Assessing the Performance and Limitations of PERSIANN-CDR: A Case Study in the São Francisco River Basin

Abstract

This study investigates the performance and limitations of the PERSIANN-CDR precipitation estimates dataset in comparison to strategically selected rain gauge stations within the São Francisco River Basin. The study covers the period from January 1989 to December 2018. The results reveal distinct patterns in the performance of PERSIANN-CDR. Notably, the dataset consistently underestimates precipitation at stations closest to the outlet, with a bias (PBIAS) of approximately -20%. Conversely, it tends to overestimate precipitation at other stations, with a bias (PBIAS) of around 40%. Furthermore, PERSIANN-CDR precipitation estimates exhibited variations in their ability to capture precipitation variability. They performed well at stations in the middle and upper São Francisco regions but fell short at stations near the sub-middle and lower São Francisco. Additionally, we highlight the tendency of PERSIANN-CDR to overestimate precipitation variability, especially during periods of heavy rainfall, such as the rainy season from October to March.