General circulation models (GCM)

Definition

The dynamics and physics that govern the atmosphere and oceans are the result of many cycles and processes interacting. A general circulation model (GCM) is a collection of mathematical equations that represent these cycles and processes using computational algorithms that can be solved on a supercomputer. In general, the components of the model, such as the atmosphere andocean, are modeled separately and interact only at their boundaries—for instance, at the interface between the atmosphere and the ocean. GCMs include equations to solve the dynamics (motions) of the atmosphere and ocean, as well as equations and parameterizations to represent atmospheric, oceanic, sea-ice, and terrestrial physics. Processes represented by parameterizations combine equations with proportionality constants, correlations, and table lookups based on observational and experimental data.

In order to be solved on a computer, the equations and variables incorporated in a GCM must be discretized, which is typically accomplished by dividing up the atmosphere and ocean on a grid and defining the key variables such as temperature and humidity at defined points on this grid. Using variable values and the governing and equations, one can compute new variable values a relatively short time in the future. This integration process is repeated to generate longer time sequences. Validation studies that compare the results of the model computation to experimental and observational data are required to acquire confidence in the accuracy of the simulation.

Significance for Climate Change

GCMs are the fundamental tools for predicting the future evolution of Earth’s climate and the impact that greenhouse gas (GHG) emissions or other environmental effects will have on future temperatures, rainfall, and other climatic conditions. GCMs arose from efforts at numerical weather prediction (NWP). The English mathematician Lewis Fry Richardson is credited with making the first NWP calculation when he attempted to predict the weather six hours into the future during World War I. Because he lacked computers, the calculations took six months to complete, but the basic approach was the same that would be used four decades later in NWP. The first working GCM is attributed to Normal Phillips, who completed a two-layer hemispheric atmospheric model in 1955. The late 1960s and the 1970s saw the introduction of and the first use of a GCM to study the effects of and pollutants in the atmosphere. From that point forward, GCMs were recognized as critical components in the study of climate change.

Four-component models (comprising atmosphere, ocean, sea ice, and land) are often referred to as atmosphere-ocean general circulation models (AOGCMs). The highest-resolution AOGCM grids have horizontal spacings of 1° (roughly 100 kilometers) and up to sixty vertical layers in the atmosphere. A typical use of these models takes an average of current conditions as the initial conditions then integrates the GCM computationally through future decades, while different parameters and boundary conditions are changed to account for different scenarios. Based on the results of these computations, the likely range of future temperatures and other future climatic conditions is determined to the extent possible given the accuracy of the model and the related inputs.

GCM simulations are also used to compute the magnitude of anthropogenic versus natural forcing of the environment in order quantitatively to isolate the human contribution to climate change. They are also used to assess the future impact of plans to reduce global GHG emissions. Key statements about global warming, such as the anticipated rise in temperature and sea level and changes in precipitation and storm patterns due to the release of anthropogenic GHGs over the next century, are the product of general circulation models.

Given their importance in the debate over global warming, GCMs are subject to much scrutiny. These models are not complete, with some physical processes generally not implemented in the models and other representations being the source of argument and uncertainty. Critical points of contention include the parameterization of cloud-radiation feedbacks and cloud microphysics, including precipitation; the impact of changes in incoming solar radiation; and the general variability of results when using different GCMs. A notable missing process in most GCMs is an explicit numerical treatment of the carbon cycle, causing most global warming studies to rely on predetermined changes in atmospheric GHG content. Ongoing work to improve modeling of physical processes, achieve higher-resolution simulations on more powerful computing platforms, and increase the use of validation studies will generate the mechanisms by which these concerns will be addressed. Even with their limitations, GCMs remain the primary tool for analyzing the future of Earth’s climate.

Bibliography

Houghton, John. Global Warming: The Complete Briefing. 4th ed. New York: Cambridge University Press, 2009.

Kochokov, Dmitrii, et al. "Neural General Circulation Models for Weather and Climate." Nature, 22 July 2024, doi.org/10.1038/s41586-024-07744-y. Accessed 17 Dec. 2024.

Leroux, Marcel. Global Warming: Myth or Reality? The Erring Ways of Climatology. New York: Springer, 2005.

Randall, David A., ed. General Circulation Model Development. San Diego, Calif.: Academic Press, 2000.

Randall, David A., et al. “Climate Models and Their Evaluation.” In Climate Change, 2007—The Physical Science Basis: Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, edited by Susan Solomon et al. New York: Cambridge University Press, 2007.