Advancing Modeling for Earth System Predictability Across Timescales (Invited Presentation)

Society is facing grand environmental challenges, including increasingly severe weather events and the pervasive impacts of climate change. There is an urgent need for more research in Earth system prediction and predictability, alongside the development of tools that can produce trustworthy and actionable information to help communities build resilience. In April 2024, NSF NCAR hosted the Earth System Predictability Across Timescales (ESPAT) Community Workshop, bringing together experts to address grand challenges, key science questions, and the necessary modeling, observational, data assimilation, and ML/AI tools to tackle these issues.

This presentation will highlight the workshop's key outcomes, focusing on the future modeling needs identified to address research gaps and meet the user community's requirements for predictive tools across weather, subseasonal to seasonal, seasonal to decadal timescales, and long-term projections. We will describe the essential modeling tools needed to tackle key science questions, including those related to scale interactions, the coupling between Earth system components, and the influence of climate change on Earth system predictability. Additionally, we will discuss the modeling needs of user communities, including urban areas, that require information at spatial scales on the order of 10 km, across timescales ranging from days to decades.

We will explore how these community-driven needs are amplifying the demand for greater coordination between weather and climate communities and increased collaboration across modeling centers. Furthermore, given the computational challenges in providing actionable information at community scales, we will discuss the integration of ML and AI into modeling efforts to accelerate the development of tools that are crucial for enhancing societal resilience.