An Ice--Free Siberia: A Clue to Carboniferous CO2 Levels

Heretofore paleo--CO₂ levels have been estimated by geochemical models or proxies (stomatal indices, fossil soil composition). A complementary approach to this problem is that of inverse modeling. In brief, for the period in question oxygen isotopic evidence is used to infer a given amount of continental ice volume, and one or more climate models with dynamic ice sheet components are run to determine which CO₂ level is compatible with the predicted ice volume. While this method holds considerable promise, uncertainties in certain model inputs (paleolatitude, topography, salinity) may result in a very wide range of hindcast CO₂ levels. However, for the late Carboniferous a constraint may be imposed on the ensemble of model runs. There is no compelling evidence for significant Siberian glaciation in the Carboniferous, although Siberia is posited to be at a relatively high paleo--latitude and isolated from other continents --an ideal configuration for the establishment of permanent ice. Thus by reducing our ensemble of model runs to those in which the Siberian ice sheet is nonexistent we can constrain the modeled paleo--CO₂ predictions. Preliminary energy balance/ice sheet model (EB/ISM) results indicate that the effectiveness of this constraint will itself depend to a degree on the paleo--topography. At present the constraint on CO₂ value seems stronger for our 320 Ma simulations than for 360 or 280 Ma. For this interval our ``best guess'' CO₂ level for little/no Siberian ice is 3X present levels -- somewhat higher than the 1X estimates from stomatal data and the Berner geochemical model.