A key application of Earth Observation (EO) imaging is monitoring the weather. Interestingly, modern machine learning methods have recently become viable alternatives to long standing physics-based solutions. Weather forecasts are of obvious immediate value. Novel insights from patterns identified by machine learning about the underlying processes, moreover, are critical for a better understanding of our environment and, ultimately, mitigating climate change. A Special Weather4cast Competition Session of the CDCEO'21 workshop (www.iarai.ac.at/ CDCEO21) presents highlights from a unique multi-sensor weather forecasting competition (www.weather4cast.ai).
The competition attracted well over 100 participants, with 17 teams eventually staying in the race for the top positions in the leaderboard. A Special Weather4cast Competition Session at the workshop presents the work of the 5 best ranked teams. The following papers describe the models of these winning teams. In a subsequent stage, the competition data set has been extended, with results expected later in the year (http://ieee.weather4cast.org/).