PALEOCLIMATE DATA AND MODEL COMPARISONS
Overview
We combined floral and lithological data to determine Permian and Jurassic terrestrial biomes. We use a classification in which the 'macroclimate' of the present-day land surface is expressed in terms of ten major biomes. It is simple and therefore more readily applicable in the geological past, when detailed knowledge of biomes is limited. However, it retains important information on temperature, precipitation, and seasonality; factors that are critical in controlling vegetation patterns. We also expressed the simulated climate (temperature, precipitation - from General Circulation Model (GCM) experiments) in terms of biomes, to compare against the observed biome distributions.
Permian
General Circulation Model (GCM) experiments were conducted for the Wordian and the results compared with the data-derived biomes. The model performs well in the tropics and northern high latitudes, reproducing reasonably well the data-derived patterns. There is no net annual snow accumulation at low elevations in the model, consistent with the observed record of ice-free conditions in the Wordian. However, even with a high CO2 level (8X Present), model summer temperatures only just rise above freezing in the southern high latitudes. This region contains the largest data/model discrepancy; cold and cool temperate biomes are indicated by the paleobotanical data, whereas tundra and cold temperate conditions are predicted by the model. Such polar discrepancies are common to other warm interval data/model comparisons, and fully resolving them remains a fundamental problem in paleoclimatology. A solution may be a 'warm polar current', which the model's configuration (a simple mixed-layer ocean) cannot resolve. This current could have arisen under the paleogeographic regime of the Wordian, where one large supercontinent lay just off the South Pole. The northward shift of Gondwana through the Permian, coupled with a rise in atmospheric CO2, may have initiated deglaciation and polar warming by allowing warm ocean currents to reach the south polar region. We are currently conducting new experiments with coupled atmosphere and ocean GCMs to investigate the potential effects of such a warm polar current.
 |
Wordian biomes derived from (a) floral and lithological data, and (b) atmospheric GCM simulation for 8X CO2 and solar luminosity reduced by 2.1% relative to the present day. Note that the glacial biome (10) is not recognised for the Wordian, either from the data or the model.
Rees, P.M., M.T. Gibbs, A.M. Ziegler, J.E. Kutzbach, and P.J. Behling (1999). Permian climates: evaluating model predictions using global paleobotanical data
Jurassic
In the Jurassic, five main biomes are recognised from the data: seasonally dry (summerwet or subtropical), desert, seasonally dry (winterwet), warm temperate and cool temperate. The boundaries between them remained at near-constant palaeolatitudes while the continents moved through them (south, in the case of Asia, and north, in the case of North America). Net global climate change throughout the Jurassic appears to have been minimal. The data-derived results can be compared with a climate simulation for the Late Jurassic (see below). The use of more detailed palaeogeography and palaeotopography has improved the overall data/model comparisons. Major discrepancies persist at high latitudes, however, where the model predicts cold temperate conditions far beyond the tolerance limits indicated by the plants.
(a) Late Jurassic biomes inferred from climate-sensitive sediment and floral data:
(b) Predicted biome/climate zones for the Late Jurassic, based on model results for temperature, precipitation and soil moisture patterns:
Overall comparison between the data and model is encouraging, maintaining the broad pattern of summerwet equatorial regions, succeeded polewards by desert then warm and cool temperate biomes. In the tropics, there are a few grid points predicting tropical rainforest type biomes, but there are no data near any of these grid points and so the model could be correct. The rest of the tropics are predicted to be summerwet, which is generally in good agreement with the data. In mid- to high latitudes, the most striking feature of the Northern Hemisphere is that the model predicts more extensive regions of winterwet climates and more restricted regions of warm temperate climates. In some senses, the distinction between these climates is relatively subtle and so perhaps the model error is correspondingly small here. Nonetheless, the model is somewhat too cool at these latitudes. The tendency of the model to be too cool is more striking in the Southern Hemisphere, where overall data and model agreement at high latitudes is poor. On the coast itself, the agreement is reasonable and it can be argued that this suggests that our choice of sea surface temperatures is acceptable. However, in the southern interior of the continent the model is predicting a mixture of cool and cold temperate climates and very few areas of warm temperate climates. As with northern high latitudes, the model is clearly predicting temperatures which are too cold.
The data/model comparison clearly suggests that the model is too cold at high latitudes in both hemispheres. Changing CO2 levels and sea surface temperatures could make some difference, but is unlikely to completely resolve the problem. It appears to be another example of the 'equable climates' issue noted for the Cretaceous and Eocene. Recent work has suggested that some of the disagreement can be reconciled by including feedbacks between the climate and vegetation. This effect is also likely to be important for the Jurassic and we are in the process of incorporating this in our simulations.
Rees, P.M., A.M. Ziegler, and P.J. Valdes, 2000. Jurassic phytogeography and climates: new data and model comparisons.