作者: Jiandong Liu , Chi Dung Doan , Shie-Yui Liong , Richard Sanders , Anh Tuan Dao
DOI: 10.1007/S11069-014-1363-5
关键词: Meteorology 、 Severe weather 、 Flood myth 、 Quantile 、 Environmental science 、 Statistics 、 Storm 、 Generalized extreme value distribution 、 Homogenization (climate) 、 Confidence interval 、 Return period
摘要: Jakarta is vulnerable to flooding and extreme rainfall events are always the main cause of occurrence heavy flood with loss life property. The in February 2007 one most devastating total during 2-week period amounted almost one-fourth at some stations. huge scale due this event as well a series big consecutive following years draw attention understand risk detail for region. study aimed (1) estimate return Jakarta, (2) develop intensity–duration–frequency (IDF) curves corresponding confidence intervals region up 7-day duration (3) uncertainty developed IDF induced at-site regional approaches. Daily data from 11 stations within record lengths ranging 15 22 years were collected. As reliability frequency analysis influenced by length available, frequencies magnitudes analyzed (RFA) based on method L-moments. One station (Darmaga) was removed after quality homogenization check. annual maximum 1, 2, 3, 4, 5, 6, 7 day 10 then aggregated investigation. Because value discordancy measure exceeded critical value, Tangerang subsequently group. Annual these 9 finally combined passed heterogeneity measures RFA. accuracy quantile estimates RFA assessed compared error bounds level 90 %, Monte Carlo simulations fitting generalized probability distributions average sample L-moment ratios. results showed that over 300 year 3-day Halim station; close 200 150-year periods Pdk Betung stations, respectively. For first time, multi-day together 90 % 10, 30, 50, 100, 200, developed. effective mapping/mitigation longer than 1-day design storm should be considered. Also frequent re-visit indeed important if severe storms occurred established. Finally, suggested more stations/data acquired so estimation robust reliable.