|
Monthly Groundwater Level Forecasting Using Deep Learning Model Based on Efficient Input Variables Selection by Partial Mutual Information Algorithm
|
Somayeh Abdi   , Hossein Fathian * , Mehdi Asadi Lour , Aslan Egdernezhad , Ali Asareh |
| Islamic Azad University of Ahvaz |
|
|
Abstract: (23 Views) |
Declining precipitation and excessive groundwater exploitation have led to significant groundwater level depletion in many regions worldwide. The Nahavand aquifer, located in Hamadan Province in western Iran, is one of the aquifers experiencing considerable groundwater level decline. In this study, a Long Short-Term Memory (LSTM) deep learning model was employed to forecast groundwater levels. The Partial Mutual Information (PMI) algorithm was utilized to identify the most influential input variables for the LSTM model. The results of the PMI analysis indicated that the current groundwater level and the groundwater level lagged by one month were the most effective predictors for monthly groundwater level forecasting. The performance of the LSTM model during the training period yielded a coefficient of determination (R²) of 0.96 and a root mean square error (RMSE) of 0.023. For the testing period, the model achieved an R² of 0.89 and an RMSE of 0.037. Overall, the results demonstrate that the LSTM model provides satisfactory performance in forecasting groundwater levels in the Nahavand aquifer. Therefore, the proposed model can be effectively applied to predict groundwater level fluctuations and support sustainable groundwater resource management in the study area.
|
Article number: 1 |
| Keywords: Groundwater Level forecasting, Nahavand aquifer, LSTM, PMI algorithm. |
|
|
|
Type of Study: Research |
Subject:
Special
Received: 2026/06/16 | Revised: 2026/07/22 | Accepted: 2026/07/30 | ePublished ahead of print: 2026/09/17
|
|
|
|
|
|
|
| Send email to the article author |
|
|
|
| Add your comments about this article |
|
|
|
