Modified Standardized Precipitation Index and Evaluation of its Effectiveness using Past Extreme Drought Cases

Abstract

The standardized precipitation index (SPI) was developed based on the fact that if precipitation decreases, a relative shortage of water would occur compared to the required water demand, resulting in drought. However, similar to other drought indexes, the relative standard distribution of the time series of precipitation or runoff is calculated. The analysis suggests that relatively high-rainfall areas occur with no actual occurrence. Hence, modified drought analysis techniques are required to mitigate this phenomenon. Therefore, in this paper, a gamma distribution, calculated for the time series of points in the SPI calculation process, was expanded nationwide, and the drought index was calculated by applying it to the standard regular distribution after the calculation. In addition, the modified SPI (M-SPI) was verified for its effectiveness using the extreme drought and drought duration in areas where there was limited water supply in the past. Consequently, the SPI recognized drought only in 1995, 2001, and 2014, while the drought duration in 2010 and 2014 remained underestimated as a result of the comparison with the year of limited water supply and the results of the SPI and M-SPI. However, the M-SPI determined droughts in 1995, 1996, 2001, and 2010. In the case of Yeosu, although the SPI did not recognize two extreme droughts, M-SPI successfully recognized one extreme drought. In the case of Goheung, both the drought indexes recognized extreme drought. Keywords: Extreme Drought, M-SPI, SPI, Space Time Series, Gamma Distribution