The sea ice emulator we featured a few months back, FloeNet, is now officially published in GRL. Led by Will Gregory, the graph neural network emulates GFDL’s global sea ice model (SIS2) while conserving mass, reproducing sea ice and snow-on-ice trends and variability with volume anomaly correlations above 0.96 in the Antarctic and 0.76 in the Arctic across a range of forcing scenarios. Read the paper .