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

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 controlled by the in situ infrastructure for the Global Terrestrial Network for Lakes (GTN-L) under the supervision of WMO (World Meteorological Organization) and GCOS (Global Climate Observing System).

Introduction

The knowledge of the lake water storage variations in time over a long period is fundamental for understanding the impact of climate change and human activities on the terrestrial water resources. Variations in air temperature and precipitation have impact on the water balance of a lake, and in an extreme case a lake can entirely disappear. There are two main types of lakes: open lakes with outflow draining rivers, and in opposite, closed lakes with no outflow. In any cases they are very sensitive to climate change.

In some regions highly ephemeral lakes provide information on the events like severe drought or inundation. Closed basin lakes are sensitive to changes in regional water balance. Therefore a Near Real Time (NRT¹) monitoring of this type of lakes is essential in the frame of such extreme events. For small lakes of this class, the sensitivity to a small change in inputs is proportionally higher, principally in terms of their level variations. In a given region covered with several lakes, if the records of their level variations are long enough, they could reveal recurrence of trends in a very reliable and accurate manner. In this way lakes can be considered as an excellent proxy for climate change.

For example the Andean chain in South America is covered with hundreds of lakes. They are located in region, which is under the influence of several climatic forcing: Southern Atlantic Oscillation, Pacific Decadal Oscillation, El Nino, Glacier melting, etc. (Garreaud and Battisti, 1999, Garreaud and Aceituno, 2001, Zola and Bengtsson, 2006). From North to South, over a distance of more than 4000 km, lakes are distributed in the Semi Arid plateau region (Peru-Bolivia) which receives extremely small rainfall and suffers from intensive evaporation, and in the boreal region (Patagonia) where inter-annual fluctuations of precipitation are high, and water release from glacier thawing has increased last century. Studying of these lakes in a continental framework would be therefore very useful for a better understanding of climate change impact on surface water resources, in particular for the big Andean cities fed in summer by the melt of water outpouring from the glaciers stocking the winter snowfalls and rainfalls. As illustration, for South America, the IPCC (International Panel for Climate Change) scenarios predict an increase of temperature going from 1 to 6 °C, and an increase of precipitation anomalies by about 20% before the end of the XXIth century (Bates et al., 2008). But the geographical distribution of these anomalies is not homogeneous: Most GCM (Global Climate Model) projections indicate larger (positive or negative) rainfall anomalies for the tropical region, and smaller for the extra-tropical part of South America. This will have an impact on the runoff of the rivers that feed the South American lakes.

Another region with dense lakes coverage is the East African rift. Understanding of the climate variability in Africa, and therefore the tropical variability is one of the key goals for the climate research. Also African lakes themselves modify greatly local climate (Nicholson and Yin, 2002). Level records analysis of the Great East African Lakes such as Victoria, Tanganyika, and Nyassa-Malawi, and the regional climatic variability show synchronicity, which can be explained by a large-scale linkage (Ropelewski and Halpert, 1996). Several studies have investigated the effect of ENSO (Mistry and Conway, 2003) but also of the Indian Ocean Dipole (Richard, 1994), which is partly responsible for driving climate variability over eastern Africa (Marchant et al., 2007).

Global warming impact on lakes has ecological consequences that are not precisely predictable. For example, warming of the lake Tanganyka has lead to a reduction in photosynthesis productivity, following by drastic decrease in fish population with obvious societal consequences. In another region, the general reduction of water level of the Great Lakes of North America has several societal and environmental consequences: pollution on the rise in coastal areas, damage in transportation system, tourism, recreational, fishery and hydroelectricity industries. Finally, large open lakes can themselves affect local climate through evaporation and albedo effects.

Hence it is essential to build up global database on lakes and reservoirs worldwide providing information about level, surface and volume variations, ice cover duration and ice cover thickness, as well as surface water temperature variations. The availability of these datasets in NRT mode would also be beneficial worldwide, as it would increase the quantity and quality of information on water management, regional network enhancement, model assimilation, climate change’s impact on the lake system and surrounding human activities, etc.

Yet, the existing gauge networks on a regional and a global scale are declining and observations are often nonexistent, particularly in developing countries. Long term change and inter-annual variability of water storage are also often lacking. When existing, data are neither harmonised nor expressed in the same reference frame. Remote sensing hence could offer an opportunity to improve this situation.

The main purpose of this article is to present the potentiality of remote sensing techniques, to fulfil the requirement of the lake data centre development in addition to the compilation of the in situ historical and current data. For this an initiative was taken in 2009 to setup such a data centre under the GCOS auspices with two main component, one, the HYDROLARE centre hosted by SHI, and two, a web portal under construction, SOLS (Service d’Observation des Lacs par Satellite: French acronym) hosted by LEGOS. Section 2 presents the different techniques that allow producing the ECVs (Essential Climate Variables) for lakes. Section 3 discusses the international institutional context and the SOLS component of the HYDROLARE data centre. Section 4 shows the prototype of the future web portal of SOLS. Conclusions are given in Section 5.