On the relationship between the GOES precipitation index and ISCCP data set variables

Abstract

The GOES Precipitation Index (GPI) is used for global, monthly rainfall estimation in the Global Precipitation Climatology Project (GPCP). Previous work has identified the existence of locally and seasonally varying bias in the GPI estimates. Most sources of bias involve cloud properties, as the GPI method uses the fraction of pixels with infrared cloud temperature below 235°K to estimate monthly rainfall totals averaged over 2.5° × 2.5° latitude/longitude grid boxes. In this work, the bias in the GPI is compared to cloud variables derived by the International Satellite Cloud Climatology Project (ISCCP). ISCCP data are used as predictor variables in regression models with the GPI estimation error as the dependent variable. The GPI estimation error is calculated using the global rain gage analysis produced by the GPCP for those locations where the rain gage network density is high. Fourteen ISCCP cloud variables were selected as the predictors in linear and nonlinear regression models. The nonlinear model explains over 60% of the variance of the GPI bias, while the linear model explains about 45% of the variance. Comparison with another method of GPI bias estimation is discussed.