华北水利水电大学学报（自然科学版）2025，Vol.46Issue(4)：1-9,9.DOI:10.19760/j.ncwu.zk.2025052
基于自组织映射神经网络的黄河流域干旱聚类分析
Drought Cluster Analysis in Yellow River Basin Based on Self-Organizing Map
摘要
Abstract
[Objective]The spatiotemporal variations of historical meteorological drought characteristics in the Yellow River Basin are investigated through cluster analysis of drought indices across multiple temporal scales.[Methods]The spatiotem-poral patterns of meteorological drought characteristics in the Yellow River Basin from 1950 to 2020 were evaluated by calcu-lating the Standardized Precipitation Index(SPI),Standardized Precipitation Evapotranspiration Index(SPEI),and Self-Calibrating Palmer Drought Severity Index(sc-PDSI)at different time scales.A drought clustering method based on Self-Or-ganizing Feature Map(SOM)was proposed to analyze the relationships and discrepancies among various drought indices.[Results]From 1950 to 2020,significant upward trends in SPI,SPEI and sc-PDSI values were observed in the source and upper reaches of the Yellow River Basin,whereas downward trends were detected in the middle and lower reaches.These findings indicated a mitigation of meteorological drought conditions in the upper reaches of the Yellow River Basin,contrasted with an intensification of drought in the middle and lower reaches.The meteorological drought patterns in the Yellow River Basin exhibited significant spatial heterogeneity.The source and upper reaches were marked by low frequency,long duration,and high intensity,whereas the middle and lower reaches were characterized by high frequency,short duration,and low in-tensity.The selection of drought indices significantly influenced the assessment outcomes of meteorological drought in the Yellow River Basin.As the number of cumulative months increased,the SPEI and SPI indices facilitated the identification of drought events characterized by low frequency,long duration,and low intensity.[Conclusion]Against the backdrop of war-ming and wetting in Northwest China,meteorological drought in the Yellow River Basin from 1950 to 2020 exhibited a pattern of mitigation in the upper reaches and intensification in the middle and lower reaches.The degree of this variation is closely related to the selection of drought indices.关键词
气象干旱/黄河流域/自组织特征映射网络/干旱指数/时空变化特征Key words
meteorological drought/Yellow River Basin/self-organizing feature map/drought index/spatiotemporal variation patterns分类
建筑与水利引用本文复制引用
于磊,李雅静,朱大炯,刘元锐..基于自组织映射神经网络的黄河流域干旱聚类分析[J].华北水利水电大学学报（自然科学版）,2025,46(4):1-9,9.基金项目
国家重点研发计划项目(2023YFC3205600) （2023YFC3205600）
河南省水利厅水利科技攻关项目(GG202125). （GG202125）
