You are here

A Bayesian multinomial regression model for paleoclimate reconstruction with time uncertainty

Authors: 
Publication Type: 
Refereed Original Article
Abstract: 
It is a pleasure to be a discussant on this paper; one that exhibits the best of both applied statistics and cutting edge methodology. The authors should be congratulated for the clarity of their expression and I hope this paper reaches a wide audience of palaeoclimate scientists, statisticians and environmental scientists more generally. Palaeoclimate is important: it provides useful constraints on the speed at which the climate changes; it helps us learn more about the fragility of our existence (all human history is encompassed in a uniquely warm and stable climate) and provides a method by which to test our physics-based models that allow us to predict future climate change. In this discussion, I hope to highlight some of the real contributions of this paper, to point out some of the important non-statistical considerations (which, as applied statisticians in this area, we should be cognisant), and to contrast with the rapidly expanding mostly- Bayesian palaeoclimate statistics literature. In particular, accounting for time uncertainty is, I suspect, almost a unique challenge for timeseries analysis in palaeoclimate science. For this reason, it has been ignored for decades. Now with tools as in this paper, they can start to draw proper inferences on climate over time with suitably quantified uncertainties
Digital Object Identifer (DOI): 
10.1002/env.2401
Publication Status: 
Published
Date Accepted for Publication: 
Tuesday, 17 June, 2014
Publication Date: 
07/06/2014
Journal: 
Environmetrics
Volume: 
2016
Issue: 
27
Pages: 
431–433
Research Group: 
Institution: 
National University of Ireland, Dublin (UCD)
Open access repository: 
No