Most of the predictions about the impacts of future global warming are derived from complex computer models that simulate a climate system too complex to interpret any other way.
A growing number of global climate models have been developed independently by different researchers around the world in recent years, but the models are not without their limitations.
Complex interactions often have to be greatly simplified for the computer, and models usually have a very "coarse" resolution - the entire Susquehanna Basin may be represented by only one or two information points on a global model.
Predictions are also imperfect because they are based on incomplete knowledge of the climate system. Nonetheless, the variety of independently produced models allow scientists to compare results and see where there is common agreement.
In 1995, Eric Barron, a climate modeler at Pennsylvania State University, organized a meeting of a large group of climate modelers to assess the levels of certainty associated with various model predictions. They ranked their conclusions in terms of probability as follows:
- The increases in carbon dioxide and the decreases in stratospheric ozone will result in a large cooling in the upper atmosphere.
- The global mean surface temperature will increase between 0.9 to 3.6 degrees Fahrenheit over the period of 1990 to 2050. The best estimate for a climate change resulting from a doubling of preindustrial carbon dioxide is a warming of 1.5 to 4.5 degrees Celsius, with 2.5 degrees being the most probable estimate. The estimated warming will be reduced if sulfur emissions, which reflect sunlight away from the planet, are not controlled. Sulfur is a prime cause of acid rain and haze, but it also reflects sunlight away from the planet. The actual temperature change may be outside the range described here if natural climate variations are large.
- A warming of the surface temperature of the globe will lead to an increase in average global precipitation because evaporation rates are closely tied to surface temperatures.
- Northern Hemisphere sea ice will be reduced. Changes to the Southern Hemisphere are less certain.
- Arctic land areas should experience amplified winter time warming associated with the reduction in land surface snow cover.
- Global warming will influence sea level. The most reasonable estimates are a sea level rise of 5 to 40 centimeters by 2050, compared with a rise of 5 to 12 centimeters if the rate of rise over the past century continues. These estimates ignore the long-term response of the ice caps or any potential catastrophic collapse of the West Antarctic ice sheet, which are the subject of considerable de-bate.
- Solar variability over the next 50 years will not induce a prolonged climate forcing that is significant in comparison with the effects of the increasing concentrations of carbon dioxide and other greenhouse gases.
- Summer Northern Hemisphere midlatitude continental dryness will increase.
- The increase in the amount of at-mospheric moisture will produce regions of higher rainfall, including the high latitudes.
- Some regions may experience warming that is smaller than the global average, particularly in oceanic regions where surface waters are mixed downward or where deeper waters upwell.
- Historical records suggest that a few explosive volcanic eruptions are likely to occur during the next half century. These transient events will result in some short-term relative cooling.
- There is the potential for multifaceted and complicated, even counter-intuitive changes in climate variability, hence predictions of changes in climate variability are uncertain. This aspect of climate change includes many possibilities, such as changes in winter storminess, changes in El Nino frequencies and variations in thunderstorms.
- At present, there is only a very limited capability to estimate how the climate of various regions will respond, although regional scale (from the scale of large metropolitan areas to the scale of states) changes are likely to be different from the global average changes.
- An increase in tropical storm intensity is plausible due to higher sea surface temperatures associated with global warming; however, there are too many unresolved issues to be certain.
- The details of climate change over the next 25 years are also uncertain. Uncertainties in the factors that control natural variability, in the model simulations and in the changes in atmospheric chemistry, make it extremely difficult to predict or even suggest decade-to-decade changes in climate. In any given decade, the changes in temperature and related variables could be substantially less than or more than the trend predicted by a climate model. Warming estimates in terms of degrees per decade and their use to analyze a single decade are unwarranted and misleading.
- Climate-biosphere feedback relationships may modify the magnitude of climate change. Whether the feedbacks will amplify or moderate climate change is not known.