A homogeneity-based feature selection algorithm for regionalization of watersheds

被引：5

作者：

Ahani, Ali ^{[1
]}

Nadoushani, S. Saeid Mousavi ^{[1
]}

Moridi, Ali ^{[1
]}

机构：

[1] Shahid Beheshti Univ, Fac Civil Water & Environm Engn, Dept Water Resources Management, Tehran, Iran

来源：

HYDROLOGICAL SCIENCES JOURNAL-JOURNAL DES SCIENCES HYDROLOGIQUES | 2020年 / 65卷 / 02期

关键词：

feature selection; homogeneity; regionalization; regional flood frequency analysis; FLOOD FREQUENCY-ANALYSIS; VARIABLE SELECTION; UNGAUGED SITES; DRAINAGE BASINS; CLASSIFICATION; REGIONS; IDENTIFICATION; SIMILARITY; STREAMFLOW;

D O I：

10.1080/02626667.2019.1686638

中图分类号：

TV21 [水资源调查与水利规划];

学科分类号：

081501 ;

摘要：

An algorithm is proposed to find an appropriate selection of available watershed features for the regionalization of watersheds. Also, the algorithm is able to determine an appropriate number of regions for regional flood frequency analysis. The proposed algorithm is applied to data related to the Karun-e-Bozorg basin in the southwest of Iran. The results show that the proposed algorithm is efficient to find an appropriate selection of watershed features and an appropriate number of regions for regionalization, in the study area. It decreased the number of assessed regionalization states from 1020 to 252. In addition, applying the proposed algorithm mostly improved the accuracy of flood quantile estimates in the study area. The performance of the algorithm was evaluated through Monte Carlo simulation experiments. The results indicate that the efficiency of the algorithm increases by increasing the number of available watershed features.

引用

页码：269 / 280

页数：12

共 55 条

[1] CLASSIFICATION OF DRAINAGE BASINS ACCORDING TO THEIR PHYSICAL CHARACTERISTICS - AN APPLICATION FOR FLOOD FREQUENCY-ANALYSIS IN SCOTLAND [J].