Comparing satellite RFE data with surface gauges for 2012 extreme storms in African East Sahel

The year 2012 was a relatively wet' year for many areas of the African Sahel. In spite of earlier predictions of normal or below normal rainfall in the East Sahel early in the season, by September of 2012, humanitarian agencies reported that flash floods in the region had affected almost a quarter million people, destroying tens of thousands of homes. Early evidence of impending conditions was supplied by satellite images largely produced from the geosynchronous EUMETSAT METEOSATs and low earth orbiting satellites of opportunity, which have become critical tools in the arsenal of emergency responders to extreme hydrologic events around the globe. This report is an assessment of the 2012 monsoon season performance of one of these satellite-based tools the U.S. Climate Prediction Center's (CPC) Rainfall Estimation (RFE) Algorithm Version 2.0 (RFE 2.0) expressly developed for purposes of drought forecasting and flood prediction in Africa. We directly compare satellite-based daily RFE 2.0 rainfall estimates with ground-based daily gauge data recorded at 9 separate but closely spaced stations in the East Sahel of Sudan for the AprilOctober, 2012 monsoon season. We find that, although the RFE 2.0 data may reasonably represent seasonal trends and annual rainfall totals when composited over substantial areas, the morphology and intensity of distinct storm events at the local level are poorly characterized, with many misses (i.e. non-detects), or alternatively, false alerts (i.e. false alarms).