中南民族大学学报（自然科学版）2026，Vol.45Issue(6)：744-753,10.DOI:10.20056/j.cnki.ZNMDZK.20260603
基于深度学习的巢湖富营养化预测
Prediction of eutrophication in Chaohu Lake based on deep learning
摘要
Abstract
As a pivotal freshwater lake in China,Chaohu Lake has persistently grappled with severe water pollution and eutrophication issues.Drawing upon water quality data spanning from 2014 to 2021,three models—CNN,LSTM,and CNN-LSTM—were utilized to forecast and contrast the trophic level index(TLI)of Chaohu Lake,with the objective of unraveling its spatiotemporal evolutionary dynamics.The findings reveal that the CNN-LSTM hybrid model exhibits the highest predictive efficacy(R2=0.8054),making it apt for water quality prediction in intricate aquatic settings.The eutrophication condition of Chaohu Lake manifests a spatial differentiation trend characterized by"the northwestern lake region>the central lake region>the eastern lake region."The northwestern area continues to be in a state of moderate eutrophication,although certain zones exhibit signs of improvement.Over the prediction period from 2022 to 2030,the TLI values across the lake are predominantly lower than historical benchmarks,implying that the prevailing watershed management initiatives have yielded tangible positive outcomes.These discoveries clarify the spatiotemporal differentiation patterns of eutrophication in Chaohu Lake,offering crucial insights for optimizing watershed management protocols and enhancing precise and effective prevention and control strategies.关键词
巢湖/富营养化/水质预测/深度学习Key words
Chaohu Lake/eutrophication/water quality prediction/deep learning分类
资源环境引用本文复制引用
龚肇青,牛宏,林晶晶,黄小龙,杨超,余明星..基于深度学习的巢湖富营养化预测[J].中南民族大学学报（自然科学版）,2026,45(6):744-753,10.基金项目
国家重点研发计划资助项目(2023YFC3205600) （2023YFC3205600）
