基于小波去噪的LSTM-Transformer模型在河流断面水质预测分析中的应用研究

许轲桐; 曾鸿滨; 陈丽萍; 郑茜匀; 陈航; 谢晓婧; 袁婧; 韦朝海; 邱光磊

doi:10.13205/j.hjgc.202603011

基于小波去噪的LSTM-Transformer模型在河流断面水质预测分析中的应用研究

doi: 10.13205/j.hjgc.202603011

许轲桐¹,
曾鸿滨¹,
陈丽萍¹,
郑茜匀¹,
陈航¹,
谢晓婧¹,
袁婧¹,
韦朝海^1,2,3,
邱光磊^1,2,3,4, ,

1. 华南理工大学环境与能源学院，广州 510006;
2. 工业聚集区污染控制与生态修复教育部重点实验室，广州 510006;
3. 固体废物污染控制与资源化广东省重点实验室，广州 510006;
4. 国家黄河流域生态保护和高质量发展联合研究中心，北京 100012

基金项目:

国家自然科学基金项目（52270035，51808297）；广东省自然科学基金项目（2021A1515010494）；广东省珠江人才计划（2019QN01L125）；广州市重点研发计划（2023B03J1334）

详细信息

作者简介:
许轲桐（2002—），男，硕士研究生，主要研究方向为水环境大数据科学与工程。xuketong@scut.edu.cn

通讯作者:
邱光磊（1984—），男，教授，主要研究方向为水污染控制理论与技术。qiugl@scut.edu.cn

计量
- 文章访问数: 12
- HTML全文浏览量: 4
- PDF下载量: 0
- 被引次数: 0
出版历程
- 收稿日期: 2026-02-01
- 网络出版日期: 2026-04-11
- 刊出日期: 2026-03-01

Application of a wavelet denoising-based LSTM-Transformer model for water quality prediction at river cross-sections

1. School of Environment and Energy，South China University of Technology，Guangzhou 510006，China;
2. Key Laboratory of Pollution Control and Ecological Restoration in Industrial Clusters，Ministry of Education，Guangzhou 510006，China;
3. Guangdong Provincial Key Laboratory of Solid Wastes Pollution Control and Recycling，Guangzhou 510006，China;
4. National Joint Research Center for Ecological Conservation and High Quality Development of the Yellow River Basin，Beijing 100012，China

摘要

摘要: 提出了一种基于小波去噪的长短时记忆网络（LSTM）-Transformer模型。以我国华南地区2个市控河流断面2021—2024年的逐小时水质监测数据（水温、浊度、pH、电导率、溶解氧）为基础，首先采用离散小波变换进行去噪，随后构建了融合LSTM与Transformer的预测模型。研究结果表明：该模型在2个站点对溶解氧（DO）未来4 h的预测中均表现出色［站点1：决定系数（R²）=0.8015，平均绝对误差（MAE）=0.5169 mg/L，均方根误差（RMSE）=0.8494 mg/L；站点2：R²=0.8873，MAE=0.4456 mg/L，RMSE=0.7143 mg/L］，性能显著均优于单一的LSTM、Transformer模型（R²平均提高5.7%，MAE和RMSE平均分别降低20.2%和10.4%）。此外，应用SHAP可解释性方法进行特征重要性分析与全局影响解读，揭示了影响DO的关键水质因子及其与DO之间复杂的非线性关系具有显著的站点异质性，强调了结合具体环境背景（如地理特征、水文条件、污染源分布等）进行机理解释的必要性。研究结果可为区域河流水质的高精度实时预测与智慧化管理提供一种有效且可解释的技术参考。
- 水质预测 /
- 深度学习 /
- 小波去噪 /
- LSTM-Transformer模型 /
- SHAP可解释性分析
Abstract: This study proposed a hybrid Long Short-Term Memory （LSTM）-Transformer model integrated with wavelet denoising for water quality prediction. Using hourly monitoring data （water temperature， turbidity， pH， conductivity， and dissolved oxygen） collected from two municipally controlled river cross-sections in South China from 2021 to 2024， the discrete wavelet transform was first applied for noise reduction. Subsequently， a predictive model combining LSTM and Transformer architectures was constructed. Experimental results demonstrated that the proposed model achieved outstanding performance in predicting dissolved oxygen （DO） concentrations for the next four hours at both sites （Site 1： coefficient of determination （R²）=0.8015， mean absolute error （MAE）=0.5169 mg/L， root mean square error （RMSE）=0.8494 mg/L； Site 2： R²=0.8873， MAE=0.4456 mg/L， RMSE=0.7143 mg/L）， significantly outperforming standalone LSTM and Transformer models （the R² of the proposed model increased by 5.7%， while MAE and RMSE decreased by 20.2% and 10.4%， respectively）.Furthermore， the SHAP interpretability method was employed for feature importance analysis and global impact interpretation， revealing that the key water quality factors influencing DO and their complex nonlinear relationships exhibited significant site-specific heterogeneity. This underscores the necessity of incorporating specific environmental contexts （e.g.， geographical features， hydrological conditions， and pollution source distribution） for mechanistic interpretation. The findings of this study provide an effective and interpretable technical reference for high-precision real-time prediction and intelligent management of regional river water quality.
- water quality prediction /
- deep learning /
- wavelet denoising /
- LSTM-Transformer model /
- SHAP interpretability analysis

HTML全文

参考文献(32)

[1]	ZHANG P，PANG Y，PAN H C，et al. Factors contributing to hypoxia in the Minjiang River estuary，southeast China[J]. International Journal of Environmental Research and Public Health，2015，12（8）：9357-9374
[2]	SHEN M H. Some reflections on China’s 15th Five-Year Plan for ecological civilization construction：Insights from studying the spirit of the Fourth Plenary Session of the 20th Central Committee of the Communist Party of China[J]. Ecological Economy，2026，42（1）：1-6. 沈满洪. 我国“十五五”生态文明建设规划的若干思考：学习党的二十届四中全会精神体会[J]. 生态经济，2026，42（1）：1-6.
[3]	曾鸿滨，龙琦，高景恒，等. 机器学习在河流断面水质预测分析中的应用[J/OL]. 环境工程，1-14[ 2026-03-25]. https：//link.cnki.net/urlid/11.2097.X.20250902.1655.012. ZENG H B，LONG Q，GAO J H，et al. Application of machine learning in water quality prediction and analysis in river segments[J/OL]. Environmental Engineering，1-14[ 2026-03-25]. https：//link.cnki.net/urlid/11.2097.X.20250902.1655.012.
[4]	JIANG Y，CHEN Z，XIANG D X，et al. Spaceborne multi-sensor combination method for water resources monitoring tasks[J]. Journal of Yangtze River Scientific Research Institute，2024，41（12）：162-170. 姜莹，陈喆，向大享，等. 面向水资源监测任务的星载多传感器组合方法[J]. 长江科学院院报，2024，41（12）：162-170.
[5]	ZHI W，APPLING A P，GOLDEN H E，et al. Deep learning for water quality[J]. Nature Water，2024，2：228-241.
[6]	TEICHERT N，BORJA A，CHUST G，et al. Restoring fish ecological quality in estuaries：Implication of interactive and cumulative effects among anthropogenic stressors[J]. Science of The Total Environment，2016，542：383-393.
[7]	ZHANG Q S，WEI Y L，HOU J，et al. AEGAN-Pathifier：a data augmentation method to improve cancer classification for imbalanced gene expression data[J]. BMC Bioinformatics，2024，25（1）：392.
[8]	SHEN C P. A transdisciplinary review of deep learning research and its relevance for water resources scientists[J]. Water Resources Research，2018，54（11）：8558-8593.
[9]	VIRRO H，AMATULLI G，KMOCH A，et al. GRQA：global river water quality archive[J]. Earth System Science Data，2021，13（12）：5483-5507.
[10]	MA X X，PENG W F，TONG B W，et al. Estimation of pollutant load in typical drainage ditches of Ningxia Yellow River Diversion Irrigation Area based on LOADEST statistical model[J]. Water，2024，16（1）：107.
[11]	ZHANG Q，BLOMQUIST J D，MOYER D L，et al. Estimation bias in water-quality constituent concentrations and fluxes：a synthesis for chesapeake bay rivers and streams[J]. Frontiers in Ecology and Evolution，2019，7：109.
[12]	ZHANG P，LIU X Y，ZHANG H R，et al. Optimized SVR model for predicting dissolved oxygen levels using wavelet denoising and variable reduction：Taking the Minjiang River estuary as an example[J]. Ecological Informatics，2025，86：103007.
[13]	LI Z L，LIU H X，ZHANG C，et al. Generative adversarial networks for detecting contamination events in water distribution systems using multi-parameter，multi-site water quality monitoring[J]. Environmental Science and Ecotechnology，2023，14：100231.
[14]	XU J H，WANG J C，CHEN L，et al. Surface water quality prediction model based on graph neural network[J]. Journal of Zhejiang University（Engineering Science），2021，55（4）：601-607. 许佳辉，王敬昌，陈岭，等. 基于图神经网络的地表水水质预测模型[J]. 浙江大学学报（工学版），2021，55（4）：601-607.
[15]	PENG L，WU H，GAO M，et al. TLT：Recurrent fine-tuning transfer learning for water quality long-term prediction[J]. Water Research，2022，225：119171.
[16]	LÓPEZ-ANDREU F J，LÓPEZ-MORALES J A，HERNÁNDEZ-GUILLEN Z，et al. Deep learning-based time series forecasting models evaluation for the forecast of chlorophyll a and dissolved oxygen in the mar menor[J]. Journal of Marine Science and Engineering，2023，11（7）：1473.
[17]	SIAMI-NAMINI S，TAVAKOLI N，NAMIN A S. The performance of LSTM and BiLSTM in forecasting time series[C]// IEEE International Conference on Big Data（Big Data），2019：3285-3292.
[18]	ALDREES A，KHAN M，TAHA A，et al. Evaluation of water quality indexes with novel machine learning and shapley additive explanation（SHAP）approaches[J]. Journal of Water Process Engineering，2024，58：104789.
[19]	SUN X，DU Z L，DING J，et al. Machine learning integrated with a causal pathway framework unravels differential mechanisms of biochar-driven soil organic carbon dynamics under cadmium stress[J]. Environmental Science& Technology，2026，60（6）：1234-1245.
[20]	LU Y，BU Y N，LI B Y，et al. From black-box prediction to probabilistic control：An explainable CNN-SHAP-Monte Carlo framework for low-carbon wastewater treatment[J]. Journal of Water Process Engineering，2025，79.
[21]	WANG C，WANG X，HARDEBERG J Y. A linear interpolation algorithm for spectral filter array demosaicking[M]// Cherbourg，France：Springer International Publishing，2014.
[22]	XU H Z，PANG G S，WANG Y J，et al. Deep isolation forest for anomaly detection[J]. IEEE Transactions on Knowledge and Data Engineering，2023，35（12）：12591-12604.
[23]	SONG C G，YAO L H. Application of artificial intelligence based on synchrosqueezed wavelet transform and improved deep extreme learning machine in water quality prediction[J]. Environmental Science and Pollution Research，2022，29（25）：38066-38082.
[24]	REES P A，LOWY R J. Optimizing reduction of western blotting analytical variations：Use of replicate test samples，multiple normalization methods，and sample loading positions[J]. Analytical Biochemistry，2023，674：115198.
[25]	HOCHREITER S，SCHMIDHUBER J. Long short-term memory[J]. Neural Computation，1997，9（8）：1735-1780.
[26]	GUO J，LIU Y，ZOU Q，et al. Study on optimization and combination strategy of multiple daily runoff prediction models coupled with physical mechanism and LSTM[J]. Journal of Hydrology，2023，624：129969.
[27]	VASWANI A，SHAZEER N，PARMAR N，et al. Attention is all you need[C]// Advances in Neural Information Processing Systems，2017：5999-6009.
[28]	HU Y K，LYU L，WANG N，et al. Long-term prediction of multiple river water quality indexes based on hybrid deep learning models[J]. Measurement Science and Technology，2024，35（12）：125803.
[29]	LUNDBERG S M，LEE S I. A unified approach to interpreting model predictions[C]// Proceedings of the 31st International Conference on Neural Information Processing Systems. Long Beach，CA：Curran Associates Inc.，2017：4768-4777.
[30]	PAN M Y，XIA B S，HUANG W B，et al. PM_2.5 concentration prediction model based on random forest and SHAP[J]. International Journal of Pattern Recognition and Artificial Intelligence，2024，38（5）：2452012.
[31]	PANDEY R，DHOUNDIYAL M，KUMAR A. Correlation analysis of big data to support machine learning[C]// Gwalior，MP，India：IEEE，2015.
[32]	SUN M，WEI S K，WANG Y J，et al. Water quality prediction model of LSTM based on wavelet decomposition[J]. Computer Systems& Applications，2020，29（12）：55-63. 孙铭，魏守科，王莹洁，等. 基于小波分解的LSTM水质预测模型. 计算机系统应用，2020，29（12）：55-63.