Climate Research Group

Permiate Climate Simulated with a Comprehensive Climate Model (CCSM3)



Key Research Findings


Papers Published



The Permian-Triassic Boundary (PTB, ~251 Ma; Bowring et al., 1998; global paleogeography is shown in Fig. 1) represents the single largest Phanerozoic mass extinction event. About 90% of all marine and about 70% of the terrestrial species disappeared (e.g. Raup, 1979; Raup and Sepkoski, 1982; Erwin, 1993, 1994; Retallack, 1995; Benton, 2003). The mass extinction was followed by a delay of about 6 Myr in recovery of marine biotas in many regions (e.g. Bottjer, 2004; Twitchett et al., 2004; Payne et al., 2006).

Figure 1: Global paleogeography for ~251 Ma (base map from Ron Blakey, Locations of study sections for geochemical proxy datasets are shown.

The causes and dynamics of the PTB mass extinction remain uncertain (for reviews see e.g. Ward, 2000; Erwin et al., 2002; Berner, 2002; Benton, 2003). Gradual deterioration of marine and terrestrial environments during the Late Permian, also reflected in changes in seawater chemistry and atmospheric composition, and persistence of inhospitable conditions through the Early Triassic suggest that long-term factors such as changing climate were important (e.g. Hardie, 1996; Ekart et al., 1999; Berner et al., 2000; Lowenstein et al., 2001, 2005; Retallack, 2001; Horita et al., 2002; Korte et al., 2003; Dickson, 2004; Tabor et al., 2004; Berner, 2005). An extinction rate peak, however, as well as abrupt lithofacies changes, and geochemical anomalies associated with the end-Permian event horizon are evidence of a catastrophic event. Massive volcanic eruptions occurred in Siberia, starting at 251.8 Ma (e.g. Campbell et al., 1992; Basu et al., 1995; Renne et al., 1995; Reichow et al., 2002; Kamo et al., 2003). Discharges of volcanic dust and aerosols into the atmosphere may have triggered a series of events, such as intense short-term climatic cooling, further leading to sea-level fall and catastrophic methane release into the ocean-atmosphere system through clathrate destabilization. The emission of volcanic CO2, methane oxidation and fallout of sulfate aerosols may have caused longer-term warming as well as groundwater and ocean acidification, hence affecting ecosystems and causing intense weathering (e.g. Eshet et al., 1995; Krull et al., 2000; Looy et al., 2001; Michaelsen, 2002; de Wit et al., 2002; Steiner et al., 2003; Visscher et al., 2004; Foster and Afonin, 2005; Gastaldo et al., 2005; Sephton et al., 2005; Watson et al., 2005; Sheldon, 2006). A contemporaneous bolide impact has also been invoked as a possible or contributing cause for the PTB mass extinction (e.g. Retallack et al., 1998; Basu et al., 2003; Becker et al., 2004), but evidence for such an event is controversial (e.g. Farley and Mukhopadhyay, 2001; Isozaki, 2001; Müller et al., 2005). Marine sections covering the PTB expose thick black shales and chert successions in deep-marine environments (e.g. Wignall and Twitchett, 1996; Isozaki, 1997; Zhang et al., 2001), suggesting protracted oceanic anoxic events for this time. Buildup of H2S in the deep ocean (Malkowski et al., 1989; Hoffman et al., 1991) and shallowing of the oceanic chemocline (Kump et al., 2005; Meyer et al., 2008) may have established conditions favorable for upwelling of sulfidic, 13C-depleted water masses into intermediate- or shallow-depth environments. Possible evidence for such large-scale oceanic overturn events comes from 32S-enriched pyrite layers, carbonate-associated sulfate, and Ce anomalies (e.g. Kajiwara et al., 1994; Nielsen and Shen, 2004; Wignall et al., 2005). Changes in seawater carbonate saturation (e.g. Heydari et al., 2001) might also be linked to these upwelling events. Marked carbon isotope shifts in δ13Ccarb and δ13Corg of -3 to -8‰ have been identified in marine and terrestrial sections (see overview in Corsetti et al., 2005, and references above). The origin of these excursions remains controversial. Possible mechanisms proposed to account for the shifts include volcanic CO2 emissions, marine productivity collapse, oxidation of terrestrial biomass and soil organic matter, methane release, and changes in the organic C burial fraction (e.g. Renne et al., 1995; Broecker and Peacock, 1999; Krull et al., 2000; de Wit et al., 2002; Payne et al., 2004).

Previous paleoceanographic modeling studies indicated that sluggish ocean circulation at the PTB may have resulted from changes in deepwater formation patterns (Hotinski et al., 2001; Kiehl and Shields, 2005), or that changes in zooplankton biomass, the ratio of siliceaous to calcareous shell production, or volcanic dust flux resulted in deeper particulate organic penetration depths (Winguth and Maier-Reimer, 2005). All of these processes have been linked to the generation of anoxic conditions in the deep sea.

To investigate the effects of various potential forcings on Permo-Triassic ocean chemistry, paleoclimate modeling is being carried out using the fully coupled climate model CCSM-3 (with T31 resolution in the atmosphere, and gx3v5 resolution in the ocean), including the ocean-carbon cycle model OCMIP (Doney et al., 2006) . Comprehensive geochemical proxy datasets for a total of 19 sections in 8 study areas, both from the former Panthalassic as well as the Tethys Ocean (Fig. 1), are being generated by the collaborating partners in this project. This high-resolution database will be used to validate and interpret paleoclimatic modeling results. Top


The focus of the proposed study is the role of oceanographic factors in the marine PTB crisis. The main questions to be addressed in this project are:

  1. How widespread and intense were Permo-Triassic deep-ocean anoxia?
  2. What may have been the patterns of upwelling of toxic deep-ocean waters onto shallow-marine shelves and platforms?
  3. How important were upward excursions of the oceanic chemocline?
  4. What was the relationship of anoxia distribution, upwelling events, and chemocline excursions to contemporaneous changes in seawater carbonate saturation and global negative C-isotope shifts?

In order to address these questions, a Permian-Triassic Boundary (PTB) reference simulation as well as a series of paleo-sensitivity experiments are being carried out using the comprehensive climate model CCSM-3 (with T31 resolution in the atmosphere, and gx3v5 resolution in the ocean) coupled on-line with the ocean-carbon cycle model OCMIP (Doney et al., 2006). The PTB simulations are be based on a previous CCSM experiment by Kiehl and Shields (2005). The strategy involves the integration of the observational database with numerical model results. Top

Key Research Findings

PTB Reference Simulation

The fully coupled run CCSM3 PTB run from Kiehl and Shields (2005) assumes a flat ocean bottom at 4000 m, with a smoothed transition from the deep ocean to the shelves, and an atmospheric CO2 level of 12x the pre-industrial value. After including the OCMIP carbon cycle model, 2,000 model years have been added to the original 2,700 years within this project. As shown by Kiehl and Shields (2005), the simulated zonal mean meridional overturning circulation is symmetrical about the equator for the southern and northern hemisphere circulation, with cells in the upper 500 m of a little over 40 Sv, and in deep water about 10 Sv (Fig. 1). The modeled temperatures (Fig. 2) are in good agreement with Late Permian climate proxies (Kiehl and Shields, 2005), with the mean surface temperatures ~8°C warmer than today, an increased ocean heat transport to the poles, no permanent land ice, and a reduced meridional temperature gradient. A warm pool of water is modeled in the eastern tropical Tethys and in western Panthalassa; a cold tongue exists in the eastern tropical Panthalassa, caused by upwelling in this region, similar to the modern Pacific. Upwelling zones are also located on the subtropical west coast of Gondwana and in mid-latitudes in the western Panthalassa (Fig. 3). Old water masses have been identified in the Tethys and mid-Panthalassic Ocean (not shown here). From these findings, Kiehl and Shields (2005) inferred that anoxic conditions were likely, due to low ocean mixing and deepwater formation mostly in southern Panthalassa.

However, when adding the carbon cycle model to the existing 12xCO2 simulation, it becomes evident that the deep sea remains well oxygenated throughout the Panthalassic as well as the Tethys Ocean (Fig. 4). An oxygen minimum zone develops in intermediate waters, similar to today.

Fig. 2: Simulated global zonal mean meridional overturning circulation for the PTB (left) and, for comparison, for the modern (right)

Fig. 3: Simulated SSTs (in °C) and circulation (in Sv) for the PTB reference run.

Fig. 4: Simulated dissolved oxygen (in μmol L-1) for the PTB reference run

PTB Sensitivity Experiments

1. Sensitivity to nutrient inventories and fluxes:

When increasing the nutrient supply from 1 to 2, 4, and 10 times the original value, the export production of particulate organic carbon increases accordingly (Fig. 5). In general, high productivity is modeled in the upwelling areas of the eastern Panthalassic Ocean as well as in mid-latitude regions in the western Panthalassic coastal areas. For the 10xPO4 experiment, higher productivity areas become also more apparent in most of the Tethys Ocean. This is consistent with the concept that basins can act as nutrient traps (e.g. Meyer & Kump, 2008).

Fig. 5: Particulate organic carbon (POC) export production (in mol C m-2 yr-1), shown here for the PTB reference run (A) and 10xPO4 (B).

Figure 6 illustrates the effect of the increased productivity due to phosphate supply on dissolved oxygen in the deep ocean, in comparison with the reference run. Higher productivity is well correlated with lower oxygen values in the deep sea, because of higher oxygen demand during the decay process. For the 2xPO4 experiment, a low oxygen zone starts building up in the eastern Panthalassa, the area of highest productivity. This low oxygen zone expands considerably for the 4xPO4 case, and in the 10xPO4 scenario, most of the deep Tethys also turns anoxic. For comparison, the approximate deep sea anoxia locations from Isozaki et al. (1997) are marked. The British Columbia location in the eastern Panthalassa lies within the low-oxygen area in all the enhanced nutrient runs, whereas for the western Panthalassa section, the oxygen levels remain above anoxic for all experiments. Even in the 10xPO4 case, the simulated value at this location is still 60-80 μmol/L-1.

Fig. 6: Dissolved oxygen (in μmol/L-1) at ~3300 m depth for the PTB reference run (A) and 10xPO4 (B). The approximate deep sea anoxia locations from Isozaki et al. (1997) are marked.

2. Sensitivity to increased dust fluxes:

Increased dust supply, parameterized by 10xFe, results in an increase in export production mainly at high latitudes. Accordingly, oxygen levels decrease throughout the ocean at these latitudes, but anoxic conditions are not achieved anywhere.

3. Sensitivity to particle flux parameterizations:

With the enhanced pump parameterization, the oxygen minimum zone extends further downward, compared to the reference run, as expected, to about 1800 m. A deep sea low oxygen area is modeled in the eastern Panthalassa, in extent comparable to the 2xPO4 simulation.

4. Summary of results to date:

The range of the modeled oxygen values is comparable to several other ocean modeling studies, e.g. the haline case by Zhang et al. (2001). As in these previous studies, no sustainable anoxia throughout the deep sea are being simulated.

The results achieved in this project in contrast to the study by Meyer et al. (2008), using the GENIE-1 earth system model, with an ocean resolution of about a third of the one used in this study, and also using the end-Permian boundary conditions from Kiehl & Shields (2005). In the Meyer et al. (2008) study, both the deep ocean as well as the surface ocean become euxinic with a 10x increased phosphate input.

The response of ocean circulation to CO2-induced warming alone can not explain the widespread presence of anoxia; probably a combination of effects is required in order to generate basin-wide anoxia.

5. Future sensitivity experiments:

In addition to the sensitivity experiment exploring the effect of a mid-oceanic ridge system on ocean circulation and deep sea anoxia, which is currently being set up, we plan to carry out a simulation combining the enhanced particle flux parameterization with increased nutrient supply (10 x PO4).

We will also test the oceanic and sedimentary response to high-latitude freshwater input scenarios and their effect on the thermohaline circulation and carbon cycling. Freshwater pulses could have originated from high-precipitation events and lakes along the west coast of Pangea.

Another set of experiments will predict the oceanic and sedimentary response to various orbital parameter settings. In a shallow-marine platform PTB section from Nhi Tao, Vietnam, Algeo et al. (2007) identified eight peaks in pyrite, associated with lower pyrite d34S values as well as with the onset of negative carbonate d13C excursions. These data might be consistent with multiple upwellings of sulfidic, 34S- and 13C-depleted deep-ocean waters into shallow waters at ~20-kyr intervals, suggesting ocean overturning controlled by orbital parameter-driven climate cyclicity. Top


Modeling strategies and results have been presented at the following meeting:

Winguth, C., and A. Winguth, 2009. Ocean Circulation at the Permian-Triassic Boundary: Modeling Anoxia and Upwelling Patterns with CCSM3. AGU Fall Meeting in San Francisco, December 14-18, 2009.

In addition, A. Winguth convened the session “Comparison of Projected Future Climate Change to Warm Intervals in Earth History” (J07) at the MOCA 2009 in Montreal, July 19-29, 2009. Top

Papers published

  1. Winguth, A., and C. Winguth, in press. Precession-driven monsoon variability at the Permian-Triassic Boundary – Implications for anoxia and mass extinctions. Global and Planetary Change, (published online Jun. 26, 2012).
    doi: 10.1016/j.gloplacha.2012.06.006

  2. Winguth, C., and A.M.E. Winguth, 2012. Simulating Permian-Triassic oceanic anoxia distribution: Implications for species extinction and recovery. Geology, 40, 127–130.
    doi: 10.1130/G32453.1

  3. Osen, A., A. Winguth, C. Winguth, and C. Scotese, 2012. Sensitivity of Late Permian climate to topographic changes and implications for mass extinctions. Global and Planetary Change, (published online Feb. 26, 2012).
    doi: 10.1016/j.gloplacha.2012.01.011

  4. Winguth, A.M.E., and E. Maier-Reimer, 2005. Changes of marine productivity associated with the Permian-Triassic boundary mass extinction: A re-evaluation with ocean general circulation models. Marine Geology, 217, 283-304.
    doi: 10.1016/j.margeo.2005.02.011

  5. Winguth, A.M.E., C. Heinze, J. Kutzbach, E. Maier-Reimer, U. Mikolajewicz, D. Rowley, A. Rees, and A.M. Ziegler, 2002. Simulated Ocean Circulation of the Middle Permian. Paleoceanography, 17 (5), 1057.
    doi: 10.1029/2001PA000646

Other References:

Payne, J. L., D. J. Lehrmann, J. Wei, M. J. Orchard, D. P. Schrag, and A. H. Knoll, 2004. Large perturbations of the carbon cycle during recovery from the end-Permian extinction. Science, 305, 506-509.

Isozaki, Y., 1997. Permo-Triassic boundary superanoxia and stratified superocean: Records from lost deep sea. Science, 276, 235-238.

Erwin, D. H., S. A. Bowring, and J. Yugan, 2002. End-Permian mass extinctions: A review, In "Catastrophic events and mass-extinctions: Impacts and beyond", Koerbl, C., and MacLeod, K. G. (eds.), Geological Society of America Special Paper, 356, 363-383. Top