Accuracy of chirps satellite rainfall data for rainfall-runoff modelling in tukad badung watershed

Abstract

The availability of rainfall station on Bali Island based on the distribution map by River Basin Organization for Bali–Penida is still limited and not evenly distributed. This study aims to test the accuracy of rainfall data from the CHIRPS Satellite in representing rainfall events in the field. Satellite rainfall data and rainfall station data are transformed into discharge through HEC-HMS modeling. The resulting discharge is continuous flow discharge and design flood discharge. To determine the accuracy of satellite rainfall data, a comparison and verification are carried out with the rainfall station data and AWLR discharge as a reference. Verification is carried out using statistical tests such as RMSE, correlation coefficient (R), and coefficient of determination (R2). A comparison of CHIRPS and rainfall station in the Tukad Badung watershed shows that the CHIRPS has a good correlation with the rainfall station and the data can represent rainfall events in the field with R=0,698 and R2=0,484 in polynomial regression. For the verification results of the HEC-HMS continuous flow discharge model, 13 out of 15 variations of the analyzed data year have RMSE values from the CHIRPS Satellite which are superior to the rainfall station. For the verification results of the design flood discharge HEC-HMS model, the peak discharge from the CHIRPS Satellite is lower than the rainfall station with a difference of 43,98% - 46,47%. The trend of the HEC-HMS design flood peak discharge graph from the rainfall station and the CHIRPS is linear with the design flood peak discharge graph from the AWLR. So it can be decided that the CHIRPS on the Tukad Badung watershed has good accuracy that can represent rainfall events in the field and can be directly used for further hydrological analysis.