What is a Climate Model?
A climate model is a scientific tool that can be used to understand the functioning of the climate system. Climate models consist of sets of mathematical expressions that describe the physical processes associated with climateÑe.g., seasonal changes in sunlight, large-scale movements of air masses, evaporation and condensation of water vapor, absorption of heat from the atmosphere into the oceans, and so forth. In most models, the atmosphere is sectioned off into "cells" roughly 500 kilometers (about 300 miles) on a side at the EarthÔs surface, and the cells are stacked about twenty layers deep. The vertical layers reach 30 kilometers (about 20 miles) or more into the atmosphere. For each cell and period of time (e.g., an hour), the sets of mathematical expressions are solved to predict such variables as temperature, humidity, air pressure, and wind speed. In simulating a century's worth of climate, the process is repeated a million or more times.
At the core of climate models are expressions of physical principles such as the conservation of energy or mass (e.g., of air or water vapor). Such "laws" govern interactions among the atmosphere, oceans, sea ice, land, and vegetation. Because many processes take place at scales smaller than are represented by the grid cells in most models, information from field experiments and observations of how the climate varies from season-to-season and year-to-year are used to develop approximations to the equations. For example, observational data on the occurrence, types, and altitudes of clouds as they relate to temperature and humidity levels are used to predict the form and character of clouds in model simulations. A potential problem with using empirical relationships rather than fundamental physical principles is the possibility that observed relationships may not hold for conditions different than those under which the observations were made.
In the climate models developed to date, atmospheric conditions have been treated more comprehensively than ocean, land, and vegetation conditions. Greater attention now is being given to the development and integration of submodels that more realistically represent non-atmospheric components of the integrated Earth system. The seasonal climate cycle provides a good test of the ability of models to simulate the short-time-scale processes involved. Additional tests of models include simulation of the effect of volcanic eruptions, El Niño events, and climate changes over historical and geological time periods.