Model for estimating rainfall from meteorological geostationary satellite data over southern China

Abstract

 In this work, a model for calculating rainfall from satellite data in the southern China was developed. Digital data from the visible and infrared channels of FY-2D meteorological geostationary satellite during a 5-year period (2008-2012) were used. The satellite - derived quantities namely, the maximum earth-atmospheric albedo (), the average earth-atmospheric albedo (), the minimum brightness temperature (), the average brightness temperature in the 25-percentile () and the number of hours with the brightness temperature less than 235 K () were used as dependent variables in the model. Rainfall data acquired from 68 rain gauge stations were used, 36 stations for model formulation and 32 stations for model validation. To establish a model relating monthly rainfall () with , , , and , the multiple linear regression was used to estimate the coefficients of the model . For model validation, the rainfall calculated from the model was compared to that obtained from rain gauge and the discrepancy in terms of root mean square difference (RMSD) and mean bias difference (MBD) was found to be 14.4% and -6.0%, respectively. ,maxEAEA,minBT,25BPT235BTNfR,maxEAEA,minBT,25BPT 235BTN