Uploaded September 2026 | Updated November 2026, 2 hours ago
Speaker:
Adam Sobel, Professor of Earth and Environmental Sciences, Columbia University.
-----
Abstract:
Information on climate risk is needed for adaptation. The information produced by traditional climate science is generally not quite fit for purpose; an additional layer of science is generally needed on top of typical Earth system model output. Downscaling, bias correction, impact modeling, and decision support are all parts of it; characterization of uncertainty is also important, including of deep uncertainties that are hard to quantify.
This “climate risk science” (so called by me because of the importance of extremes and the inherently probabilistic nature of the information) has been practiced in academia and government, but the private sector, so far, has arguably been more organized and focused about it, at least in the U.S. Private sector products, however, are generally expensive, proprietary, and in many cases opaque or “black box.” The result is a situation in which many potential or actual users of the information are confused and adaptation is inhibited.
Adam argues for a more organized effort to develop usable climate risk science, especially for public sector applications, and outline one concrete proposal towards this end. A few recent research results on hurricane risk under climate change are used to illustrate specific points, with particular focus on current uncertainty around trends in the tropical Pacific.
-----
Further information:
bsc.es/research-and-development/research-seminars/sors-usable-climate-risk-science-motivation-examples-and-proposal
Speaker:
Adam Sobel, Professor of Earth and Environmental Sciences, Columbia University.
-----
Abstract:
Information on climate risk is needed for adaptation. The information produced by traditional climate science is generally not quite fit for purpose; an additional layer of science is generally needed on top of typical Earth system model output. Downscaling, bias correction, impact modeling, and decision support are all parts of it; characterization of uncertainty is also important, including of deep uncertainties that are hard to quantify.
This “climate risk science” (so called by me because of the importance of extremes and the inherently probabilistic nature of the information) has been practiced in academia and government, but the private sector, so far, has arguably been more organized and focused about it, at least in the U.S. Private sector products, however, are generally expensive, proprietary, and in many cases opaque or “black box.” The result is a situation in which many potential or actual users of the information are confused and adaptation is inhibited.
Adam argues for a more organized effort to develop usable climate risk science, especially for public sector applications, and outline one concrete proposal towards this end. A few recent research results on hurricane risk under climate change are used to illustrate specific points, with particular focus on current uncertainty around trends in the tropical Pacific.
-----
Further information:
bsc.es/research-and-development/research-seminars/sors-usable-climate-risk-science-motivation-examples-and-proposal










