REGIONAL LOW FLOW FREQUENCY ANALYSIS USING BAYESIAN REGRESSION AND PREDICTION AT UNGAUGED CATCHMENT IN KOREA
This study employs the Bayesian multiple regression analysis using ordinary least square method for regional low flow frequency analysis. Especially the estimates of confidence interval using the Bayesian multiple regression analysis are compared with those of confidence interval using t-distribution. In these comparative studies, it is suggested that the mean values with t-distribution and with the Bayesian analysis at each return period are not different seemingly. However, the difference between upper and lower limits using the Bayesian multiple regression is more reduced remarkably than those of conventional method using t-distribution. Therefore, from the point of view of uncertainty analysis, Bayesian multiple regression analysis is a more attractive method than conventional method based on a t-distribution because the sample size of low flow at the site of interest is mostly not enough to perform the at-site low flow frequency analysis. Also, the low flow prediction including confidence interval at two ungauged catchments by the developed Bayesian multiple regression model is performed. Finally, it can be suggested that the Bayesian prediction is effective to infer the low flow characteristic at ungauged catchment.
regional low flow frequency analysis Bayesian multiple regression uncertainty confidence interval using t-distribution ungauged catchment
Sang Ug Kim Kil Seong Lee Kyungshin Park
Post-doc., BK21 SIR Group, Seoul National University Dept.of Civil & Environmental Eng., Seoul National University
国际会议
第16届亚太地区国际水利学大会暨第3届水工水力学国际研讨会(16th IAHR-APD Congress and 3rd Symoposium of IAHR-ISHS)
南京
英文
314-319
2008-10-20（万方平台首次上网日期，不代表论文的发表时间）