Nonstationary drought index analysis of wet-season rainfall in Taiwan

Abstract

Drought is a natural disaster that has devastating effects on agriculture water resouces ecology and socirty Global climate change and social development would induce changes of the rate of hydrologic cycles Intensified frequency and intensity of hydrologic extreme events lead to the hypothesis of stationarity in traditional hydrologic analysis no longer applicable Nonstationary analysis of hydrologic characteristics thus becomes one of the important issues in water resources management recently In this study the generalized additive models for location scale and shape (GAMLSS) is adopted for the nonstationary drought analysis in Taiwan The standardized precipitation index (SPI) is used in this study to characterize droughts The wet-season (May-October) rainfall and annual rainfall at eight rainfall stations including Taipei Chengkung Ilan Taichung Tawu Sun Moon Lake Kaohsiung and Hengchun located in the north central south and east regions of Taiwan are used in this study The data length covers the period of 1897-2017 1947-2017 1947-2017 1947-2017 1940-2017 1942-2017 1932-2017 and 1947-2017 for these eight stations respectively In this study the Mann-Kendall trend test and the stationary and nonstationary drought index analysis were used for the annual wet-season rainfall and annual rainfall in each stations The nonstationary standardized precipitation index (NSPI) can be established by using time-varying location and scale parameters in the gamma distribution The results show that the temporal variations of the stationary and the nonstationary standardized precipitation index of the annual wet-season rainfall and annual rainfall are similar Differences between the stationary and the nonstationary SPI values is mainly caused by the change of the probability density function over time

Date of Award	2019
Original language	English
Supervisor	Jenq-Tzong Shiau (Supervisor)