TERTIARY ICE-SHEET DYNAMICS - THE SNOW GUN HYPOTHESIS

We observe strong negative correlation between Tertiary low- to mid-latitude planktonic foraminiferal delta-18-O and the difference between these data and coeval benthic foraminiferal delta-18-O. Late Quaternary data do not show this correlation. Coupling statistical model/delta-18-O comparisons and evidence for Antarctic ice and ocean temperature variation, we infer that Tertiary ice volume, recorded by tropical planktonic delta-18-O, increased as the deep ocean warmed. Because the isotopic signatures of deepwater temperature variation and ice volume change were of opposite sign, the sum of these signals in Tertiary benthic delta-18-O became lost in the noise. This renders low correlation between Tertiary planktonic and benthic delta-18-O time series compared to late Quaternary data. We contend that Tertiary ice sheet growth was commonly driven by warming of deep water from low-to mid-latitude marginal seas (snow gun hypothesis). In contrast, late Quaternary ice sheets grew as deep water, formed at high latitude, cooled. Because tectonic forcing and orbital forcing at low-latitude primarily controlled production and temperature variations of this Warm Saline Deep Water, these influences largely dictated Tertiary ice volume fluctuations. Through the Tertiary, we infer ice volume fluctuations to be an important component of sea level history on timescales between 10(3) and 10(7) years.