SSA-UNet: Advanced Precipitation Nowcasting via Channel Shuffling

--Weather forecasting is essential for facilitating diverse socio-economic activity and environmental conservation initiatives. Deep learning techniques are increasingly being explored as complementary approaches to Numerical Weather Prediction (NWP) models, offering potential benefits such as reduced complexity and enhanced adaptability in specific applications. This work presents a novel design, Small Shuffled Attention UNet (SSA-UNet), which enhances SmaAt-UNet's architecture by including a shuffle channeling mechanism to optimize performance and diminish complexity. T o assess its efficacy, this architecture and its reduced variant are examined and trained on two datasets: a Dutch precipitation dataset from 2016 to 2019, and a French cloud cover dataset containing radar images from 2017 to 2018. Three output configurations of the proposed architecture are evaluated, yielding outputs of 1, 6, and 12 precipitation maps, respectively. T o better understand how this model operates and produces its predictions, a gradient-based approach called Grad-CAM is used to analyze the outputs generated. The analysis of heatmaps generated by Grad-CAM facilitated the identification of regions within the input maps that the model considers most informative for generating its predictions. Weather forecasting is an indispensable domain, deemed crucial for various operations, including aviation safety, emergency response, agricultural planning, maritime navigation, and outdoor event management, in addition to improving public safety. Furthermore, accurate weather forecasting can significantly help mitigate the pollution from heavy-vehicle traffic. The author in [5] showed that severe weather can significantly increase vehicle utilization and traffic congestion. Consequently, accurate precipitation nowcasting could help people avoid superfluous vehicle journeys, thus alleviating traffic congestion and its related impacts on the environment.