Numerical models of earth's climate provide a means of calculating the seasonal cycle of temperature, precipitation and wind. These models are based upon physical laws, such as the first law of thermodynamics (heating of a gas causes expansion and warming), and F = ma (forces (F) cause masses (m) to accelerate (a)). Given these laws governing the internal workings of the atmosphere (and land surface and ocean), it is then only necessary to specify certain so-called external boundary conditions in order to calculate the climate. Examples of such external conditions include the seasonal and latitudinal distribution of solar radiation, the composition of the atmosphere, the spatial distribution of land and ocean, and the location and height of mountains.
How do these external boundary conditions and governing laws or equations combine to yield climate? An example. In summertime, solar radiation (externally imposed) raises the temperature of the air more over continents (low heat capacity) than over oceans (high heat capacity). This produces air of relatively low density over land and relatively high density over ocean. The density difference leads to slight differences in the overall weight of a column of air over ocean, compared to land. And this difference represents a force that accelerates air from ocean to land. Here, oversimplified, is the explanation for the so-called summer monsoon circulations that carry maritime air masses onshore in Asia, Africa, and southeastern North America every summer producing abundant rains.
With the aid of large computers, these repetitive calculations of temperature, wind and precipitation are carried out over what can be visualized as large horizontal and vertical grids involving tens of thou- sands of locations. It is necessary to perform these calculations at intervals on the order of hours and extending over many years, because the climate of the model, like the real climate, is a synthesis of the ever-changing atmospheric conditions on the scale of hours, days, weeks, months and years.
How good are these models? Simulations of present-day climates depict the continent-scale circulations and the associated thermal and hydrologic regimes fairly well. However, if one examines regional fea- tures, one readily finds temperatures too high or low, and precipitation too much or too little, compared to observations; in some cases the climatic patterns are simulated correctly, but shifted from their correct location. Climate models should produce increasingly accurate regional simulations when calculations are made on a finer grid. Current models use a coarse grid (generally around 500 km by 500 km) because of computer limitations. Other needed improvements include coupling the atmospheric portion of climate models to ocean general circulation models.
Models of present-day climate can rather easily be modified to simulate past or future climates by appropriate changes in external boundary conditions. By changing the composition of the model's atmosphere to include increased amounts of carbon dioxide, one can study scenarios of future climate. By changing earth's orbital parameters and modifying the surface "topography" by inserting ice sheets, one can simulate climates of the past.
The COoperative Holocene MApping Project (COHMAP) produced a series of paleoclimate simulations for 18-, 15-, 12-, 9-, 6- and 3000 years ago (COHMAP, 1988; Wright et al., 1993; Kutzbach and Guetter, 1986). These climatic "snapshots" consisted of averages of January and July conditions (based on three Januarys and three Julys). Details of the experiments are reported in the above-mentioned publications. The simulated values of wind, temperature, precipitation and other variables are summarized for the entire earth on grid boxes of 4.4° latitude by 7.5° longitude. This computational grid is large compared to the space scale of local paleoenvironmental studies. Moreover, it is prudent to average the model results over several grid boxes to obtain representative estimates of regional paleoclimates. In spite of this scale mismatch between model grid and observations, the model simulations can be of value for comparison with paleoclimate observations for several reasons. First, the model-simulated climate is the result of changes in specific external boundary conditions, such as solar radiation or prescribed ice sheets. Therefore it is usually possible to understand cause and effect relationships and assess both the direction and magnitude of the climatic changes produced by imposed changes in boundary conditions. Second, the model provides a spatial context for relating changes in a particular area to a broader regional and continental perspective. Third, the model provides a temporal context for comparing the simulated changes in the climate of a particular region, say, for the past 18,000 years, to the observed paleoenvironmental history of the region.
If you are interested in comparing these simulations with your
paleoenvironmental observations, you may access the COHMAP simulations
through the NGDC computer:
ftp.ngdc.noaa.gov (IP address 192.149.148.109).
We plan on having gridded ASCII data sets available at NGDC by the end of
August. These data sets will include a three year ensemble average of
the CCM0 perpetual January and July runs for the COHMAP experiment time
periods mentioned above, plus the model modern climate. There will be
two files per experiment, (January and July). Each file will contain
ASCII grids (40° latitude x 48° longitude grid points) for the complete
list of output variables. A readme file will contain formatting
information and an explanation of the output variables.
Within the next year, we hope to make available a new set of simulation experiments covering the same time period but based upon an improved climate model. We will also be participating in a Paleoclimate Modeling Intercomparison Project (PMIP) that aims to intercompare the simulations of several modeling groups for 6000 yr BP and for the Last Glacial Maximum.
References.
COHMAP (1988). Climatic changes of the last 18,000 years: Observations and Model Simulations. Science 241, 1043-1052.
Kutzbach, J.E., and P.J. Guetter (1986). The influence of changing orbital parameters and surface boundary conditions on climate simulations for the past 18,000 years. Journal of the Atmospheric Sciences 43(16), 1726-1759.
Wright, H.E., J.E. Kutzbach, T. Webb III, W.F. Ruddiman, F.A. Street-Perrott, and P.J. Bartlein, eds. (1993). Global Climates since the Last Glacial Maximum, University of Minnesota Press, Minneapolis, MN, 569 pp.