| 545 | 0 | 43 |
| 下载次数 | 被引频次 | 阅读次数 |
【目的】准确、快速的预测地下水水位动态变化特征，对地下水科学管控尤为重要，但针对多因素影响下的地下水位中长期预测研究较少。【方法】为提高地下水位中长期预测能力，构建了一种基于PatchTST的多变量地下水位中长期预测模型(PatchTST-GWL),利用互相关函数分析了地下水开采量、地表补水量、降雨量、侧向补给等影响因素间的多元相关性及地下水水位变化的滞后性，对北京市西郊地区4眼典型浅层监测井地下水水位进行了中长期预测，采用纳什系数(NSE)、均方根误差(RMSE)、平均绝对误差(MAE)评估了模型性能及准确性，并基于控制变量法分析了模型可解释性。【结果】结果显示：随着预测期增加，PatchTST-GWL模型预测精度逐渐提升，各监测井90 d、180 d预测期地下水水位模拟结果NSE系数均提升至0.9以上，MAE、MSE、RMSE较Attention-Bi-LSTM、SVM常用深度学习模型降低10%～80%。【结论】PatchTST-GWL模型对中长期地下水水位预测性能优势明显，模型通过引入互相关函数计算地下水开采、降雨、地表补水、侧向补给变化过程对地下水水位变化过程影响的滞后期，有效提升了模型预测精度，同时模型预测结果与各影响因子变化的响应规律符合客观物理规律，呈现出较好的可解释性。该模型可准确、快速的进行地下水水位预测，为科学评估与合理利用地下水提供有效支撑。
Abstract：[Objective]Accurate and rapid prediction of dynamic groundwater level changes is crucial for scientific groundwater management, yet long-term predictions influenced by multiple factors remain insufficiently researched.[Methods]To enhance the capability of long-term groundwater level forecasting, a multivariate long-term prediction model based on PatchTST(PatchTST-GWL) was developed. This model utilized cross-correlation functions to analyze the multivariate correlations and lag effects among factors like groundwater extraction, surface water recharge, rainfall, and lateral recharge. Long-term forecasts were conducted on groundwater levels in four typical shallow monitoring wells in the western suburbs of Beijing. The model performance and accuracy were assessed using the Nash-Sutcliffe Efficiency(NSE), Root Mean Square Error(RMSE), and Mean Absolute Error(MAE). The model's interpretability was also analyzed using the controlled variable method.[Results]The result showed that the prediction accuracy of the PatchTST-GWL model improves with the extension of the forecasting period. For 90 and 180 days forecasting periods, the NSE coefficients of groundwater level simulations exceeded 0.9 across all monitoring wells, with MAE, MSE, and RMSE reductions of 10% to 80% compared to commonly used deep learning models like Attention-Bi-LSTM and SVM.[Conclusion]The PatchTST-GWL model exhibits a significant advantage in the performance of long-term groundwater level predictions. By incorporating cross-correlation functions to calculate the lag effects of groundwater extraction, rainfall, surface water recharge, and lateral recharge changes on groundwater levels, the model significantly enhances prediction accuracy. Furthermore, the predictions align with the response patterns of various influencing factors, consistent with objective physical laws, demonstrating good interpretability. This model can accurately and swiftly predict groundwater levels, effectively supporting scientific assessments and rational utilization of groundwater resources.
[1] SUN A Y.Predicting groundwater level changes using GRACE data[J].Water Resources Research,2013,49(9):5900-5912.
[2] KIM N W,CHUNG I M,WON Y S,et al.Development and application of the integrated SWAT-MODFLOW model[J].Journal of Hydrology,2008,356(1/2):1-16.
[3] USMAN M,QAMAR M U,BECKER R,et al.Numerical modeling and remote sensing-based approaches for investigating groundwater dynamics under changing land use and climate in the agricultural region of Pakistan[J].Journal of Hydrology,2020,581:124408.
[4] ZHOU Y,LI W.A review of regional groundwater flow modeling[J].Geoscience Frontiers,2011,2(2):205-214.
[5] DALIAKOPOILOS I N,COULIBALY P,TSANIS I K.Groundwater level forecasting using artificial neural networks[J].Journal of Hydrology,2005,309(1/4):229-240.
[6] SAHOO S,RUSSO T A,ELLIOTT J,et al.Machine learning algorithms for modeling groundwater level changes in agricultural regions of the US[J].Water Resources Research,2017,53(5):3878-3895.
[7] YI S,KONDOLF G M,SANDOVAL SOLIS S,et al.Groundwater level forecasting using machine learning:A case study of the Baekje weir in Four Major Rivers Project,South Korea[J].Water Resources Research,2024,60(5):e2022WR032779.
[8] DASH N B,PANDA S N,REMESAN R,et al.Hybrid neural modeling for groundwater level prediction[J].Neural Computing and Applications,2010,19:1251-1263.
[9] SAHOO S,JHA M K.Groundwater-level prediction using multiple linear regression and artificial neural network techniques:A comparative assessment[J].Hydrogeology Journal,2013,21(8):1865.
[10] WAMG X,LIU T,ZHENG X,et al.Short-term prediction of groundwater level using improved random forest regression with a combination of random features[J].Applied Water Science,2018,8:1-12.
[11] YOON H,JUN S C,HYUN Y,et al.A comparative study of artificial neural networks and support vector machines for predicting groundwater levels in a coastal aquifer[J].Journal of Hydrology,2011,396(1/2):128-138.
[12] 薛萍，廖丽莎，廖卫红，等.南水北调中线工程水位的水动力-神经网络耦合预测模型[J].南水北调与水利科技(中英文),2023,21(6):1116-1125.XUE P,LIAO L S,LIAO W H,et al.Hydrodynamic-neural network coupling prediction model of water level for the Middle Route of South-to-North Water Transfers Project[J].South-to-North Water Transfers and Water Science & Technology,2023,21(6):1116-1125.
[13] 徐淑琴，乔厚清，王雅君，等.基于LSTM水稻生育期地下水水位预测：以黑龙江省查哈阳灌区为例[J].东北农业大学学报，2021,52(2):70-78.XU Shuqin,QIAO Houqing,WANG Yajun,et al.Prediction of groundwater level based on LSTM:Take Chahayang Irrigation District in Heilongjiang Province as an example [J].Journal of Northeast Agricultural University,2021,52(2):70-78.
[14] VU M T,JARDANI A,MASSEI N,et al.Reconstruction of missing groundwater level data by using Long Short-Term Memory (LSTM) deep neural network[J].Journal of Hydrology,2021,597:125776.
[15] KHAN J,LEE E,BALOBAID A S,et al.A comprehensive review of conventional,machine leaning,and deep learning models for groundwater level (GWL) forecasting[J].Applied Sciences,2023,13(4):2743.
[16] ROY D K,SARKAR T K,KAMAR S S A,et al.Daily prediction and multi-step forward forecasting of reference evapotranspiration using LSTM and Bi-LSTM models[J].Agronomy,2022,12(3):594.
[17] GUO F,YANG J,LI H,et al.A convLSTM conjunction model for groundwater level forecasting in a karst aquifer considering connectivity characteristics[J].Water,2021,13(19):2759.
[18] MUNIAPPAN A,JARIN T,SABITHA R,et al.Bi-LSTM and partial mutual information selection-based forecasting groundwater salinization levels[J].Water Reuse,2023,13(4):525-544.
[19] SAMANTARAY S,SAHOO A,SATAPATHY D P.Prediction of groundwater-level using novel SVM-ALO,SVM-FOA,and SVM-FFA algorithms at Purba-Medinipur,India[J].Arabian Journal of Geosciences,2022,15(8):723.
[20] HOSSEINI Z,GHARECHELOU S,NAKHAEI M,et al.Optimal design of BP algorithm by ACO R model for groundwater-level forecasting:A case study on Shabestar plain,Iran[J].Arabian Journal of Geosciences,2016,9:1-16.
[21] 邓志勇，黄海峰，李清清，等.基于Swin Transformer目标全景分割的三峡库首土质滑坡识别[J].水利水电技术(中英文),2024,55(4):176-185.DENG Zhiyong,HUANG Haifeng,LI Qingqing,et al.Identification of soil landslides at the head of the Three Gorges Reservoir based on swin transformer target panoramic segmentation[J].Water Resources and Hydropower Engineering,2024,55(4):176-185.
[22] 杨阳蕊，朱亚萍，刘雪梅，等.知识驱动的南水北调工程巡检信息推荐方法[J].南水北调与水利科技(中英文),2024,22(2):368-377.YANG Y R,ZHU Y P,LIU X M,et al.Knowledge-driven recommended method for inspection information of South-to-North Water Transfers Project[J].South-to-North Water Transfers and Water Science & Technology,2024,22(2):368-377.
[23] YANG Z,JIAJUN Z,CHENGQING Z.Transformer:A general framework from machine translation to others[J].Machine Intelligence Research,2023,20(4):514-538.
[24] 周从亨，高凌霄.基于两阶段优化的多能源系统双层调度策略研究[J].水利水电技术(中英文),2023,54(11):15-28.ZHOU Congheng,GAO Lingxiao.Two-stage optimization-based study on bi-level scheduling strategy of multi-energy system [J].Water Resources and Hydropower Engineering,2023,54(11):15-28.
[25] TIAN M,WANWAN W,YU C.Attention is all you need:An interpretable transformer-based asset allocation approach[J].International Review of Financial Analysis,2023,90(3):1-17.
[26] 常启帆，杨煦旻，朱方辉，等.自然语言方法解决气井修井专业性强术语信息收集准确性差难题[J].石油钻采工艺，2024,46:1-14.CHANG Q F,YANG X M,ZHU F H,et al.A natural language processing data collection method for weakly regulated professional petroleum documents[J].Oil Drilling & Production Technology,2024,46:1-14.
[27] HUANG X,TANG J,SHEN Y.Long time series of ocean wave prediction based on PatchTST model[J].Ocean Engineering,2024,301:117572.
[28] YI H,ZHANG S,AN D,et al.PatchesNet:PatchTST-based multi-scale network security situation prediction[J].Knowledge-Based Systems,2024,299(5):112037.
[29] 朱晓彤，林培光，孙玫，等.基于APDFinformer模型的金融数据的多元时序预测[J].南京大学学报(自然科学),2024,60(6):930-939.ZHU X T,LIN P G,SUN M,et al.Multivariate time series forecasting of financial data based on the APDFinformer model[J].Journal of Nanjing University(Natural Science),2024,60(6):930-939.
[30] KHAN Z N,AHAMD J.Attention induced multi-head convolutional neural network for human activity recognition[J].Applied Soft Computing,2021,110:107671.
[31] ZHANG J,XING L,TAN Z,et al.Multi-head attention fusion networks for multi-modal speech emotion recognition[J].Computers & Industrial Engineering,2022,168:108078.
[32] WU D,YE M,LIN G,et al.Person re-identification by context-aware part attention and multi-head collaborative learning[J].IEEE Transactions on Information Forensics and Security,2021,17:115-126.
[33] ZENG P,HU G,ZHOU X,et al.Muformer:A long sequence time-series forecasting model based on modified multi-head attention[J].Knowledge-Based Systems,2022,254:109584.
[34] YOO J C,HAN T H.Fast normalized cross-correlation[J].Circuits,Systems and Signal Processing,2009,28:819-843.
[35] SEIFI A,EHTERAM M,SINGH V P,et al.Modeling and uncertainty analysis of groundwater level using six evolutionary optimization algorithms hybridized with ANFIS,SVM,and ANN[J].Sustainability,2020,12(10):4023.
[36] CAO Y,YIN K,ZHOU C,et al.Establishment of landslide groundwater level prediction model based on GA-SVM and influencing factor analysis[J].Sensors,2020,20(3):845.
[37] ALI A S A,EBRAHIMI S,ASHIQ M M,et al.CNN-Bi LSTM neural network for simulating groundwater level[J].Environmental Science,Engineering,Computer Science,2022,8(1):1-7.
[38] KHAN J,LEE E,BALOBAID A S,et al.A comprehensive review of conventional,machine leaning,and deep learning models for groundwater level (GWL) forecasting[J].Applied Sciences,2023,13(4):2743.
[39] 吕琳.北京西郊砂石坑蓄洪对地下水影响的模拟研究[D].北京：中国地质大学(北京),2014.LYU LIN.Simulation of Flood Storage by Grit Pit Effeets on Groundwater at the Western Suburb of Beijing[D].Beijing:China University of Geosciences,2014.
[40] 刘娣，余钟波，吕海深，等.基于SVM-EnKF双向数据同化的地下水水位变化预测[J].水文，2023,43(6):58-65.LIU Di,YU Zhongbo,LV Haishen,et al.Groundwater level changes forecast using support vector machine and ensemble kalman filtering dual-data assimilation technique[J].Journal of China Hydrology,2023,43(6):58-65.
[41] 赵文举，樊新建，范严伟.工程水文学与水利计算[M].北京：中国水利水电出版社，2016.ZHAO Wenju,FAN Xinjian,FAN Yanwei.Engineering Hydrology and Hydraulic Calculation[M].Beijing:China Water Power Press,2016.
[42] 束龙仓，张彤豪，澈丽木格，等.降水和地下水开采对通辽市科尔沁区潜水位的影响[J].吉林大学学报(地球科学版),2024,54(3):954-966.SHU Longcang,ZHANG Tonghao,CHE Limuge,et al.Effects of precipitation and groundwater exploitation on the phreatic water table in Khorqin District,Tongliao City[J].Journal of Jilin University (EarthScienceEdition),2024,54(3):954-966.
基本信息:
DOI：10.13928/j.cnki.wrahe.2025.11.007
中图分类号:P641.8
引用信息:
[1]程帅,张娟,杨默远,等.基于PatchTST的地下水位预测模型[J].水利水电技术(中英文),2025,56(11):83-97.DOI:10.13928/j.cnki.wrahe.2025.11.007.
基金信息:
北京市自然科学基金项目(8232032); 国家自然科学基金项目(52209005); 水利部重大科技项目(SKS-2022044); 北京学者培养经费资助项目(GZ-2024-002-SZY)