Walter Orr Roberts Lecture: Climate Variability and Society: Are We Fooling Ourselves?

CLIMATE VARIABILITY AND SOCIETY: ARE WE FOOLING OURSELVES?

Jonathan Overpeck, Institute for the Study of Planet Earth and Department of Geosciences, University of Arizona, Tucson, AZ; jto@u.arizona.edu

The current best estimate for global warming by the end of this century is 1.5 to 4.5°. However, many in the climate science field, and probably more outside our field, are under the impression that the warming will be at the low end of this range, and thus relatively easy to deal with. Is it possible that the scientific community, and society in general, may be fooling themselves with regard to possible future change? My concern stems from the paleoclimatic record. Over the last several years there have been a number of "multi-proxy" (e.g., tree-ring, ice core, sediment, and coral) reconstructions of surface air temperature for the last 400 to 1000 years. These reconstructions all agree that late 20th century temperatures were the warmest of the whole period (in part due to humans), but they disagree with colder borehole-based reconstructions of pre-industrial temperatures. New 3D climate model simulations for the last 500 years indicate that the multi-proxy reconstructions may get interannual to decadal variability correct (perhaps excepting the response to large volcanic eruptions), but that the boreholes provide the best reconstructions of century-scale variability. If this is true, the implication is that climate sensitivity to a doubling of CO2 could be at least 4°C. If climate sensitivity is this high, it could help explain why sea level was 4-6m higher than today during the last interglacial. Climate forcing of 125ka was not that different from recent pre-industrial times, and yet it appears that the most of the Greenland Ice Sea may have melted. Perhaps this also makes sense in light of observations suggesting that Arctic Sea ice cover has shrunk and thinned substantially over the last several decades. Let's now consider that all of the above is unwarranted concern; global warming turns out to be modest. From a paleoenvironmental perspective, this is not that comforting. Even in the absence of warming, we know that the climate system can generate substantial climatic extremes. On one hand the climate community might be telling society that future climate change isn't much to worry about, when in fact we know that a 5-7 year drought hits North America once or twice per century, and multi-decadal "mega-droughts" happen at least once or twice per millennium. The really troubling point is that we don't have a predictive understanding of these events, nor do we have an understanding of what caused other types of events (e.g., increases in flood or hurricane frequency) that could be as effective as prolonged drought in wreaking havoc on human life and livelihood. Finally, the paleoclimatic record might also indicate that the frequency of difficult-to-predict shifts in climate variability (e.g., to periods of greater drought frequency and duration) may increase as a function of climate mean state change. From a paleoclimatic perspective, 2°C global warming in 100 years is still an unusually rapid change in mean state. One way or the other, we could be in for undesirable climate surprises, and it is our responsibility not to underestimate this possibility.