Monitoring and evaluation of flooding extent by using the Getis statistic of the SSM/I data

Abstract

Observation by the DMSP SSM/I over flooding area along the Yangzi River, China during July-August 1998 is studied. A flooding index FI defined by the difference of horizontally-polarized brightness temperature at 37 and 85 GHz channels is proposed to detect flooding. Spatial and temporal variation of the FI index during flooding July-August is demonstrated. The Getis statistics, as a local autocorrelation function, is discussed to show spatial clustering pattern and correlation of high value FI, and is applied to monitoring and evaluation of flooding extent and temporal evolution.