Geological evaluation of multiple general circulation model simulations of Late Jurassic palaeoclimate

General circulation models (GCMs) are currently used to predict future global change. However, the robustness of GCMs should be evaluated by their ability to simulate past climate regimes. Their success in 'retrodiction' can then be assessed by reference to the geological record. Geological evidence provides a database that can be used in the estimation of sea surface temperatures, orography and other proxy data useful in palaeoclimatic studies, These data can then be used to refine the prescribed boundary conditions for running GCMs themselves. Results from a series of modelling experiments, run with Late Jurassic (Kimmeridgian) boundary conditions, and using the UK University Global Atmospheric Modelling Programme (UGAMP) GCM and the UK Meteorological Office GCM are presented. Simulations from these two quite independently generated models, although subtly different, confirm a generally warmer Jurassic Earth with arid zones over the Tethys and SW USA, parts of Gondwana dominated by 'monsoonal' systems and convective rainfall generally higher over the oceans than at present. Circum-polar wetlands are also indicated. These results generally conform well to the distributions of known facies in these regions. Modelled cloudiness is also higher in the Jurassic, and although unconfirmed geologically, such conditions would have contributed to greenhouse conditions at high latitudes and could have influenced both terrestrial biomes and marine ecosystems. Using one of the GCMs (UGAMP) we have also investigated the role of orbital parameters for high latitude climate. At times of 'minimum seasonal forcing' (comparable with an orbital geometry affecting the Earth at 115 ka BP) parts of Antarctica could have sustained a modest ice sheet over areas exceeding 1 km elevation, but such modelled sheets would have been ephemeral features and very dynamic in character. (C) 2000 Elsevier Science B.V. All rights reserved.