Forecasting of Groundwater Quality by Using Deep Learning Time Series Techniques in an Arid Region

Groundwater is regarded as the primary source of agricultural and drinking water in semi-arid arid regions. However, toxic substances released from sources such landfills, industries, insecticides, fertilizers previous year exhibited extreme levels groundwater contamination. As a result, it crucial to assess quality for activities, both its current use potential become reliable supply individuals. The critical Egypt’s Sohag region because serves major alternative activities residential supplies, addition providing water, residents there frequently have issues with water’s suitability human consumption. This research assesses future forecasting using Deep Learning Time Series Techniques (DLTS) long short-term memory (LSTM) Sohag, Egypt. Ten parameters (pH, Sulfate, Nitrates, Magnesium, Chlorides, Iron, Total Coliform, TDS, Hardness, Turbidity) at seven pumping wells were used analysis create index (WQI). model was tested trained actual data over nine years high quantities iron magnesium samples produced WQI. proposed provided good performances terms average mean-square error (MSE) root-mean-square (RMSE) values 1.6091 × 10−7 4.0114 10−4, respectively. WQI model’s findings demonstrated that could assist managers policymakers better managing resources areas.