Evaluation of Quantile Mapping Methods in Bias Correction of the TRMM Satellite’s Estimated Precipitation Data During Vegetation Growth Period (Apr-Oct) in Kermanshah Province

Global space coverage of the TRMM satellite imagery has provided a good opportunity to use the precipitation data estimated by this satellite in the country. Various studies have been conducted in the country to evaluate the accuracy of the above data in comparison to the measured data at ground stations. However, few studies have examined the efficacy of postprocessing methods in correcting TRMM precipitation data. The purpose of this study was to evaluate the performance of quantile mapping methods in improving TRMM precipitation data compared to ground data. For this purpose, 10 quantile mapping methods were applied to the gridded TRMM precipitation data in Kermanshah province on a monthly time scale (Apr-Oct) from 2005 to 2012. The ground precipitation data for the same time periods were collected from 13 synoptic weather stations and 82 rain gauges. The results showed that non-corrected estimations of TRMM precipitation data in the elevated (lowland) areas higher (lower) than the ground data. In addition, it was found that the difference between satellite- and ground-based estimates of precipitation in high-precipitation months was much greater than low-precipitation ones. The Parametric transformation of scale method with the least error, among the others, was introduced as the most appropriate quantile correction method. The results of data post-processing showed that the mentioned method could improve the accuracy of TRMM precipitation data. In addition, the correlation coefficient between ground measurements and satellite precipitation data varied in the range of 0.73 to 0.94 (significant at the 5% level), with the highest correlations obtained compared to the precipitation of synoptic and TRMM stations.