Earth science

Baked Alaska

Article metrics

The warming of the Earth's climate more than 50 million years ago is as yet unexplained. Now the finger points to the heating of sediment in the Gulf of Alaska as an important source of the greenhouse gas methane.

Between about 58 and 52 million years ago, during the late Palaeocene and early Eocene epochs, the Earth's climate warmed considerably. What was the cause of this unusually long warming trend? Writing in Geology, Hudson and Magoon1 put forward an idea that adds to thinking on the subject.

The marine geological record shows that Earth's climate has experienced swings of cooling and heating over various timescales. Change over thousands to hundreds of thousands of years seems to be related to orbital variations, but climate trends lasting millions of years have been harder to explain. The interval between the late Palaeocene and early Eocene has attracted particular interest, not only because of the exceptional warming but also because this was a time of widespread and intense tectonic activity at the Earth's surface.

Attempts to find out whether the two are connected have usually centred on the release of greenhouse gases, especially CO2, generated by tectonic activity. In some models the CO2 produced is related to volcanism, especially in the northeast Atlantic2. In others, the suggestion is that CO2 was released from limestones originally deposited on the northern coast of the Indian subcontinent, which were buried, heated and then emitted CO2 on a large scale when this landmass collided with the rest of Asia3.

Hudson and Magoon1 propose an alternative tectonic trigger for the global warming of 58–52 million years ago, one that invokes methane, not CO2, as the key greenhouse gas. Their model deals with the Gulf of Alaska, where they suggest that large volumes of sediment, eroded from freshly uplifted mountains, were deposited in deep water off the 2,200-km coast of Alaska. In this region the oceanic Pacific plate is subducting below Alaska but the overlying sediments are largely scraped off against the edge of the continent to form a 'subduction–accretion complex' (Fig. 1). As sediment becomes buried and granite pushes up into the sedimentary layer, the heat leads to the production of hydrocarbons, including methane, from organic matter in the sediment. The large volumes of methane released to the atmosphere, increasing greenhouse-gas concentrations, might explain the higher global temperatures 58–52 million years ago.

Figure 1: The Alaskan continental margin during the late Palaeocene epoch.

Sediment washed down from newly lifted mountains collects at the sea floor. The subduction of the oceanic plate carries that sediment down into the Earth's crust, where, below the 150 °C level, the organic matter in it is converted into methane and other hydrocarbons. Hudson and Magoon1 propose that the release of methane — a greenhouse gas — from this 'accretionary prism' of rock contributed to the global climate warming that happened over 50 million years ago.

Dating of events is crucial in assessing possible connections between tectonic events and greenhouse-gas emissions in the late Palaeocene and early Eocene. Take the case of the models invoking CO2. Although volcanism in the northeast Atlantic began about 63 million years ago, the greatest volumes of lava did not erupt until about 56 million years ago4 — too late to trigger the initial warming. Analysis of the sedimentary rocks in the 'collision zone' between India and Asia5 shows that the earliest collision between these land masses happened only about 50 million years ago. In assessing Hudson and Magoon's hypothesis, a similar caveat might apply. Age constraints indicate that heating of the eastern part of Alaska6 might have occurred too late to have been important in generating methane.

Another aspect of the new model1 deserving scrutiny is the estimated total methane production. Hudson and Magoon's estimate might be too high because it is based on a total sediment thickness of 25 km, assuming that 10 km has been removed by erosion. The eroded thickness has been estimated from radioisotope dating of the cooling of rocks through 300 °C, assuming a typical temperature gradient in the crust of 30 °C km−1 (see, for example, ref. 6). However, such cooling could also be caused by the removal of the overlying rocks by faulting, not erosion. In this case, the 'missing' 10 km of sediment has not been lost but merely displaced sideways. In practice this would reduce the amount of organic material available to be cooked into methane. Nonetheless, so great is the amount of methane predicted by Hudson and Magoon1 that even if production were halved this would still have resulted in a methane concentration in the atmosphere of several times the modern level.

Despite the apparent efficiency of Alaska as a generator of greenhouse methane, there is reason to wonder if this gas is really the likely climate culprit here. The popularity of methane as an agent in climate change has grown since its implication in the brief and dramatic Palaeocene–Eocene Thermal Maximum7 (around 55 million years ago), when temperatures in the deep oceans and at high latitude increased by 5–7 °C in less than 10,000 years. Here, the abrupt release of methane, previously stored in frozen form beneath the sea floor, provides a viable explanation for observed changes in marine and terrestrial carbon isotope records. But the direct climatic effects of methane that escapes to the atmosphere remain a matter of debate. Other new work8 shows that, because of its much longer residence time in the ocean–atmosphere system, CO2 might be responsible for the formation of polar stratospheric clouds (whose insulating effects could explain the especially elevated temperatures at high latitudes1,9). The uncertainty over this issue is all the more severe because most climate models currently lack the chemical detail and the resolution to predict the occurrence and radiative effects of these clouds.

There is also a long-standing problem in understanding the climate in continental interiors, for which models predict lower winter temperatures than those inferred from fossil evidence. Neither the direct radiative effects of increased greenhouse gases nor polar stratospheric clouds can account for this discrepancy. The cause of model underpredictions of temperatures both in polar regions and in continental interiors might lie in inadequate sensitivity of the modelled feedbacks from water vapour and cloud to increased greenhouse-gas concentrations. If so, then these same models are probably underestimating the strength of the hydrological cycle and the warming effects of the current increases in greenhouse-gas levels.

Hudson and Magoon's work1 is notable in highlighting a possible overlooked influence on climate change. In attempting to understand past climate trends, episodes of faster or slower carbon recycling at subduction– accretion complexes require more attention, as Hudson and Magoon have shown that in certain cases this process can influence the global greenhouse-gas budget to the same degree as does volcanic activity. Sedimentation and tectonic activity in the Gulf of Alaska is probably inadequate to explain the entire warming trend during the late Palaeocene and early Eocene by itself, although it might have contributed. Despite this advance in relating solid-Earth evolution to climate change, we are still struggling to understand the disparity between climate-model temperature estimates and those inferred from the fossil record, for polar regions and continental interiors during past warm intervals. A satisfactory, comprehensive mechanism has yet to be proposed. But when it emerges, we can hope that it will solve both model problems simultaneously, and apply more generally to warming episodes in the Earth's history.


  1. 1

    Hudson, T. L. & Magoon, L. B. Geology 30, 547–550 (2002).

  2. 2

    Eldholm, O. & Thomas, E. Earth Planet. Sci. Lett. 117, 319–329 (1993).

  3. 3

    Kerrick, D. M. & Caldeira, K. Chem. Geol. 108, 201–230 (1993).

  4. 4

    Saunders, A. D., Fitton, J. G., Kerr, A. C., Norry, M. J. & Kent, R. W. in Large Igneous Provinces (eds Mahoney, J. J. & Coffin, M. L.) 45–94 (Am. Geophys. Union Monogr. 100, 1997).

  5. 5

    Clift, P. D., Carter, A., Krol, M. & Kirby, E. in The Tectonic and Climatic Evolution of the Arabian Sea Region (eds Clift, P. D., Kroon, D., Craig, J. & Gaedicke, C.) (Geol. Soc. Lond. Spec. Publ. 195, in the press).

  6. 6

    Sisson, V. B., Hollister, L. S. & Onstott, T. C. J. Geophys. Res. B 94, 4392–4410 (1989).

  7. 7

    Dickens, G. R., O'Neil, J. R., Rea, D. K. & Owen, R. M. Paleoceanography 10, 965–971 (1995).

  8. 8

    Kirk-Davidoff, D. B., Schrag, D. P. & Anderson, J. G. Geophys. Res. Lett. 29, 10.1029/2002GL014659 (2002).

  9. 9

    Sloan, L. C. & Pollard, D. Geophys. Res. Lett. 25, 3517–3520 (1998).

Download references

Author information

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Clift, P., Bice, K. Baked Alaska. Nature 419, 129–130 (2002) doi:10.1038/419129a

Download citation

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.