Devnath, Maloy KumarChakraborty, SudipJaneja, Vandana2024-12-112024-12-112024-11-06Devnath, Maloy Kumar, Sudip Chakraborty, and Vandana P. Janeja. “Deep Learning for Antarctic Sea Ice Anomaly Detection and Prediction: A Two-Module Framework.” Proceedings of the 1st ACM SIGSPATIAL International Workshop on Geospatial Anomaly Detection, GeoAnomalies ’24, November 6, 2024, 90–93. https://doi.org/10.1145/3681765.3698457.https://doi.org/10.1145/3681765.3698457http://hdl.handle.net/11603/37109GeoAnomalies '24: 1st ACM SIGSPATIAL International Workshop on Geospatial Anomaly Detection, Atlanta GA USA, 29 October 2024- 1 November 2024The Antarctic sea ice cover plays a crucial role in regulating global climate and sea level rise. The recent retreat of the Antarctic Sea Ice Extent and the accelerated melting of ice sheets (which causes sea level rise) raise concerns about the impact of climate change. Understanding the spatial patterns of anomalous melting events in sea ice is crucial for improving climate models and predicting future sea level rise, as sea ice serves as a protective barrier for ice sheets. This paper proposes a two-module framework based on Deep Learning that utilizes satellite imagery to identify and predict non-anomalous and anomalous melting regions in Antarctic sea ice. The first module focuses on identifying non-anomalous and anomalous melting regions in the current day by analyzing the difference between consecutive satellite images over time. The second module then leverages the current day's information and predicts the next day's non-anomalous and anomalous melting regions. This approach aims to improve our ability to monitor and predict critical changes in the Antarctic sea ice cover.4 pagesen-USAttribution 4.0 International CC BY 4.0 Deedhttps://creativecommons.org/licenses/by/4.0/deed.enUMBC Cybersecurity InstituteText