From !khure

Project B6: Paleomagnetic study of South African Paleozoic and Precambrian formations : ancient ice ages and geodynamics

French pi: J. Besse
South African pi: M. de Wit (with R. Domoney, UWC)


Project Participants

  • France: J. Besse, F. Fluteau, collaboration, Y. Donnadieu (LSCE)
  • South Africa: M de Wit, R. Domoney


The climatic history of the Earth is marked by the alternation of hot periods and glacial eras (Fig1). The causes of climatic variability on long timescales are numerous: paleogeographic changes, evolution of atmospheric chemistry, evolution of the solar constant, etc.... Understanding the causes of these climatic changes is thus an essential stake in the current context of global warming and sustainable development.

Fig 1 Main glacial periods (from Hoffman and Shrag)

The purpose of this project is to study the glaciations of Precambrian and EarlyPalaeozoic. South Africa is a good target since most of main glaciations (PermoCarboniferous, Ordovician, Neoproterozoic, 2.2Ga and even a recently discovered 3.3 Ga, de Wit, personal communication) have been recorded. We plan first to focus on the glacial episode at the end of Ordovician, which happens during a greenhouse period.

The Ordovician glaciation resulted in the formation of an ice cap over a broad part of Gondwana. Contrary to the other glaciations, which lasted several tens of million years, this one may not have exceeded a few million years, even less. The causes of this glaciation are far from being elucidated. In addition to its volume and its duration, this glaciation occurs during a period known for its high atmospheric CO2 content. This period is also marked by deep upheavals of the carbon cycle which are marked by d13C anomalies, observed on Baltica and Laurentia). The acquisition of new paleomagnetic poles specifying the drift of this continent during this period is essential. For that, we propose to sample the Cape fold belt glaciogenic neighboring sediments in the Pakhuis and Cedarberg formations.

These data will constitute an essential basis to tackle the numerical modelling of the Late Ordovician glaciation and its validation. This project will also rely upon several numerical models: a coupled ocean-atmosphere GCM: FOAM, a geochemical model: COMBINE and a model of ice-cap: GRIZZLY.

The objectives are to understand the influence of the paleogeographic changes on the climate, the consequences on the carbon cycle and the pCO2, as well as the consequences on the formation of the Gondwanian ice-cap. Finally we propose to compare the glaciations (Precambrian, Late Ordovician, Permo-Carboniferous) occurring within distinct paleogeographic and environmental frameworks in the context of the interactions climate/geodynamic/carbon cycle.


Figure 2: Polar wander path of Gondwana during the Early Paleozoic.

Palaeogeography is a key element to better understand the climatic evolution during the Early Paleozoic. The Neoproterozoic glacial events occurred in a paleogeography dominated by the amalgamation and disintegration of a supercontinent, Rodinia, sedimentary glacial deposits being observed at low paleolatitudes (Evans 2003), and in a environmental context in which the biologic activity was restricted with respect to Palaeozoic. The Permo-Carboniferous glacial event occurred during a ice-house period, also related to a supercontinent amalgamation, and followed the development of an abundant continental vegetation. The late Ordovician glacial event probably occurred in a fragmented paleogeography, where continental mass dispersed, in a general greenhouse age. The figure 2 shows the South pole position for Gondwana during the Palaeozoic (McElhinny et al., 2003). Upper Cambrian and lower Ordovician poles are relatively well constrained in Western Africa, whereas the rapid Northward drift of Gondwana (more than 6000km in less than 100Ma) is only constrained by two poorly determined paleopoles between 455 and 405 Ma. Whether this rapid drift represents a continuous, plate-related movement, or a local high shift linked to a true polar wander remains unknown. A better knowledge of the location of Gondwana all along the Early Paleozoic is thus required to improve our understanding of the climate changes during the Ordovician. The uncertainties on pole position preclude any reliable modeling of the climatic evolution, as the surface of continent close to the pole is by itself a critical parameter for the inlandsis location. Moreover, the general plate dynamic may also influence the glacial mechanisms. At last, critical parameters to be modelled are the oceanic temperature gradients, the calculation of which requires a precise knowledge of the paleolatitude of the sampling sites.

In the cape Fold belt, Pakhuis and Cedarberg formation describe well the Ordovician glaciations, with the presence of glaciogenic sediments and fossils. Several spots may allow parallel sampling of this important section, already partially sampled by Bachtadse and colleagues (1987), but with an unsufficient number of sites.

Climate Modelling

Numerical tools are required to investigate the influences of forcing factors. As an example, different causes have been advanced to explain the Hirnantian glaciation: a drawdown of CO2 induced by an enhanced weathering of silicate rocks due to the Taconic orogeny (Kump et al., 1999) or a major change in ocean circulation induced by paleogeographic changes (Hermann et al., 2004). Climate modelling has already been used to simulate the Ordovician ice age (Crowley and Baum, 1991; Crowley and Baum, 1995; Gibbs et al., 1997, 2000). Poussart et al. (1999) used a coupled atmosphere-ocean-sea ice model to simulate the climate during the Latest stage of Ordovician. Most of the experiments have focused on the pCO2 permitting the inception of an ice cap over Gondwana. Accounting for a lower solar constant at 430 Ma, a 8 to12 x pre-industrial atmospheric level is needed to simulate perennial snow over Gondwana. Despite the fact that the previous studies favoured a better knowledge of the Late Ordovician glaciation, the evolution of the Early Paleozoic climate remains largely unkown.

We propose to simulate climate, using a fully coupled ocean-atmosphere GCM FOAM1.5 (Fast Ocean Atmosphere Model). The ocean and atmospheric models are linked by a coupler, which implements the land and sea ice models and calculates and interpolates the fluxes of heat and momentum between the atmosphere and ocean models (Jacob, 1997). FOAM successfully simulates many aspects of the present-day and past climate (Donnadieu, 2005; Poulsen et al., 2001). The FOAM GCM will run on a parallel supercomputer at the Institut de Physique du Globe de Paris. We also propose to determine the atmospheric pCO2 using a geochemical COMBINE model (Goddéris and Joachimski, 2004). This is an ocean-atmosphere box-model including the mathematical description of the global biogeochemical cycles of carbon, phosphorus, alkalinity and oxygen. In order to calculate the silicate weathering, the COMBINE model will be forced with simulated climatic variables in a suite of FOAM GCM experiments performed at different pCO2. Fixing the CO2 degassing to a given constant value, COMBINE is run until a steady-state PCO2 is reached. This method has been employed by Donnadieu et al. (2003) to simulate the inception of Sturtian ice age during the Neoproterozoic. An ice-sheet model will be used to estimate the volume of the ice-cap. This model includes the dynamic and thermodynamic of land ice. The FOAM GCM climatic variables will be used to force the ice-sheet model.


  • Bachtadse, V., R. Van der Voo, and I. Haelbich (1987), Paleozoic paleomagnetism of the western Cape Fold Belt, South Africa, and its bearing on the Paleozoic apparent polar wander path for Gondwana, Earth and Planetary Science Letters, 84, 487-499.
  • Crowley, T.J., SK. Baum, KY. Kim 1991. General circulation model sensitivity studies experiments with pole-centered supercontinents. Journ. Geophys. Res. 96, 597-610
  • Crowley, T.J. and SK. Baum. 1995. Towards reconciling Late Ordovician (440 Ma) glaciation with very high CO2 levels. Journ. Geophys. Res. 100, 1093-1101.
  • Donnadieu, Y., Fluteau, F., Ramstein, G., Ritz, C. et Besse, J., 2003. Is there a conflict between the Neoproterozoic glacial deposits and the snowball earth interpretation : an improved understanding with numerical modeling. Earth Planet. Sci. Let., 208, 101-112.
  • Donnadieu, Y., Goddéris, Y., Ramstein, G., Nédelec, A. and Meert, J.G., 2004a. Snowball Earth triggered by continental break-up through changes in runoff. Nature, 428: 303-306.
  • Evans, D. A. D. (2003). "A fundamental Precambrian-Phanerozoic shift in earth's glacial style?" Tectonophysics 375: 353-385.
  • Gibbs, M.T. et al. 1997. An Atmospheric pCO2 threshold for glaciation in the Late Ordovician. Geology. 25, 447-450.
  • Gibbs, MT. et al. 2000. Glaciation in the Early Paleozoic “greenhouse”: the roles of paleogeographies and atmospheric pCO2. in : Huber; B.T. et al. (eds). Warm Climates in Earth history. Cambridge. Univ. Press. 386-422.
  • Goddéris, Y. and Joachimski, M.M., 2004. Global change in the late Devonian: modelling the Frasnian-Famennian short-term carbon isotope isotope excursions. Palaeogeography, Palaeoclimatology, Palaeoecology, 202: 309-329.
  • Herrmann, A.D., et al. (2004) Response of a Late Ordovician paleoceanography to changes in sea level, continental drift, and atmospheric pCO2: potential causes for long-term cooling and glaciation. Palaeogeography, Palaeoclimatology, Palaeoecology 210, 387-401.
  • Hoffman, P., and Shraag, D.P, (2002) The snowball earth Hypothesis: testing the limits of the global change, Terra Nova review article, 51p.
  • Jacob, R., 1997. Low frequency variability in a simulated atmosphere ocean system, Univ. Wisconsin-Madison, Madison
  • McElhinny, M.W., Powell, C.M., et Pisarevsky, S.A., 2003. Paleozoic terranes of eastern Australia and the drift history of Gondwana. Tectonophys., 362, 41-65.
  • Poulsen, C.J., Pierrehumbert, R.T. and Jacob, R.L., 2001. Impact of ocean dynamics on the simulation of the Neoproterozoic "snowball Earth". Geophysical research letters, 28: 1575-1578.
  • Poussart, C.J. et al. 1999. Late Ordovician glaciation under high atmospheric CO2: a coupled model analysis. Paleoceanography 14, 542-558.