Designing a deep learning-based framework for the prediction of lake surface closed curves

Predicting the surface area and shape of lakes is critical for ecological, hydrological, and climatic studies. Accurate predictions enhance the understanding of lake dynamics, facilitate water resource management, and support environmental change assessments. Traditional methods, while foundational, often lack the efficiency, accuracy, and scalability required to handle complex lake systems, necessitating modern, technology-driven approaches. This study introduces a novel methodology for predicting changes in lake surface area and shape, including shoreline dynamics. The approach employs advanced remote sensing techniques and mathematical modeling, integrating Mathematica simulations with Net-Encoder and Deconvolution models. The framework achieved a high accuracy rate of 93% in water pixel extraction. Results indicate significant reductions in surface area, with Lake Eucumbene shrinking by 4.3% and the Salton Sea by 14.54%. The most notable shoreline changes occurred in the southwestern and northern regions of Lake Eucumbene and the southwestern region of the Salton Sea. This research highlights the effectiveness of a remote sensing-based approach for monitoring lake surface dynamics, offering a low-cost, high-accuracy tool for environmental monitoring and climate change impact assessment. Compared to existing methods, the proposed approach provides equivalent accuracy while delivering enhanced operational simplicity and flexibility.