This study compares seven global gridded daily precipitation datasets against gauged precipitation to evaluate their accuracy for capturing extreme precipitation in Iran. We evaluated the performance of satellite-based (CHIRPS and MSWEP-V220), reanalysis-based (CFSR and MERRA-2), ensemble-based (MRE3ensemble), and “bias-correction”-based (MRE3ensemble, EWEMBI, and W5E5) precipitation datasets for the period of 1980–2016. The extreme precipitation indices that we examined consist of intensity indices [the maximum consecutive 1-day precipitation (Rx1day) and simple precipitation intensity (SDII)], duration indices [the consecutive dry days (CDD) and the consecutive wet days (CWD)], and frequency indices [heavy precipitation events (R10mm) and very heavy precipitation events (R20mm)]. The results showed that MSWEP-V220 had the best performance in Iran and Bias-Correction W5E5 was the second-best dataset to estimate precipitation in Iran. Although RMSE and MBE statistics showed high error and bias for all precipitation datasets in northern Iran, the evaluation of the PBIAS showed the share of bias value in the northern regions of Iran compared to the total precipitation in the climate zone of Iran is less than 5%. In contrast, most datasets showed a high percentage of bias in the eastern and interior regions of Iran. The results showed that all the studied datasets in the rainy areas of Iran (Cfa, Csa, and Dsa) underestimate maximum one-day precipitation (Rx1day), precipitation intensity (SDII), and heavy and very heavy precipitation (precipitation > 10 and 20 mm). In addition, MERRA-2 and CFSR overestimate the indices related to intensity and frequency in the most desert (BW) and semi-desert (BS) climates of Iran, respectively. CHIRPS data in all climate zones of Iran—except the CWD index in Cfa climate zone – overestimate the CDD index and underestimated the CWD. Accordingly, CHIRPS data show a drier climate for Iran unrealistically.

Keywords