RAINFALL CORRECTION USING SATELLITE DATA TOWARDS SURFACE OBSERVATIONS (CASE STUDY: BANDUNG, WEST JAVA)

• Main Author: Rinaldi, Yari
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/66769
• Institution: Institut Teknologi Bandung

Satellite-based rainfall products are needed in the Bandung area but the accuracy of satellite products needs to be evaluated. this study corrects and evaluates rainfall using satellite data on surface observations in the Bandung area which are scattered as many as 19 surface observation stations as reference corrections. This research was started by collecting CHIRPS, TRMM and GSMAP satellite data in accordance with the surface observation points from 2010 - 2016. After that initial evaluation of raw satellite data on surface observations was indicated by scatter plots, averages and standard deviations of surface observations. The results of the initial evaluation as a reference for comparison after the satellite data is corrected. Satellite data were corrected using the Linear Scaling and Quantile Mapping method. After the corrected satellite data is evaluated the rainfall correction results in increasing the satellite estimate close to the surface observation value as evidenced by the increase in the average and standard deviation using the LS method with a similarity value of 70-80% or QM with a similarity value of 80-90%. Correction also increases the correlation of 10-20% for the QM method. Correction also increases QPOD by 5% so that it is more accurate to detect extreme rainfall both LS and QM methods. The QM correction method is better than the LS correction method because in the LS method the correction factor is only the average monthly rainfall whereas in the QM method the correction factor is based on the quantile of rainfall.