March 18 Martin Tingley Department of Earth and Planetary Sciences, Harvard University A hierarchical Bayesian analysis of a high latitude, high resolution multiproxy data set over the last 600 years.
Add to my calendar:
I will present a hierarchical Bayesian approach to reconstructing climate fields from incomplete (in space and time) instrumental and proxy data sets, and apply the approach to a Northern high latitudes multiproxy data set. The end result of the analysis is an ensemble of draws of the spatially and temporally complete climate field consistent with the data and model assumptions. This ensemble can be used to calculate many quantities: point estimates, uncertainties, and, for example, the probability that a given year was the warmest in the reconstruction. I will then discuss how this basic Bayesian approach to climate reconstruction problems can be improved and expanded upon to incorporate different data types, and address more complex problems.