Arctic Sea Ice Drift Predicted Using Machine Learning

The movement of sea ice is influenced by a number of factors, from winds to ocean currents. As climate change continues to advance, sea ice drift in the Arctic is a key parameter to understanding the effects of rising temperatures in the region and around the world. Research has shown that the Arctic and the Antarctic are warming faster than the rest of the planet, which raises questions regarding climate justice, as most of the carbon emissions causing anthropogenic climate change are produced in other regions. To analyze this impact, we employ artificial intelligence to predict sea ice drift velocity based on external features. Machine learning is the process of computers gaining insights by seeing patterns and correlating large quantities of data. Using external parameters, including wind speed, and drift velocity ground truth as the inputs of the model, we train multiple different architectures and compare the results. In particular, we experiment with a convolutional neural network (CNN), a random forest (RF), and a support vector machine (SVM). We also experiment with various model specifications. The ultimate aim is to achieve a greater understanding of the Arctic's response to climate change.