Use of satellite altimeter data for validating large scale hydraulic models
Matt Wilson(1) , Doug Alsdorf(2) , Paul Bates(3) , and Matt Horritt(3)
University of Exeter,
Penryn TR10 9EZ,
(2) Ohio State University, 275 Mendenhall Laboratory, 125 South Oval Mall, Columbus OH 43210-1308, United States
(3) University of Bristol, University Road, Bristol BS8 1SS, United Kingdom
Over large, remote basins very few ground observations of surface water elevation are available to validate the output of the distributed hydraulic models which are used to understand better the fluxes of water, sediment and nutrients in such basins. Those observations that are available are subject to considerable uncertainty as (a) elevation levels are frequently not tied in to any datum and (b) records are often incomplete due to instrument failures. Moreover, the number of ground observation stations is in serious decline, especially in the developing world. Given these constraints the data on river water surface elevation available from satellite radar altimeters may be useful source of model validation data. In this paper we examine the potential of such data to assist in the validation of a two-dimensional hydraulic model of a 280x200km reach of the Solimoes and Purus rivers, Amazonia, Brazil. This region includes the extensive confluence plain of these two rivers and provides a challenging problem in hydraulic modeling. Model simulations are conducted at a spatial resolution of 270m with ground topography parameterized from NASA SRTM data that have been corrected for vegetation effects on the basis of vegetation height surveys in the region conducted in June/July 2005. River bathymetry is parameterized from sonar data acquired during the same field campaign. The model is run for an annual flood cycle based on the average of 20 years of flow data and the predicted water surface elevations compared to observations of the same derived from TOPEX-POISIDON. The comparison allows us to better understand the strengths and weaknesses of the hydraulic model and leads to an improved understanding of the space and time scales of data required to validate such codes. The analysis shows that while point data of the type derived from existing altimeters is a contribution to model validation, they do not allow a complete validation of the distributed output of numerical models. Data of the type to be acquired in the proposed WatER mission will, potentially, allow this more complete analysis.