are allocated to identifying threats and reducing risks. The public and elected officials have consistently supported such risk-based allocations of resources.

Unlikely Extreme Events in an Unprecedented Climate

From the perspective of security, events that could disrupt the social and political systems of importance to U.S. national security are particularly important. Certain climate events—most likely, rare and extreme events— could meet this criterion. These are, by definition, events on the long tails of the probability distributions of climate events. Unfortunately, the climate science community is much less confident in its skill at projecting the rate of change in the frequency or magnitude of events at the extremes of such distributions than it is in projecting averages or even the likelihoods of events in the middle 80 to 90 percent of the distributions. One reason for caution is, as already noted, that there is a very limited population of such events to use for validating models. Moreover, the spatial resolution of climate change models is often coarser than is required to resolve the spatial structure of many extreme events to the degree needed for security analysis.

Another factor limiting confidence in the projections of extreme climate events is that the fundamental attributes of Earth’s climate system have moved or very soon will move beyond the bounds of experience on which models are based. For example, the concentrations of greenhouse gases (GHGs) in the atmosphere are now greater than they have been for at least 800,000 years (National Research Council, 2010a), and the current rate of carbon dioxide accumulation in the atmosphere is at least an order of magnitude greater than the natural rate that prevailed prior to the rise of human civilizations (see [accessed November 14, 2012]). As climate moves outside the range of experience, models of the effects of higher GHG concentrations cannot be validated against the kinds of high-resolution observational data that provide the most desirable basis for model testing. Global average temperature already is or soon will be higher than it has been at any time in recorded human history, and it is increasing at an unprecedented rate (Intergovernmental Panel on Climate Change, 2007; National Research Council, 2009). Moreover, the variance in temperature indicators has been increasing over the past half century (see Figure 1-1). All of these phenomena—higher temperatures, higher variation in temperatures, and rapid change—increase the likelihood of the occurrence of extreme temperature events. In addition, as the climate moves outside the bounds of the experience on which existing models are validated regarding the averages and ranges of variation of climate parameters and their rates of change, scientists may attach less confidence to projected extremes simulated by the models.

The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement