1 Introduction

The Cooperative Holocene Mapping Project (COHMAP) aims at producing global maps of paleoclimatic conditions for 3, 6, 9, 12, and 15 ka BP (Reference WebbWebb 1984).

These maps, along with the Climate: Long-range Investigation Mapping and Prediction (CLIMAP) reconstruction of climatic conditions for 18 ka BP (CLIMAP Project Members 1981), will provide “snapshots” of past climates at intervals of 3 ka covering a period of major climatic change: full glacial, deglaciation, full interglacial, present. Diagnostic studies of the changes in the large-scale climatic patterns should advance our knowledge of the dynamics of climate.

To complement the sequence of global paleoclimatic maps, COHMAP is preparing a sequence of general circulation model experiments for the same times: 3,6, 9, 12, 15, and 18 ka BP. Whereas zonal-average climate models are well suited for a wide range of paleoclimate sensitivity experiments involving orbital parameter changes, lower boundary condition changes, and internal feedback parameterizations (for example, Reference Suarez and HeldSuarez and Held 1976, Reference Suarez and Held1979, Reference Schneider and ThompsonSchneider and Thompson 1979 Reference Birchfield, Weertman and LundeBirchfield and others 1982), the general circulation model provides the unique opportunity for detailed regional- and continental-scale intercomparisons of model simulations with the observed paleoclimate.

The paleoclimate simulations described here are for 6, 9, and 18 ka BP and involve studies of the response of a general circulation model to both lower boundary condition changes (sea-surface temperature, glacial ice, etc.) and changes of the seasonal cycle of solar radiation. This summary is in the form of a brief progress report because many aspects of the computational work planned by COHMAP remain incomplete. The experiments for 3, 12, and 15 ka BP must be finished, and further experimentation is planned for all of the designated times using improved versions of the general circulation model. Moreover, detailed intercomparison of the model results and the paleoclimatic observations remains to be accomplished.

2 Review of Previous Work

Gates (Reference Gates and Berger1980, Reference Gates1982) has reviewed paleoclimate modeling studies and summarized the problems and approaches. In order to simulate the past climate with a general circulation mortel, it is necessary to specify certain external conditions, such as the magnitude and seasonal distribution of solar radiation, and certain lower boundary conditions, such as the location and size of ice sheets. Ice sheets are, of course, an integral part of the climate system and would not be specified as a boundary condition in a completely general climate model. However, the time required for the growth and decay of ice sheets is long compared to the time-scale of processes normally included in atmospheric general circulation models, and it is therefore convenient for some purposes to prescribe the ice sheet or other lower boundary conditions and study a restricted (less general) climate system.

The sequence of paleoclimate experiments planned by COHMAP involves specification of both external and lower boundary conditions and combinations thereof.The sensitivity of a low resolution general circulation model to changes of Earth’s orbital parameters has been studied by Reference KutzbachKutzbach (1981) and Reference Kutzbach and Otto-BliesnerKutzbach and Otto-Bliesner (1982). The seasonal cycle of solar radiation at 9 ka BP was strikingly different from the present because the perihelion was in July (it is now in January) and the axial tilt of the Earth was greater then. There was an amplified seasonal cycle of solar radiation in the northern hemisphere compared to the present (more radiation in July, less in January) and the general circulation model simulated stronger summer and winter monsoon circulations over Africa and southern Asia. In one version of the 9 ka BP experiment, the location and size of the remnant North American ice sheet was specified in order to isolate both the combined and separate model response to 9 ka BP radiation and ice-sheet conditions (Reference Kutzbach and Otto-BliesnerKutzbach and Otto-Bliesner 1982).

The general circulation model experiments with 9 ka BP boundary conditions have been repeated with a different general circulation model, the National Center for Atmospheric Research Community Climate Model (NCAR CCM) with similar results (Reference Kutzbach, Guetter, Imbrie and BergerKutzbach and Guetter in press). The NCAR CCM has higher spatial resolution and more detailed physical parameterizations compared to the low resolution model used in the earlier experiments, and therefore the results can be compared more readily with paleoclimatic evidence.

At around 18 ka BP, the external solar radiation conditions were similar to those at present (in July and January), but there were major changes, compared to present, in sea-surface temperature patterns, sea-ice extent, land albedo, and, of course, ice-sheet location and size. A number of experiments with general circulation models have used the lower boundary conditions specified by CLIMAP (1976) in order to simulate the atmospheric circulation for the time of glacial maximum (Reference GatesGates 1976[a], Reference Gates[b], Reference Manabe and HahnManabe and Hahn 1977).

3 Description of Model and Experiments

3(b) Experiments

The changes of solar radiation and lower boundary conditions for 18, 9, and 6 ka BP are summarized in Table I. The reasons for prescribing the lower boundary conditions for 6 and 9 ka BP at modern values are detailed in Reference Kutzbach and Otto-BliesnerKutzbach and Otto-Bliesner (1982). The one exception was the North American ice sheet, which, at 9 ka BP, still covered 30 grid points in the model, i.e. an area of 5.4 × 10⁶ km .The simulated climate of 9 ka BP is mainly a response to the changed orbital parameters; the North American ice sheet has a major effect on the simulated climate only in the vicinity of the ice sheet itself (Reference Kutzbach and Otto-BliesnerKutzbach and Otto-Bliesner 1982, Reference Kutzbach, Guetter, Imbrie and BergerKutzbach and Guetter in press). At 18 ka BP, the solar radiation values for January and July are within 1% of modern values at all latitudes, but there are major changes in lower boundary conditions. For example the North American ice sheet covers 77 model grid points (14.4 × 10⁶ km²) and the European ice sheet covers 56 model grid points (7.9 × 10⁶ km²). The 18 ka BP conditions for sea-ice extent, sea-surface temperature, and land albedo are found in CLIMAP (1981). The lower boundary conditions from CLIMAP (1981) differ somewhat from CLIMAP (1976) estimates.

Table I

Orbital Parameter Values and Lower Boundary Conditions for General Circulation Experiments for Present Conditions (0) and for 6, 9 and 18 ka BP.

4 Results

A detailed description of the results of the 6, 9, and 18 ka BP simulations is in preparation. This preliminary report summarizes certain area-average results for the land surface and for the region of the monsoon climates of the northern hemisphere, and certain global-average statistics (land plus ocean) (Table II).

Table II

Area-Average Values of Incoming Solar Radiation (S), Simulated Surface Temperature (T) and Precipitation(P)

For July, the increased solar radiation at 6 and 9 ka BP (compared to present) is associated with higher northern hemisphere land temperature (higher by 1.5 and 1.8 K, respectively) and increased monsoon rains (greater by 1.1 and 1.4 mm d⁻¹, or about 20%). In contrast, the changed lower boundary conditions at 18 ka BP are associated with lower northern hemisphere land temperature (lower by 5.7 K) and decreased rains (about 10% less for the global average, the northern hemisphere land average, and the monsoon-region average).

For January, the decreased solar radiation at 6 and 9 ka BP (compared to present) is associated with lower temperatures. Precipitation changes are small. With boundary conditions for 18 ka BP, the northern hemisphere land temperature 1s lowered by 7.4 K and precipitation is decreased.

Viewed as a sequence of climatic “snapshots”, these area-averaged results indicate that, compared to present, the climate of 18 ka BP was colder-and drier than now in both July and January; seasonality (the July to January range) was not greatly different than now. By 9 ka BP, and continuing to 6 ka BP, the climate had warmed dramatically compared to conditions at 18 ka BP. Moreover, the amplified seasonal cycle of solar radiation in the northern hemisphere produced an increased range of the seasonal temperature extremes (warmer northern hemisphere summers and colder northern hemisphere winters) and increased summer monsoon rains. During the period since 6 ka BP, the seasonal extremes of northern hemisphere land temperature have been reduced and monsoon rainfall has decreased.

With this overview of the broad-scale features of the simulated climate of the past 18 ka, a more detailed analysis of regional climatic sequences {for North America, for polar regions, etc.) will be prepared and compared with the paleoclimatic record.