Data driven of underground water level using artificial intelligence hybrid algorithms.

Area of Science:

• Environmental Science
• Water Resource Management
• Artificial Intelligence

Background:

• Growing populations and industrial/agricultural development necessitate robust quantitative and qualitative water resource management.
• Monitoring underground water storage through water level fluctuations is vital, especially in arid regions like Khuzestan province.
• Artificial intelligence (AI) has emerged as a powerful tool for global groundwater resource prediction and management.

Purpose of the Study:

• To develop and evaluate a novel hybrid artificial intelligence model for predicting underground water levels in the Qale-Tol area of Khuzestan province.
• To assess the efficacy of the DL-FF-KNN-ABC-MLP hybrid model in managing and predicting groundwater resources.
• To investigate the data noise reduction capabilities of the proposed hybrid model.

Main Methods:

• A hybrid model combining Feed Forward-K-Nearest Neighbors (FF-KNN), Artificial Bee Colony-K-Nearest Neighbors (ABC-KNN), and a novel DL-FF-KNN-ABC-MLP architecture was developed.
• The DL-FF-KNN-ABC-MLP model utilizes a two-block approach: classification via FF-DWKNN and prediction via ABC-MLP.
• Data from wells 1-5 were used for training and testing AI hybrid models, with wells 6-8 used for model development and validation.

Main Results:

• The DL-FF-KNN-ABC-MLP model achieved low statistical Root Mean Square Error (RMSE) values: 0.0451 (test), 0.0597 (train), and 0.0701 (total).
• The hybrid model demonstrated a high performance accuracy in predicting underground water levels.
• The novel architecture effectively reduced data noise, enhancing prediction reliability.

Conclusions:

• The DL-FF-KNN-ABC-MLP hybrid model is highly effective for predicting underground water levels in Khuzestan province.
• This AI-driven approach offers a significant advancement in groundwater resource management and prediction accuracy.
• The study highlights the potential of advanced hybrid AI models for addressing water scarcity challenges in arid regions.