Earth System Predictability across Time Scales for a Resilient Society: A Research Community Perspective

Abstract With extreme weather events becoming more frequent and severe, accelerating progress in Earth system predictability is urgently needed to deepen fundamental understanding, improve predictive tools, and provide reliable, actionable information for societal resilience. Building on prior and ongoing efforts by the broader community and informed by discussions at a National Science Foundation’s National Center for Atmospheric Research (NSF NCAR) workshop on Earth System Predictability Across Time Scales, this essay articulates a perspective on the scientific and structural priorities needed to advance Earth system predictability from short-range weather forecasts to century-scale projections, underscoring the urgency of a comprehensive, integrative approach capable of meeting emerging societal needs. Three scientific grand challenges are highlighted: understanding interactions across spatial and temporal scales, across interconnected Earth system components, and the influence of external forcing on predictability. To address these grand challenges, we identify potential implementation priorities across five key areas: 1) enhancing observations and data accessibility, 2) advancing data assimilation techniques, 3) improving modeling frameworks, 4) developing artificial intelligence (AI) and machine learning (ML) methods, and 5) applying convergence research. To support these areas, we outline four intersecting pillars of an integrated strategy: (i) a multiscale and multidisciplinary approach; (ii) closer coordination across modeling, observations, data assimilation, and AI/ML; (iii) intentional convergence research; and (iv) codevelopment of science with users. We also propose a collaborative path forward focused on strengthening scientific and technical connections, rewarding interdisciplinary and team-based science, expanding support for engagement with users, and investing in relationship building, shared language, and trust across scientific and societal domains. Significance Statement This perspective highlights the need to advance Earth system predictability to enhance societal resilience amid increasingly frequent and severe extreme weather events. Grounded in broader community insights, including those from a National Science Foundation’s National Center for Atmospheric Research (NSF NCAR)–hosted workshop, it outlines a strategy to deepen understanding, improve predictive tools, and enable actionable outcomes across time scales from days to decades. This perspective identifies key scientific grand challenges: understanding scale interactions, Earth system component interactions, and the influence of external forcing. It also outlines corresponding implementation needs—including improved observations, data assimilation, modeling tools, and artificial intelligence and machine learning. Presented ideas emphasize an integrative strategy built on four intersecting pillars: a multiscale and multidisciplinary approach, coordination across tools, intentional convergence research, and codevelopment with users.