Precipitation and the Potential for Extreme Temperature Change
Sunlight penetrates the Earth's "blanket" of air to heat the ground, but some of the gases in the air do not permit heat from the ground to escape back into space. This heat-trapping, warming influence of the blanket of air over the Earth's surface is called the greenhouse effect, and it will become even stronger as greenhouse gases such as carbon dioxide, methane and water vapor increase in concentration. To predict changes in surface temperatures in the coming decades, scientists use global climate models (GCMs), which offer a general overview of future trends in surface temperature, but not reliable regional details that can cause extreme temperatures. For example, Figure 1 shows that one GCM simulation underestimated the observed average maximum surface air temperature over the eastern US during five summers by 4.6°C (8.3°F).
Recently at the Goddard Institute for Space Studies, we used a regional mesoscale model (RMM), which monitors climate variables at much smaller spatial intervals than the typical GCM, to bring GCM results into better focus. The regional model's predictions of surface temperature changes over the eastern United States were compared to parallel forecasts made by the same GCM. Figure 1 shows that this model simulated the same maximum temperatures only 2.4°C (4.3°F) higher than observed.
Our study determined that the GCM likely underestimates future air temperatures near the ground because it simulates too many rainy days on which clouds block sunlight and on which the wet ground is additionally cooled by evaporation (Figure 2). Note that the statistics of rain frequency inherently depend on the size of the area being monitored, since it rains more often somewhere within a large area than somewhere within a much smaller area.
However, even accounting for the relatively large area of the GCM's computational elements (400 km × 500 km), Figure 2 shows that the GCM still overestimates precipitation frequency. By comparison, the corresponding percentage of rainy days predicted by the regional mesoscale model for the same summers was lower and much more realistic (compare to the observed frequency within the mesoscale model's smaller computational elements, 36 km × 36 km).
Figure 3 shows the relationship between observed surface air temperature and observed precipitation frequency for 28 summers, demonstrating that summers were hotter when it rained less often. The GCM's overestimation of precipitation frequency interferes with its skill to simulate the warmest summers characterized by relatively low precipitation.
The consequences for models' predictions of the future temperature can be seen in Figure 4, which shows that the mesoscale model's projections of mean maximum summertime temperatures over the eastern US for July 2085 soar into the 95-110°F range, while the corresponding predictions for the GCM range between 75-95°F. Similarly conflicting temperature projections occurred for other summers as well.
The scientific literature indicates that other GCMs are also flawed by computations of too frequent precipitation and unrealistic morning showery precipitation. We are not aware of any other study that has documented the impact of the precipitation simulation imperfections on GCMs' predictions of surface air temperature, but the ability of such flawed models to predict global warming and its extremes could be compromised. This study suggests that climate change will cause more extreme temperatures than implied by previous GCM studies.
Science Brief: NASA Study Suggests Extreme Summer Warming in the Future
Lynn, B.H., R. Healy, and L.M. Druyan (2007), An analysis of the potential for extreme temperature change based on observations and model simulations, J. Climate, 20, 1539-1554, doi:10.1175/JCLI4219.1.