Simulations of Permian climate, and sedimentary proxies
The Early Permian floras of Prince Edward Island, Canada


The Paleozoic era is fascinating from botanical, geographic and climate perspectives. It encompasses the evolution of land plants, with their impact on global climate, atmosphere and animal evolution. Geographically, it was very different to today - there was one supercontinent (Pangea) in the late Paleozoic. In terms of global climate and vegetation, however, the late Paleozoic Carboniferous Period was broadly analogous to today. There were polar icecaps and equatorial rainforests (albeit populated by very different kinds of plants). However, a transition occurred from a Carboniferous 'icehouse' world to one with no polar ice sheets and no equatorial rainforests. This new 'hothouse' world persisted for ~200 million years, to at least the late Mesozoic. The effects of this transition were first apparent in the Permian, which has been the focus of my Paleozoic research.


A knowledge of the paleogeographic context is essential to properly understand global patterns of climate change, faunal and floral distributions, ecology, evolution and extinctions. For my Permian work, I've used maps produced by my colleague Fred Ziegler and others at the Paleogeographic Atlas Project (PGAP), The University of Chicago. They remain the most detailed set of Permian global paleogeographic maps available.

Tatarian (Late Permian)

Wordian (Middle Permian)
Sakmarian (Early Permian)
Artinskian (Early Permian)

The maps include mountain ranges that were being produced in the Permian and remnant ranges from earlier periods. They are the first attempts to characterize paleotopography at this detailed level. The distribution of land and sea as well as land surface elevations are basic inputs for paleoclimate modeling. Note the southern polar icecap in the Sakmarian and the change to ice-free conditions on younger maps. Click any map to enlarge. Contact David Rowley for details.


I've used fossil plants, specifically leaf remains, to intepret global patterns of climate for different geologic intervals. Two Permian stages, the Sakmarian and Wordian, were examined in detail, using sedimentary as well as floral data. I chose these intervals because they represent glaciated and ice-free worlds, as evidenced most markedly by the loss of glacial tillites between the Sakmarian and Wordian (see maps, above). In addition, there was a loss of equatorial 'rainforests' between the the two stages, apart from some remnants on islands that today form part of China. Instead, the temperate regions were heavily-vegetated by the Wordian, and the equatorial region only seasonally wet.

Sakmarian and Wordian floral localities, expressed as scaled pie diagrams and showing the morphological categories and numbers of genera present in each flora.

Click here or on any of these floral and biome maps for details and larger images.

The combined floral and lithological data were used to determine terrestrial climate zones, or biomes. I used a classification in which the 'macroclimate' of the present-day land surface is expressed in terms of ten major biomes (Walter, 1985). The classification is simple and therefore more readily applicable in the geologic past, for which detailed knowledge of biomes is limited (e.g. Ziegler, 1990).

Collaboration with John Kutzbach and other climate modelling colleagues enabled us to make detailed comparisons between the data- and model-derived biomes for the Sakmarian and Wordian. We compared these for a range of CO2 levels (1x, 4x, and 8x Present). The simulations with 4xCO2 match the observations better than the simulations with 1xCO2 and, at least in some areas, the simulations with 8xCO2 match slightly better than those for 4xCO2. Overall, the 4xCO2 and 8xCO2 biome simulations match the data reasonably well in the equatorial and mid latitudes as well as the northern high latitudes. However, even these highest CO2 levels fail to produce the temperate climates in high southern latitudes indicated by the data.

This figure shows modelled biomes for the Sakmarian and Wordian, derived from the 8xCO2 circular orbit experiments.

The lack of sufficient ocean heat transport into polar latitudes may be one of the factors responsible for this cold bias of the climate model. Another factor could be the treatment of land surface processes, and the lack of an interactive vegetation module.

Plant Diversity and Evolution

Using the frameworks of paleogeography and paleoclimate developed for the Permian and other intervals, I've turned my attention to patterns of land plant diversity and extinction through time. What makes the results significant is their spatial context. Clearly, it is not enough to provide global summations through time of whatever happens to have been preserved and collected. The following images illustrate some of my results for the Permian and Triassic, addressing the end-Permian mass extinction and it's effects on vegetation. Details can be viewed here.

These paleogeographic maps show the extent of continental motion during the Permian and Triassic. A: Late Permian (~255 Ma); major geographic regions are highlighted. B: Early Permian (~285 Ma). C: Early Jurassic (~190 Ma). The southern polar region was covered by land and the north was open ocean in the Early Permian, but the opposite was true by the Early Jurassic. Also, most of the smaller landmasses present in the Permian had joined major ones by the Jurassic.

In addition to these geographic changes, global climate changed markedly such that a glaciated southern pole and equatorial rain forests in the Early Permian gave way to ice-free poles and an equatorial region that was only seasonally wet later in the Permian, a situation that lasted through the Mesozoic. Permian examples are shown above, and Jurassic ones can be seen on my Mesozoic web page. Global compilations of plant genera for the 16 Permian and Triassic stages are shown here:

(A). Temporal patterns of diversity and plant localities, oldest to youngest from left to right. This is really just a count of everything that has been documented for each stage and contains no spatial information. The vertical line is the Permian-Triassic boundary. Note how diversity decreases towards the boundary, consistent with marine faunal evidence for a mass extinction at this time.

(B). This is the same plant data as used in (A) but shows a markedly different pattern now that diversity has been considered in a geographic context. The plot shows spatial-diversity patterns, scaled in six size categories within 10 degree latitudinal bins (largest squares, >75 genera; smallest, 1-15 genera). It's intended to provide an idea of where the diversity highs and lows were located geographically in each geologic stage. This should be viewed in conjunction with the paleogeography and paleoclimate maps shown above. One striking feature is the absence of data at top left and bottom right as well as between ~ 0 and 30 degrees south in the Triassic. These are ocean (see the paleogeogaphic maps, above). The most obvious feature is the latitudinal shift in the locus of maximum plant diversity, from equatorial regions in the Early Permian to mid-latitude temperate regions by the Late Triassic.

If this plot was extended backwards in time, maximum diversity would remain at the equator in the Carboniferous (as evidenced by the 'rainforests' of Europe and North America during that time). Forward in time, the Jurassic and Cretaceous were times of high diversity in the mid- and high-latitudes while the equatorial regions were sparsely vegetated and of low diversity. So, it is the combined effect of continental motion through climate zones, as well as the change from an icehouse to a hothouse world, that has caused the diversity pattern shown here. For further details, view a pdf of my publication.

It simply isn't all that informative to bin everything from a given time interval and call it global diversity. A better approach would be to calculate diversity within a given latitudinal range and then compare the same range through time. Another would be to calculate 'maximum diversity' for a given interval by choosing the most diverse latitudinal bin, and then compare through time.


Walter, H. (1985). Vegetation of the Earth and ecological systems of the geo-biosphere (3rd edition): Springer-Verlag, New York, 318pp.

Ziegler, A.M. (1990). Phytogeographic patterns and continental configurations during the Permian Period. In: McKerrow, W.S. & Scotese, C.R. (eds), Palaeozoic Palaeogeography and Biogeography. Geological Society of London Memoir 12: 363-379.

Ziegler, A.M., Hulver, M.L. & Rowley, D.B. (1997). Permian world topography and climate. In: Martini, I.P. (ed.), Late Glacial and Post-Glacial Environmental Changes - Quaternary, Carboniferous-Permian and Proterozoic. Oxford University Press: 111-146.

View PDF files of my related publications: