<![CDATA[Precipitation, particularly rainfall, is vital in understanding weather and climate. In Indonesia, the uneven distribution of in situ rainfall observations poses a challenge to accurately measuring surface rainfall. Remote sensing systems and cloud computing technologies, such as Google Earth Engine (GEE), offer potential solutions. This study evaluates the spatial distribution and performance of four multi-satellite rainfall estimates available in GEE, namely CHIRPS, GSMAP, GPM-IMERG, and PERSIANN-CDR, before and after calibration using BMKG rain gauge data in South Sulawesi during the 2018–2023 period. The original multi-satellite data revealed significant discrepancies, with an annual RMSE of 1534 mm/year, a yearly RSQ value of 0.3, and an annual RBIAS of 27% compared to observational data. Among the datasets, O_CHPS demonstrated the best spatial similarity visually. Calibration using the Geographical Differential Analysis (GDA) method effectively enhanced the accuracy, reducing the annual RMSE to 807 mm/year, increasing the yearly RSQ to 0.5, and lowering the annual bias to 1.6%. Improvements were also noted in monthly and daily rainfall estimates. After calibration, C_PRSN exhibited the most favorable spatial distribution and performance, achieving a 26% reduction in annual RMSE, a 105% increase in annual RSQ, and a 101% decrease in annual bias compared to its initial data. Furthermore, sensitivity to elevation and rainfall intensity was enhanced, with improved detection indicators, particularly for heavy to extreme rainfall events. This included a 43% increase in POD, a 262% increase in CSI, and a 42% reduction in FAR.]]>
