The University of Texas logo
\
Beating the mean: Using models and observations together to improve our confidence in climate model output

Beating the mean:
Using models and observations together to improve our confidence in climate model output

By:
Ben Sanderson
National Center for Atmospheric Research (NCAR)


When: Friday, November 12, 2010, 10:30 a.m. to 11:30 a.m.
Join us for coffee beginning at 10:00 a.m.
Where: Seminar Conference Room, 10100 Burnet Road, Bldg 196-ROC, Austin, Texas 78758
Host: Charles Jackson, UTIG

Click for Live Broadcast

Photo of Ben Sanderson

Abstract
In recent IPCC reports, model predictions of future climate change are often been presented as a simple mean using output from a number of different climate models. In this seminar, we examine why the multi-model mean is so successful at reducing systematic biases in individual models, and also some of the lost information and incorrect interpretations which one might draw from over-interpreting averaged output. We examine current Bayesian and regression-based techniques which might be used to incorporate additional information from observations and other sources such as perturbed physics ensembles into synthesis projections from multiple climate models, and how these techniques might be used to produce probabilistic forecasts of future climate change.