event
Tuesday 03 Mar 2015: Experiences of Hydrological Frequency Analysis in Taiwan – Coping with Extraordinary Extremes and Climate Change
Professor KE-SHENG CHENG, - National Taiwan University???
Harrison 170 14:00-15:00
Taiwan is among the countries which are most vulnerable to natural disasters. Heavy rainfalls of typhoons are the major cause of most natural disasters in Taiwan. Increasing number of record breaking heavy rainfalls have been observed at different locations in Taiwan in recent years. Such extraordinary extremes are considered outliers and pose difficulties in hydrological frequency analysis. In addition, many studies have found that rainfall intensities of typhoons are likely to increase under climate change scenarios. Thus, assessing the impact of climate change on design rainfall depths is crucial in coping with climate change and forming adaptation measures. In this talk, two focal issues will be addressed:
(1) How to deal with outliers and how to estimate the return periods of concurrent occurrences of multi-site rainfall extremes (i.e. multi-site extreme events)?
(2) How can the GCM outputs and stochastic modeling be coupled to assess the impact of climate change on design rainfalls?
Stochastic modeling and simulation plays an important role in tackling these problems. For the first issue, we developed a multi-site event-maximum typhoon rainfalls simulation model which can generate many realizations of event-maximum typhoon rainfalls at different locations. Using these samples, the effect of site-specific outliers can be reduced and the return periods of multi-site extreme events can be estimated. For the second issue, a stochastic storm rainfall simulation model (SSRSM) was developed. The SSRSM considers storm occurrences, storm duration, event-total rainfalls and time distribution of event-total rainfalls as random processes or random variables. Changes in the expected values of these storm characteristics were estimated using high resolution GCM outputs. By coupling GCM outputs and SSRSM, a large number of annual sequences of hourly rainfalls under certain climate change scenarios were simulated and used as the basis for assessing the impact of climate change on design rainfalls. Example applications in Taiwan will be demonstrated in this talk.
