Investigating Trends in Regional Hydrological Characteristics Under Climate Change Using Long-term Rainfall Observation Data: A Case Study of Taipei and Tainan, Taiwan

With the increasing impacts of climate change in recent decades, numerous studies have reported a rising frequency of extreme rainfall events worldwide, accompanied by intensified droughts and floods. Consequently, traditional hydrological analyses require further consideration of climate change effects.This study analyzes more than 125-year record of daily rainfall observations (1900–2025) from the Taipei and Tainan meteorological stations in Taiwan. The dataset is divided into a historical period (1900–1990) and a recent period (1991–2025) to investigate long-term variations in extreme rainfall and drought characteristics. Three hydrological indicators are examined: (1) changes in return periods of extreme rainfall, (2) trends in the number of non-rainy days, and (3) months of extreme precipitation occurrence.In the return period analysis, the Annual Maximum Series (AMS) method combined with the Weibull plotting position formula was applied. The results reveal a decreasing trend in return periods at both the Taipei and Tainan stations, indicating an increased frequency of extreme rainfall events.In terms of drought characteristics, long-term variations in non-rainy days were examined based on the concept of an accelerated hydrological cycles. The results show that the number of non-rainy days has increased at a rate of approximately 0.3 days per year in both northern and southern Taiwan. Furthermore, the average annual number of non-rainy days in the recent period increased by approximately 23 days compared to the historical period, reflecting climate characteristics associated with reduced light rainfall and intensified drought–flood extremes.Regarding seasonal variability, the probability distribution of extreme precipitation occurrence by month was analyzed. The Taipei station exhibits an expansion of the flood season, with extreme precipitation events primarily occurring in June and October, forming a bimodal distribution. In contrast, the Tainan station shows a pronounced concentration of extreme precipitation during the wet season, with approximately 43.8% of events occurring in August.Based on these findings, it is recommended that flood control design standards be upgraded to account for shortened return periods of extreme rainfall. In addition, water resource allocation and management strategies should be strengthened to mitigate the increasing risk of water shortages associated with the rise in non-rainy days. Flood warning systems and construction planning should also be dynamically adjusted in response to shifts in the occurrence months of extreme precipitation.