execution of SEVIR required several novel ideas and insights, including recognition of a gap in ML-ready weather