Comment on S. H. Schneider's Editorial 'Can Modeling of the Ancient Past Verify
Prediction of Future Climates?' (Climatic Change 8, 117-119) Dr Schneider addressed the question of verification of climate model results, and, in particular, climate model sensitivity. Although we agree with his basic premise, that paleoclimate simulations provide a useful test of the ability of climate models to respond to alterations in climate forcing, the model simulations he chose were not comparable. If the different models show different responses to the same climate forcings, comparison with paleoclimatic evidence might produce an assessment of which model, or which model parameterization, is most realistic. However, if the models employ different boundary conditions their responses will reflect the boundary conditions as well as any differences in model sensitivity; such is the case in the comparisons cited in the above editorial.
Schneider compared the Younger Dryas simulation of 11 000 yr ago (1 lk) using the GISS model reported by Rind et aL (1986), with the 9k simulation of Kutzbach and Guetter (1984J'-made with the NCAR CCM. Each run used the orbital parameters appropriate to the particular time, and the difference in insolation between 9k and 1 lk was small -Northern Hemisphere summer at 9k had 1.3% more solar insolation than at 1 lk. In both time periods the Northern Hemisphere summer insolation was greater than that of today, and both models showed increased warming over Eurasia in summer compared to the current climate simulation. However, the GISS model, with boundary conditions appropriate for 1 lk, produced warming of up to 10 ~ during this season, while the NCAR model, with a North American ice sheet and the solar insolation change, fiad warming on the order of 4 ~ This difference is not the result of different model sensitivities: instead, it reflects the different boundary conditions used in the simulations.
The GISS 1 lk simulation included changes of other elements of the climate system appropriate for the late-glacial time, in addition to the orbital parameters. In particular, this simulation included the presence of the 1 lk land ice, which was still relatively extensive (Denton and Hughes, 1981). The 9k simulation with the NCAR CCM used today's land ice distribution in Eurasia. How would this difference affect the results? Kutzbach and Wright (1985) discussed the influence of elevated land ice on the local meridional circulation. Regions to the northwest and southeast of the ice sheets tend to experience subsidence (op. cir., Figure 2), as a result of their location vis-a-vis the jet stream. The eastern edge of the ice sheet marks the termination of the zone of increased latitudinal temperature gradient associated with the land ice, and is thus the region in which the jet stream decelerates. The deceleration is accomplished by the effect of the coriolis force on a local indirect circulation, the circulation which produces the subsidence to the southeast. Wanning to the southeast associated with this subsidence is the natural response to an ice sheet in a west wind regime, although the degree of warming would be affected by additional processes (advection, radiation, etc.).
The GISS 1 lk experiment included the remaining Western European land ice at 1 lk, with an elevation about 2/3 of the full ice age topography. As expected from the above Climatic Change 9 (1986) 357-360. 9 1986 by D. ReiclelPublishing Company.