Validasi data hujan satelit IMERG terkalibrasi dengan metode geographically weighted regression terhadap data hujan stasiun

Abstract

Rainfall monitoring is a crucial aspect of climate modeling and water resource management. The widespread availability of satellite rainfall data, such as Integrated Multi-satellitE Retrievals for Global Precipitation Measurement (IMERG), offers the potential to obtain high-resolution global precipitation information. However, it is essential to evaluate the validity of IMERG satellite rainfall data against ground-based station measurements. The aim of this study is to evaluate the validity of IMERG satellite rainfall data in representing ground station rainfall data. In this study, IMERG satellite rainfall data with a 10 km resolution is downscaled to a resolution of 250 m using the Geographically Weighted Regression (GWR) method. The GWR model incorporates the Normalized Difference Vegetation Index (NDVI) data as an environmental variable. Then, the downscaled data is calibrated with the ground station rainfall data. The calibration ensures optimal alignment between the downscaled data and the ground observations. The calibrated rainfall data is then compared to ground station rainfall data using various validation metrics, including R-squared for linear regression, Bias, Root Mean Squared Error (RMSE), and Mean Absolute Error (MAE). In general, the validation results indicate that the calibrated IMERG provides more accurate results (R2=0,95; RMSE=148,80 mm; MAE=124,91 mm; and Bias=0,05) compared to the original IMERG and downscaled results. High correlation and low prediction errors suggest that the calibrated IMERG is sufficiently reliable and can serve as a good alternative rainfall data to represent the actual rainfall occurring in the field in 2021.