Climate model sensitivity to atmospheric CO2 levels in the Early–Middle Paleogene

https://doi.org/10.1016/S0031-0182(02)00718-6Get rights and content

Abstract

This study examines the sensitivity of the slab ocean version of the National Center for Atmospheric Research Climate System Model with revised Eocene geography, orography, and vegetation to changing carbon dioxide (CO2) levels. We compare model results with temperature proxies from the geologic record for the Early–Middle Paleogene. We ran three modeling experiments with CO2 levels at 500, 1000, and 2000 ppm, and all with atmospheric methane levels of 3.5 ppm. Surface temperatures in the two higher CO2 scenarios are warmer than those of the 500 ppm scenario. The largest warming with increasing CO2 occurred in the high latitudes, particularly in the Northern Hemisphere, during the wintertime. Compared to the 500 ppm case, Arctic wintertime temperatures increased by ∼10°C for the 1000 ppm scenario, and ∼20°C for the 2000 ppm scenario. The 1000 and 2000 ppm scenarios produced mean annual and cold month mean temperatures in mid- and high latitudes that are much more compatible with the climate interpretations from Eocene flora, especially for data from the Southern Hemisphere. Tropical sea surface temperatures (SSTs) in the 2000 ppm scenario, however, are still ∼4°C higher than the warmest temperatures inferred from proxy data. The better match between temperatures in the high CO2 modeling scenario and high latitude climate interpretations is consistent with the idea that the CO2 levels during the Eocene were high, at least 3–4 times the pre-industrial value of 280 ppm, but the discrepancies in the tropics suggest that SST estimates from proxies are too low or that the models lack some tropical cooling mechanism that was important at this time.

Introduction

The wide variation in carbon dioxide (CO2) estimates for the Early Paleogene (50–60 Ma), a period that proxies suggest was the warmest of the Cenozoic, have added considerable difficulty to the already challenging effort to understand the driving mechanisms of past climate states. Pearson and Palmer (2000) used boron isotope ratios to reconstruct the pH of seawater, from which they inferred very high atmospheric CO2 levels (>2000 ppm) for the Late Paleocene and Early Eocene, and values that may have been as high as 3000 ppm at the Paleocene–Eocene boundary. This is approximately ten times the pre-industrial level of atmospheric CO2 (280 ppm). Retallack (2001), Kürschner et al. (2001), and Royer et al. (2001a) all used stomatal abundances from fossil leaf cuticles to infer ancient CO2 levels, but their estimates for the Paleogene vary widely (∼2000 ppm, ∼500 ppm, and ∼300–400 ppm, respectively). Estimates from paleosol carbonates also vary considerably for the Early–Middle Paleogene, from 0 to ∼1000 ppm (Cerling, 1991, Ekart et al., 1999). In addition, mass balance models of CO2 suggest that CO2 has been decreasing steadily over the past 170 m.y. (with a slight rise at ∼50 Ma), but was anywhere from 300 ppm to 1000 ppm during the Early Paleogene (Tajika, 1998, Berner and Kothavala, 2001).

Fossil leaf assemblages support the picture of a warm, wet climate at middle and high latitudes during the Late Paleocene and Early Eocene (e.g. Wing and Greenwood, 1993, Greenwood and Wing, 1995), and data from oxygen isotope paleothermometry suggest a fairly shallow latitudinal sea surface temperature (SST) gradient (Zachos et al., 1994, Tripati et al., 2001).

Paleoclimate modeling studies generally have had difficulty reproducing the shallow gradients of the greenhouse world depicted by the proxy climate data no matter what forcing mechanisms are invoked (e.g. Sloan and Rea, 1995, Sloan and Pollard, 1998, Huber and Sloan, 2001). Understanding the mechanisms that kept the planet warm during the Early Paleogene is particularly important because, superimposed on a long-term warming through the Late Paleocene, there is a period of abrupt warming (∼56 Ma) that lasted ∼100 k.y. (Zachos et al., 1994, Bralower et al., 1995, Zachos et al., 2001). The most plausible mechanism for this abrupt warming, known as the Paleocene–Eocene Thermal Maximum (PETM), is the release of a large amount of methane gas (CH4) from clathrates in ocean sediments (Dickens et al., 1995, Katz et al., 1999). The rate at which carbon was added to the atmosphere at the PETM approximates the rate at which carbon is presently being added to the atmosphere. We do not know how sensitive a warm climate, such as that of the Late Paleocene or Early Eocene, would be to subsequent additions of CO2 or CH4.

In this study we examine the sensitivity of the slab ocean version of the National Center for Atmospheric Research (NCAR) Climate System Model (CSM) (v.1.2) with Eocene geography to a range of CO2 values, from 500 to 2000 ppm. Our primary goal is to determine whether the newer NCAR CSM can produce a better match between the model results and the inferred proxy climate of the Early Paleogene than the older GENESIS model used by Sloan and Rea (1995). The CSM is a more recently developed model that includes improvements in cloud parameterizations, long wave radiation, and land surface processes. Of particular importance for this study, the atmosphere component of CSM, the NCAR Community Climate Model (v.3.6.6) (CCM3), includes long wave radiative effects of CH4, and two additional weak CO2 bands at 9.4 and 10.4 μm, which become important at very high levels of atmospheric CO2 (Kiehl et al., 1998, Kothavala et al., 1999). This new atmospheric physics in the model may provide us with new information about the role of greenhouse gases in past climate states.

This is the first study to examine the sensitivity of Paleogene climate to a CO2 level of 2000 ppm and a CH4 level of 3.5 ppm. Additionally, we have updated Eocene geography and vegetation with that of Sewall et al. (2000), and in order to account for a slight reduction in solar luminosity at 50–60 Ma, we have reduced the present-day solar constant by 6 Wm−2 to 1361 Wm−2.

We take both a quantitative and a qualitative approach to comparing model results with proxy data. While we recognize that there are other mechanisms affecting surface temperature, for this study we choose to focus only on CO2.

Section snippets

Experimental design

We use the NCAR CSM (v.1.2) as described by Boville and Gent (1998). Our boundary and initial conditions are from recent modeling studies of Early Eocene climate (Sewall et al., 2000, Huber and Sloan, 2001). This is the first study to use the CSM with Eocene boundary conditions coupled to land surface, sea ice, and slab ocean model components to examine the sensitivity of the global ocean response to massive increases in CO2.

The CSM divides the climate into component models for the atmosphere,

Surface temperature

As expected, global mean surface temperature rose with increasing CO2. From 20°C in the LoCO2, it increased by 2°C in the MidCO2 scenario and by 4°C in the HiCO2 scenario relative to LoCO2. Fig. 1 shows the mean annual, zonally averaged combined ocean and land surface temperature for each case plotted with Eocene mean annual temperature (MAT) estimates from marine and terrestrial proxy data. The most dramatic increases in surface temperatures with increasing CO2 were in the high latitudes

Discussion

Overall, our results for the higher CO2 experiments (1000–2000 ppm) are in stronger agreement with proxy data for the Late Paleocene and Early Eocene than the results for the lower pCO2 experiments. Of the three experiments, the HiCO2 scenario is the only one that produces mean annual and CMM temperatures in mid-latitudes and high latitude coastal regions compatible with climate interpretations from Eocene flora. Below we provide a more detailed comparison of our experimental output with proxy

Conclusions

The Mid- and HiCO2 experiments produce modeled climates more consistent with that depicted by proxy data than the LoCO2 scenario. The higher CO2 cases are in agreement with proxy data in the following key areas:

• Antarctic coastal temperatures in the Mid- and HiCO2 cases are warm enough to support observed fossil flora in coastal regions.

• CMM land temperatures in the Mid- and HiCO2 cases fall within the mean error of CMM estimates from proxy data in the Southern Hemisphere.

• Fossil plants

References (46)

  • T.J. Bralower et al.

    High-resolution records of the late Paleocene thermal maximum and circum-Caribbean volcanism: Is there a casual link?

    Geology

    (1995)
  • T.E. Cerling

    Carbon dioxide in the atmosphere: Evidence from Cenozioc and Mesozoic paleosols

    Am. J. Sci.

    (1991)
  • G.R. Dickens et al.

    Dissociation of methane hydrate as a cause of the carbon isotope excursion at the end of the Paleocene

    Paleoceanography

    (1995)
  • D.D. Ekart et al.

    A 400 million year carbon isotope record of pedogenic carbonate: Implications for paleoatmospheric carbon dioxide

    Am. J. Sci.

    (1999)
  • D.R. Greenwood et al.

    Eocene continental climates and latitudinal temperature gradients

    Geology

    (1995)
  • D.R. Greenwood

    Eocene monsoon forests in central Australia?

    Aust. Syst. Bot.

    (1996)
  • D.R. Greenwood et al.

    Climate change and vegetation responses during the Paleocene and Eocene in southeastern Australia: GFF

    Geol. Soc. Swed.

    (2000)
  • K.J. Hennessy et al.

    Changes in daily precipitation under enhanced greenhouse conditions

    Clim. Dyn.

    (1997)
  • Houghton, J.T., Ding, V., Griggs, D.J., Noguer, M., van der Linden, P.J., Xiaosu, D. (Eds.), 2001. Climate Change 2001:...
  • S.A. Hovan et al.

    Paleocene/Eocene boundary changes in the atmospheric and oceanic circulation. A Southern Hemispheric record

    Geology

    (1992)
  • M. Huber et al.

    Heat transport, deep waters, and thermal gradients: Coupled simulation of an Eocene ‘greenhouse’ climate

    Geophys. Res. Lett.

    (2001)
  • V.M. Kattsov et al.

    Twentieth century trends of Arctic precipitation from observational data and a climate model simulation

    J. Clim.

    (2000)
  • M.E. Katz et al.

    The source and fate of massive carbon input during the latest Paleocene Thermal Maximum

    Science

    (1999)
  • Cited by (108)

    • The Arctic: A Barometer of Global Climate Variability

      2021, The Arctic: A Barometer of Global Climate Variability
    • Was the Arctic Eocene 'rainforest' monsoonal? Estimates of seasonal precipitation from early Eocene megafloras from Ellesmere Island, Nunavut

      2015, Earth and Planetary Science Letters
      Citation Excerpt :

      In addition, multiple studies have analyzed the climate of this fossil environment using paleontological and other available proxies from Arctic sediments (e.g., Greenwood and Wing, 1995; Jahren and Sternberg, 2003, 2008; Greenwood et al., 2010; Eldrett et al., 2009, 2014). Although some climate models have shown low to moderate Arctic paleoprecipitation during the early to middle Paleogene (Shellito et al., 2003), the majority of studies have shown high precipitation (annual precipitation >150 cm/yr), mesothermal conditions (i.e., mean annual temperature ∼12–15 °C), and moderate winter temperatures (cold month mean temperature >0 °C) characterized climates of high latitude Arctic rain forests during the late Paleocene through the PETM and into the middle Eocene (Greenwood et al., 2010; Eberle and Greenwood, 2012; Huber and Goldner, 2012). Notably, Huber and Goldner (2012) reconstructed global precipitation patterns of Eocene, identifying the existence of a robust global monsoonal precipitation pattern.

    • An early-middle Eocene Antarctic summer monsoon: Evidence of 'fossil climates'

      2014, Gondwana Research
      Citation Excerpt :

      That the Fossil Hill flora does not plot within the cloud of modern Chinese monsoon sites can be explained by 1) the Chinese sites with a strong monsoon signal are in subtropical regions with high temperatures (Jacques et al., 2011), whereas the Fossil Hill flora comes from cooler high latitudes (the proximity of some sites in north China confirms this), and 2) the Fossil Hill flora is situated on King George Island, part of the Southern Shetland archipelago, and its monsoon regime is likely to have been tempered by the influence of the surrounding ocean similar to that for Japan today. Climate models, including coupled models and general circulation models, usually reconstruct cold and dry climates for Antarctica during the Eocene (e.g., Shellito et al., 2003; Huber and Goldner, 2012). Proxies, either biological or chemical, indicate that the climate was warmer than that indicated by the model reconstructions (Korty et al., 2008).

    View all citing articles on Scopus
    View full text