Researchers have developed a deep learning model to create high-resolution maps of surface meltwater on the Greenland ice sheet. The model combines data from remote sensing observations and physics-based models to produce daily maps at 100m resolution. In testing, the model outperformed existing methods by over 10 percentage points in accuracy, with some approaches also showing notable results without relying on deep learning. The team has released a benchmark dataset called MeltwaterBench for further comparison and development. --- Why it matters: This work matters to researchers in AI because it demonstrates the potential of deep learning to improve the accuracy of complex tasks such as spatiotemporal downscaling, which is crucial for understanding climate-related phenomena like ice sheet melting. The model's performance also highlights the importance of data fusion and multi-source information integration in achieving high-resolution predictions. Source: https://arxiv.org/abs/2512.12142