Variability of terrestrial freshwater storage in the Tropics from multi-satellite observations
Terrestrial water is critical to sustaining life on Earth and plays a primary role in the global water cycle and the global climate. However basic questions still remain open related to the land water budget such as: Where, how and why do the different components of the terrestrial water storage vary at time scales from a few days to years, from regional to continental and global scale?
We propose to work towards answering these crucial questions by developing new data products and innovative analysis techniques integrating a combination of multi-satellite observations, in-situ measurements and modeling outputs to infer the spatial and temporal variability of global components of the terrestrial hydrological cycle, including surface water storage, soil moisture and groundwater. Our analysis will first focus on three large tropical river basins (the Amazon basin, the Congo basin and the Ganges-Brahmaputra basin) but with an ultimate goal to study the whole Tropical regions (and possibly at the global scale).
The foundation of our analysis is a global, multi-year dataset quantifying the monthly distribution of surface water extent at ~25km resolution and generated from multiple satellite observations: passive microwave (SSM/I), scatterometer (ERS) and visible and near-IR (AVHRR) optimized specifically for surface water detection. We will first produce estimates of the change in surface water volume stored in rivers, floodplains, lakes and wetlands by combinations with other observations. This surface water volume changes dataset will be derived in two parallel ways. The first technique uses the combination of the surface water extent products with altimeter-derived river/floodplains water level heights and the second technique combines the surface water extent products with Digital Elevation Model (DEM such as ASTER), using a hypsographic curve approach. Over the three basins, the results will be cross-checked and evaluated against related hydrological parameters (in-situ river discharge, rainfall data, etc). The latter technique using DEM can be extended to the whole tropical band with uncertainty estimates based on our 3 basin studies. These multi-year, monthly surface water volume data sets will then be jointly analyzed over the tropical area with related hydrological information: AMSR-E soil moisture products (and SMOS from 2010), simulated soil moistures from models and GRACE measurements, which provides the total terrestrial water storage change (the integrated variations of surface water storage, soil moisture, snow and the ground water). Together with soil moisture products, the surface water storage can be used to separate the integral GRACE signal into the contributions of individual storage components and isolate the variations of groundwater. Bringing together these variables with river discharge measurements, rainfall and surface evaporation estimates in an integrated approach will yield improved overall knowledge of the different components of the terrestrial water budget, their link with natural climate variability, but also the influence of human activities (ground water mining, irrigation, dam building on rivers...). In addition, our results will help to validate/evaluate large-scale hydrological models and better prepare and validate future hydrology-oriented missions such as SWOT.