The properties predicted by this simulation are probably incorrect in a number of respects, but this section will not be used to hedge and stipulate because that would be counter to the goal of having a straw man. It is where the model and data meaningfully disagree that the most new information will be learned. I will discuss the most likely weaknesses and strengths of the model and some ways in which the model predictions might be compared productively with the climate proxies and the arguments embodied in the Leg 199 drilling proposal and the "Leg 199 Scientific Prospectus."
The most important weakness of this model and all CGCMs is the tendency to have a split ITCZ, that is, an overly strong Southern Hemisphere ITCZ in the central and eastern Pacific. In the modern day, the Northern Hemisphere preference of the ITCZ leads to a cross-equatorial asymmetry in flow and water mass properties, with a strong NECC but no substantial SECC in the eastern Pacific. The atmospheric GCM used in this study accurately reproduces the ITCZ position and seasonal movement when driven by modern SSTs. It has been proposed that the bias in the coupled simulation is due to some error in the feedback between the ocean currents, upwelling, stratus clouds, and wind fields in the eastern Pacific (more below).
This brings us to an important point, with regard to both the deficiency of CGCMS in modeling the eastern Pacific ITCZ and an underlying assumption in the Leg 199 drilling proposal. The theory, widely used in the Leg 199 drilling proposal and the "Leg 199 Scientific Prospectus," postulates that the Northern Hemisphere preference for the ITCZ is due to the larger land area in the Northern Hemisphere, which drags the "thermal equator" across the geographic equator (espoused originally by Flohn, 1981). The theory that the ITCZ responds to hemisphere-wide land-sea fraction differences is not currently accepted by the climate community and may not be the best conceptual model for paleoclimate investigations either.
In fact, there is no general theory that is completely successful in explaining the location of the ITCZ. Current work on this subject, exemplified by "Why is the ITCZ Mostly North of the Equator" (Philander et al., 1996), takes into account the fact that the Northern Hemisphere preference for the ITCZ is a regional feature (i.e., in the eastern Pacific and Atlantic) and caused by regional features. For example, the western Pacific warm pool and the largely ignored South Pacific Convergence Zone do not have a Northern Hemisphere preference. Thus, the key question is, "What sets the location of the ITCZ in the eastern Pacific?"
Although no full explanation currently exists, the orientation of the coast of South and Central America with respect to the western Pacific warm pool appears to be very important because of the impact of feedback involving near-coastal wind-driven upwelling and downwelling. As summarized in Thuburn and Sutton (2000), the asymmetries thus created are enhanced by and, in turn, enhance the regional warming of the ocean north of the equator and ocean cooling south of the equator. Some role for the Central American summer monsoon has also been suggested (Mitchell and Wallace, 1992). Stratus cloud albedo feedback further enhances the cross-equatorial temperature gradient and associated large-scale circulation patterns. Because of the complexity of the feedback and the fact that we still do not have a full theory for the ITCZ, CGCMs do not reproduce the seasonal cycle of the ITCZ in the eastern Pacific correctly. These factors are just as likely to have controlled the eastern Pacific ITCZ behavior in the past as they do today. Indeed, there is support for such linkages already in the early Neogene record (Tsuchi, 1997). Hopefully, a major contribution of Leg 199 will be in increasing our understanding of these ITCZ processes, but that requires dealing with all of the possible factors involved.
In a related vein, although frequent attempts have been made, there is some difficulty in attributing a special relationship between the location of the ITCZ and a sediment bulge. This difficulty becomes especially apparent when trying to make specific predictions about the proxy record from a quantitative climate model, as I do in Figure F7. Clearly, there are many aspects of productivity and accumulation and their relationship to ocean circulation that are still poorly understood (e.g., Ragueneau et al., 2000; Friedrichs and Hofmann, 2001), and these present practical difficulties but also exciting opportunities for Leg 199 researchers. However, there are specific questions about variations in accumulation rates in the Pacific that may need to be considered. Today there is enhanced sedimentation at the equator in the western Pacific warm pool (Higgins et al., 1999; Broecker et al., 1999), which is very poorly understood, and it raises questions about our understanding, in general, of the mechanisms that might focus (or unfocus) sediment in the equatorial Pacific, even in the absence of an ITCZ. Furthermore, the presence of enhanced accumulation rates at the equator in the western and central Pacific during much of the Cretaceous (Ogg et al., 1992) (another extreme "greenhouse" climate interval) is difficult to reconcile with the conceptual picture laid out in the Leg 199 drilling proposal and the "Leg 199 Scientific Prospectus." It should also be remembered that the warm pool is also a region of productivity and upwelling (although the upwelling arrives at a deeper level [Helber and Weisberg, 2001]).
In general, the model results do not predict dramatic changes in the major features of the tropical ocean or atmosphere circulation in equilibrium with realistic early Paleogene conditions. It should also be acknowledged that the winter-season extratropical continental temperatures are colder than terrestrial proxies, so the simulation does not reproduce "equable climates." Other model results indicate ~2000 ppm pCO2 would be required to produce equable climates, although mean tropical SSTs would then be at or above 34°C (C.J. Shellito, L.C. Sloan, and M. Huber, unpubl. data).
The fact that simulated equator-to-pole and surface-to-deep temperature gradients are substantially less than modern gradients has little impact on the results shown here because the tropical thermocline, ITCZ precipitation, and tropical wind fields are all controlled by the tropical-subtropical temperature gradient (Huber and Sloan, 2000). The tropical-subtropical meridional temperature gradient produced by the simulation is very close to modern (e.g., Huber and Sloan, 2001). There are, however, many important regional features of the eastern Pacific that would have been difficult to predict without the modeling framework used in this study. According to the model, the mean location of the ITCZ should be in about the same place as today, as should the major upwelling regions. The seasonal cycles of temperature, precipitation, and current variations are similar to those simulated for today but with potentially important second-order differences, including somewhat broader seasonal swings of these characteristics (~3° of latitude) and a stronger Walker cell. Thus, on a seasonal basis, the region of upwelling (and productivity?) is broader than in modern cases. If major differences are found between the model results presented here and the results of Legs 198 and 199, a unique opportunity might present itself for understanding the failure of the model with regards to both key climatic features.
It bears mentioning that the response of tropical circulations in CGCMs to increased greenhouse gas concentrations (in modern-day simulations) is still very sensitive to details of model parameterizations. There is currently a significant debate as to whether or not increased greenhouse-gas concentrations should lead to a mean state in the tropics that bears more of a resemblance to La Niņa or ENSO conditions or whether there should be no change in the mean state at all (Cane et al., 1997; Timmerman et al., 1999; Fedorov and Philander, 2000; Jin et al., 2001). Although the simulation presented here produces a tropical system that is similar to the modern system, it is slightly skewed toward a permanent La Niņa state. If the "correct" state is more skewed toward a permanent ENSO state, as has been suggested by Sun (2000), Liu and Huang (1997), and Fedorov and Philander (2000), the difference may be noticeable in the Leg 199 results. The notable difference, according to Sun (2000), Liu and Huang (1997), and Fedorov and Philander (2000), is that the eastern equatorial upwelling zone should not exist in past warm climates. Instead, it should be replaced by a broad, warm, and well-stratified region more similar to the warm pool of today.
If data from Leg 198 and 199 can either confirm or rule out the presence of a warm-pool cold-tongue (downwelling-upwelling) dichotomy pattern in the Pacific, the impact on controversies in modern and future climate change theories will be substantial. Romero et al. (2001) present clear evidence from the early 1990s that ENSO largely controls the temporal and spatial variability in export production in parts of the Pacific, with up to 75% decreases in production during ENSO events and a tendency for diatoms to be much more sensitive to such fluctuations than radiolarians. On the face of things, it appears that the pattern of extreme radiolarian productivity in the East Pacific (found during Leg 199) is in keeping with the model results presented here, contrary to prediction that there should be no thermocline tilt and much reduced upwelling in past warm climates.