Assessing confidence in Pliocene sea surface temperatures to evaluate predictive models
Author: ["Harry J. Dowsett","Marci M. Robinson","Alan M. Haywood","Daniel J. Hill","Aisling M. Dolan","Danielle K. Stoll","Wing-Le Chan","Ayako Abe-Ouchi","Mark A. Chandler","Nan A. Rosenbloom","Bette L. Otto-Bliesner","Fran J. Bragg","Daniel J. Lunt","Kevin M. Foley","Christina R. Riesselman"]
Publication: Nature Climate Change
CITE.CC academic search helps you expand the influence of your papers.
Abstract
In light of mounting empirical evidence that planetary warming is well underway, the climate research community looks to palaeoclimate research for a ground-truthing measure with which to test the accuracy of future climate simulations. Model experiments that attempt to simulate climates of the past serve to identify both similarities and differences between two climate states and, when compared with simulations run by other models and with geological data, to identify model-specific biases. Uncertainties associated with both the data and the models must be considered in such an exercise. The most recent period of sustained global warmth similar to what is projected for the near future occurred about 3.3–3.0 million years ago, during the Pliocene epoch. Here, we present Pliocene sea surface temperature data, newly characterized in terms of level of confidence, along with initial experimental results from four climate models. We conclude that, in terms of sea surface temperature, models are in good agreement with estimates of Pliocene sea surface temperature in most regions except the North Atlantic. Our analysis indicates that the discrepancy between the Pliocene proxy data and model simulations in the mid-latitudes of the North Atlantic, where models underestimate warming shown by our highest-confidence data, may provide a new perspective and insight into the predictive abilities of these models in simulating a past warm interval in Earth history. This is important because the Pliocene has a number of parallels to present predictions of late twenty-first century climate. Sea-surface-temperature proxy data for a period of natural climate warming during the Pliocene are used in this study to show how palaeoclimatic data can help ‘ground truth’ numerical models, increasing the confidence in these same models for projecting future climate.
Cite this article
Dowsett, H., Robinson, M., Haywood, A. et al. Assessing confidence in Pliocene sea surface temperatures to evaluate predictive models. Nature Clim Change 2, 365–371 (2012). https://doi.org/10.1038/nclimate1455
