To improve our understanding of historical and modern extreme temperatures, researchers at Harvard University will analyze instrumental and temperature-sensitive proxies of past climate using new statistical approaches. Specifically, the researchers will explore the relationships between instrumental and proxy indicators of extremes, reconstruct temperature extremes in space and time, and provide a better understanding of how changes in mean climate relate to the occurrences of extreme events. The analysis strategy is to extend an existing Bayesian methodology for inferring past climate to more directly target extremes. Definition of "extremes" will meet the standards set forth by the National Research Council in 2006 for accurate quantification of the probability that quantities such as maxima, minima, or rates-of-change are truly exceptional.
Extremes are by definition rare, making the need for records longer than are afforded by instruments alone perhaps even more acute than for evaluation of mean climate. Successful completion of this project will provide insight into the co-variability between background climate and extremes over a time period long enough to establish baseline conditions. The work will also provide important context in which to interpret and assess modern changes. Given that disparate mechanisms are responsible for warming and cooling over the last 600 hundred years, this research will also indicate the extent to which past relationships between mean climate and extremes can be used as analogues for future changes.
In addition to the societal relevance of better understanding extreme events, other broader impacts include training of a graduate student and support of an early career researcher with no prior NSF support. All software tools developed during this research will be made freely and publicly available to the paleoclimate community, as will any compilations of publicly available data sets.