Perbandingan Metode Koreksi Data Radar Cuaca Dengan Penakar Hujan Wilayah Bengkulu

Abstract

Weather radar data can provide spatio temporal rainfall forecast at high- resolution. Radar covers a large area and captures the spatial variability of rainfall that increases a short-term predictability of rain. However, the rainfall derived from radar is inherently an estimated value that may contain bias. Also, the estimated value is derived indirectly from the radar reflectivity equation. Therefore, adjusting the estimated rainfall value based on the local observed rain gauge remains a research challenge. This study employed Bengkulu weather radar data and combined with the rainfall from automatic weather station at the Bengkulu Climatology Station. Observed rainfall datasets from 34 rain gauges over Bengkulu for 2015 to 2020 were used for the analysis, we selected December-February as representative of wet season, whereas June-August was for the dry season. Correction of radar data with rain gauge is calculated using the Mean Field Bias (MFB), Local Bias Correction (LBC), and Quantile Mapping (QM) approaches. MFB is a method for estimating the overall bias (average) for radar in a large domain. The LBC correction is a correction by considering the distance factor of the rain gauge to the radar location. Correction of bias QM corrects the data distribution using the probability approach. The Conversion of reflectivity to intensity using three different Z-R relationship, namely Marshall Palmer, Rosenfeld, and the new Z-R equation refers to previous research on the island of Sumatra. The results of analyses in the present study strongly indicated an improvement in radar data quality after the correction method was applied. The Local Bias Correction (LBC) method gives the best results compared to the MFB and QM methods for the Bengkulu region, with the highest correlation of 0.51. The best Z-R relationship is obtained from the Marshall Palmer equation. Bias correction performance in the wet season (DJF) has better results than in the dry season (JJA).