|
Accession number;07A0041882
|
| Title;Frequency Analysis of Larger Samples of Hydrologic Extreme-Value Data-How to estimate the T-year quantile for samples with a size of more than the return period T |
| Author;
TAKARA KAORU
(Kyoto Univ., Disaster Prev. Res. Inst., JPN)
|
Journal Title;Abstracts for Annuals. Disaster Prevention Research Institute, Kyoto University (CD-ROM)
|
Journal Code:S0431B
|
ISSN:
|
|
VOL.;NO.49;PAGE.NO.49B,7-12(2006)
|
| Figure&Table&Reference;FIG.2, REF.3 |
| Pub. Country;Japan |
| Language;Japanese |
| Abstract;This paper describes how to estimate the T-year events (quantiles) for hydraulic systems design and planning in river basins. Traditional frequency analysis methods, which have been providing T-year quantile estimates, include the graphical analysis with probability papers for smaller samples and the parametric approach with a number of probability distribution functions (PDFs) especially for extreme hydrologic variables for larger samples. Recently many hydrological observatories are getting longer historical data and the length of record there is exceeding the return period T. This paper recommends the usage of non-parametric approach for such larger samples, because this approach avoids the difficulty of selecting the best PDF and parameter estimation method from many candidates. The bootstrap method, which is one of the resampling methods, is used for bias correction and quantification of quantile estimation errors. A basic idea and its application results are illustrated here to indicate a direction of hydrologic frequency analysis in the future where more observatories will have longer datasets. (author abst.) |
|
|
|
Related Articles;
|
