Akademik Veri Yönetim Sistemi
CARIBBEAN JOURNAL OF SCIENCE, cilt.55, sa.2, ss.414-441, 2025 (SCI-Expanded)
The aim of this study is to evaluate the performance of four different Satellite Precipitation Products (SPPs) (GPM-IMERG, PERSIANN-CDR, CHIRPS, and MSWEP) in southwestern Haiti. For this purpose, rain gauge data for the period 2015–2018 were used and compared with SPPs data. For a comparison, quantitative (KGE and its components, RMSE) and qualitative (POD, FAR, CSI) statistical parameters were calculated on a daily and monthly basis. The same parameters were also calculated using each SPP for different precipitation classes. In order to perform the analysis, satellite pixel values corresponding to the coordinates of the rainfall stations provided for the study area were extracted with computer software. The results of the study showed that GPM-IMERG, PERSIANN-CDR, CHIRPS, and MSWEP have correlations with rain gauge data of 0.52, 0.42, 0.27, and 0.52 on a daily basis, and 0.74, 0.71, 0.42, and 0.79 on a monthly basis, respectively. As the amount of precipitation increases, the RMSE value tends to increase. GPM-IMERG, PERSIANN-CDR, CHIRPS, and MSWEP satellite products have POD and FAR values of 0.65, 0.58, 0.32, and 0.96, and 0.53, 0.58, 0.49, 0.62 on a daily basis and 1, 0.96, 1, 1, and 0.06, 0.00, 0.04, and 0.06 on a monthly basis, respectively. All SPPs overestimated precipitation for no rain (0–1 mm) and light rain (1–5 mm) classes. Overall, it was determined that MSWEP had the best performance, while CHIRPS had the lowest performance. Similar studies should be carried out in other regions of the world with different climatic characteristics and in other regions of Haiti in order to determine the SPP to be used in future studies. The rainfall satellite data are an important input for determining the spatial distribution of precipitation, especially in large areas where rain gauge data are not available or are scarce.