High latitude temperatures have declined by as much as 15
C since the
early Eocene, about 50 million years ago [ e.g., Kennett and Stott,
1991; Zachos et al., 1993]. There has been little corresponding change
in tropical temperatures. The Eocene climatic optimum is just the
most recent and best observed of warm periods in Earth history
[ Barron, 1987]. The Cretaceous is another example. In fact, warm
poles appear to characterize much of Earth history. The climate we
know best, with ice at high latitudes and a large difference in
temperature between equator and poles, may be anomalous. But we are
far from understanding how the climate system could have maintained
warm poles [ Sloan et al., 1992]. Although it is quite likely that
carbon dioxide concentrations have decreased since the Eocene [ Dean et
al., 1986; Popp et al., 1989; Freeman and Hayes, 1992], the
temperature differences between then and now do not look like a
greenhouse effect. A well-mixed greenhouse gas tends to warm all
latitudes equally, but the Eocene was warmer only at high
latitudes [ Crowley, 1991; Walker and Sloan, 1992].
Therefore, paleoclimatologists have frequently invoked increased
transport of heat from low to high latitudes during the period of
warmth [ Rind and Chandler, 1991]. On this topic I disagree
significantly with many of my colleagues. In my opinion, these
invocations lack any convincing explanation of why the
transport of heat in either ocean or atmosphere would have been large
when the temperature differences that drive circulation were small.
Most hydrodynamic systems circulate faster and transport heat faster
when temperature differences are large. Geophysical fluid dynamics may
be different, but I suggest that we should be slow to accept an
explanation that goes against intuition without a clear physical
description of mechanism.
As an alternative, I suggest that high latitude cloud forcing was much reduced during the Eocene. I make this suggestion because I suspect that the cloud cover of today, in excess of 80%, is not possible over warm oceans and that clouds over warm oceans cause less negative forcing than clouds over cold oceans [ Liou and Ou, 1989]. Cloud microphysical processes related to the higher saturated vapor pressure in warm air might decrease cloud cover or albedo [ Baker and Charlson, 1990; Ackerman et al., 1993, 1994]. Alternatively, cloud cover may decrease over warm water because of thermodynamical and dynamical processes affecting subsidence and the formation of temperature inversions.
The suggestion is consistent with the data. First, cloud cover decreases with decreasing latitude (see Figure 2). This may be an effect of the global circulation, but it may also result from the increase of sea surface temperature with decrease in latitude. Second, satellite data show that cloud optical thickness decreases with increasing temperature [ Tselioudis et al., 1992, 1993]. Third, high altitude clouds are more frequent at low, warm latitudes [ Warren et al., 1986, 1988], and high altitude clouds cause less cooling because their tops are cold. So the Eocene may have been warm at high latitudes because of a strong positive feedback associated with clouds. Clouds over warm oceans may cause less cooling.