GEOPHYSICAL RESEARCH LETTERS, VOL. 19, NO. 8,PAGES 793-796, APRIL 24,1992 MILANKOVITCH FLUCTUATIONS ON SUPERCONTINENTS Thomas J. Crowley and Steven K. Baum Applied Research Corporation, College Station, Texas William T. Hyde Dalhousie University, Halifax, Nova Scotia, Canada Abstract. Observational and modeling studies support the concept of significant Milankovitch fluctuations in the pre- Pleistocene. Many of these studies havefocused on themid- Cretaceous, a time of high sea level and moderately dispersed land-sea distribution. In thispaper we utilize a two-dimen- sional energy balance climate model to examine the potential effects of Milankovitch fluctuations for a supercontinent con- figuration (Early Jurassic, 195 Ma). We calculate that Milankovitch variations can modulate themagnitude of sum- mer warming by asmuch as 14-16øC on Pangaea, with large changes occurring in boththe northern andsouthern hemi- spheres. These values arecomparable to or slightly larger than the calculated rangefor the present Eurasian landmass forthe Pleistocene. In addition, mean monthly summer tem- peratures of 25øC reach 65 øpaleo!afitude for "hot orbit" con- figurations. This latter result suggests thatsome organisms may have beenable to migrate into higher latitudes during precession half-cycles of 10,000 years duration, and a poten- tialbias should be considered whenevaluating evidence for past high4atitude warmth. Ourresults indicate thatgeneral circulation model(GCM) runs for past time periods which use only one orbital configuration maybemissing a substan- tial amount of thepotential variance onlarge land masses. Introduction There is a growing body of evidence forMi!ankovitch cy- cles in thepre-Pleistocene [e.g., Olsen, 1986; Herbert and Fischer, 1986], and climate modeling studies indicate that for some geographic configurations significant variations can occur in the system response toorbital forcing [e.g., Crowley and Baum, 1991; Park and Oglesby, 1991].Although some ofthese simulations involved large landmasses, and one has studied variations in equatorial regions on the Triassic- Jurassic supercontinent [Crowley et al., 1992], none have addressed therelative magnitude, vis4t-vis thepresent, of Milankovitch fluctuations in mid- and high-latitudes of the Triassic-Jurassic landmass. Thisis theregion where we would expect the largest amplitude Milankovitch fluctuations -~ the seasonal cycle of forcing is greatest in higher latitudes and there is a verylarge land area that should amplify the temperature response. In this paperwe conduct such an exercise and explore the implications. Modeland Boundary Conditions We use atwo-dimensional, nonlinear seasonal energy bal- ance model (EBM) in our simulations [Noahet al., 1983; Copyright 1992 bythe American Geophysical Union. Paper number 92GI.1)0561 0094-8534/92/92 GL-00561 $03.00 Hyde et al, 1990]. This is a thermodynamic model which resolves theseasonal cycle in terms of land-sea distribution (described in terms of differential heat capacities), with solar forcing balanced by outgoing longwave radiation, and heat transport modeled asa diffusive process. The single non- linear feedback involves higher albedo for areas with sim- u!ated snow cover. Extensive comparisons with observations and general circulation models (GCMs)indicate that the EBM has a sensitivity comparable to GCMs for seasonal changes over land [Crowley andBaum, 1991; Hyde et al., 1989, 1990; Kim and North, 1991; North et al., 1992], and that this sensitivity implies a near-linear response of the system to solar forcing over !and. We utilizethegeography of Parrish et al. [1982] for the Early Jurassic (Pliensbachian, 195Ma), a time interval just prior to or coincident withinitiation of breakup of theland- mass. Because the solar output has varied over Earth history, we utilized a solar constant 2.25%less than thepresent [of. Crowley and Baum, !991, Figure 5]. Simulations were conducted forboth present and higher levels of CO2. Berner [ 1991]has estimated CO2levels for this timeintervalasfour times greater than thepresent. This estimate is subject to considerable uncertainty but receives some general support from initial proxy-CO2 approaches [Cerling, 1991]. Because of these uncertainties, and because thesensitivity of ourcli- mate model to CO2 changes is onthelow end of the range in GCMs [Hydeet al., 1990,Table 3], we utilizevalues of both 4X and 8X CO2 in oursimulations. Incorporation of the CO2effect required adjusting theoutgoing infrared emission termin the EBM to radiative changes as calculated from a model by Kiehl and Dickinson [ 1987]. Simulations were conductedfor two different levels of so- lar forcing -- "hot" and "cold" summerorbits (HSO and CS O). Theseare definedas maximum and minimum orbital insolation values as determined from Pleistocene fluctuations ascalculated by Berger [1978]. Although orbital periods were 5-10% less in the Jurassic [Berger et al., 1989],we do not have any information onchanges in magnitude of forcing; thususeof Pleistocene valuesis justified. The two orbital configurations result in seasonal insolation values asmuchas 160 W/m 2 different formid-latitude sites. These changes should bemanifested in terms of a significant temperature re- sponse on large landmasses. Results and Discussion The EarlyJurassic response to orbital insolation variations is illustrated in Figures 1 and2. For comparison, the differ- ence between hotand coldsummer orbits for thepresent ge- ography is also shown.Note thatfor both hemispheres, the "hot minus cold" maximum range is ~14-16øC -- comparable to or slightly larger than themodeled range on the present 793