[1]韩尚,杨娜,聂青,等.基于K-means聚类算法的中小流域雨型推求[J].江苏水利,2025,(10):30-33.
 HAN Shang,YANG Na,NIE Qing,et al.Derivation of rainfall patterns in small and medium-sized watersheds based on the K-means clustering algorithm:a case study of the hilly areas in Xuyi, Yizheng and Liuhe[J].JIANGSU WATER RESOURCES,2025,(10):30-33.
点击复制

基于K-means聚类算法的中小流域雨型推求()

《江苏水利》[ISSN:1006-6977/CN:61-1281/TN]

卷:
期数:
2025年10期
页码:
30-33
栏目:
水利工程管理
出版日期:
2025-10-01

文章信息/Info

Title:
Derivation of rainfall patterns in small and medium-sized watersheds based on the K-means clustering algorithm:a case study of the hilly areas in Xuyi, Yizheng and Liuhe
文章编号:
1007-7839(2025)10-0030-0004
作者:
韩尚12杨娜3聂青1陆小明1纪小敏1
(1.江苏省水文水资源勘测局,江苏 南京 210029;2.河海大学 水文水资源学院,江苏 南京 210098;3.南京信息工程大学 水文与水资源工程学院,江苏 南京 210044)
Author(s):
HAN Shang12 YANG Na3 NIE Qing1 LU Xiaoming1 JI Xiaomin1
(1.Jiangsu Province Hydrology and Water Resources Investigation Bureau, Nanjing 210029, China; 2.College of Hydrology and Water Resources, Hohai University, Nanjing 210098, China; 3.School of Hydrology and Water Resources, Nanjing University of Information Science and Technology, Nanjing 210044, China)
关键词:
雨型K-means聚类聚类数确定
Keywords:
rainfall pattern K-means clustering determination of clustering number
分类号:
TV125
文献标志码:
B
摘要:
山丘区水文资料相对匮乏,雨型研究对于产汇流的计算具有关键意义。以盱眙仪征六合山丘区为例,选取其中24个雨量站在1960—2023年汛期发生的大暴雨场次中的最大24 h降水数据作为样本,使用K-means聚类算法对其进行聚类分析并概化出各聚类的雨型。结果表明,K-means聚类的最佳聚类数为4,聚类后第四类包含的降雨样本最多,且四类降雨的概化雨型之间差异明显。
Abstract:
Hydrological data in hilly areas are relatively scarce, and rainfall pattern research is of crucial significance for the calculation of runoff generation and confluence. Taking the hilly areas in Xuyi, Yizheng and Liuhe as an example, the maximum 24-hour precipitation data from heavy rain events during the flood season (1960—2023) at 24 rainfall stations in the region were selected as samples. The K-means clustering algorithm was used to conduct clustering analysis on these samples and generalize the rainfall pattern of each cluster. The results show that the optimal number of clusters for K-means clustering is 4; the fourth cluster contains the largest number of rainfall samples after clustering, and there are significant differences between the generalized rainfall patterns of the four types of rainfall.

参考文献/References:

[1]李致家,朱跃龙,刘志雨,等. 中小河流洪水防控与应急管理关键技术的思考[J]. 河海大学学报(自然科学版),2021,49(1):13-18.
[2]白云鹏,郑爱民,王文川,等. 基于深度学习的山丘区中小河流洪水预报误差校正方法[J]. 水文,2025,45(1):37-44.
[3]章可颖,江爱秋,刘桢义. 中小河流淹没分析及防洪情景库构建[J]. 水利信息化,2025(2):45-49.
[4]杨俊闯,赵超. K-Means聚类算法研究综述[J]. 计算机工程与应用,2019,55(23):7-14,63.
[5]刘媛媛,刘洪伟,霍风霖,等. 基于机器学习短历时暴雨时空分布规律研究[J]. 水利学报,2019,50(6):773-779.
[6]LIN W,LIU Y,LI N,,et al. Study of the spatiotemporaldistribution characteristics of rainfall using hybrid dimensionality reduction-clustering model:a case study of Kunming City,China[J]. Atmosphere,2024,15(5):534.

备注/Memo

备注/Memo:
收稿日期:2025-05-20
基金项目:江苏省水利科技项目(2024008)
作者简介:韩尚(2001—),男,硕士研究生,研究方向为水文水资源。E-mail:185109728@qq.com
通信作者:杨娜(1983—),女,副教授,博士,研究方向为水资源系统优化配置与管理、气候变化与水循环。E-mail:yangna@nuist.edu.cn
更新日期/Last Update: 2025-10-01