Spatial downscaling of TRMM precipitation using vegetative response on the Iberian Peninsula

Precipitation data with accurate, high spatial resolution are crucial for improving our understanding of basin scale hydrology. We explore the relation between precipitation estimates derived from the Tropical Rainfall Monitoring Mission (TRMM) and the normalized difference vegetation index (NDVI) for different spatial scales on the Iberian Peninsula in southern Europe, using time series from 2001 to 2007 Analysis shows that NDVI is a good proxy for precipitation. On an annual basis an exponential function best describes the relation between NDVI and precipitation. The optimum relation between NDVI and precipitation is found at an approximate scale of 75-100 km. This is an intermediate scale and it is likely that at smaller scales NDVI is determined primarily by anthropogenic land use and at larger scales factors such as geology, soils, and temperature play an increasingly important role. The fact that both TRMM and NDVI are subject to bias due to orbital deviations, atmospheric conditions and imperfect retrieval algorithms could also influence the scale dependency. The derived relation between NDVI and precipitation is used to develop a new downscaling methodology that uses coarse scale TRMM precipitation estimates and fine scale NDVI patterns. The downscaled precipitation estimates are subsequently validated using an independent precipitation dataset. The downscaling procedure resulted in significant improvements in correlation, bias, and root mean square error for average annual precipitation over the whole period, for a dry year (2005), and a wet year (2003). (C) 2008 Elsevier Inc. All rights reserved.