Return Levels of Rainfall Extremes during the Typhoon Season in Taiwan under a Changing Climate

This study investigates changes in return levels of rainfall extremes during the typhoon season in Taiwan based on the non-stationary Generalized Extreme Value (NGEV) analysis. A nonparametric Mann-Kendall method and Sen’s test is used to detect the trend. Data used are the annual maximum 24-hr rainfall induced by typhoons for 20 stations and the Ocean Nino Index (ONI) compiled by the Climate Prediction Center of the National Centers of Environmental Prediction, NOAA from 1958 to 2013. Three return periods (2-yr, 20-yr, and 100-yr) are chosen. The return period, also known as recurrence interval, is interpreted to be the average time between occurrence of events of that magnitude or greater. It is commonly used for engineering design, risk analysis, environmental regulation, and flood insurance policy.

The NGEV model is governed by three parameters: the location, the scale, and the shape parameters. The return level can be expressed as a function of the aforementioned parameters and the return period. It has the same unit as rainfall itself. Because the variability of the shape parameter is small, we focus on the influence of the location and scale parameters on three return levels.

Results show that a majority of stations (18 out of 20) experienced increasing trends in return levels over the last 56 years. For some stations, the return-level threshold values are found to increase with time considerably. For example, at Alishan in the Central Mountain Range of Taiwan, an event with a 20-yr return level in 1958 is about 1,000 mm. By 2013, it rises to approximately 1,300 mm. This amounts to an increase of 30% over the last 56 years. Independently, rainfall extremes during the typhoon season are co-varied with the ONI of the preceding winter. A positive relationship is found between the rainfall extremes during the typhoon season and ONI for northern and eastern Taiwan, implying greater extreme events following El Nino winters and the opposite for La Nina years