Warm and dry downslope winds called Foehn winds (Dutch: hairdryer winds) occur all over the world. These winds cause warm and dry conditions that can lead to explosive wildfires and also melt glaciers and ice sheets. Foehn winds are prominent on the Antarctic Peninsula and can lead to surface melt that can destabilize ice shelves in the region. Our work uses in situ Automatic Weather Stations, realaysis, and model output, along with Machine Learning techniques to identify when, where, and how foehn winds cause melt. Results from these experiments have helped to provide a clearer picture of the causes of melt and stability of the existing ice shelves on the Antarctic Peninsula, and how melt is evolving due to climate change. This research also provides another example of how Machine Learning can be a useful tool to understand physical Earth systems and their relation to climate change. Our future research aims to use Machine Learning to better understand the role foehn winds played in the collapse of Larsen A and B Ice shelves in 1995 and 2002. To learn more about this project click here.