SEARCH

SEARCH BY CITATION

Keywords:

  • Carbon cycle;
  • snowball Earth;
  • stable isotopes;
  • Neoproterozoic;
  • methane;
  • glaciation

[1] The Snowball Earth hypothesis explains the development of glaciation at low latitudes in the Neoproterozoic, as well as the associated iron formations and cap carbonates, in terms of a runaway ice-albedo feedback leading to a global glaciation followed by an extreme greenhouse climate. The initiation of a snowball glaciation is linked to a variety of unusual perturbations of the carbon cycle operating over different timescales, as evidenced by unusual patterns in the carbon isotopic composition of marine carbonate. Thus a theory for why multiple glaciations happened at this time, and not in the Phanerozoic nor earlier in the Proterozoic, requires a reexamination of the carbon cycle and the controls on climate stability. We propose that the concentration of continental area in the tropics was a critical boundary condition necessary for the onset of glaciation, both because the existence of substantial continental area at high latitudes may prevent atmospheric carbon dioxide from getting too low and because a tropical concentration of continental area may lead to more efficient burial of organic carbon through increased tropical river discharge. Efficient organic carbon burial sustained over tens of millions of years, required by the high carbon isotopic compositions of preglacial carbonate, may lead to the buildup of enormous quantities of methane, presumably in hydrate reservoirs. We examine how the slow release of this methane may explain the drop in δ13C values immediately before the glaciation. Moreover, the accumulation of methane in the atmosphere coupled with the response of silicate weathering to the additional greenhouse forcing can lead to a climate with methane as the major greenhouse gas. This situation is unstable because methane is not buffered by a large ocean reservoir like carbon dioxide, and so the collapse of the methane source may provide a trigger for the onset of a runaway ice-albedo feedback. A simple model of the carbon cycle is used to explore the boundary conditions that would allow this to occur.