Quantitative analysis of input data uncertainty for SPI and SPEI in Peninsular Malaysia based on the bootstrap method

Abstract

ABSTRACT Drought assessment has attracted attention in the research community, especially regarding the accuracy of drought indices due to input data uncertainty. This study addressed the impacts of input data uncertainty on the estimation of the Standardized Precipitation Index (SPI) and Standardized Precipitation Evapotranspiration Index (SPEI) for drought assessment in Peninsular Malaysia. The Kolmogorov-Smirnov test recommended the gamma distribution function for the SPI and the log-logistic distribution function for the SPEI. The bootstrap method was used to estimate SPIU and SPEIU, which account for input data uncertainty, and provided estimates for SPI and SPEI values. However, the standard deviation indicated significant input data uncertainty, with values ranging from 0.1038 to 0.1378. The two drought indices exhibited similar classifications of drought categories, but SPIU showed greater uncertainty for very dry and extremely dry events. The findings emphasize the importance of input data uncertainty, especially when dealing with extreme drought events.