Science Briefs

Uncertainties in Understanding Low- and High-Latitude Climate Sensitivity Affect Ability to Predict Climate Change Impacts

Our need to limit future greenhouse warming is proportional to how severe we think the impacts of climate change will be. Skeptics feel the warming itself, and therefore the consequences, will be small. Most scientists, however, feel that significant warming is pretty much assured unless we limit the growth of greenhouse gases. But even researchers in the field cannot specify with confidence impacts that will arise, outside of providing some generalizations.

One reason for our inability to better define climate change consequences is explored in a recent article in the Bulletin of the American Meteorological Society, and that is our inability after some 30 years of research to understand the likely climate response in the tropics and in polar regions. For the same scenario of future greenhouse gas increases, climate models differ by a factor of two in terms of their predicted warming in both regions. This has obvious implications for our ability to predict events in the tropics, such as hurricanes and drought, and at high latitudes, such as sea ice and ice sheet melting (with sea level rise).

Figure 1, at right. Surface air temperature change in winter and summer when using doubled CO2 sea surface temperatures as calculated in the GISS (DBL CO2) and GFDL (ALT) models circa early-mid-1980s. The same type of distinctions still exist in model simulations today.

In addition, the temperature gradient between low and high latitudes governs many features of atmospheric dynamics, the movement of winds and storms. Without understanding how this gradient will change, we cannot predict how the dynamics will respond, and therefore what will happen to important meteorological parameters such as regional rainfall. In addition, the pattern of sea surface temperatures at low latitudes is extremely important for regional climate variations (shown, for example, by the increased likelihood of heavy winter rainfall in California when the eastern tropical Pacific warms in El Niño events). Our inability to understand the low latitude response includes a lack of consensus on how tropical regional patterns will change, adding to our uncertainty concerning regional climates.

Ironically, when we look at paleoclimates, both the cold climates of the Ice Ages, and the warm climates of the Tertiary (from 65 to 2.5 million years ago), the same uncertainty exists. We do not know how cold/warm the tropics were at these times, nor can we properly reproduce the high latitude responses from these climates in our models. The tropical paleoclimate proxies are conflicting and may be misinterpreted; the high latitude responses may be arising under different circumstances. So we cannot use paleo-observations to determine which, if any, of our models has the proper sensitivity in these regions — and in fact, models cannot reproduce what at face value seem to be the extreme changes in low-to-high latitudinal gradients suggested by paleo-data.

We can, however, investigate what parameters are responsible for the differences among our models. In the tropics, the prime reason for the diversity of response is associated with the models' cloud cover changes as climate warms. This is the result of different cloud and convection formulations used in the models, as well as differing atmospheric and oceanic dynamical responses associated with changing temperature gradients. At high latitudes a number of features play a role — different changes in sea ice and snow cover and their reflectivity; varying atmospheric and ocean dynamical changes; as well as cloud cover effects. A lot of these changes are interactive with temperature — once the models' predictions start diverging, these feedbacks respond and help create even larger differences.

Figure 2, above. Sea level pressure changes in the two solstice seasons from a GCM simulation with increased sea surface temperature gradient minus a simulation with a decreased gradient in the Atlantic (top row), in the Pacific (middle row), and increased gradient in the Atlantic along with a decreased gradient in the Pacific minus the reverse (bottom row). Where sea level pressure decreases there is generally more rain, with the reverse happening when sea level pressure increases. Note how the regional tendencies vary with the gradient change. While the canonical view is that latitudinal temperature gradients will decrease as climate warms (greater warming at high latitudes), increased gradients in the Pacific could result from semi-permanent El Niño conditions as simulated by some models; increased gradient in the North Atlantic could arise from decreased ocean circulation (i.e., North Atlantic Deep Water production) as also occurs in some models.

NASA has launched several satellites such as CloudSat and Calipso with the goal of better understanding cloud formulations. This year also represents the International Polar Year (IPY) organized through the International Council for Science and the World Meteorological Organization; it is aimed at improving our understanding of polar processes in both the Arctic and Antarctic. The hope is that efforts such as these will lead to gradual improvement in our understanding of climate sensitivity at low and high latitudes, and therefore a better ability to predict the likely consequences of climate warming. Until this is achieved, it will be hard to be specific about the societal impacts of future greenhouse gas emissions, an uncertainty that, it can be argued, should make us even more cautious about disturbing the system.


Rind, D., 2008: The consequences of not knowing low- and high-latitude climate sensitivity. Bull. Amer. Meteorol. Soc., 89, 855-864, doi:10.1175/2007BAMS


Please address all inquiries about this research to Dr. David Rind.