Assessment of the reliability of three gauged-based global gridded precipitation datasets for drought monitoring

Abstract

This study evaluates the reliability on three global gridded precipitation dataset including CRU-TS-V.4.01, GPCC-V.7 and UDEL-V.4.01 to drought monitoring over Lake Urmia basin in Iran using standardised precipitation index (SPI). Assessing the precipitation data of global datasets indicate that the major error in precipitation of global datasets is bias error which means that they can recognise the behaviour of precipitation time series but there is a difference in estimation of precipitation amount. To diminish the bias error, bias correction (BC) was done before evaluating the datasets in drought monitoring. The results indicated that GPCC is the most suitable dataset over the study area for drought studies. In addition, results revealed that there is no significant difference between original and BC data in drought monitoring. For example, in nine-month SPI, the critical success index (CSI) for original CRU and CRU-BC were equal to 71 and 76%, respectively, and the GPCC and UDEL in both original and BC versions were equal to 88 and 55%, respectively. Also, it can be claim that the bias error is not important in drought studies due to the fact that during calculation of SPI, standardisation process will remove bias significantly.