Hybrid Glacio‐Hydrological Modeling Reveals Contrasting Runoff Changes in Western Patagonia Over the 21st Century

Climate change poses a serious risk for the freshwater ecosystem of Western Patagonia and threatens glacial and non‐glacial water resources. Here, we model the historical glacio‐hydrology of 2,236 catchments across the Western Patagonia, and project climate change impacts through the 21st century. To this end, we develop a novel modeling framework that combines Long Short‐Term Memory (LSTM) neural networks with ice‐dynamical glacier modeling using the Open Global Glacier Model (OGGM). We evaluate the ability of this hybrid framework to predict streamflow in ungauged basins (PUB) and regions (PUR) through 10‐fold cross‐validation and compare the results with those obtained with a LSTM model without a glacier component, and two process‐based coupled glacio‐hydrological models. The hybrid modeling approach outperforms all other approaches in 38% and 44% of the catchments considering PUB and PUR evaluations, respectively. Using our new hybrid approach, we estimate an average regional freshwater flux of 19,815 m 3 s −1 for the period 2000–2019, with glacier melt contributing 29% during the summer season. Under a high‐emission scenario (Shared socioeconomic pathways 5‐8.5), the northern region (>46°S) is projected to experience the largest reductions in runoff, with dry season runoff decreasing by almost 50% by the end of the century. In contrast, runoff increases are projected for glacierized basins in the southern regions, with average relative changes of 10%–25% and a marked seasonality shift. The results highlight the potential of hybrid modeling in glacio‐hydrology and provide important information for climate change adaptation in Western Patagonia.