• Barnola, J. M., D. Raynaud, Y. S. Korotkevich, and C. Lorius, 1987:Vostok ice core provides 160 000-year record of atmospheric CO2. Nature,329, 408–418.

  • Berger, A., 1978: Long-term variation of daily insolation and quaternary climate change. J. Atmos. Sci.,35, 2362–2367.

  • Boer, G. J., and M. Lazare, 1988: Some results concerning the effect of horizontal resolution and gravity-wave drag on simulated climate. J. Climate,1, 789–806.

  • ——, and B. Denis, 1997: Numerical convergence of the dynamics of a GCM. Climate Dyn.,13, 359–374.

  • Chervin, R. M., and S. H. Schneider, 1976: On determining the statistical significance of climate experiments with general circulation models. J. Atmos. Sci.,33, 405–412.

  • Coe, M. T., and G. B. Bonan, 1997: Feedback between climate and surface water in northern Africa during the middle Holocene. J. Geophys. Res.,102, 11 087–11 101.

  • COHMAP members, 1988: Climate changes of the last 18,000 years:Observations and model simulations. Science,241, 1043–1052.

  • Delworth, T., and S. Manabe, 1989: The influence of soil wetness on near-surface atmospheric variability. J. Climate,2, 1447–1462.

  • de Noblet, N., P. Braconnot, S. Joussaume, and V. Masson, 1996: Sensitivity of simulated Asian and African summer monsoons to orbitally induced variations in insolation 126, 115, and 6 kyr BP. Climate Dyn.,12, 589–603.

  • de Vernal, A., A. Rochon, C. Hillaire-Marcel, J.-L. Turon, and J. Guiot, 1993: Quantitative reconstruction of sea-surface conditions, seasonal extent of sea–ice cover and meltwater discharges in high-latitude marine environments. Ice in the Climate System, W. R. Peltier, Ed., Pergamon, 611–621.

  • Foley, J. A., 1994: The sensitivity of the terrestrial biosphere to climate change: A simulation of the middle Holocene. Global Biogeochem. Cycles,8, 505–525.

  • Guiot, J., R. Cheddadi, I. C. Prentice, and D. Jolly, 1996: A method of biome and land surface mapping from pollen data: Application to Europe 6000 years ago. Paleoclimates,1, 311–324.

  • Hall, N. M. J., and P. J. Valdes, 1997: A GCM simulation of the climate 6000 years ago. J. Climate,10, 3–17.

  • Hewitt, C. D., and J. F. B. Mitchell, 1996: GCM simulations of the climate of 6 kyr BP: Mean changes and interdecadal variability. J. Climate,9, 3505–3529.

  • Joussaume, S., and P. Braconnot, 1997: Sensitivity of paleoclimate simulation results to seasonal definitions. J. Geophys. Res.,102, 1943–1956.

  • Kutzbach, J. E., and B. L. Otto-Bliesner, 1982: The sensitivity of the African–Asian monsoonal climate to orbital parameter changes for 9000 years B. P. in a low-resolution general circulation model. J. Atmos. Sci.,39, 1177–1188.

  • ——, and P. J. Guetter, 1986: The influence of changing orbital parameters and surface boundary conditions on climate simulations for the past 18 000 years. J. Atmos. Sci.,43, 1726–1759.

  • ——, and R. G. Gallimore, 1988: Sensitivity of a coupled atmosphere/mixed layer ocean model to changes in orbital forcing at 9000 years B. P. J. Geophys. Res.,93, 803–821.

  • Laprise, R., and C. Girard, 1990: A spectral general circulation model using a piecewise-constant finite-element representation of a hybrid vertical coordinate system. J. Climate,3, 32–52.

  • Laskar, J., F. Joutel, and F. Boudin, 1993: Orbital, precessional, and insolation quantities for the Earth from −20 Myr to +10 Myr. Astron. Astrophys.,270, 522–533.

  • Legates, D. R., and C. G. Willmott, 1990: Mean seasonal and spacial variability in gauge-corrected, global precipitation. Int. J. Climatol.,10, 111–127.

  • Li, C., and M. Yanai, 1996: The onset and interannual variability of the Asian summer monsoon in relation to land-sea thermal contrast. J. Climate,9, 358–375.

  • McFarlane, N. A., G. J. Boer, J.-P. Blanchet, and M. Lazare, 1992: The Canadian Climate Centre second-generation general circulation model and its equilibrium climate. J. Climate,5, 1013–1044.

  • Meehl, G. A., 1994: Influence of the Land Surface in the Asian summer monsoon: External conditions versus internal feedbacks. J. Climate,7, 1033–1049.

  • Mitchell, H. L., C. Charette, S. J. Lambert, J. Halle, and C. Chouinard, 1993: The Canadian global data assimilation system: Description and evaluation. Mon. Wea. Rev.,121, 1467–1492.

  • ——, C. Chouinard, C. Charette, R. Hogue, and S. Lambert, 1996: Impact of a revised analysis algorithm on an operational data assimilation system. Mon. Wea. Rev.,124, 1243–1255.

  • Mitchell, J. F. B., 1993: Modelling of paleoclimates: Examples from the recent past. Philos. Trans. Roy. Soc. London B,341, 267–275.

  • ——, N. S. Grahame, and K. J. Needham, 1988: Climate simulations for 9000 years before present: Seasonal variations and effect of the Laurentide Ice Sheet. J. Geophys. Res.,93, 8283–8303.

  • Peltier, W. R., 1994: Ice age paleotopography. Science,265, 195–201.

  • Peng, C. H., J. Guiot, E. Van Campo, and R. Cheddadi, 1995: Temporal and spatial variations of terrestrial biomes and carbon storage since 13 000 yr bp in Europe: Reconstruction from pollen data and statistical models. Water, Air Soil Pollut.,82, 375–390.

  • Phillipps, P. S., and I. M. Held, 1994: The response to orbital perturbations in an atmospheric model coupled to a slab ocean. J. Climate,7, 767–782.

  • Prell, W. L., and J. E. Kutzbach, 1992: Sensitivity of the Indian monsoon to forcing parameters and implications for its evolution. Nature,360, 647–652.

  • Prentice, K. C., and I. Fung, 1990: The sensitivity of terrestrial carbon storage to climate change. Nature,346, 48–51.

  • Quinn, T. R., S. Tremaine, and M. Duncan, 1991: A three million year integration of the Earth’s orbit. Astro. J.,101, 2287–2305.

  • Rasmusson, E. M., and T. H. Carpenter, 1983: The relationship between the eastern Pacific sea surface temperature and rainfall over India and Sri Lanka. Mon. Wea. Rev.,111, 517–528.

  • Ruddiman, W. F., and A. C. Mix, 1993: The Northern and equatorial Atlantic at 9000 and 6000 yr bp. Global Climates Since the Last Glacial Maximum, H. E. Wright et al., Eds., University of Minnesota, 94–124.

  • Saurez, M. J., and I. M. Held, 1979: The sensitivity of an energy balance model to variations in the orbital parameters. J. Geophys. Res.,84, 4825–4836.

  • Sperber, K. R., S. Hameed, G. L. Potter, and J. S. Boyle, 1993: Simulations of the Indian and East-Asian summer monsoon in the ECMWF model: Sensitivity to horizontal resolution. PCMDI Rep. 12, 35 pp.

  • Street, F. A., and A. T. Grove, 1979: Global maps of lake-level fluctuations since 30 000 yr bp. Quat. Res.,12, 83–118.

  • Street-Perrott, F. A., D. S. Marchand, N. Roberts, and S. P. Harrison, 1989: Global lake-level variations from 18 000 to 0 years ago: A paleoclimatic analysis. U.S. Dept. of Energy Tech. Rep. 46, 44 pp.

  • ——, J. F. B. Mitchell, D. S. March, and J. S. Brunner, 1990: Milankovitch and albedo forcing of the tropical monsoon: A comparison of geological evidence and numerical simulations for 9000 ybp. Philos. Trans. Roy. Soc. Edinburgh,81, 407–427.

  • Tibaldi, S., T. N. Palmer, C. Brankovic, and U. Cubasch, 1990: Extended-range prediction with ECMWF models: Influence of horizontal resolution on systematic error and forecast skill. Quart. J. Roy. Meteor. Soc.,116, 835–866.

  • Vance, R. E., A. B. Beaudoin, and B. H. Luckman, 1995: The paleoecological record of 6 ka BP climate in the Canadian prairie provinces. Geogr. Phys. Quat.,49, 81–98.

  • Yan, Z., and N. Petit-Maire, 1994: The last 140 ka in the Afro-Asian arid/semi-arid transitional zone. Paleogeo. Paleoclimatol. Paleoecol.,110, 217–233.

  • Yu, G., and S. P. Harrison, 1996: An evaluation of the simulated water balance of Eurasia and northern Africa at 6000 ybp using lake status data. Climate Dyn.,12, 723–735.

  • Zoltai, S. C., 1995: Permafrost distribution in peatlands of west-central Canada during the Holocene warm period 6000 years BP. Geogr. Phys. Quat.,49, 45–54.

  • Zwiers, F. W., 1987: A comparison of climates simulated by a general circulation model when run in the annual cycle of perpetual modes. Mon. Wea. Rev.,115, 2626–2644.

  • ——, 1993: Simulation of the Asian summer monsoon with the CCC GCM-1. J. Climate,6, 470–486.

  • View in gallery

    Incoming solar radiation anomaly at the top of the atmosphere (insolation anomaly) for the difference between 6 kyr BP and modern in W m−2 as a function of latitude and time of year. The contour interval is 4 W m−2.

  • View in gallery

    Approach to equilibrium in the computed SST model simulation, for both the present-day and 6-kyr BP time periods. Modern global average JJA 2-m surface air temperature (°C) vs model time (yr) for present day is denoted by the solid line, while that for the 6-kyr BP simulation is denoted by the dashed line. The solid line traversing both curves denotes the 10-yr running mean for the time series. The inset displays the 6-kyr BP JJA global average sea–ice amount in kg m−2. The 6-kyr BP simulation was extended from the equilibrium configuration of the modern control at year 52 of simulation time.

  • View in gallery

    Latitude vs time of year for the 2-m surface air temperature anomaly (°C) between 6 kyr BP and modern. The contour interval is at 0.5°C. Panels (a) and (b) are for the fixed SST experiment zonally averaged over land and ocean, respectively. Panels (c) and (d) are the same as for (a) and (b), but for the computed SST experiment.

  • View in gallery

    Global distributions of DJF and JJA 2-m surface air temperature anomalies (°C) for the fixed and computed SST experiments. The contour interval is at 1°C with shaded regions designating changes below −1°C for DJF and changes above 1°C for JJA.

  • View in gallery

    Insolation forcing and SST response in the computed SST experiment over the tropical ocean area averaged from 24°S to 24°N. Insolation anomalies at 6 kyr BP as compared to modern are designated by the dashed line while the solid line denotes changes in surface air temperature at 2 m between 6 kyr BP and present. Thick solid and dashed horizontal lines denote the annual mean surface air temperature and annual mean insolation anomalies in this tropical region, respectively.

  • View in gallery

    Zonally averaged vertical temperature structure anomalies as a function of latitude (6 kyr BP − modern) (a) for the DJF-fixed SST experiment, (b) the JJA fixed SST experiment, (c) the DJF computed SST experiment, and (d) the JJA computed SST experiment. The contour interval is 0.5°C.

  • View in gallery

    (a) Sea-ice mass anomalies (6 kyr-modern) within the Arctic and Antarctic circles expressed as monthly means throughout the annual cycle (solid lines) and corresponding anomalies in insolation (dashed lines). (b) Polar stereographic projection of the Arctic sea-ice mass anomaly in September (the contour interval is 300 kg m−2).

  • View in gallery

    Surface air temperature at 2 m: the modern August fixed SST surface air temperature (a) and the observed August surface temperature (b) are in units of °C with a contour interval of 10°C. Precipitation: the modern August fixed SST precipitation (c) and the observed August precipitation (d) are in units of mm day−1 with a contour interval of 2 mm day−1. The 850-hPa winds: the modern JJA-fixed SST 850-hPa wind (e) and the observed JJA 850 hPa wind (f) are in units of m s−1 with a contour interval of 5 m s−1.

  • View in gallery

    Surface air temperature at 2 m: the 6 kyr BP − modern August anomalies for the (a) fixed and (b) computed SST experiments are in contour intervals of 1°C, with positive regions shaded at the same interval. Precipitation: the 6 kyr BP − modern August anomalies for the (c) fixed and (d) computed SST experiments are contoured at ±1, ±2, ±4, ±8, and ±16 mm day−1, with positive regions shaded at the same interval. The 850-hPa winds: The 6 kyr BP − modern August anomalies for the (e) fixed and (f) computed SST experiments are in contour intervals of 2 m s−1 with positive regions shaded at the same interval.

  • View in gallery

    Grid-cell analysis of the major anomalies in the previous figure. The letters denoting the actual grid cells investigated are displayed upon the model spectral topography with a contour interval of 200 m (a). The annual cycle of the variations in 850-hPa wind (m s−1) (b), surface air temperature at 2 m (°C) (c), and the precipitation (mm day−1) (d) for the fixed and computed SST experiments denoted by the solid and dashed lines, respectively, at each grid cell.

  • View in gallery

    Lake-level variations for the African–Asian monsoon region with the corresponding JJA PE anomalies (mm day−1) in the (a) fixed and (b) computed SST experiments. Lake-level anomalies range from much lower to much higher at 6 kyr BP as compared to present. The PE anomalies are contoured at ±0.5, ±1.0, ±1.5, ±2.0, ±4.0, and ±8.0 mm day−1.

  • View in gallery

    Lake-level variations for the North American region with the corresponding annual PE anomalies (mm day−1) in the (a) fixed and (b) computed SST experiments. Lake-level anomalies range from much lower to much higher at 6 kyr BP as compared to present. The PE anomalies are contoured at ±0.5, ±1.0, ±1.5, ±2.0, ±4.0, and ±8.0 mm day−1.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 167 150 0
PDF Downloads 66 53 0

Simulations of Mid-Holocene Climate Using an Atmospheric General Circulation Model

View More View Less
  • 1 Department of Physics, University of Toronto, Toronto, Ontario, Canada
  • | 2 Canadian Centre for Climate Modelling and Analysis, University of Victoria, Victoria, British Columbia, Canada
© Get Permissions
Full access

Abstract

The authors describe a first paleoclimatological application of the Canadian Centre for Climate Modelling and Analysis atmospheric general circulation model (AGCM) to simulate the climate state 6000 calendar years before present (6 kyr BP). Climate reconstructions for this period are performed with both fixed SSTs and with the AGCM coupled to mixed layer ocean and thermodynamic sea–ice modules. The most important difference between this epoch and the present involves the increased surface heating and cooling of the continental land masses in the Northern Hemisphere during summer and winter, respectively, which are a consequence of the modified orbital configuration. A comparison of a fixed SST experiment with a calculated SST experiment, incorporating a thermodynamic representation of oceanic response, is performed to assess the impact on the mid-Holocene climate. The results are also contrasted with those obtained on the basis of proxy climate reconstructions during this mid-Holocene optimum period. Of interest in this calculated SST experiment is the impact on the seasonal cycle of sea–ice distribution due to the increased insolation at high latitudes during Northern Hemisphere summer. Also important is the fact that the mixed layer ocean in the simulation is found to further enhance the monsoon circulation beyond the enhancement found to occur due to the influence of modified orbital forcing alone. This increased response is found to be a consequence of the sensitivity of tropical SST to the amplification of the seasonal cycle due to the change in insolation forcing that was characteristic of the mid-Holocene period.

Corresponding author address: Dr. W. R. Peltier, Department of Physics, University of Toronto, 60 St. George Street, Toronto, ON M5S 1A7, Canada.

Email: peltier@rainbow.physics.utoronto.ca

Abstract

The authors describe a first paleoclimatological application of the Canadian Centre for Climate Modelling and Analysis atmospheric general circulation model (AGCM) to simulate the climate state 6000 calendar years before present (6 kyr BP). Climate reconstructions for this period are performed with both fixed SSTs and with the AGCM coupled to mixed layer ocean and thermodynamic sea–ice modules. The most important difference between this epoch and the present involves the increased surface heating and cooling of the continental land masses in the Northern Hemisphere during summer and winter, respectively, which are a consequence of the modified orbital configuration. A comparison of a fixed SST experiment with a calculated SST experiment, incorporating a thermodynamic representation of oceanic response, is performed to assess the impact on the mid-Holocene climate. The results are also contrasted with those obtained on the basis of proxy climate reconstructions during this mid-Holocene optimum period. Of interest in this calculated SST experiment is the impact on the seasonal cycle of sea–ice distribution due to the increased insolation at high latitudes during Northern Hemisphere summer. Also important is the fact that the mixed layer ocean in the simulation is found to further enhance the monsoon circulation beyond the enhancement found to occur due to the influence of modified orbital forcing alone. This increased response is found to be a consequence of the sensitivity of tropical SST to the amplification of the seasonal cycle due to the change in insolation forcing that was characteristic of the mid-Holocene period.

Corresponding author address: Dr. W. R. Peltier, Department of Physics, University of Toronto, 60 St. George Street, Toronto, ON M5S 1A7, Canada.

Email: peltier@rainbow.physics.utoronto.ca

1. Introduction

The application of general circulation model (GCM) capability to the understanding of paleoclimate regimes has often been motivated by the availability of specific proxy-databased inferences of climate state for well-constrained epochs. For example, it has been long realized that lake levels across northern Africa, the Middle East, and India during mid-Holocene time (8–5 kyr before present; BP) were much higher than present (Street and Grove 1979). Early modeling experiments (see, e.g., Kutzbach and Otto-Bliesner 1982) demonstrated that the amplified seasonal cycle that results from low-latitude solar radiation changes does, in fact, lead to an enhanced monsoon circulation in Northern Hemisphere summer in this epoch and thus provided a satisfactory explanation of the proxy-derived inferences of paleoclimate state that strongly suggested that such had, in fact, occurred. An important motivation for continuing to perform and to improve such GCM-based paleoclimate analyses lies in the expectation that by comparing a range of model-based paleoclimate predictions with proxy data–based inferences it is possible to test the robustness of the parameterization schemes through which subgrid-scale processes are represented in the model. Additional analyses of the mid-Holocene epoch (Kutzbach and Gallimore 1988; Mitchell et al. 1988) that employed an atmospheric general circulation model coupled (AGCM) to a mixed layer ocean have been employed to study the sensitivity of the model climate to insolation forcing at 9 kyr BP. Sensitivity analyses with the same models were later performed in order to better understand the feedback effects due to decreased surface albedo (increased vegetation) on the tropical monsoons and these results were also employed as a basis for comparisons with geological evidence (Street-Perrott et al. 1990, Foley 1994). For a summary of these and other experiments, the interested reader should consult Mitchell (1993).

Also of interest from the general perspective of paleoclimate model data intercomparison are the climate simulations performed with the Laboratoire de Meteorologie Dynamique (LMD) model for three time slices within the last glacial period by de Noblet et al. (1996). These simulations focused on the sensitivity of the African–Asian summer monsoon to variations in insolation and demonstrated the fact that the climate at 6 kyr BP (the mid-Holocene) and at 126 kyr BP (during the last interglacial) were characterized by stronger monsoon circulations, while that at 115 kyr BP (during glacial inception) was characterized by a weaker monsoon circulation. More recent simulations of the 6-kyr BP time slice using fixed SSTs include a study by Hewitt and Mitchell (1996) that focused upon the issue of moisture availability on the land surface and included analysis of the interannual and decadal variability based upon long time series simulations employing the Hadley Centre’s AGCM. An important outcome of this work was the suggestion that longer than 10-yr average climatologies may be necessary to produce an adequate statistical basis for study of the impact of changes in orbital configuration. Hall and Valdes (1997), also using fixed SST analyses, have employed the U.K. Universities’ Global Atmospheric Modelling Programme (UGAMP) GCM to study the 6-kyr BP time slice, and concentrated upon understanding the impact on synoptic-scale atmospheric dynamics. They investigated the impact of the modified radiation regime characteristic of the mid-Holocene period on, for example, midlatitude storm tracks and transient eddy activity.

In this paper our focus will also be on the time slice centered upon 6000 calendar years before present (6 kyr BP), a period within the so-called Holocene optimum. This period of maximal Northern Hemisphere summer warming is considered to have occurred in the range of 8.5 kyr BP–6.5 kyr BP depending on the region of the planet being considered or the type of proxy indicator employed to measure the climate extreme (e.g., Yan and Petit-Marie 1994). The 6-kyr BP time slice has been selected as a first fiducial epoch in a project called the Paleoclimate Model Intercomparison Project (PMIP). This period, while not an insolation extremum, experienced larger Northern Hemisphere warming during the summer season as compared to present. Although the time of the Northern Hemisphere summer insolation extremum occurred closer to 9 kyr BP when the summer solstice occurred at perihelion, at that time there were still significant remnants of the continental ice sheets remaining from the time of peak glaciation at last glacial maximum (LGM), 21 kyr BP (Peltier 1994). By 6 kyr BP, deglaciation was almost complete and there is furthermore an abundant supply of terrestrial data that is often well dated (e.g., COHMAP 1988), which can be employed to test model predictions. SSTs are also considered to have been close to those characteristic of present day (Ruddiman and Mix 1993) and there is, therefore, apparently no a priori need in addressing the 6-kyr BP epoch to introduce major changes in surface boundary conditions into the AGCM. The only significant impact on 6-kyr BP climate will therefore be due to the modified orbital insolation regime (Berger 1978).

In this paper we will focus upon the nature of the climate state at this time using two versions of the global climate model, one employing fixed SSTs and another that incorporates both mixed layer ocean and thermodynamic sea–ice modules and that will allow for the investigation of changes in mid-Holocene SSTs and sea–ice cover. We will then be able to test, in the context of a simple thermodynamic “slab” representation of the oceans, previous suggestions (e.g., Ruddiman and Mix 1993) to the effect that these quantities were much the same as present. However, it must be borne in mind that the incorporation of a simple mixed layer description of the oceans into the dynamical structure really constitutes an attempt to include the feedbacks associated with changes in the water vapor content of the atmosphere and sea–ice cover, rather than to accurately simulate SSTs. Changes in tropical ocean surface temperatures will be investigated in what follows in order to measure the sensitivity of the mixed layer ocean surface to orbital insolation change and to provide an indication of how these changes might modify the mid-Holocene monsoon response. In particular, these analyses will assess the impacts of an interactive ocean on the strength of the African–Asian monsoon circulation and the relationship of these impacts to the proxy-inferred monsoon strength and spatial distribution. Also of interest in comparing the two sets of experiments that we will describe is the surface temperature response over land and ocean separately and the demonstration of the impact that changing CO2 levels have on the vertical atmospheric temperature profile. The manner in which the experiments have been designed will allow us to separate the impact of the modified radiative forcing due to CO2 changes from that due to the orbital insolation change itself.

A description of the AGCM to be employed in our analyses and of experimental design is provided in the following section 2. Section 3 describes the results obtained from the two main paleoclimate experiments that were performed. In section 4 we present a preliminary comparison of model results from both experiments to appropriate proxy climate data for the 6-kyr BP epoch. Our conclusions are summarized in section 5.

2. Description of the climate model and discussion of experimental design

The model employed in this work is the second generation AGCM of the Canadian Centre for Climate Modelling and Analysis (CCCMA). As described in McFarlane et al. (1992), the dynamics of this model are implemented in terms of a spectral decomposition of the horizontal variations of the prognostic variables on a basis of spherical harmonics. The version of this model that we employ has a horizontal resolution defined by a triangular truncation to degree and order 32 (T32). The governing equations are written in terms of a generalized terrain-following coordinate with vertical discretization in terms of rectangular finite elements following Laprise and Girard (1990). Vertical structure is defined on 10 levels in this coordinate system.

As previously discussed, we will present the results of two paleoclimate experiments, the major difference between them being that the first uses prescribed SSTs and sea–ice while the latter calculates SSTs and sea–ice using the mixed layer ocean and sea–ice modules described in McFarlane et al. (1992). The complete set of experiments then consists of four simulations, two modern control experiments and two paleoclimate simulations. Both of the latter experiments are intended to reveal the influence on climate state of the change in orbital configuration from present day to 6 kyr BP. The first set of experiments, of course, uses prescribed SSTs and sea–ice for both the modern control and 6-kyr BP experiments. The present-day control simulation has the atmospheric CO2 concentration set to a “modern” level of 345 ppmv. At 6 kyr BP the CO2 concentration had returned to the average interglacial level of approximately 280 ppmv that was characteristic of preindustrial time (Barnola et al. 1987). This value will be employed for each of the paleoclimate simulations. In the mixed layer ocean calculations the CO2 levels are held constant at the preindustrial level of 280 ppmv for both the present day and paleosimulations. This will allow for an unambiguous investigation of the influence of modified insolation forcing at 6 kyr BP, and the effect of this change on SST, sea–ice mass, and sea–ice distribution. The PMIP (fixed SST) experiment, which had CO2 levels reduced by 20% in the 6-kyr BP simulation, does not provide a clear indication of the role that these separate forcings have on the climate, although the impact of the CO2 forcing on the surface is expected to be minor since the ocean surface temperatures are held fixed. As previously mentioned, by 6 kyr BP the LGM continental ice sheets had almost completely disappeared and the surface topography was therefore much the same as that which exists at present. There being no significant changes in land–ice and land–ocean distribution, the topographic boundary conditions are fixed to modern values. For the purpose of these analyses, vegetation patterns and soil type and all other internal boundary conditions have also been fixed to modern values for the sake of simplicity.

A comparison of the differences in the insolation at the top of the atmosphere between 6 kyr BP and present (Fig. 1) illustrates the expected consequence of a shift in the position of the summer solstice to a position closer to perihelion at 6 kyr BP. The Northern Hemisphere June–August (JJA) average insolation increase is 20.7 W m−2, which amounts to a change of 4.7% from the current situation. The Southern Hemisphere December–February (DJF) average insolation decrease is 19.0W m−2, which amounts to a change of −4.8% from present. The dynamic processes that control the earth’s orbital insolation regime and the theory required to model them have been well studied and the interested reader should consult the following for detailed discussion (e.g., Berger 1978; Saurez and Held 1979; Mitchell et al. 1988;Quinn et al. 1991; Laskar 1993; Phillipps and Held 1994; Hall and Valdes 1997).

In comparing model climatologies for two different time periods, the date of the vernal equinox (1200 LT 21 March in this case) along with the manner in which the seasons are defined for the two time periods must be consistent. The seasons can be defined either as they are at present, based on the astronomical calendar or based on the angular positions of the earth in its orbit. The use of astronomical positions provides for a better phasing of the insolation patterns between the two time periods rather than defining the seasons as having the same duration as today. Since the first of the two sets of experiments (PMIP mandate) described in this paper employs present-day seasonal definitions, the second set was performed subject to the same assumption. Thus, intercomparison of climate state for a particular climatological month or season from the 6-kyr BP time period and the present day will not be entirely consistent. For a review of this problem the interested reader might usefully consult Joussaume and Braconnet (1997).

In the first set of AGCM experiments based upon the assumption of fixed SSTs, both the modern and 6-kyr BP simulations were run for 12 annual cycles from observed present day January initial conditions. The process that generally takes longest to reach equilibrium, under fixed SST boundary conditions, is the ground water adjustment, which was found to take approximately two annual cycles. The first two annual cycles of these simulations were, therefore, discarded and the remaining 10 annual cycles were monthly, seasonally, and annually averaged to produce a model climatology for both the modern control and 6-kyr BP experiments. For the computed SST mixed layer ocean experiments, adjustment to an equilibrium climate was found to require significantly longer integration times. The globally averaged JJA surface temperature and JJA sea–ice mass were employed in these analyses to investigate the spinup to equilibrium because annually averaged values of the surface temperature are perturbed less due to the nature of the orbital forcing. Changes in position of the seasons with respect to perihelion of the earth’s orbit are expected to enhance the seasonality of climate but not the annually averaged climate. In these mixed layer ocean analyses, the modern preindustrial control experiment was initially spun up for 50 annual cycles with the last 10 annual cycles being averaged to produce the modern climatology. Orbital parameters were then changed to those appropriate to 6 kyr BP and the paleosimulation was initialized from the climate of the modern day control. Although surface temperatures appeared to adjust within approximately 15 annual cycles, the summer sea–ice distribution took at least 30 annual cycles to adjust to equilibrium (Fig. 2); beyond that point the model was run for 10 additional annual cycles to produce the 6-kyr BP 10-yr-averaged climatology for the mixed layer ocean simulation.

3. Results: The climate state near Holocene optimum

In the following sections we present analyses of a selected set of properties of the 6-kyr BP climate state revealed by the series of simulations that we have performed. The discussion of these results will culminate with discussion of the sensitivity of the monsoon circulation to the incorporation of a mixed layer ocean and the relation to proxy inferences.

a. Surface air temperature at screen height (2 m)

One of the main features of the climate response forced by the 6-kyr BP insolation pattern is the change in surface temperature resulting from the difference in absorbed radiation as compared to present. The response to the reduced insolation in Northern Hemisphere fall, and globally during December–April, is readily seen in the zonally averaged surface temperature over land in both the fixed and computed SST experiments (Fig. 3a,c). Likewise, there is an increased heating over land from June to October that is evident in both experiments and that is consistent with the insolation change (Fig. 1). These temperature changes, which are statistically significant, are approximately 1°–2°C in magnitude and occur primarily in high and midlatitudes. Statistically significant changes are at the 95% confidence level according to a Student’s two-sided t-test. For detailed discussions of the statistical issues that arise in interpreting the results obtained with global climate models see, for example, Chervin et al. (1976) and Zweirs (1987). These temperature changes are consistent in both magnitude and spatial pattern with those obtained in previous numerical experiments in which the annual cycle of the zonally averaged surface air temperature response to mid-Holocene insolation forcing was investigated (e.g., Mitchell et al. 1988; Hewitt and Mitchell 1996).

The Southern Hemisphere land surface temperature response appears to lag the insolation anomaly by approximately 1 month in the two different experiments, but the lag appears to be largest in southern midlatitudes in the computed SST experiment. This may be a result of the combined influence of the insolation anomaly and the computed SST anomalies. Examination of the ocean response, including sea–ice, in the computed SST experiment reveals a cooling of approximately 0.5°C of the ocean surface during the Northern Hemisphere spring season in the northern midlatitudes and in the Tropics (Fig. 3d). This response is consistent with the proxy-derived temperature changes for February, inferred from deep sea sedimentary cores in the tropical Atlantic at 6 kyr BP (Ruddiman and Mix 1993). Such cooling may be the primary cause of the slightly delayed temperature response over land in the computed SST experiment. The heating anomaly over the Southern Ocean from September through October, due to the insolation change, may result in the heating anomaly over land in southern midlatitudes extending into December (Fig. 3c). This explains the discrepancy between the phase lag of the response in the fixed SST and computed SST experiments during the annual cycle in this Southern midlatitude geographical region. The changes in surface air temperature over the ocean in the fixed SST experiment (Fig. 3b) are negligible, as must be the case.

The common feature present in the global DJF and JJA difference plots (Fig. 4) for both the fixed and computed SST experiments are, respectively, the cold and warm anomalies. In this figure, shading denotes cool (warm) anomalies for DJF (JJA) that are less than (greater than) 1°C in magnitude, and which are distributed more or less over the interiors of the major continental land masses. The regions in which the resulting anomalies are on the order of 1°C or more have a statistically significant change at the 95% confidence level.

The surface temperature anomalies in the DJF period for the fixed SST experiment (Fig. 4a) reveal changes on the order of −1° to −2°C over much of North America, North Africa, the Middle East, and Asia. Accompanying these decreases in Northern Hemisphere winter temperatures are increases on the order of 1°–2°C over the same regions in the Northern Hemisphere summer (Fig. 4b). The surface temperature anomalies in the computed SST experiment are similar in their spatial structure to those obtained in the fixed SST experiment, but are smaller in magnitude by an amount on the order of 1°C over land in both DJF and JJA (Figs. 4c,d). The structure of the warming over the continents also appears to have changed in the mixed layer ocean simulation from the experiment in which SSTs were held fixed. This increased seasonality is still present in the computed SST experiment but the total response has decreased over the continents. The signals over land in this computed SST experiment are less statistically significant than those that are obtained in the fixed SST experiment. There are, however, statistically significant portions of the southern midlatitude oceans that are characterized by a warm anomaly on the order of 0.5°C in DJF compared to present day in this computed SST experiment. In JJA, changes of similar magnitude in the tropical ocean surface temperatures, which are statistically significant, are characterized by a cooling trend in the computed SST experiment.

Changes in the tropical SSTs may be further investigated by comparing the average temperature in the section from 24°S to 24°N with the insolation anomaly at the top of the atmosphere in the same region as a function of the time of year (Fig. 5). The solid curve illustrates the deviation of the tropical ocean surface temperature at 6 kyr BP from that for the modern control simulation, a difference that reaches a minimum of −0.55°C in early May and a maximum of +0.25°C in late October. The deviations in temperature are all statistically significant at the 95% level except for times of the year when the anomalies are within 0.1°C of zero. The dashed curve illustrates the annual cycle of the anomaly in insolation over the same zonally averaged section of the globe. The insolation variation reaches a minimum of −23 W m−2 in mid-February and a maximum of +23 W m−2 in mid-August. The tropical ocean surface response is out of phase with the insolation distribution by more than two months. The most interesting characteristic of this cycle lies in the fact that the anomaly in the surface temperature of the tropical ocean does not average to zero but has an annually averaged value of −0.16°C while the annually averaged insolation anomaly is −0.88 W m−2. In a comparison between the normalized annually averaged surface temperature anomaly (mean value = 0.16°C/amplitude = 0.40°C) and that for the normalized insolation variation (0.88 W m−2/23.0 W m−2), the annually averaged ocean surface temperature response is a factor of 10 greater than the insolation forcing. The large ocean surface temperature response, which lags and is highly correlated with the insolation forcing, demonstrates that the climate system may be characterized by extreme sensitivity to a small decrease in the annually averaged insolation in the Tropics. However, investigations with more complex fully three-dimensional ocean models will be necessary to validate conclusions derived from this analysis that is based on a simple thermodynamic slab ocean model.

b. The zonally averaged temperature field

The DJF and JJA fixed SST zonally averaged temperature anomalies between 6 kyr BP and present are displayed in Fig. 6a,b, respectively. Common to both fields is a decrease in the tropical lower stratospheric temperature of almost 3°C in DJF and 2°C in JJA. The anomaly fields differ seasonally in that there is a predicted warming at the model top above about 30 hPa in JJA for both the fixed SST and computed SST experiments. These results are shown on Figs. 6b and 6d. Both stratospheric fields, however, are characterized by an increasing temperature anomaly with height in the upper levels of the model. The troposphere, on the other hand, responds to the imposed insolation anomaly with rather minor changes between 1000 and 200 hPa. We note that the DJF zonally averaged temperature (Fig. 6c) is characterized by a lower stratospheric cooling of about 1°C globally, with no hint of the minimum observed in the fixed SST experiment just above the tropopause. The JJA computed SST zonally averaged temperature field (Fig. 6d) is characterized by a stratosphere that is almost a mirror image of the DJF result.

The most obvious differences between the fixed and computed SST experiments are readily apparent by noting the symmetries present in Fig. 6. These reflect the differences in the nature of the boundary conditions and the trace gas-related radiative forcing employed in the two experiments. While the computed SST experiment employed an interactive mixed layer ocean, it is the CO2 amount, with a uniform mixing ratio throughout the atmosphere, which is the most likely cause of the discrepancy between the zonally averaged temperature fields obtained in the two sets of experiments. In the fixed SST experiment, the CO2 amount was changed from a modern value of 345 ppmv in the control to the 6 kyr BP value of 280 ppmv while in the computed SST experiment the CO2 levels were fixed to the value of 280 ppmv. In the computed SST experiment this preindustrial value was used for both the control and 6 kyr BP simulation. Thus, for the fixed SST experiment, the added radiative effect associated with the 19% reduction of CO2 must be kept in mind when seeking an explanation for the simulated changes in the zonally averaged temperature fields.

The effect of the change in orbital configuration between 6 kyr BP and the present time is more readily apparent in the computed SST experiment. The warming, in the lower stratosphere and at high latitudes of the Northern Hemisphere for the 6-kyr BP simulation during the JJA period, correlates well with the enhanced insolation in that region and season. The warming in the lower stratosphere is consistent with the enhanced absorption of solar radiation by the ozone layer in that region. The reduced insolation in the DJF season is not accompanied by an associated reduction of tropospheric temperatures. However, there is a broad region of cooling in the lower stratosphere. The lack of cooling in the lower troposphere of the Southern Hemisphere in this season may seem somewhat counterintuitive. However, it must be borne in mind that the large expanse of open ocean in the Southern Hemisphere mitigates against and retards in time the response to insolation anomalies of either sign.

c. The sea–ice distribution at 6 kyr BP

The orbital configuration at 6 kyr BP is characterized primarily by the increased obliquity and the change in the longitude of the perihelion. The change in obliquity at 6 kyr BP from modern results in a larger meridional gradient of insolation and more insolation being received at the Northern Hemisphere summer pole. The change in the longitude of the perihelion at 6 kyr BP compared to modern increases the seasonality of the insolation forcing between winter and summer. At 6 kyr BP the insolation forcing is therefore characterized by an increase in both the polar insolation and the seasonality at high northern latitudes during the Northern Hemisphere summer season. In Southern Hemisphere summer there are also increases expected in polar insolation anomalies due to the influence of obliquity, but this effect is masked by the decreased Southern Hemisphere summer seasonal insolation. In a previous section we demonstrated the way in which changes of the orbital configuration result in the differences observed in the model surface temperature at 6 kyr BP. In the computed SST experiment the changes in insolation resulted in the surface temperature of the ocean being generally cooler through most of the year while a warm anomaly over the oceans at 6 kyr BP was observed to onset roughly at the beginning of September and to reach a maximum in November (Fig. 3b). As a result, we might expect this trend to influence and possibly delay the formation of sea–ice in the Northern Hemisphere (Mitchell et al. 1988).

Ongoing work within the Canadian Climate System History and Dynamics program (CSHD) to reconstruct the extent of sea–ice cover and the seasonal distributions of sea surface salinity and temperature through the subpolar basins of the North Atlantic for the 6-kyr BP time slice [and for LGM (last glacial maximum)] is based upon isotopic measurements in dinoflagellate cysts (de Vernal et al. 1993). These data will provide a strong constraint on the actual changes that may have occurred in both sea–ice and the snow line at 6 kyr BP, at least in this area of the globe. Changes in sea–ice thickness and distribution in the computed SST experiment are, however, predicted to have been significant at 6 kyr BP although the sea–ice margins remain essentially fixed. A simulation of 9-kyr BP climate by Mitchell et al. (1988) also revealed that year-round sea–ice thickness was significantly reduced. Ruddiman and Mix (1993) speculated that warm anomalies at 6 kyr BP in the midlatitude North Atlantic SSTs, which were inferred on the basis of isotopic information from deep sea sedimentary cores, could conceivably be the result of reduced sea–ice.

The periods of the year for which the sea–ice distribution was investigated in detail were September and March, periods during the annual cycle in which sea–ice amount is at its maximum (minimum) and minimum (maximum) extent for the Southern (Northern) Hemisphere, respectively. Both the south and north polar regions were investigated to determine the response of sea–ice to the 6-kyr BP forcing. The average monthly polar sea–ice distributions were all characterized by anomalies corresponding to decreases in sea–ice mass (Fig. 7a). The South Polar region responded to the orbital insolation anomaly with roughly one-half the variation of sea–ice thickness at 6 kyr BP from the modern control compared to that at the North Pole. The dominant changes in sea–ice thickness are predicted by the model to occur in the North Polar regions where the 6-kyr BP insolation anomaly is especially large and, thus, sea–ice mass would be expected to exhibit a much larger diminution. Near the North Pole, the predicted anomalies are indeed on the order of −1200 kg m−2 and −900 kg m−2 for September and March, respectively, amounting to a decrease of approximately 1 m in sea–ice thickness throughout the entire year at 6 kyr BP at the pole. The entire Arctic Circle in September is characterized by a uniformly decreasing anomaly from 0 to −1200 kg m2 from about 75°N to the North Pole (Fig. 7b). A similar result was obtained for March, suggesting that the role of solar heating was more dominant in this hemisphere at 6 kyr BP.

The increased Northern Hemisphere summer insolation anomaly at 6 kyr BP possibly accelerated sea–ice melting prior to the time that the modern minimum extent was achieved, thereby forcing the summer sea–ice cycle more strongly than occurs in the modern climate system. It is expected that, although Northern Hemisphere winter may have been colder at 6 kyr BP than at present, the formation rate of sea–ice would be constrained by the conductivity properties of ice and the freezing point of saline water, whereas during the spring–summer melting of sea–ice the melting rate is not buffered by such influences. Based upon a 9-kyr BP mixed layer ocean simulation, Mitchell at al. (1988) suggest that the mixed layer absorbs large amounts of heat throughout the summer months that delays the regrowth of ice in Northern Hemisphere fall. Thus, the 6-kyr BP − present anomaly in the annual cycle of sea–ice is not expected to be sinusoidal about zero but rather to have a mean value well below the modern value obtained from the control simulation throughout the entire annual cycle. This behavior is consistent with that documented by Mitchell et al. (1988) and is evident in Fig. 7a, which displays anomalies of the annual variability of sea–ice within the Arctic and Antarctic Circles. The asymmetric behavior in the two hemispheric sea–ice anomalies correlates well with expectations based upon the interhemispheric insolation anomalies, as denoted by the dashed lines in Fig. 7a. Also of note is that the phase lag between maximum insolation and minimum sea–ice anomalies is on the order of 1 month.

d. The African–Asian monsoon

Monsoon circulations are driven by the differential response of the land and oceans to solar insolation reaching the surface. It is therefore to be expected that the major monsoon circulations of the globe will vary on long timescales in a way that is consistent with the evolution of the insolation patterns. The change in the strength of this monsoon system has been the focus of many previous paleoclimate analyses (e.g. Kutzbach and Otto-Bliesner 1982; Kutzbach and Guetter 1986; Prell and Kutzbach 1992; de Noblet et al. 1996). The land–sea temperature contrast that determines the strength of the monsoon circulation is itself dependent on the combined influences of external conditions and internal feedbacks that arise during the monsoon season. Based on previous analyses of these various influences affecting modern monsoon behavior (see, e.g., Meehl 1994; Zweirs 1993; Delworth and Manabe 1989; Rasmusson and Carpenter 1983) it is expected that the nature of the paleo-African–Asian monsoon would have been influenced by a number of factors, both internal and external. It is, however, beyond the scope of this study to attempt to investigate all of the contributing factors that influence modern monsoon behavior in the context of paleoclimate change. Only the main variables that define the summer monsoon, such as temperature, precipitation, and winds, between 6 kyr BP and the present will be discussed here. While the African–Asian summer monsoon season occurs from May to September, with the most intense variability occurring in JJA, it is the month of August that has been selected to illustrate the anomalies between 6 kyr BP and the present. At 6-kyr BP it is the month of August in which the most statistically significant deviations from the conditions that are characteristic of the present day occur. Also, the maximum temperature variation (Fig. 3) as a result of the 6-kyr BP insolation anomaly (Fig. 1) in these tropical and extratropical latitudes would occur in late August or early September over land and in late fall over the oceans. An analysis of this period will therefore provide an indication of the maximum anomalies simulated by the model.

The modern August surface air temperature, August precipitation, and JJA 850-hPa winds predicted by the fixed SST experiment over the region from 30°W to 170°E longitude and 40°S to 60°N latitude are displayed in Figs. 8a, 8c, and 8e. The observed August surface air temperature and precipitation from Legates and Willmott (1990) over the same region and for the same month are displayed in Figs. 8b and 8d. The observed JJA 850-hPa field (Fig. 8f) was extracted from a 5-yr climatology that was prepared using the Canadian Meteorological Centre’s (CMC) operational analyses for the period December 1990–November 1995. The CMC data assimilation system is described in Mitchell et al. (1993) and Mitchell et al. (1996). The modern August temperature, precipitation, and wind for the computed SST experiment is not appreciably different from that which is characteristic of the fixed SST experiment and therefore is not shown; rather a comparison is made with the observed modern data.

Although both simulations have deficiencies, the observed and model surface temperature fields are characterized by a high degree of correlation with one another. Over the Sahel in Northern Africa, the precipitation maximum is too weak in the model prediction and too strong over the horn of Africa. The model is also unable to resolve the intensity of the precipitation over the western Ghats and the rain shadow effect in the lee of the mountains. The distributions over Nepal and Burma are fairly well simulated in both extent and magnitude; however, farther to the east there is excessive precipitation over the tropical Pacific. The simulation of the 850-hPa jet in the fixed SST experiment is only slightly modified from that which is observed. In particular, the westerly flow over equatorial Africa is slightly stronger than that implied by the observational data. Furthermore, the strength of the Somalia jet, while similar in magnitude to the observed, is directed more to the northwest. The main problem with the predicted low-level flow occurs over Southeast Asia and the Philippines, where the simulated low-level westerly flow is excessively strong.

A comparison between the August 2-m surface air temperature, precipitation, and 850-hPa wind variations predicted by the fixed and computed SST experiments for 6 kyr BP are displayed in Fig. 9. Both experiments are characterized by common positive and negative anomalies in this African–Asian monsoon region. The surface temperature anomalies (Figs. 8a and 8b) demonstrate that areas over the oceans are slightly colder in the computed SST experiment. At first glance is it obvious that the common temperature anomaly in both experiments consists of a cool band to the south of a more widespread northerly warm band originating over Northern Africa and extending northeast-east through the Middle East, India, and into Mongolia.

The 6-kyr BP anomalies in precipitation for both the fixed and computed SST experiments are displayed in Figs. 9c and 9d, respectively. The most consistent anomaly patterns in both the fixed and computed SST experiments are found over the African Sahel, the horn of Africa, the Middle East, Nepal, and Burma. These anomalies are closely correlated with those in the temperature field but with positive anomalies in precipitation occurring in regions in which there were negative anomalies in temperature. This is consistent with the fact that these increases in precipitation also correlate with increases in evaporation at the surface and cloud cover in the atmosphere (not shown). These precipitation anomalies are consistent with those recently obtained in a lower-resolution gridpoint model experiment on 6-kyr BP climate using the LMD model (de Noblet et al. 1996) in which their JJA precipitation anomalies are very similar to those obtained in this study over North Africa, India, and Southeast Asia.

The August 850-hPa wind anomaly for the fixed SST experiment (Fig. 9e) and that for the computed SST experiment (Fig. 9f) are again characterized by a structure that is consistent with those previously described in the surface temperature and precipitation fields. These two experiments deliver a positive anomaly over Africa that extends from the Sahel into a maximum over the Middle East–Arabian Sea. To the south of this positive anomaly over the Arabian Sea are negative anomalies in both the fixed and computed SST experiments, respectively. The main change in westerly wind strength appears downwind of the precipitation anomalies obtained in both the fixed and computed SST experiments that provide the diabatic forcing that drives the wind. Most significantly, there is a strengthening (weakening) on the northwest (southeast) flank of the Somalia jet, increased westerly winds inland, and increased easterly winds north of the equator over the Arabian Sea. The nature of this anomaly indicates a northward shift in both the easterly winds to the south and the westerly winds inland to the north but not necessarily a change in wind strength. These anomalies are, however, much more pronounced in the computed SST experiment and, therefore, must be attributed to variations in the SST.

The main feature of the African–Asian summer monsoon consists of the low-level wind patterns that mainly originate from the southwest and extend northeast inland. This flow is a result of the low-level convergence of air inland in regions of the monsoon where there is a large thermal contrast between the land and the surrounding ocean. In the fixed SST experiment the near-zero surface temperature anomalies over the ocean are readily apparent. One of the more interesting results obtained in this study consists of the cooler sea surface temperature predicted over the tropical oceans in the computed SST experiment. The tropical southeastern Atlantic is characterized by a negative anomaly of 0.6°C. The southern Indian Ocean from the equator to about 40°S latitude is also characterized by a similar anomaly but of about half the magnitude. The western Pacific from 40°S to 40°N latitude is also characterized by a negative anomaly on the order of half a degree. This consistent negative anomaly, although not large in magnitude, would act to strengthen the meridional and longitudinal gradients of the isentropes and pressure isobars over land and ocean in the monsoon regions of Africa and Asia at 6 kyr BP. It is therefore expected that the strength of the monsoon circulation, as characterized by precipitation and wind intensities, would be stronger in the computed SST experiment than in the fixed SST experiment. A study of the effects of sea surface temperature anomalies on the present day monsoon circulation from 1979–92 (Li and Yanai 1996) has demonstrated a strong correlation between episodes of strong versus weak African–Asian monsoons and La Niña and El Niño SST anomalies, respectively. This strong correlation, which involves SST anomalies that are only on the order of 0.1°–0.2°C, demonstrates that the SST variations in our computed SST experiment, which are on the order of a factor of 3 greater in some regions, are certain to influence the strength of the monsoon circulations at 6 kyr BP. In particular we note the consistency between the negative anomalies described in Li and Yanai over the Indian Ocean and parts of the western Pacific that are associated with strong monsoon years. Inconsistencies between our paleoresults and those of Li and Yanai occur over the tropical South Atlantic and tropical South Pacific Oceans.

The influence of spectral resolution on GCM-simulated climatology has been investigated in a number of studies (e.g., Boer and Lazare 1988; Tibaldi et al. 1990) and, in particular, its affect on the African–Asian monsoon was thoroughly investigated in the European Centre for Medium-Range Weather Forecasts (ECMWF) model by Sperber et al. (1993). Generally, it was found that large-scale features such as those characteristic of the precipitation distributions were not significantly different at resolutions higher than T42 in a typical GCM. Our model, GCMII, which is based upon T32 resolution may therefore be subject to some error in the large-scale climatological features, but these are not expected to be significant (Boer and Denis 1997). In Fig. 10a we show the spectral topography over the Indian monsoon region of the planet along with the locations of six grid cells (true dimensions) in which the major August anomalies in surface temperature, precipitation, and 850-hPa wind were obtained in the analyses described previously. Here we will employ the grid-cell data to gauge the intensity and onset/duration of the 6-kyr BP Indian monsoon for both the fixed and computed SST experiments.

The 850-hPa wind anomalies situated at the southeast (grid cell A) and northwest (grid cell B) flank of the Somalia jet are shown in Fig. 10b. The annual cycle of the variation in Somalia jet wind strength is characterized by a symmetric departure from zero from June to October with more random variations throughout other times of the year. The surface air temperature anomalies are displayed in Fig. 10c, from locations with the same elevation of about 500 m above sea level. The positive temperature anomaly (grid cell C) is associated with diabatic heating while that of the negative anomaly (grid cell D) is associated with evaporative cooling. The precipitation anomalies at grid cells E and F are displayed in Fig. 10d. The maximum (grid cell E) is, however, much more intense than the corresponding minimum (grid cell F) to the southeast. The maximum is located directly over the largest topographic gradient (Fig. 10a) in this region that is situated over Nepal and that leads to the Tibetan Plateau. It is not surprising that large anomalies in precipitation occur at this point as the warm moist air is advected upslope by the westerly arm of the Indian monsoon jet. The anomalies in the fixed SST experiment are denoted by the solid lines and the anomalies in the computed SST experiment are those denoted by the dashed lines.

The Indian summer monsoon is generally characterized as occurring from May to September under modern climate conditions, and it would therefore not be surprising to find a lag of up to a month in the onset and duration of the monsoon at 6 kyr BP even though the mean June wind is characterized by a near-zero anomaly at this point. The difference in intensity between the fixed and computed SST experiments is revealed by the increased positive slope in the wind maxima and the increased negative slope in the wind minima in July. Both experiments appear to deliver a similar behavior as the monsoon season progresses into September. The lower surface temperatures observed in the winter season are determined by the cooler land masses that are a consequence of the negative insolation anomaly at 6 kyr BP in these latitudes. The positive anomaly (grid cell C), however, appears to have an onset and duration that is different from that in grid cell D that is at a lower-latitude location. This is primarily a consequence of the difference in the thermal heat capacities between land and ocean where the temperature at grid cell D, although over land, is being influenced more directly by ocean temperatures and lags the variation at grid cell C in both the fixed and computed SST experiments. Once more the computed SST experiment is characterized by a stronger temperature response in both the positive and negative anomalies at grid cells C and D. The 6-kyr BP wind anomaly implies that an increase in wind velocity over the maximum precipitation point (Figs. 9e,f) carries more moisture-laden air up into the higher elevation region of the Himalayan mountain range. The amplitude of the negative anomaly (grid cell F) to the southeast, being of about half the magnitude of the positive anomaly, may simply be associated with a more northerly shift in the jet, but also appears to be related to the nonlinear dependence of the saturation vapor pressure of water on temperature. This grid-cell analysis suggests that the Indian winter monsoon at 6 kyr BP was not much different from that which occurs at present. Anomalies in the winter monsoon fields for temperature, precipitation and low-level winds in this region (not shown) are much smaller in magnitude than those that are characteristic of the summer season.

e. Model comparisons with implications derived on the basis of proxy data

Of particular interest in the context of the PMIP project is the generation of globally uniform datasets for the LGM and 6-kyr BP epochs (e.g., Guiot et al. 1996). The most abundant and best dated source of proxy climate data for the 6-kyr BP period is that from lake-level records from closed-basin lakes (e.g., Yu and Harrison 1996). Also of interest for such global model intercomparisons is the use of biome models (Prentice and Fung 1990; Guiot et al. 1996) as tools for comparing the results of such climate simulations to the climate states inferred on the basis of pollen spectra. The goal in analyses of this kind is to produce terrestrial biome distributions based upon the climate model output that can be compared to those inferred on the basis of the paleodata, such as the northward expansion of boreal forests during the warmer climate regime of the mid-Holocene (Foley 1994; Peng 1995).

Information on the long-term changes of lake-level status, with time resolution on the order of hundreds to thousands of years throughout the late quaternary, is provided primarily by the information contained in three generally accessible databases. For a description of the methodologies employed in the compilation of these databases, see Yu and Harrison (1996). For present purposes, we will focus our analyses on the Oxford Lake Level Database (hereafter OLLDB) (Street-Perrott et al. 1989) that provides good area coverage over the main regions of interest in the context of this paper, namely, the African–Asian monsoon region and the North American continent. For the purpose of comparison with our model predictions we also employ a relative indicator to represent changes from 6 kyr BP to the present. Changes in lake status from that characteristic of 6 kyr BP to that characteristic of the present are divided into five categories consistent with the methodology that has been employed to perform such comparisons in the past (Street-Perrott et al. 1990; de Noblet et al 1996; Yu and Harrison 1996). The categories range from much drier to much wetter (Fig. 11) at 6 kyr BP.

In this section we map the changes in precipitation minus evaporation as predicted by the model to the changes in lake level inferred from the lake status data to derive a more spatially precise comparison. The year to year variations of the buildup or decay of runoff in an individual basin is most usefully described by inferring changes in runoff through changes in the annual precipitation minus evaporation (PE) climatology. The annually averaged PE, 6-kyr BP present anomalies (not shown) are, however, rather weak in both the fixed and computed SST experiments. Rather than investigate this weak signal, we choose to focus upon the PE signal for the summer season (JJA), which is characterized by a much stronger and more coherent response. We further argue that because the annual cycle monsoon precipitation anomaly observed in Fig. 10d (over Nepal) reveals almost no change between 6 kyr BP and the present over the course of the other seasons, annual averaging would degrade the strength and possibly the coherence of the PE signal that actually exists during this summer monsoon season. Second, it is clear that it is the summer monsoon precipitation itself that is instrumental to the changes observed in lake levels throughout the mid-Holocene climatic optimum. For all of these reasons we employ JJA runoff to correlate with signals observed in the lake-level database (Fig. 11).

The region spanning 20°W–100°E in longitude and 35°S–45°N in latitude is a region that rather faithfully includes the present-day African–Asian monsoon region in this model (as was discussed in the last section). The JJA PE anomaly between 6 kyr BP and present for the fixed SST experiment over this region is displayed in Fig. 11a in which three areas exist that are characterized by a significant response in the runoff produced during the summer monsoon. The first region extends over the African Sahel and is characterized by anomalies of over 3 mm day−1 in excess of present. The second anomalous region, which extends from south-central Africa northeast into the Middle East is characterized by a positive–negative–positive anomaly of the same order of amplitude. The final region of interest is over Nepal where the largest positive anomaly is observed and that reaches over 10 mm day−1. The computed SST-based prediction (Fig. 11b) is very similar to that from the fixed SST experiment but is slightly stronger in magnitude. What is most readily apparent from the observations are the much higher lake levels that span most of Northern Africa. The latitudinal band of much wetter conditions from about 20° to 40°N latitude indicates that the Sahara was much wetter at 6 kyr BP than it is at present. This band does appear to be well simulated by the model but is of reduced latitudinal extent. This result is consistent with a previous study carried out by PMIP participants (Yu and Harrison 1996) in which the predictions of five models were compared with the same lake-level database. These results demonstrate that the model-predicted northward expansion of the African monsoon is too small relative to observations. The positive runoff anomaly over east India that is simulated in both the fixed and computed SST experiments is also very well correlated with the observed signal in that region although the data are sparse.

There may be several factors that are not accounted for in these model simulations of the mid-Holocene, many of which may need to be incorporated in the model in order to properly reconcile the proxy date–based inferences. For example, a study of the sensitivity of the mid-Holocene North African monsoon climate to prescribed changes in surface water (Coe and Bonan 1997) has demonstrated that there are feedbacks between the surface water and local hydrologic and energy budgets in these regions that are on the same order as the changes induced by orbital forcing alone. This study demonstrates that ocean surface temperatures in the African–Asian monsoon region might play a role in the mid-Holocene monsoon climate. The above influences combined with feedback effects from vegetation and soil need to be incorporated in future model simulations of mid-Holocene climate to further enhance the monsoon response to orbitally induced changes.

Very recent contributions of the Canadian scientific community toward paleoenvironmental reconstruction (CSHD) in the context of the PMIP project may also be brought to bear on the simulations described in this paper. One reconstruction of 6-kyr BP temperature and precipitation anomalies in the prairie provinces of Canada (southwest-central Canada) (Vance et al. 1995) has demonstrated that mid-Holocene conditions were warmer and drier than present in this region. The area investigated in this study, composing the sector extending roughly from 45° to 60°N and from 90° to 135°W, consisted of an investigation of both the plains and mountainous regions.

Quantitative estimates of the climatic changes responsible for some of the inferred changes to the vegetative and hydrological characteristics of this region are varied. In general, these changes suggest that the mean annual surface temperature was between 0.5° and 1.5°C warmer at 6 kyr BP than at present, while summer temperatures are inferred to have been 0.5°–3.0°C warmer than the present. The reduction of precipitation at 6 kyr BP was such that mean annual changes were 65 mm lower than at present, while that in the growing season was lower by approximately 50 mm. In the two experiments conducted in this paper the simulated changes varied in both temperature and precipitation in this prairie region. The summer temperature anomalies, at 6 kyr BP, were 0.9°C warmer than present in the fixed SST experiment while those of the computed SST experiment were characterized by a 0.6°C warming. The annually averaged temperature change was near zero in the fixed SST experiment while that from the computed SST experiment was characterized by being 0.3°C warmer than present. These changes, while modest, are comparable with the minimum inferences derived from the paleoenvironmentally reconstructed temperature data. The annual and summer season precipitation changes in the two 6-kyr BP simulations were small. The fixed SST experiment delivered a negative precipitation anomaly of about 7 mm in the prairie region for the 6-kyr BP summer season, while the computed SST experiment was characterized by a positive precipitation anomaly of 12 mm for the summer season at 6 kyr BP. The annually averaged anomalies in this region were 3 mm in the fixed SST experiment while that in the computed experiment showed no change from the present at 6 kyr BP.

The OLLDB reconstruction provides rather good coverage over the North American continent (Fig. 12). The annually averaged PE anomaly between 6 kyr BP and present in this region is displayed for both the fixed and computed SST experiments in Figs. 12 and 12b, respectively. There are three regions that have consistent and relatively abundant observations of lake level in this North American region. The first, which is in western Canada, is characterized by conditions that were much drier at 6 kyr BP than present, consistent with Vance et al. (1995). To the south, in the western United States, the observed data is characterized by little change in lake levels from modern. To the west of the Great Lakes there is a further region in which the lake-level data indicates that conditions were much drier than at present. The PE anomaly in both the fixed SST and computed SST experiment are inconclusive with respect to the changes inferred from the OLLDB reconstruction. Changes in both experiments do not appear to correlate with one another or with the changes inferred from the lake-level data.

In general, the precipitation changes between 6 kyr BP and the present are not consistent with, or even of the same magnitude as, those inferred from the paleoenvironmental reconstructions just as was found to be the case for the African continent. The temperature anomalies are, however, of the correct sign, though somewhat reduced in magnitude compared with those inferred from the proxy data reconstruction. A model with significantly higher spatial resolution, perhaps obtained by nesting, will possibly be required to adequately resolve these regional signals.

Other proxy climate reconstructions on the basis of which model predictions may be assessed, would include the distribution of permafrost in the peatlands of west-central Canada during the mid-Holocene at 6 kyr BP. The distribution of permafrost peatlands at 6 kyr BP is inferred to have had a southern boundary that was 300–500 km north of its present location (Zoltai 1995). These findings have been construed to suggest that the mean annual surface temperature in these regions was on the order of 5°C warmer than present. In comparison with the two experiments conducted in this paper, the annually averaged surface temperature anomalies between 6 kyr BP and present, in both the fixed and computed SST experiments, are characterized by near-zero temperature changes in the permafrost region of the paleoenvironmental reconstruction. The JJA 6-kyr BP anomalies in both the fixed and computed SST experiment do, however, indicate changes of the order of 1°–2°C warmer than present. The fixed SST experiment is characterized by an even stronger positive anomaly with maxima reaching 4°C in some areas. The increased seasonality in the model as a response to the 6-kyr BP insolation changes, however, results in colder winter surface temperature anomalies of the same magnitude as those observed in summer. Thus, the annual surface temperature distributions in the model do not correlate well with the changes inferred from those reconstructed from the paleoclimatic proxy data. Again, the problem may be connected to the limited spatial resolution of the model. Of equal concern, however, is the possibility that the inference based upon the proxy data may be significantly in error. Overriding both of these possible contributions to the misfit between model prediction and paleodatabased inference, however, is the issue of the extent to which feedbacks associated with changes in biome type, which have been entirely ignored in the work reported herein, may be contributing to these discrepancies, most notably in Africa but also in North America. This issue will be the subject of an investigation that will be reported elsewhere.

4. Conclusions

The CCCMA AGCM was employed herein to investigate changes in climate under mid-Holocene boundary conditions characteristic of those at 6000 yr BP. By conducting two experiments, one with prescribed sea surface temperatures and sea–ice distributions and the other using the model coupled to mixed layer ocean and thermodynamic sea–ice modules, the impact of computed sea surface temperatures and sea–ice extent was assessed. Also, the changes in model boundary conditions, these being both external (insolation) and internal (CO2 concentration), along with SST and sea–ice, directly influence both the surface response of the earth and the response of the free atmosphere.

The response at the surface was found to be characterized by significant changes in both surface temperature and precipitation consistent with those expected from the changes in the insolation distribution. The simulation employing the mixed layer model of the oceans revealed a further sensitivity of four major subcomponents of the climate system, namely, the following. 1) The seasonal cycle of sea–ice was modified in both the Northern and Southern Hemispheres, with the Northern Hemisphere displaying a much stronger response consistent with expectations based upon the orbital insolation change at 6 kyr BP. 2) The tropical SST anomalies, as simulated with the simple mixed layer ocean, were characterized by a high degree of phase correlation and a strong sensitivity to the annual cycle of the insolation anomaly between 6 kyr BP and the present. 3) The differences in CO2 concentration between the fixed SST experiment and the computed SST experiment also caused changes in the zonally averaged tropospheric and stratospheric thermal structure. These changes indicate that the simulated temperature fields are likely to be a purely radiative response to the insolation conditions that were present during mid-Holocene time. In particular, reversed “CO2 doubling experiment” effects were present in the fixed SST simulation. 4) The SST anomaly at 6 kyr BP was characterized by colder tropical oceans in JJA as a result of a lagged response of these oceans to the increased seasonality. This resulted in increased land–sea contrast, further enhancing the African–Asian monsoon circulation and precipitation. In particular, the warmer conditions present during the summer tropical monsoon season that are simulated by the model at 6 kyr BP predict the northward extension of the Afro–Asian transitional zone located between the northern monsoon regions and the southern fringes of the deserts (i.e., the Sahara–Sahel, the Arabian Peninsula, and northwestern India).

As mentioned at the end of the last section, the results that we have obtained in the 6-kyr BP simulations fail to entirely account for the signals inferred on the basis of a variety of paleodata. For example, the strength of the enhancement of the African monsoon, evidenced by the inferred moistening of the entire Sahara region, is not adequately explained, nor is the significant drying of the Canadian prairie. Although these discrepancies may be partly explained as a consequence of lower than necessary spatial resolution of the model, they are of a kind that one might expect to be ameliorated by the incorporation of feedbacks associated with changes in biome type. This is a subject that we intend to fully address in subsequent work.

REFERENCES

  • Barnola, J. M., D. Raynaud, Y. S. Korotkevich, and C. Lorius, 1987:Vostok ice core provides 160 000-year record of atmospheric CO2. Nature,329, 408–418.

  • Berger, A., 1978: Long-term variation of daily insolation and quaternary climate change. J. Atmos. Sci.,35, 2362–2367.

  • Boer, G. J., and M. Lazare, 1988: Some results concerning the effect of horizontal resolution and gravity-wave drag on simulated climate. J. Climate,1, 789–806.

  • ——, and B. Denis, 1997: Numerical convergence of the dynamics of a GCM. Climate Dyn.,13, 359–374.

  • Chervin, R. M., and S. H. Schneider, 1976: On determining the statistical significance of climate experiments with general circulation models. J. Atmos. Sci.,33, 405–412.

  • Coe, M. T., and G. B. Bonan, 1997: Feedback between climate and surface water in northern Africa during the middle Holocene. J. Geophys. Res.,102, 11 087–11 101.

  • COHMAP members, 1988: Climate changes of the last 18,000 years:Observations and model simulations. Science,241, 1043–1052.

  • Delworth, T., and S. Manabe, 1989: The influence of soil wetness on near-surface atmospheric variability. J. Climate,2, 1447–1462.

  • de Noblet, N., P. Braconnot, S. Joussaume, and V. Masson, 1996: Sensitivity of simulated Asian and African summer monsoons to orbitally induced variations in insolation 126, 115, and 6 kyr BP. Climate Dyn.,12, 589–603.

  • de Vernal, A., A. Rochon, C. Hillaire-Marcel, J.-L. Turon, and J. Guiot, 1993: Quantitative reconstruction of sea-surface conditions, seasonal extent of sea–ice cover and meltwater discharges in high-latitude marine environments. Ice in the Climate System, W. R. Peltier, Ed., Pergamon, 611–621.

  • Foley, J. A., 1994: The sensitivity of the terrestrial biosphere to climate change: A simulation of the middle Holocene. Global Biogeochem. Cycles,8, 505–525.

  • Guiot, J., R. Cheddadi, I. C. Prentice, and D. Jolly, 1996: A method of biome and land surface mapping from pollen data: Application to Europe 6000 years ago. Paleoclimates,1, 311–324.

  • Hall, N. M. J., and P. J. Valdes, 1997: A GCM simulation of the climate 6000 years ago. J. Climate,10, 3–17.

  • Hewitt, C. D., and J. F. B. Mitchell, 1996: GCM simulations of the climate of 6 kyr BP: Mean changes and interdecadal variability. J. Climate,9, 3505–3529.

  • Joussaume, S., and P. Braconnot, 1997: Sensitivity of paleoclimate simulation results to seasonal definitions. J. Geophys. Res.,102, 1943–1956.

  • Kutzbach, J. E., and B. L. Otto-Bliesner, 1982: The sensitivity of the African–Asian monsoonal climate to orbital parameter changes for 9000 years B. P. in a low-resolution general circulation model. J. Atmos. Sci.,39, 1177–1188.

  • ——, and P. J. Guetter, 1986: The influence of changing orbital parameters and surface boundary conditions on climate simulations for the past 18 000 years. J. Atmos. Sci.,43, 1726–1759.

  • ——, and R. G. Gallimore, 1988: Sensitivity of a coupled atmosphere/mixed layer ocean model to changes in orbital forcing at 9000 years B. P. J. Geophys. Res.,93, 803–821.

  • Laprise, R., and C. Girard, 1990: A spectral general circulation model using a piecewise-constant finite-element representation of a hybrid vertical coordinate system. J. Climate,3, 32–52.

  • Laskar, J., F. Joutel, and F. Boudin, 1993: Orbital, precessional, and insolation quantities for the Earth from −20 Myr to +10 Myr. Astron. Astrophys.,270, 522–533.

  • Legates, D. R., and C. G. Willmott, 1990: Mean seasonal and spacial variability in gauge-corrected, global precipitation. Int. J. Climatol.,10, 111–127.

  • Li, C., and M. Yanai, 1996: The onset and interannual variability of the Asian summer monsoon in relation to land-sea thermal contrast. J. Climate,9, 358–375.

  • McFarlane, N. A., G. J. Boer, J.-P. Blanchet, and M. Lazare, 1992: The Canadian Climate Centre second-generation general circulation model and its equilibrium climate. J. Climate,5, 1013–1044.

  • Meehl, G. A., 1994: Influence of the Land Surface in the Asian summer monsoon: External conditions versus internal feedbacks. J. Climate,7, 1033–1049.

  • Mitchell, H. L., C. Charette, S. J. Lambert, J. Halle, and C. Chouinard, 1993: The Canadian global data assimilation system: Description and evaluation. Mon. Wea. Rev.,121, 1467–1492.

  • ——, C. Chouinard, C. Charette, R. Hogue, and S. Lambert, 1996: Impact of a revised analysis algorithm on an operational data assimilation system. Mon. Wea. Rev.,124, 1243–1255.

  • Mitchell, J. F. B., 1993: Modelling of paleoclimates: Examples from the recent past. Philos. Trans. Roy. Soc. London B,341, 267–275.

  • ——, N. S. Grahame, and K. J. Needham, 1988: Climate simulations for 9000 years before present: Seasonal variations and effect of the Laurentide Ice Sheet. J. Geophys. Res.,93, 8283–8303.

  • Peltier, W. R., 1994: Ice age paleotopography. Science,265, 195–201.

  • Peng, C. H., J. Guiot, E. Van Campo, and R. Cheddadi, 1995: Temporal and spatial variations of terrestrial biomes and carbon storage since 13 000 yr bp in Europe: Reconstruction from pollen data and statistical models. Water, Air Soil Pollut.,82, 375–390.

  • Phillipps, P. S., and I. M. Held, 1994: The response to orbital perturbations in an atmospheric model coupled to a slab ocean. J. Climate,7, 767–782.

  • Prell, W. L., and J. E. Kutzbach, 1992: Sensitivity of the Indian monsoon to forcing parameters and implications for its evolution. Nature,360, 647–652.

  • Prentice, K. C., and I. Fung, 1990: The sensitivity of terrestrial carbon storage to climate change. Nature,346, 48–51.

  • Quinn, T. R., S. Tremaine, and M. Duncan, 1991: A three million year integration of the Earth’s orbit. Astro. J.,101, 2287–2305.

  • Rasmusson, E. M., and T. H. Carpenter, 1983: The relationship between the eastern Pacific sea surface temperature and rainfall over India and Sri Lanka. Mon. Wea. Rev.,111, 517–528.

  • Ruddiman, W. F., and A. C. Mix, 1993: The Northern and equatorial Atlantic at 9000 and 6000 yr bp. Global Climates Since the Last Glacial Maximum, H. E. Wright et al., Eds., University of Minnesota, 94–124.

  • Saurez, M. J., and I. M. Held, 1979: The sensitivity of an energy balance model to variations in the orbital parameters. J. Geophys. Res.,84, 4825–4836.

  • Sperber, K. R., S. Hameed, G. L. Potter, and J. S. Boyle, 1993: Simulations of the Indian and East-Asian summer monsoon in the ECMWF model: Sensitivity to horizontal resolution. PCMDI Rep. 12, 35 pp.

  • Street, F. A., and A. T. Grove, 1979: Global maps of lake-level fluctuations since 30 000 yr bp. Quat. Res.,12, 83–118.

  • Street-Perrott, F. A., D. S. Marchand, N. Roberts, and S. P. Harrison, 1989: Global lake-level variations from 18 000 to 0 years ago: A paleoclimatic analysis. U.S. Dept. of Energy Tech. Rep. 46, 44 pp.

  • ——, J. F. B. Mitchell, D. S. March, and J. S. Brunner, 1990: Milankovitch and albedo forcing of the tropical monsoon: A comparison of geological evidence and numerical simulations for 9000 ybp. Philos. Trans. Roy. Soc. Edinburgh,81, 407–427.

  • Tibaldi, S., T. N. Palmer, C. Brankovic, and U. Cubasch, 1990: Extended-range prediction with ECMWF models: Influence of horizontal resolution on systematic error and forecast skill. Quart. J. Roy. Meteor. Soc.,116, 835–866.

  • Vance, R. E., A. B. Beaudoin, and B. H. Luckman, 1995: The paleoecological record of 6 ka BP climate in the Canadian prairie provinces. Geogr. Phys. Quat.,49, 81–98.

  • Yan, Z., and N. Petit-Maire, 1994: The last 140 ka in the Afro-Asian arid/semi-arid transitional zone. Paleogeo. Paleoclimatol. Paleoecol.,110, 217–233.

  • Yu, G., and S. P. Harrison, 1996: An evaluation of the simulated water balance of Eurasia and northern Africa at 6000 ybp using lake status data. Climate Dyn.,12, 723–735.

  • Zoltai, S. C., 1995: Permafrost distribution in peatlands of west-central Canada during the Holocene warm period 6000 years BP. Geogr. Phys. Quat.,49, 45–54.

  • Zwiers, F. W., 1987: A comparison of climates simulated by a general circulation model when run in the annual cycle of perpetual modes. Mon. Wea. Rev.,115, 2626–2644.

  • ——, 1993: Simulation of the Asian summer monsoon with the CCC GCM-1. J. Climate,6, 470–486.

Fig. 1.
Fig. 1.

Incoming solar radiation anomaly at the top of the atmosphere (insolation anomaly) for the difference between 6 kyr BP and modern in W m−2 as a function of latitude and time of year. The contour interval is 4 W m−2.

Citation: Journal of Climate 11, 10; 10.1175/1520-0442(1998)011<2607:SOMHCU>2.0.CO;2

Fig. 2.
Fig. 2.

Approach to equilibrium in the computed SST model simulation, for both the present-day and 6-kyr BP time periods. Modern global average JJA 2-m surface air temperature (°C) vs model time (yr) for present day is denoted by the solid line, while that for the 6-kyr BP simulation is denoted by the dashed line. The solid line traversing both curves denotes the 10-yr running mean for the time series. The inset displays the 6-kyr BP JJA global average sea–ice amount in kg m−2. The 6-kyr BP simulation was extended from the equilibrium configuration of the modern control at year 52 of simulation time.

Citation: Journal of Climate 11, 10; 10.1175/1520-0442(1998)011<2607:SOMHCU>2.0.CO;2

Fig. 3.
Fig. 3.

Latitude vs time of year for the 2-m surface air temperature anomaly (°C) between 6 kyr BP and modern. The contour interval is at 0.5°C. Panels (a) and (b) are for the fixed SST experiment zonally averaged over land and ocean, respectively. Panels (c) and (d) are the same as for (a) and (b), but for the computed SST experiment.

Citation: Journal of Climate 11, 10; 10.1175/1520-0442(1998)011<2607:SOMHCU>2.0.CO;2

Fig. 4.
Fig. 4.

Global distributions of DJF and JJA 2-m surface air temperature anomalies (°C) for the fixed and computed SST experiments. The contour interval is at 1°C with shaded regions designating changes below −1°C for DJF and changes above 1°C for JJA.

Citation: Journal of Climate 11, 10; 10.1175/1520-0442(1998)011<2607:SOMHCU>2.0.CO;2

Fig. 5.
Fig. 5.

Insolation forcing and SST response in the computed SST experiment over the tropical ocean area averaged from 24°S to 24°N. Insolation anomalies at 6 kyr BP as compared to modern are designated by the dashed line while the solid line denotes changes in surface air temperature at 2 m between 6 kyr BP and present. Thick solid and dashed horizontal lines denote the annual mean surface air temperature and annual mean insolation anomalies in this tropical region, respectively.

Citation: Journal of Climate 11, 10; 10.1175/1520-0442(1998)011<2607:SOMHCU>2.0.CO;2

Fig. 6.
Fig. 6.

Zonally averaged vertical temperature structure anomalies as a function of latitude (6 kyr BP − modern) (a) for the DJF-fixed SST experiment, (b) the JJA fixed SST experiment, (c) the DJF computed SST experiment, and (d) the JJA computed SST experiment. The contour interval is 0.5°C.

Citation: Journal of Climate 11, 10; 10.1175/1520-0442(1998)011<2607:SOMHCU>2.0.CO;2

Fig. 7.
Fig. 7.

(a) Sea-ice mass anomalies (6 kyr-modern) within the Arctic and Antarctic circles expressed as monthly means throughout the annual cycle (solid lines) and corresponding anomalies in insolation (dashed lines). (b) Polar stereographic projection of the Arctic sea-ice mass anomaly in September (the contour interval is 300 kg m−2).

Citation: Journal of Climate 11, 10; 10.1175/1520-0442(1998)011<2607:SOMHCU>2.0.CO;2

Fig. 8.
Fig. 8.

Surface air temperature at 2 m: the modern August fixed SST surface air temperature (a) and the observed August surface temperature (b) are in units of °C with a contour interval of 10°C. Precipitation: the modern August fixed SST precipitation (c) and the observed August precipitation (d) are in units of mm day−1 with a contour interval of 2 mm day−1. The 850-hPa winds: the modern JJA-fixed SST 850-hPa wind (e) and the observed JJA 850 hPa wind (f) are in units of m s−1 with a contour interval of 5 m s−1.

Citation: Journal of Climate 11, 10; 10.1175/1520-0442(1998)011<2607:SOMHCU>2.0.CO;2

Fig. 9.
Fig. 9.

Surface air temperature at 2 m: the 6 kyr BP − modern August anomalies for the (a) fixed and (b) computed SST experiments are in contour intervals of 1°C, with positive regions shaded at the same interval. Precipitation: the 6 kyr BP − modern August anomalies for the (c) fixed and (d) computed SST experiments are contoured at ±1, ±2, ±4, ±8, and ±16 mm day−1, with positive regions shaded at the same interval. The 850-hPa winds: The 6 kyr BP − modern August anomalies for the (e) fixed and (f) computed SST experiments are in contour intervals of 2 m s−1 with positive regions shaded at the same interval.

Citation: Journal of Climate 11, 10; 10.1175/1520-0442(1998)011<2607:SOMHCU>2.0.CO;2

Fig. 10.
Fig. 10.

Grid-cell analysis of the major anomalies in the previous figure. The letters denoting the actual grid cells investigated are displayed upon the model spectral topography with a contour interval of 200 m (a). The annual cycle of the variations in 850-hPa wind (m s−1) (b), surface air temperature at 2 m (°C) (c), and the precipitation (mm day−1) (d) for the fixed and computed SST experiments denoted by the solid and dashed lines, respectively, at each grid cell.

Citation: Journal of Climate 11, 10; 10.1175/1520-0442(1998)011<2607:SOMHCU>2.0.CO;2

Fig. 11.
Fig. 11.

Lake-level variations for the African–Asian monsoon region with the corresponding JJA PE anomalies (mm day−1) in the (a) fixed and (b) computed SST experiments. Lake-level anomalies range from much lower to much higher at 6 kyr BP as compared to present. The PE anomalies are contoured at ±0.5, ±1.0, ±1.5, ±2.0, ±4.0, and ±8.0 mm day−1.

Citation: Journal of Climate 11, 10; 10.1175/1520-0442(1998)011<2607:SOMHCU>2.0.CO;2

Fig. 12.
Fig. 12.

Lake-level variations for the North American region with the corresponding annual PE anomalies (mm day−1) in the (a) fixed and (b) computed SST experiments. Lake-level anomalies range from much lower to much higher at 6 kyr BP as compared to present. The PE anomalies are contoured at ±0.5, ±1.0, ±1.5, ±2.0, ±4.0, and ±8.0 mm day−1.

Citation: Journal of Climate 11, 10; 10.1175/1520-0442(1998)011<2607:SOMHCU>2.0.CO;2

Save