G31A-1093
Investigating GRACE Range-Rate Observations over West Africa with respect to Small-Scale Hydrological Signals

Wednesday, 16 December 2015
Poster Hall (Moscone South)
Anne Springer1, Annette Eicker1, Juergen Kusche1 and Laurent Longuevergne2, (1)University of Bonn, Institute of Geodesy and Geoinformation, Bonn, Germany, (2)Géosciences Rennes, CNRS - University of Rennes, Rennes Cedex, France
Abstract:
Here, GRACE K-band range rate (KBRR) observations are analyzed for the effects from small-scale hydrological signals over West Africa including water level changes in reservoirs, extreme weather events, and water storage variability predicted by hydrological models. The presented approach, which is based on level 1B data, avoids the downward continuation and filtering process required for computing monthly gravity field solutions and, thus, enables to assess hydrological signals with a high temporal resolution and at small spatial scales.

In a first step, water mass variations derived from tide gauges, altimetry, and from hydrological model output are converted into simulated KBRR observations. Secondly, these simulated observations and a number of geophysical corrections are reduced from the original GRACE K-band observations to obtain the residuals for a time span of ten years. Then, (i) the residuals are used to validate differently modeled water mass variations and (ii) extreme weather events are identified in the residuals.

West Africa represents an interesting study region as it is increasingly facing exteme precipitation events and floodings. In this study, monthly and daily output from different global hydrological models is validated for their representation of long-term and short-term (daily) water storage variability over West Africa. The daily RMS of KBRR residuals ranges between 0.1 μm/s and 0.7 μm/s. Smaller residuals imply that the model is able to better explain the observations. For example, we find that in 2007 the Land Surface Discharge Model (LSDM) better agrees with GRACE range-rate observations than the Water-GAP Global Hydrology Model (WGHM) and the GLDAS-Noah land surface model.

Furthermore, we confirm previous studies and show that the signal from Lake Volta is distinctly contained in the residuals. Finally, we investigate variations of other smaller reservoirs and the floodings over West Africa in June 2009 and over Benin in October 2010.