Forecasting of groundwater level of Dhaka city in Bangladesh with different climate factors using Deep Neural Networks (LSTM, GRU, LSTM+GRU) and Machine Learning Algorithms (SVR, RF, KNN) via univariate and multivariate time series analysis.
Rapid urbanization of Dhaka city, Bangladesh, has resulted in a significant decline in groundwater levels, causing severe environmental and socio-economic challenges. This study focuses on groundwater level forecasting using deep learning techniques, long short-term memory (LSTM), gated recurrent units (GRU) and hybrid LSTM+GRU models, as well as machine learning algorithms such as support vector regression (SVR), random forest (RF) and K-nearest neighbors (KNN). The models are applied to both univariate and multivariate time series analysis to incorporate various climatic factors to assess their impact on groundwater variability. The results demonstrate the effectiveness and forecasting accuracy of deep learning models compared to traditional machine learning approaches, especially in capturing long-term dependencies and complex patterns in multivariate scenarios. Comparative analysis reveals that LSTM and LSTM+GRU models are the most accurate groundwater-level forecasting models. This study will provide policymakers and urban planners with a reliable framework for effectively managing groundwater resources in Dhaka. The findings of this study will provide a robust framework for managing groundwater resources in Dhaka, enabling policymakers and stakeholders to practice sustainable water use and mitigate future water scarcity issues.
Key words: Groundwater Level, Climate Factors, Deep Learning, Machine Learning, LSTM, GRU, SVR, Multivariate Time Series, Dhaka City.
