required in order to understand the continental ice volumes that are inferred to have been characteristic of the LGM epoch based on the observed sea level depression. There is still no entirely satisfactory mechanistic model that can correctly predict the observed variation of [CC>2_] as a combined effect of the.solubility pump and the biological pump. The magnitude of the problems that continue to bedevil our understanding of biospheric feedbacks on climate is also clearly evident in the fundamental disagreements that have arisen in connection with the identification of particular continental surfaces as being cither sources or sinks of this trace gas in the modern climatc system (see in particular Chapter 4 by I icimann in this volume).

An important issue in paleoclimatology remains outstanding, providing a cogent reminder of the extent to which the climatc system continues to provide challenges to the scientific imagination. This issue relates to the theory of ice-age occurrence in the more remote past. Recent analysis has focused on the late pre-Cambrian glaciation that occurred approximately 700 million years ago. Paleomagnetic evidence has been widely construed to suggest that this glaciation occurred not at high latitude but rather at extremely low latitude. The debate concerning mechanism has recently come to be polarized between two extreme views. In one, the so-called snowball Earth hypothesis, the entire surface of the Earth is imagined to have become ice-covered (e.g., as most recently discussed by Hoffman et al, 1998), In the other view, the obliquity of the orbit is imagined to have been so large at that time that only the lowest latitudes were susceptible to glacial advance (Williams et aL, 1998; see also Kasting, Chapter 13, in this volume). Neither of these hypotheses has yet been subjected to particularly rigorous test, although the snowball hypothesis is gaining intellectual momentum (e.g., see I lyde et a!., 2000), This most recent work, however, which was based on the neoproteriozoic continental reconstruction of Dalziel (1997), strongly suggests the fully glaciated uhard snowball" scenario of HafTman et al. (1998) to be as unlikely as the high obliquity scenario of Williams et al. is unnecessary. The detailed GCM reconstructions of the climate of that era discussed in Hyde et al. do lead to glaciation of the continental fragments that then existed at low latitude, but this occurs in the presence of a substantial equatorial refugium of open w ater.

0 0

Post a comment