Page 109 - Earth's Climate Past and Future

CHAPTER 5 • Greenhouse Climate 85

This concept has been proposed in different forms by several degrees centigrade than even the maximum target
several climate scientists and can be called the ocean signal values shown in Figure 5-4.
heat transport hypothesis. After the plankton formed their shells in warm sur-
A related idea was that deep water might have face waters, they died and fell to the seafloor where
formed in the northern subtropics in regions of very water temperatures were 10º–15ºC cooler than those at
high ocean salinity (>37%), rather than in polar regions the surface. As the shells lay on the seafloor or in the
like today. The concept was that high salinities could uppermost sediments before being completely buried,
have made surface waters dense enough to sink into the they were bathed by cooler bottom waters. Some of
deep ocean as warm, saline bottom water. A strong these shells were partly dissolved and recrystallized.
flow of this warm deep water from the tropics to the Scientists found that the more pristine the shells they
poles might then have contributed to the poleward heat examined, the warmer the temperatures their oxygen
flux needed to warm polar regions and resolve the data- isotopic ratios indicated, whereas the shells that were
model mismatch. more heavily altered yielded cooler temperature esti-
More recent experiments with improved ocean mod- mates. This trend indicated that chemical alteration on
els have not supported the ocean heat transport hypoth- the seafloor had reset the temperatures toward those
esis. The models do not show significant increases of typical of the cooler bottom waters. The best-preserved
poleward transport of heat by the ocean in a climate shells indicated that Cretaceous temperatures were
warmer than that today. warmer by as much as 5ºC than those used in the initial
A second kind of data-model mismatch occurred in target signal shown in Figure 5-4.
climate simulations for 100 Myr ago, and this mismatch This explanation appears to resolve much of the
persists in model simulations of the warm climates that data-model mismatch shown in Figure 5-4. The
continued over the next several tens of millions of years. warmer tropical temperatures match the model simula-
Data from warm-adapted vegetation (early palmlike tion with altered geography and higher CO levels rea-
2
trees) and from fossil reptiles suggest that continents at sonably closely. The simulation is consistent with a CO
2
high and middle latitudes had moderate climates and concentration above the one initially chosen (four times
did not freeze in winter. In contrast, all the GCM the natural modern level). Levels of CO higher than
2
experiments described here (those with altered geogra- this amount would also prevent the interiors of mid-lat-
phy, higher CO levels, and increased ocean heat trans- itude continents from freezing in winter.
2
port, alone or in combination) simulated hard freezes in
winter across the interiors of the northern hemisphere IN SUMMARY, higher CO levels appear to be a major
2
continents. cause of the warmer climate 100 Myr ago.
As a further test on the possible role of the ocean, an
experiment was run in which warm waters were specified 5-3 Relevance of Past Greenhouse Climate to
(imposed) in the Arctic Ocean as an initial boundary the Future
condition before running the simulation. One purpose
of this experiment was to find out whether a very warm The results from studies of Cretaceous climate fit into a
Arctic ocean could explain the warm interiors of the larger picture of the effect of CO on Earth’s climate. Cli-
2
northern continents. The simulation showed that even a mate scientists have run a series of GCM sensitivity tests
warm Arctic Ocean failed to keep the interiors of the using Earth’s present geography as common boundary
northern continents warm enough in winter to prevent condition input to all simulations but allowing the level of
freezing because the winter heat losses were too large. atmospheric CO to vary from as low as 100 ppm to as
2
Problems with the Data? A second path of investi- high as 1000 ppm. The preindustrial or natural modern
gation of the data-model mismatch focused on the possi- value of 280 ppm lies in the lower part of this range.
bility that the proxy data used to reconstruct past climate This set of simulations shows that global average
might have been giving an incorrect “target signal” for temperature rises with increasing CO levels, but the
2
comparison to the models. Several studies have con- relationship is not linear (directly proportional). Instead,
verged on evidence that this explanation is promising. Earth’s temperature reacts strongly to CO changes at
2
The tropics have been the main focus of these reassess- the lower end of the range but much less so to changes at
ments. Most of the data used to reconstruct Cretaceous the high end of the range.
temperatures across lower latitudes have come from geo- One reason for the shape of this curve is positive feed-
chemical analyses of ocean plankton using oxygen isotope back effects from snow and sea ice. At low (<200 ppm)
ratios (Appendix 1). The shells of the plankton record CO values, sea ice and snow advance well past their
2
oxygen-isotopic ratios that reveal changes in the tempera- average limits today and cover a relatively large frac-
ture of the ocean water in which they form. It now seems tion of Earth’s high and middle latitudes. These bright
likely that tropical temperatures were actually warmer by surfaces reflect incoming solar radiation back to space,