Figure 7
From: Exploring the link between microseism and sea ice in Antarctica by using machine learning
Sea ice concentration prediction. Examples of prediction based on microseism features for different patterns of sea ice concentration. In (c,f,i) the spatial distribution of the prediction error (computed as the difference between the true sea ice concentration and the predicted one) is plotted. The used model is obtained by Extremely Randomized Trees techniques applied on RMS amplitude data sequentially post-processed by PIT, LDA and SMT, with the addition of the time-related feature. The red areas in (g) and the grey areas in (i) represent two extended coastal regions, where the sea ice satellite data were missing on 19 June 2016. The maps were created by Matplotlib package for Python61.
