Investigation of optimal probability distribution of Standard Precipitation Index for daily precipitation time series in Southern Korean Peninsula

The Standardized Precipitation Index (SPI) is applied worldwide for drought assessment. In general, in many studies, SPI was estimated from a two-parameter gamma distribution. However, in other climatic regions, there are also studies that suggest that distributions other than the Gamma distribution are more suitable. In addition, as the frequency of drought events increases, the need for daily SPI calculated with relatively short time-scales for immediate drought response is increasing. In this study, the optimal probability distribution for estimating SPI using daily precipitation in the southern part of the Korean Peninsula was explored. Gumbel, Gamma, GEV, Loglogistic, Lognormal, and Weibull are applied as candidate distributions, and optimal distributions for each season, region, and time-scale are investigated. The Chi-square test was applied to investigate the probability distribution function appropriate to the cumulative daily precipitation time series for various time-scales. In the process of calculating the SPI, when the cumulative daily precipitation has a value of 0, the cumulative probability value was calculated by reflecting the probability of having a value of 0. Then, by applying the candidate distribution, it was verified whether the estimated SPI conformed to the standard normal distribution. Finally, a more precise drought assessment could be performed by determining the optimal probability distribution for each region, season, and time-scale. It is also expected to increase the applicability of daily SPI by reducing problems that occur in a short time-scale.

 

Acknowledgement

This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (No. NRF-2022R1A2B5B01001750).