An Explainable Bayesian TimesNet for Probabilistic Groundwater Level Prediction

Abstract Reliable groundwater level (GWL) prediction is essential for sustainable water resources management. Despite recent advances in machine learning (ML) methods for GWL prediction, further improvements may be made in uncertainty quantification and model interpretability. This study proposes Ba...

Full description

Saved in:
Bibliographic Details
Main Authors: Zechen Peng, Shaoxing Mo, Alexander Y. Sun, Jichun Wu, Xiankui Zeng, Miao Lu, Xiaoqing Shi
Format: Article
Language:English
Published: Wiley 2025-06-01
Series:Water Resources Research
Subjects:
Online Access:https://doi.org/10.1029/2025WR040191
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849421275861090304
author Zechen Peng
Shaoxing Mo
Alexander Y. Sun
Jichun Wu
Xiankui Zeng
Miao Lu
Xiaoqing Shi
author_facet Zechen Peng
Shaoxing Mo
Alexander Y. Sun
Jichun Wu
Xiankui Zeng
Miao Lu
Xiaoqing Shi
author_sort Zechen Peng
collection DOAJ
description Abstract Reliable groundwater level (GWL) prediction is essential for sustainable water resources management. Despite recent advances in machine learning (ML) methods for GWL prediction, further improvements may be made in uncertainty quantification and model interpretability. This study proposes Bayesian TimesNet (BTimesNet), a novel deep learning model for probabilistic and explainable GWL prediction. BTimesNet transforms 1D time series data into 2D matrices based on periodicity, enhancing the capture of temporal patterns through convolutional filters. A Bayesian framework using Stein Variational Gradient Descent is implemented to quantify predictive uncertainties. For model interpretability, SHapely Additive exPlanations (SHAP) is utilized to quantify predictor contributions. The efficacy of BTimesNet for multi‐step‐ahead GWL prediction is evaluated using monthly data collected from 19 monitoring wells across three hydroclimatic regions in the U.S., and compared against the widely used long short‐term memory (LSTM) and Autoformer models. Results show that BTimesNet consistently outperforms LSTM and Autoformer, providing more accurate predictions up to 4 months ahead. SHAP analysis reveals that historical GWLs are the most informative features, with meteorological predictors making secondary contributions. BTimesNet's superior performance stems from its ability to extract both short‐ and long‐term temporal features. This approach represents a valuable advancement for accurate GWL prediction and risk‐informed decision‐making, providing critical lead time for proactive groundwater and ecosystem management and agricultural irrigation planning. Its data‐driven nature also facilitates broader applications across hydrological and environmental prediction domains.
format Article
id doaj-art-fbab882d3ea04f878def6713b0cd98d6
institution Kabale University
issn 0043-1397
1944-7973
language English
publishDate 2025-06-01
publisher Wiley
record_format Article
series Water Resources Research
spelling doaj-art-fbab882d3ea04f878def6713b0cd98d62025-08-20T03:31:30ZengWileyWater Resources Research0043-13971944-79732025-06-01616n/an/a10.1029/2025WR040191An Explainable Bayesian TimesNet for Probabilistic Groundwater Level PredictionZechen Peng0Shaoxing Mo1Alexander Y. Sun2Jichun Wu3Xiankui Zeng4Miao Lu5Xiaoqing Shi6Key Laboratory of Surficial Geochemistry of Ministry of Education School of Earth Sciences and Engineering Nanjing University Nanjing ChinaKey Laboratory of Surficial Geochemistry of Ministry of Education School of Earth Sciences and Engineering Nanjing University Nanjing ChinaBureau of Economic Geology Jackson School of Geosciences University of Texas at Austin Austin TX USAKey Laboratory of Surficial Geochemistry of Ministry of Education School of Earth Sciences and Engineering Nanjing University Nanjing ChinaKey Laboratory of Surficial Geochemistry of Ministry of Education School of Earth Sciences and Engineering Nanjing University Nanjing ChinaCollege of Hydraulic Science and Engineering Yangzhou University Yangzhou ChinaKey Laboratory of Surficial Geochemistry of Ministry of Education School of Earth Sciences and Engineering Nanjing University Nanjing ChinaAbstract Reliable groundwater level (GWL) prediction is essential for sustainable water resources management. Despite recent advances in machine learning (ML) methods for GWL prediction, further improvements may be made in uncertainty quantification and model interpretability. This study proposes Bayesian TimesNet (BTimesNet), a novel deep learning model for probabilistic and explainable GWL prediction. BTimesNet transforms 1D time series data into 2D matrices based on periodicity, enhancing the capture of temporal patterns through convolutional filters. A Bayesian framework using Stein Variational Gradient Descent is implemented to quantify predictive uncertainties. For model interpretability, SHapely Additive exPlanations (SHAP) is utilized to quantify predictor contributions. The efficacy of BTimesNet for multi‐step‐ahead GWL prediction is evaluated using monthly data collected from 19 monitoring wells across three hydroclimatic regions in the U.S., and compared against the widely used long short‐term memory (LSTM) and Autoformer models. Results show that BTimesNet consistently outperforms LSTM and Autoformer, providing more accurate predictions up to 4 months ahead. SHAP analysis reveals that historical GWLs are the most informative features, with meteorological predictors making secondary contributions. BTimesNet's superior performance stems from its ability to extract both short‐ and long‐term temporal features. This approach represents a valuable advancement for accurate GWL prediction and risk‐informed decision‐making, providing critical lead time for proactive groundwater and ecosystem management and agricultural irrigation planning. Its data‐driven nature also facilitates broader applications across hydrological and environmental prediction domains.https://doi.org/10.1029/2025WR040191groundwater level predictionBayesian TimesNetexplainable deep learningprobabilistic prediction
spellingShingle Zechen Peng
Shaoxing Mo
Alexander Y. Sun
Jichun Wu
Xiankui Zeng
Miao Lu
Xiaoqing Shi
An Explainable Bayesian TimesNet for Probabilistic Groundwater Level Prediction
Water Resources Research
groundwater level prediction
Bayesian TimesNet
explainable deep learning
probabilistic prediction
title An Explainable Bayesian TimesNet for Probabilistic Groundwater Level Prediction
title_full An Explainable Bayesian TimesNet for Probabilistic Groundwater Level Prediction
title_fullStr An Explainable Bayesian TimesNet for Probabilistic Groundwater Level Prediction
title_full_unstemmed An Explainable Bayesian TimesNet for Probabilistic Groundwater Level Prediction
title_short An Explainable Bayesian TimesNet for Probabilistic Groundwater Level Prediction
title_sort explainable bayesian timesnet for probabilistic groundwater level prediction
topic groundwater level prediction
Bayesian TimesNet
explainable deep learning
probabilistic prediction
url https://doi.org/10.1029/2025WR040191
work_keys_str_mv AT zechenpeng anexplainablebayesiantimesnetforprobabilisticgroundwaterlevelprediction
AT shaoxingmo anexplainablebayesiantimesnetforprobabilisticgroundwaterlevelprediction
AT alexanderysun anexplainablebayesiantimesnetforprobabilisticgroundwaterlevelprediction
AT jichunwu anexplainablebayesiantimesnetforprobabilisticgroundwaterlevelprediction
AT xiankuizeng anexplainablebayesiantimesnetforprobabilisticgroundwaterlevelprediction
AT miaolu anexplainablebayesiantimesnetforprobabilisticgroundwaterlevelprediction
AT xiaoqingshi anexplainablebayesiantimesnetforprobabilisticgroundwaterlevelprediction
AT zechenpeng explainablebayesiantimesnetforprobabilisticgroundwaterlevelprediction
AT shaoxingmo explainablebayesiantimesnetforprobabilisticgroundwaterlevelprediction
AT alexanderysun explainablebayesiantimesnetforprobabilisticgroundwaterlevelprediction
AT jichunwu explainablebayesiantimesnetforprobabilisticgroundwaterlevelprediction
AT xiankuizeng explainablebayesiantimesnetforprobabilisticgroundwaterlevelprediction
AT miaolu explainablebayesiantimesnetforprobabilisticgroundwaterlevelprediction
AT xiaoqingshi explainablebayesiantimesnetforprobabilisticgroundwaterlevelprediction