Fu et al. (2026) Analysis of Driving Factors and Trend Prediction of Groundwater Levels in the West Liao River Basin Based on the STL-LSTM Model

⚠️ Warning: This summary was generated from the abstract only, as the full text was not available.

Identification

• Journal: Water
• Year: 2026
• Date: 2026-04-06
• Authors: Sutong Fu, Liangping Yang, J Liu, Pengfei Hao, Fuli Wang, Jianmin Bian
• DOI: 10.3390/w18070876

Research Groups

Not explicitly stated in the provided text.

Short Summary

This study characterized groundwater dynamics in the West Liao River Basin from 2000-2016, revealing a persistent decline accelerated post-2009, primarily driven by soil moisture and climatic factors, and developed an accurate STL-LSTM hybrid model for forecasting.

Objective

• To characterize groundwater dynamics, analyze trends, quantify driving factors, and develop a forecasting model for sustainable water management in the West Liao River Basin.

Study Configuration

• Spatial Scale: West Liao River Basin
• Temporal Scale: 2000–2016

Methodology and Data

• Models used: Seasonal-Trend decomposition using Loess (STL), change-point detection, random forest, SHapley Additive exPlanations (SHAP), STL–Long Short-Term Memory (STL-LSTM) hybrid model.
• Data sources: Groundwater level data (2000–2016).

Main Results

• Groundwater levels declined persistently, with a significant change point detected in 2009.
• The post-2009 decline rate accelerated to −0.749 m/yr, representing a 55.7% increase compared to the pre-2009 period.
• Statistical attribution identified soil moisture (43.5%) and climatic factors (29.0%) as the primary predictors of groundwater variability, highlighting the significant role of agricultural irrigation.
• The developed STL-LSTM hybrid model achieved optimal predictive performance with an R² of 0.8805 and a Root Mean Square Error (RMSE) of 0.7081 m, demonstrating enhanced accuracy for non-stationary sequences.

Contributions

• Developed an integrated framework combining trend diagnosis, driver interpretation, and hybrid modelling for groundwater management.
• Introduced a novel STL-LSTM hybrid model demonstrating enhanced accuracy for forecasting non-stationary groundwater level sequences.
• Provided scientific support for precise groundwater management strategies in semi-arid agricultural basins by quantifying key drivers and forecasting future trends.

Funding

Not explicitly stated in the provided text.

Citation

@article{Fu2026Analysis,
  author = {Fu, Sutong and Yang, Liangping and Liu, J and Hao, Pengfei and Wang, Fuli and Bian, Jianmin},
  title = {Analysis of Driving Factors and Trend Prediction of Groundwater Levels in the West Liao River Basin Based on the STL-LSTM Model},
  journal = {Water},
  year = {2026},
  doi = {10.3390/w18070876},
  url = {https://doi.org/10.3390/w18070876}
}