General circulation (GCM) modelling of palaeoclimates: a critique.

Abstract: The best method so far devised for testing the ability of a general circulation model (GCM) to simulate climates very different from today is to conduct ’snapshot ’ experiments for specific times in the past at which the climatic forcing can be prescribed in detail. This paper focuses on the period from the last glacial maximum through the Holocene. During this interval, changes in atmospheric circulation were driven primarily by variations in the seasonality of insolation and secondarily by variations in surface-boundary conditions accompanying deglaciation. Changes in atmospheric transparency are usually neglected, apart from lower CO2 at 18 ka BP. Palaeoclimatic simulations have yielded important insights into the dynamics of glacial/interglacial climate change. The results have been tested in three main ways: 1) qualitative comparison of the spatial patterns of climate anomalies; 2) comparison of the simulated climate at individual grid points or groups of grid points with quantitative estimates derived from geological data; and 3) further processing of the model output to yield simulated values of palaeoclimatic indicators such as pollen spectra. The first approach makes effective use of qualitative data, for example, aeolian evidence. The second is largely confined to data-rich areas such as Europe and North America. The third technique, which capitalises on the relative strengths of both approaches, deserves much greater attention