西北地质2025，Vol.58Issue(3)：236-245,10.DOI:10.12401/j.nwg.2024118
基于GM(1,1)与BP神经网络模型的西安市地下水位动态特征及趋势预测研究
Dynamic Characteristics and Trend Prediction of Groundwater Level in Xi'an City,China Based on GM(1,1)and BP Neural Network Models
摘要
Abstract
Groundwater is exteremely important in arid and semiarid regions,and the core of its effective pro-tection and rational utilization lies in accurate prediction and evaluation of groundwater dynamics,based on which protection,utilization,and planning strategies are formulated.Based on groundwater level monitoring da-ta from 2010 to 2020 in Xi'an City,this study systematically analyzed the inter-annual and intra-annual dynamic changes in groundwater levels,investigated the main factors influencing groundwater dynamics,and conducted a correlation analysis using SPSS on the two primary factors affecting groundwater dynamics:precipitation and extraction volume.Furthermore,the study utilized the GM(1,1)grey prediction model and the BP neural net-work model to forecast the trend of groundwater level changes.The results indicate that:① From 2010 to 2016,the groundwater level showed an overall decreasing trend.However,from 2016 to 2020,due to the yearly reduc-tion in extraction volume and continuous optimization and improvement of water supply facilities,the ground-water level exhibited a rising trend.② Both precipitation and human extraction significantly impact the ground-water level fluctuations in Xi'an.The depth of the groundwater level is a crucial factor determining the degree of influence from precipitation,with river floodplains being the most sensitive,followed by terraces,and loess plateaus showing the weakest response.The correlation between groundwater extraction volume and groundwa-ter depth is stronger,highlighting its dominant role in regulating groundwater level dynamics.③ Groundwater level predictions suggest that as groundwater extraction continues to decline annually,the overall groundwater in the study area is on a fluctuating upward trend.This study has conducted research on the influencing factors and prediction trends of groundwater dynamics in Xi'an,which has important reference value for groundwater re-source management and sustainable development.关键词
地下水位动态/主导因素/回归分析/灰色模型/BP神经网络预测Key words
groundwater level dynamics/dominant factors/regression analysis/grey model/BP neural net-work prediction分类
天文与地球科学引用本文复制引用
李培月,梁豪,杨俊岩,田艳,寇晓梅..基于GM(1,1)与BP神经网络模型的西安市地下水位动态特征及趋势预测研究[J].西北地质,2025,58(3):236-245,10.基金项目
国家重点研发计划项目课题"土壤-地下水污染时空演化规律及主控因子"(2023YFC3706901),国家自然科学基金面上项目"大型灌区地下水多场协同作用下典型农业污染物迁移转化机制研究"(42472316)联合资助. （2023YFC3706901）
