Best ML Innovation Award for "Deep Learning for Climate Model Output Statistics"

For the paper "Deep Learning for Climate Model Output Statistics" Michael Steininger, Daniel Abel, Katrin Ziegler, Anna Krause, Heiko Paeth and Andreas Hotho were presented with the Best ML Innovation Award on the Tackling Climate Change with Machine Learning Workshop at NeurIPS 2020. Tackling Climate Change with Machine Learning is an interdisciplinary Workshop bridging the gap between Machine Learning and other fields related to Climate Change. From 74 accepted papers the workshop awarded 3 papers including ours.

In our joint work with Heiko Paeth’s group, we propose a novel deep learning based approach to Climate Model Output Statistics (MOS) which aims to reduce the deviation between the climate model REMO's simulated precipitation and the observed precipiation. We find that our architecture ConvMOS provides mostly better performance compared to existing approaches.

We developed this architecture in our project BigData@Geo in order to improve the accuracy of climate models outputs.