Submitted to: American Geophysical Union
Publication Type: Abstract Only
Publication Acceptance Date: November 5, 2006
Publication Date: December 11, 2006
Citation: Bindlish, R., Crow, W. Jackson, T. 2006. Regional-scale flood detection using AMSR-E observations [abstract]. EOS Transactions, American Geophysical Union, Fall Supplements. 87(52):H21B-1372. Technical Abstract: Remote sensing observations provide spatially distributed information that can be exploited to improve flood forecasting and risk mitigation. These observations provide potential tools for improving the detection and monitoring of flooding events - particularly within data poor regions of the world lacking extensive ground-based rainfall observations. Of particular importance is information concerning antecedent soil moisture conditions. Soil moisture determines the partitioning between infiltration and overland flow and the efficiency of surface runoff generation during rainfall. If properly interpreted, passive microwave observations from the NASA’s Advanced Microwave Scanning Radiometer (AMSR-E) can provide large-scale estimates of antecedent surface soil moisture conditions and potentially enhance the forecasting and monitoring or regional-scale flooding events. Such microwave observations provide an all weather monitoring system - capable of penetrating dense cloud cover typically associated with large-scale flooding events. Here, AMSR-E observations along with satellite-based precipitation observations are used to detect flooding in large-scale watersheds in Africa during 2002-2005. Two limitations of AMSR-E soil moisture products are their relatively coarse horizontal resolution (~ 60 km) and shallow vertical measurement depth (1-3 cm). While this may limit their value for certain applications, it is important to note that regional-scale flooding events are typically associated with saturated near-surface conditions over very large areas. A decision system based on change in AMSR-E observations and GPCP precipitation is developed and validated over a multi-year period using known historical flood monitoring maps within Africa.