Research Objectives:

The climatic, biological and geochemical characteristics of the PETM have been the focus of many recent high-profile papers and popular articles. To a large degree this reflects the community's excitement at discovering an extraordinary perturbation in biogeochemical systems that was unimaginable ten years ago (e.g., Dickens, 1999). Suddenly, it now seems possible to examine, at both global and regional scales, the ecological and biogeochemical impacts of massive carbon injection and abrupt greenhouse warming at rates and magnitudes similar to predicted anthropogenic effects. Further, this examination can be framed within long-term, more gradual changes in climate and ecosystems. The PETM probably serves as our best natural experiment in which to evaluate how biogeochemical cycles respond to carbon imbalances in the ocean, atmosphere, and terrestrial systems in the time domain.

However, as with most scientific discoveries, our initial and current understanding of the PETM has revealed a series of fundamental issues that require new and truly interdisciplinary research. Consequently, the goals and objectives of this project are multifold :

1) Determine the dynamics of carbon (methane) release, redistribution and climate impacts.

A variety of evidence points to a massive input of carbon to the ocean or atmosphere at the onset of the PETM. The most satisfactory explanation for this perturbation is an order of magnitude increase in CH₄ fluxes from the seafloor via thermal dissociation of gas hydrate. But this raises some crucial issues regarding currently available data and models. Was the CH₄ oxidized to CO₂ in the atmosphere or in the ocean (via bacteria)? If the latter, as might be expected from studies of modern CH₄ seeps (e.g., Masuzawa et al., 1992; Boetius et al., 2000), why did the atmosphere warm, especially at high latitudes, and is there evidence for an expanded deep sea biosphere (Bralower and Thomas, 2000)? If the former, how does CH₄ pass through the water column and what causes the significant drop in deep sea O₂ (below)?

These questions can be resolved by establishing the timing and magnitude of the initial d¹³C excursion, carbonate dissolution and O₂ deficiency in different carbon reservoirs, all in relation to evidence for warming (Dickens, 2000). In principle, a rapid and massive CH₄ input to any reservoir of the ocean or atmosphere should cause a negative d¹³C excursion in all reservoirs because of mixing. However, this excursion will occur first, most rapidly and most prominently in the reservoir where CH₄ is oxidized. Carbonate dissolution should also be most pronounced in the deep-sea reservoir that is closest to the location of carbon input along the path of thermohaline circulation. A slow release of CH₄ into the ocean should cause relatively similar d¹³C excursions in all reservoirs but significant deep-sea carbonate dissolution and O₂ deficiency. By contrast, a catastrophic release of CH₄ should result in isotopic disequilibrium and a more prominent excursion in the atmosphere and upper ocean. Available records are inconclusive with regards to the location of carbon input and no record shows compelling evidence for requisite warming before carbon input.

We need: a) advanced models to quantify expected differences between CO₂ and CH₄ inputs to the ocean and to the atmosphere. For example, climate modeling might distinguish between CO₂ and CH₄ forcing on planetary radiative balance; b) detailed coupled d¹³C and d¹⁸O records of multiple phases (planktonic and benthic foraminifera; terrestrial and marine organic matter) in expanded records to get time offsets between reservoirs and forcings; c) new proxy records of carbonate dissolution and O₂, particularly along depth transects; d) detailed records of multiple parameters prior to the PETM to document potential forcing mechanisms.

2) Identify and quantify the primary processes, biologic and geochemical, that remove excess carbon after the PETM.

Despite existing constraints on large-scale perturbations of the global carbon cycle, uncertainties remain concerning the primary carbon sources and sinks (permanent and transient) on long and short time scales. The 2000 Gt of CH₄-derived carbon were eventually removed from the ocean-atmosphere system by biogeochemical and inorganic chemical feedbacks. According to existing models, this excess carbon would have been sequestered mostly through inorganic processes: dissolution of silicates on land would release bicarbonate to rivers that would eventually precipitate as marine carbonate. However, according to ecosystem models higher pCO₂ and more humid conditions should have cultivated increased biomass (Beerling, 2000). A large portion of the excess carbon thus may have been sequestered through biotic processes, namely enhanced biomass production and organic carbon burial

Moreover, warmer temperatures, greater precipitation, and higher pCO₂ should have increased dissolved nutrient fluxes to the oceans, particularly near the coast (e.g., Berner, 1994; , thereby fueling production. The dramatic proliferation of heterotrophic dinoflagellates (i.e., the Apectodinium acme) (Crouch et al., 2001) as well as the occurrence of 'black shales' across the PETM in shallow marine sediments may reflect this process. On grounds of a 100 fold increase in Ba accumulation rates during the PETM at several deep-sea locations, Bains et al. (2000) have further argued that runoff derived nutrients stimulated pelagic productivity as well. However, the ocean residence time of Ba (~10 kyr) is too short to explain a prolonged Ba spike at multiple locations unless continental weathering rates increased by orders of magnitude. Available nannofossil and planktic foraminiferal assemblage data from open ocean sites suggest a marked decline in productivity during the PETM . Given these somewhat contradictory observations, should we consider other mechanisms to regulate or redistribute nutrients? For example, did a reduction in latitudinal thermal gradients reduce wind stress and vertical mixing in the open ocean? Did a drop in deep ocean dissolved O₂ (see discussion below) alter rates of nutrient burial in marine sediments

The shape and timing of the d¹³C excursion after carbon injection might be used to discriminate between these scenarios because the isotopic composition of the two potential outputs differ significantly (e.g., Kump and Arthur, 1999; Dickens, 2000). However, lower and variable inputs from gas hydrates after injection (i.e., capacitor recharge) will significantly modify the d¹³C response (Dickens, 2001). Moreover, there are few (if any) records appropriate for such analysis (e.g., Norris and Röhl, 1999 presented a high-resolution bulk carbonate d¹³C record in the time domain but the curve shape differs from the low resolution foraminifera d¹³C record). The increased kaolinite deposition immediately after carbon input may signify enhanced weathering (REF?), but the response seems too fast.

We require: a) carbon cycle models where organic carbon fluxes and seafloor gas fluxes are fully integrated. For example, Beerling (2000) modeled the biomass response during the PETM with a single, constant gas input while Dickens (2000) modeled the CH₄ input (Fig. X) with a constant biomass; b) high-resolution d¹³C records in the time domain from greatly expanded sections; c) constrained intensities and patterns of continental silicate weathering (e.g., kaolinite fluxes), d) quantified fluxes of organic and carbonate carbon to the rock cycle, e) models that integrate atmospheric conditions, ocean circulation and nutrient cycling, f) quantified accumulation rates of key nutrients (e.g., P, N, trace metals), and g) detailed assemblage data for benthic foraminifera, dinoflagellates, nannoplankton and planktic foraminifera.

3) Determine the cause and consequences of deep ocean O₂ deficiency.

Benthic foraminifera assemblages across the PETM suggest that O₂ concentrations dropped significantly in the deep ocean as well as along many continental margins . This interpretation is consistent with the presence of laminated sediments at the PETM in some sediment cores (e.g., Bralower et al., 1997). But what is the source of this O₂ deficiency? Is it caused by a change in ocean circulation, particularly a rise in deep ocean temperature? Alternatively, is it due to oxidation of substantial amounts of CH₄ in the water column? If so, is the apparent decrease in O₂ consistent with the amount of CH₄ released as inferred from carbon isotopes? In any case, how does this sudden O₂ deficiency affect cycles of redox sensitive elements, for example Mn? Recent papers (e.g., Hesselbro et al., 2000; Jahren et al., 2001) have suggested that certain ocean anoxic events (OAEs) in the Mesozoic are related to rapid and massive input of CH₄ to the ocean and atmosphere. Interestingly, agglutinated benthic foraminifera show very similar ecological community structures after OAEs and the PETM (Kaminski et al., 1999). Are OAEs similar to the PETM except that O₂ became entirely depleted? The issue of O₂ variations during this period is central to understanding aspects of key biotic and biogeochemical responses.

We need: a) ocean circulation/chemistry models that include O₂ but with potential removal through CH₄ oxidation; b) semi-quantified records of dissolved O₂.

4) Establish how marine and terrestrial biota responded to, and contributed to, the climatic and biogeochemical changes in their environments at the PETM.

The two most pronounced biotic effects at the PETM are the extinction in deep-sea benthic foraminifera and the first appearances of many mammal orders on Holarctic continents . What caused the extinction in benthic protists? Was it a response to food supply, increased temperature, decreased dissolved O₂, or some other currently unrecognized factor? Whatever the cause, why did multicellular eukaryotic benthos (e.g., ostracodes) survive the event with minor and transient faunal changes? Among other marine organisms, how does the biotic response in groups such as planktonic foraminifera, calcareous nannoplankton, and dinoflagellates differ between open ocean and near shore regions? Are these responses primarily evolutionary, ecological, or biogeographic? How are they related to bottom-up forcing (e.g., changes in nutrient delivery and productivity) vs. top-down forcing (e.g., the introduction of new consumers and predators)?

The continents present a different set of questions. There were major changes in mammalian faunal diversity, composition, body size, and trophic adaptations across the PETM, but changes in floral diversity and composition were muted (Clyde & Gingerich 1998, Wing & Harrington 2001). Why does response differ so dramatically between plants and animals? Although pre- and post-PETM floras in mid-latitudes are similar, at present the PETM itself has not been sampled. Were there major changes in floral composition and diversity during the PETM itself? If not, why didn’t the flora respond to this rapid warming as we might expect from analogy with Quaternary floras (Wing & Harrington 2001)? Were floras in different climatic zones and latitudinal bands differentially affected by the PETM? Are the first appearances of plants and animals on Holarctic continents controlled by evolutionary constraints on origination or by biogeographic barriers? If the later is the case, is one continent the source of these plants and animals or is the exchange multidirectional?

We need a) detailed analyses of changes in faunal and floral composition in different environments across the PETM, b) quantitative analysis of evolutionary and ecological shifts in flora and fauna, c) a means of correlating between deep ocean, shallow water, and continental systems, d) a method for monitoring changes in primary production on land (if possible), and e) quantitative data and analysis of biogeographic patterns, and f) a modeling approach to explore the possible climatic responses (e.g., temperature vs. moisture thresholds) that might have differentially influenced floral and faunal communities during and after the PETM

5) Establish whether the PETM is a unique event in the Paleogene The PETM occurred within a gradual, long-term warming trend that culminated in a period of global warmth (the EECO) (Figure 1).

While warming at the EECO was not as rapid, generally warm conditions persisted for much longer (~1 million years). These contrasts in magnitude and duration raise many interesting questions. For example, are the PETM and the EECO reflecting the same phenomenon (i.e., discharge from a gas hydrate capacitor, FIG. X), or do they differ fundamentally in origin? The evidence for CH₄ release at the PETM is difficult to refute, but could CH₄ play a role in the long-term warming and carbon isotope trends as well? Were there, for instance, several repeated or protracted CH₄ release events during the EECO? Do the spectrum of short-term (10³-10⁴ yr) biotic responses and biogeochemical feedbacks at the PETM influence trends over long time (10⁶ yr) scales as well? Is the mode of biotic response different when forced by a sudden, short event (i.e., the PETM) vs. a more gradual, longer event (i.e., the EECO)? This certainly seems to be true for plant communities in North America, which change much more conspicuously at the EECO than at the PETM (Wing 1998, Wing et al. 2000). Is this because of the longer duration of the event? Are there modes of biotic response that have long recurrence intervals, prohibiting response at similar, closely spaced events like the PETM and EECO? Such a mechanism might explain why benthic marine protists suffered greatly at the PETM but were largely unaffected by the EECO.

To address the questions above, which are largely of a comparative nature, we need to collect data similar to that described for points 1-5 for the EECO. Because of its long duration, and the high temporal resolution required to answer some of these questions, using current low resolution records, we plan to target specific windows in the EECO for high-resolution analysis, rather than the entire event.