Cambridge Publications |
|
The Conference Company |
April 8th - 12th, 2001 |
|
European Union of Geosciences |
Cambridge Publications |
|
The Conference Company |
April 8th - 12th, 2001 |
|
European Union of Geosciences |
(Tuesday April 10th 2001 at 12:15 in Room GO: Schweitzer Auditorium)
Department of Earth and Planetary Sciences, Harvard University, Cambridge, MA 02138, USA
In the 1960's, simple climate models indicated that if ice caps grew to cover half the Earth's surface area, runaway ice-albedo feedback would drive ice lines to the equator and global mean surface temperature would drop to ~225K. The resulting ice-albedo catastrophe would not be irrevocable, as originally believed, but would last for millions of years, until atmospheric CO2 (built up by volcanic outgassing in the absence of silicate weathering) exceeded 0.1 bar. The resulting greenhouse forcing would raise tropical surface temperature to the melting point, after which deglaciation would proceed violently due to reverse ice-albedo feedback. As the planetary albedo falls, the CO2rich atmosphere would create a transient ultra-greenhouse environment. Surface temperatures ~325K in the tropics would drive a strong hydrologic cycle. Carbonic acid rain would react with emergent shelf carbonates, unaltered syn-glacial volcanics and comminuted glacial debris. The resultant alkalinity flux would cause rapid inorganic carbonate sedimentation in warming surface waters. Atmospheric CO2 would be gradually drawn down by silicate weathering, leading to eventual reestablishment of small ice caps. The global 'freeze-fry' cycle (essentially due to a hysteresis in the path of atmospheric CO2) has the virtue of making many predictions that can be tested geologically.
Meanwhile, geologists had long been puzzled by many features of the late Neoproterozoic (730-580 Ma) sedimentary record. The salient features were observed world wide, and the match between the observations and the climate-model predictions is remarkable. Glacial deposits are widespread on every continent. Robust paleomagnetic data shows that ice lines descended to sea level near the equator, and remained there through several magnetic polarity reversals. Banded iron-formations (with ice-rafted dropstones) appear for the only time in the last 1.8 Ga, consistent with an ice-covered ocean, consequent deep-water anoxia, and lack of riverine sulfate input. Glacial deposits are overlain sharply by 'cap' carbonates with unusual textures and negative delta 13C values, implying rapid carbonate production in the glacial aftermath. Unusually high delta 34S values for post-glacial marine sulfate suggests prolonged attenuation of riverine input to the ocean during glaciation. None of these features was previously explained, but all follow readily from the snowball hypothesis. Various arguments have been raised against the hypothesis, involving sea-level change, thickness of glacial deposits, 87Sr/86Sr records, timing of iron-formation, general circulation models, algal survival, and the fossil record of early metazoa (which appear abundantly only after the last of perhaps four late Neoproterozoic snowball events). Only the last of these arguments has yet to be reconciled with the hypothesis. Joe Kirschvink, who originated the hypothesis, and my colleague Dan Schrag, with whom I have worked closely on every aspect of it, believe that the timing of snowball events in Earth history is ultimately related to unusual distributions of continents and oceans on the globe. The development of the snowball Earth hypothesis is a fine example of Wegener's maxim that "all earth sciences must contribute evidence towards unveiling the state of our planet in earlier times, and that the truth of the matter can only be reached by combining all this evidence"[italics added].
