A conceptual metaheuristic-based framework for improving runoff time series simulation in glacierized catchments

Glacio-hydrological modeling is a key task for assessing the influence of snow and glaciers on water resources, essential for water resources management. The present study aims to enhance a conceptual hydrological model (namely Glacial Snow Melt (GSM)) by data-driven and swarm computing for enhancing the accuracy of rainfall runoff prediction. The proposed framework combines the conceptual hydrological model (i.e. GSM) with the time series predictor model (SVR) and optimization-driven parameter tuning of the firefly algorithm (SVR-FFA). This integration uniquely captures the complex interplay between meteorological variables, glacier processes, and hydrological responses. Applying the hybrid framework proved better results than the standalone GSM and ordinary SVR in simulating runoff time series. The performance of the proposed conceptual integrated metaheuristic-based framework (W-SG-SVR-FFA) demonstrated several enhancements over the standalone GSM model. During the calibration (validation) period, the evaluation metric coefficient of determination (R²) was 0.77 (0.77) for the standalone GSM model and 0.98 (0.91) for the W-SG-SVR-FFA model. The Kling-Gupta Efficiency (KGE) values were 0.81 (0.77) and 0.97 (0.87), respectively. Applying the method in glacierized catchments underscores its importance in areas undergoing swift climate change and glacial melting. This approach enables readers to witness the intricate equilibrium between the model's complexity and the accuracy of simulation outcomes.