Evaluation of GPM IMERG Product Against Ground Station Rainfall Data in Semi-Arid Region

Benanain River is the longest and largest river on Timor Island, with a length of 132 km and an area of 6,460.12 km². In this region, a significant factor affecting the presence of surface water sources is rainfall. To compensate for the lack or unavailability of automatic Rainfall Data (RD) in the Benanain River Basin (BRB), Global Rainfall Measurement (GPM) data from 1998 to 2018 (20 years) were used. The accuracy of GPM rainfall analysis was obtained when parameter conformity and compatibility with data recorded at Rainfall Station (RS) were maintained. The difficulty of predicting rainfall values, spatially and temporally, in the field led to data gaps and unreliable data for analysis needs. Additionally, RD obtained from observation stations contributed to measuring rainfall because there was insufficient RD for analysis in a few regions. The challenge of accurately predicting rainfall values in the field led to differences in data, rendering it unreliable for analysis. To address this issue, satellite data was required as an alternative method to estimate rainfall. Among a total of 7 RS, only 2 passed rainfall characteristic tests. Following this discussion, Lahurus station showed a correlation coefficient of 0.7046, an RMSE of 25.89, and an NSE of 0.476. In addition, the rainfall characteristic test result for Haliwen Station was 1.66 (R100/R2). The second station that passed was Kaubele Station, signifying a correlation coefficient of 0.7907, RMSE of 25.28, and NSE of 0.604. Additionally, the rainfall characteristic test result for Haliwen Station was 3.04 (R100/R2) and the daily performance of the GPM product in the rainy season with low rainfall (≤ 50 mm) was better compared to extreme rainfall (≥ 100 mm). In this study, corrected GPM daily RD in the range >100 mm was underestimated. This analysis implied that the GPM IMERG Final Run product on daily and monthly rainfall timescales had strong detection capabilities and provided data support for long-time series investigations on Timor Island.