Presented by: Andrew Gettelman, Pacific Northwest National Laboratory, University of Colorado, Boulder

Abstract: Earth’s climate is changing quickly enough that relying on historical weather statistics to estimate the risk of extreme events is no longer reliable. Communities, infrastructure planners, and policymakers increasingly need information about how the likelihood of extreme weather may change over the next 10–15 years in order to plan and adapt. However, existing tools do not adequately address this need. Weather forecasts focus on timescales of days to weeks, while climate projections are typically aimed decades into the future. Even initialized decadal climate predictions face major scientific and practical challenges and are unlikely to fully bridge this gap.
In this talk, I will argue for a different approach: climate nowcasting. Climate nowcasting aims to estimate near-term climate risk by combining observations, physical climate models, and data-driven methods to better characterize the current state of the climate system and how it is likely to evolve over the coming decade. Rather than focusing on global averages, this approach emphasizes specific regions, types of extreme events, and risks at human-relevant spatial scales of a few kilometers.
Beyond its practical value, a climate nowcasting framework also offers a new way to study predictability in the Earth system. By focusing on conditional risk—how future extremes depend on the present climate state—it may help improve our understanding of what aspects of climate are predictable on decadal timescales and how to better quantify uncertainty in a rapidly changing system. The presentation will be illustrated with examples of typical (and topical) weather extremes relevant to Boulder: extreme wind, fire weather and variability in seasonal snowpack. 

Host: Ivy Tan