As climate change increases the frequency and severity of extreme weather events, communities are facing growing vulnerability. Addressing this urgent challenge requires deepening our understanding of Earth system processes, uncertainties, and predictability across timescales. It also depends on new, collaborative approaches that directly engage user communities and leverage coordination within the research community. To contribute scientific leadership in this domain, the NSF NCAR has launched the Earth System Predictability Across Timescales (ESPAT) initiative. As part of this effort, NCAR hosted a community workshop in 2024 to gather input for developing a roadmap. Over 500 participants discussed grand challenges, key science questions, and advancements needed in modeling, observations, data assimilation, machine learning, and AI. The workshop also explored ways to bridge fundamental science with societal needs through co-creation and enhanced partnerships. This report synthesizes the key messages and input from the workshop, without added interpretation, to inform the ESPAT roadmap development.
