H13H-1667
Merging PERSIANN-CCS Satellite-based Rainfall Data and Pointwise Gauge Measurements over Chile
Monday, 14 December 2015
Poster Hall (Moscone South)
Zhongwen Yang1, Kuolin Hsu2, Soroosh Sorooshian2, Xinyi Xu1, Dan Braithwaite2 and Koen M J Verbist3,4, (1)College of Water Sciences, Key Laboratory of Water and Sediment Sciences of Ministry of Education, Beijing Normal University, Beijing, China, (2)University of California Irvine, Civil and Environmental Engineering, Irvine, CA, United States, (3)UNESCO-IHP, Hydrological Systems and Global Change section, Santiago, Chile, (4)nternational Centre for Eremology, Department of Soil Management, Ghent University, Ghent, Belgium
Abstract:
Satellite-based Precipitation Estimates (SPEs) are promising alternative rainfall data for climatic and hydrological applications, due to their ability to estimate precipitation for regions where ground-based observations are limited. However, existing satellite-based rainfall estimations are subject to systematic biases. It is widely recommended that the ground-based gauge can provide reliable precipitation measurement, while uncertainty from gauges increases when the precipitation estimation is extended from the point scale to a spatial coverage. This study aims to provide a high-quality rainfall data set over Chile, by merging the Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks–Cloud Classification System (PERSIANN-CCS) rainfall retrievals and point-wise gauge observations. The PERSIANN-CCS (0.04°×0.04°) and gauged rainfall data (daily) are merged together for the period of 2009-2014. Two steps are associated in this work: 1) the non-parametric Quantile Mapping approach and Gaussian Weighting (GW) interpolation are used to remove the biases in the satellite rainfalls; 2) the gauge measurement is gridded with the GW interpolation, and the bias-corrected satellite and gridded gauge data sets are combined according to weighting factors estimated at each satellite grid. The weighting factors indicate the accuracy of the two data sets, respectively. The results show that the merged precipitation associates high data quality. The spatial patterns of the merged satellite rainfall have good consistency to the gauge observations, which is evidenced with significantly reduced root-mean-square errors (RMSEs) and mean biases (BIASs). At both monthly and daily scales, the performance of the merged rainfall time series, are significantly improved, in terms of the correlation coefficients (CORRs), RMSEs, and BIASs. This study serves as a valuable reference for the bias-correction of existing SPEs with gauge observations worldwide.
