手法を比較

	Peaks-Over-Threshold Flood Analysis	Flood Frequency Analysis
分野	Disaster Studies	Disaster Studies
系統	Process / pipeline	Process / pipeline
提唱年	1999	2018
提唱者	M. Lang, T. B. M. J. Ouarda & B. Bobée (operational POT guidelines); Pickands–Balkema–de Haan theory	Emil J. Gumbel; J. R. M. Hosking & J. R. Wallis (GEV/PWM); USGS Bulletin 17C
種類	Threshold-exceedance extreme-value frequency pipeline	At-site extreme-value frequency estimation pipeline
原典	Lang, M., Ouarda, T. B. M. J., & Bobée, B. (1999). Towards operational guidelines for over-threshold modeling. Journal of Hydrology, 225(3-4), 103-117.	England, J. F., Jr., Cohn, T. A., Faber, B. A., Stedinger, J. R., Thomas, W. O., Jr., Veilleux, A. G., Kiang, J. E., & Mason, R. R., Jr. (2018). Guidelines for Determining Flood Flow Frequency — Bulletin 17C. U.S. Geological Survey Techniques and Methods, book 4, chap. B5, 148 p.
別名	POT Flood Analysis, Partial Duration Series Analysis, Generalized Pareto Flood Modeling, Threshold Exceedance Flood Frequency	At-Site Flood Frequency Analysis, Annual Maximum Flood Frequency, Extreme Value Flood Analysis, Design Flood Estimation
関連	3	3
概要	Peaks-over-threshold (POT) flood analysis models every independent flood peak that exceeds a chosen high threshold, rather than only the single largest peak in each year. The number of exceedances in time is treated as a Poisson process and the amounts by which peaks exceed the threshold are modeled with the Generalized Pareto distribution — the extreme-value limit for threshold exceedances given by the Pickands-Balkema-de Haan theorem. Because a wet year may contain several damaging floods and a dry year none, POT (also called the partial duration series) uses the data more efficiently than the annual-maximum approach, which is why Lang, Ouarda, and Bobée's 1999 operational guidelines and USGS Bulletin 17C both treat it as a key complement to annual-maximum frequency analysis. The method delivers the same design-flood quantiles for chosen return periods, often with lower variance at short return periods.	Flood frequency analysis estimates how often floods of a given magnitude occur at a river site by fitting an extreme-value probability distribution to the record of annual maximum discharges and then inverting it to read off design floods for specified return periods. The classical approach uses the Gumbel distribution, the limiting form for maxima of light-tailed variables; the more general Generalized Extreme Value (GEV) distribution adds a shape parameter that lets the tail be lighter or heavier, while the log-Pearson Type III distribution is the U.S. federal standard codified in USGS Bulletin 17C. Hosking, Wallis, and Wood's 1985 work on probability-weighted moment estimation of the GEV made robust at-site fitting practical, and Bulletin 17C (England et al., 2018) sets out the modern operational procedure. The output — the 100-year flood, the 500-year flood — underpins dam design, floodplain mapping, and infrastructure standards worldwide.