supported by the Centre National Biodiversity. Backhuys, Publ., Leiden, 1998.
d’Etudes Spatiales (CNES) in the frame of the TOSCA Kouraev, A.V., Semovski, S.V., Shimaraev, M.N., Mognard, N.M.,
program. It also benefits from Envisat data made avail- Legresy, B., Remy, F. Observations of lake Baikal ice from
satellite altimetry and radiometry. Remote Sens. Environ. 108, 240–
able at ESA, Frascati. The altimetry data are downloaded 253, 2007.
from the Centre de Topographie des Océans et de Kouraev, A.V., Shimaraev, M.N., Buharizin, P.I., et al. Ice and snow
l’Hydrosphère (CTOH) of Legos. We therefore gratefully cover of continental water bodies from simultaneous radar
acknowledge the CNES, the ESA and the CTOH. We altimetry and radiometry observations. Surv. Geophys. – Thematic
Issue “Hydrology from Space” 2008, doi:10.1007/s10712-008-9042-2,
also acknowledge Stephan Bojinski from GCOS and
2008.
Wolfgang Grabbs from WMO for their constant support Livingstone, D. Ice break-up on southern Lake Baikal and its relationship
in the development of the SOLS contribution to the to local and regional air temperatures in Siberia and the North
HYDROLARE data centre. Atlantic Oscillation. Limnol. Oceanogr. 44, 1486–1497, 1999.
 Author's personal copy

J.-F. Crétaux et al. / Advances in Space Research 47 (2011) 1497–1507 1507

Mackay, A.W., Battarbee, R.W., Flower, R.J., Granin, N.G., Jewson, Richard, Y. Variabilite pluviometrique en Afrique du Sud-Est. La
D.H., Ryves, D.B., Sturm, M. Assessing the potential for developing Meteorologie 8, 11–22, 1994.
internal diatom-based inference models in Lake Baikal. Limnol. Ropelewski, C.F., Halpert, M.S. Quantifying southern oscillation–precip-
Oceanogr. 48, 1183–1192, 2003. itation relationships. J. Clim. 9, 1043–1059, 1996.
Mackay, A.W., Ryves, D.B., Battarbee, R.W., Flower, R.J., Jewson, D., Peng, D.Z., Xiong, L., Guo, S., et al. Study of Dongting Lake area
Rioual, P., Sturm, M. 1000 years of climate variability in central Asia: variation and its influence on water level using MODIS data. Hydrol.
assessing the evidence using Lake Baikal diatom assemblages and the Sci. 50 (1), 31–44, 2005.
application of a diatom-inferred model of snow thickness. Global Sakamoto, T., Nguyen, N.V., Kotera, A., et al. Detecting temporal
Planet. Change 46, 281–297, 2005. changes in the extent of annual flooding within the Cambodia and the
Marchant, R., Mumbi, C., Behera, S., Yamagata, T. The Indian Ocean Vietnamese Mekong Delta from MODIS time series imagery. Remote
dipole – the unsung driver of climatic variability in East Africa. Afr. J. Sens. Environ. 109 (3), 295–313, doi:10.1016/j.rse.2007.01.011, 2007.
Ecol. 45 (1), 4–16, 2007. Shum, C., Yi, Y., Cheng, K., Kuo, C., Braun, A., Calmant, S., Chambers,
Medina, C., Gomez-Enri, J., Alonso, J., Villares, P. Water level D. Calibration of Jason-1 Altimeter over lake Erie. Mar. Geod. 26
fluctuations derived from ENVISAT Radar altimetry (RA-2) and (3–4), 335–354, 2003.
in situ measurements in a subtropical water body: lake Izabal Swenson, S., Wahr, J. Monitoring the water balance of Lake Victoria,
(Guatemala). RSE, doi:10.1016/J.rse.2008.05.001, 2008. East Africa, from Space. J. Hydrol. 370, 163–176, doi:10.1016/
Mercier, F., Cazenave, A., Maheu, C. Interannual lake level fluctuations j.hydrol.2009.03.008, 2009.
in Africa from TOPEX–Poseidon: connections with ocean-atmosphere Töyrä, J., Pietroniro, A., Martz, L.W. Multisensor hydrologic assessment
interactions over the Indian ocean. Global Planet. Change 32, 141– of a freshwater wetland. Remote Sens. Environ. 75 (2), 162–173, 2001.
163, 2002. Töyrä, J., Pietroniro, A. Towards operational monitoring of a northern
Mistry, V., Conway, D. Remote forcing of East African rainfall and wetland using geomatics-based techniques. Remote Sens. Environ. 97,
relationships with fluctuations in levels of Lake Victoria. Int. J. 174–191, 2005.
Climatol. 23 (1), 67–89, 2003. Wüest, A., Ravens, T.M., Granin, N.G., Kocsis, O., Schurter, M., Sturm,
Nicholson, S.E., Yin, X. Mesoscale patterns of rainfall, cloudiness and M. Cold intrusions in Lake Baikal: direct observational evidence for
evaporation over the Great lakes of East Africa. The East African deep-water renewal. Limnol. Oceanogr. 50 (1), 184–196, 2005.
great lakes: limnology, paleolimnology and biodiversity. Advance in Zola, R.P., Bengtsson, L. Long-term and extreme water level variations of
Global Change Research, vol. 12, 2002. the shallow Lake Poopo, Bolivia. Hydrol. Sci. J. 51 (1), 98–114, 2006.

View publication stats