Category: Climate change and palaeoclimatology

There is a large body of opinion, supported by plenty of circumstantial evidence, that the end of the last glacial maximum around 20 ka was controlled by processes that operated in the North Atlantic and its seaboard.  A favoured mechanism is the re-establishment of thermohaline circulation involving North Atlantic deep water that dragged surface water northwards from the tropics, to set up the Gulf Stream.  Temporary shut-down of thermohaline flux, probably by massive release of freshwater to the North Atlantic from melting of ice sheets, is widely understood to have triggered the sudden reversal to frigid conditions in the Younger Dryas around 11.5 ka.  The largest warming pulse in the northern hemisphere, between 14.6 to 14.0 ka, is recorded by a sudden increase in d¹⁸O of ice in the Greenland cores, and is known as the Bølling-Allerød warm interval.  Around that time, sea level rose by 20 m in a few hundred years, and that involved production of fresh glacial meltwater at a rate equivalent to the continual flow of five rivers the size of the Amazon.  Such rapid sea-level rise drowned coastlines and in some areas killed coral reefs.  On such drowned reef in the Caribbean gave a date of 14.2 ka, which since 1989 has been the only indicator of precise timing for the massive influx of meltwater to the oceans.  The date is within the Bølling-Allerød, hence the link between warming and events around the North Atlantic.  That central hypothesis is now under threat, following the dating of drowned coral reefs on the Sunda Shelf at 14.7 ka, and a re-evaluation of the Caribbean data. (Weaver, A.J. et al. 2003.  Meltwater pulse 1A from Antarctica as a trigger of the Bølling-Allerød warm interval.  Science, v. 299, p. 1709-1713).

Using the revised ages and climate modelling, Andrew Weaver and colleagues from the Universities of Victoria and Toronto, Canada and Oregon State University see the massive ice-melting as the precursor to the Bølling-Allerød warm interval and deglaciation of lands around the North Atlantic.  A more plausible source of freshwater influx is a major melting event in Antarctica, so warming in the south may well have driven that of the northern hemisphere.

See also: Kerr, R.A. 2003.  Who pushed whom out of the last ice age.  Science, v. 299, p. 1645.

For about a decade it has been suggested that the Tibetan Plateau, which rises to more than 5000 metres, has a profound effect on climate.  This may be partly due to the way such a high and enormous area deflects regional wind patterns, but largely to its profound interconnection with the South Asian monsoon.  When such a circulation barrier arose is critical to understanding how it relates to climate evolution in the latter part of the Cenozoic.  There are various suggestions, based on aspects of its structural and magmatic evolution.  Theory suggests that the southern part came into being in Eocene times, possibly because a segment of the lithosphere beneath broke off to subside into the mantle – there are volcanic rocks whose chemistry does suggest such a mechanism.  About 8 Ma ago the southern Plateau began to spread laterally, producing a series of N-S extensional basins, which suggests that by then sufficient gravitational potential had accumulated to make the thickened crust unstable.  About that time various signatures arose in foraminifera of the Indian Ocean and sediments derived by erosion, which suggest that the monsoon increased in intensity.

When the Plateau attained sufficient elevation above sea level to start spreading sideways and affect atmospheric circulation largely rests on these theoretical judgements.  For the ideas to firm up needs some means of estimating topographic elevation, which is not easy to do.  One way is to use plant remains that can give clues, either because the species involved are sensitive to elevation today, or the morphology of their leaves shows signs of physiological adaptation to elevation.  The first is ruled out in old sediments, simply because the species present are now extinct..  Plants metabolism is dependent on diffusion of water and CO₂ into their leaves during photosynthesis, and features, such as stomata density, give clues to the conditions for such diffusion.  Luckily, sediments from southern Tibet do contain well-preserved plants, and a multinational group led by Bob Spicer of the British Open University have attempted to assess palaeo-elevation for the time at which they were deposited (Spicer, R.A. and 7 others 2003.  Constant elevation of southern Tibet over the last 15 million years.  Nature, v. 421, p. 622-624).  Their method relies on linking leaf morphology to a property of the atmosphere, known as moist static energy (MSE), through estimates of atmospheric enthalpy from the leaves.  That is not the end of the estimation, because MSE needs to be related to elevation and the only way is to use climatic modelling for the past.  Whatever, Spicer and colleagues reckon that 15 Ma ago their sampling site was more or less at the same elevation as today, around 4.5 km above sea level.  If true, they have established that the south part of the Plateau was already in existence during the Middle Miocene.  Being so convoluted, despite its apparent precision, the leaf analysis method does need independent confirmation.  There is a much easier and arguably more reliable method, based on the change in the size of bubbles formed by gas escaping from lavas, according to atmospheric pressure (see Cunning means of estimating uplift in November 2002 issue of Earth Pages News).  There are lavas in southern Tibet that date from Cretaceous times, including some about a million years younger than the plant remains.

Precambrian warmth and methane

Methane is a more efficient “greenhouse” gas than CO₂, but it soon oxidises in the presence of oxygen.  During the Phanerozoic there have been several massive releases of methane, probably from gas hydrates in deep-ocean sediments, which produced warming spikes that decayed away quickly in geological terms.  Before there was much, if any, oxygen in the atmosphere, methane could linger and add to the retention of heat by carbon dioxide and water in the atmosphere.  One of the longest running disputes in environmental geochemistry concerns when oxygen levels became significant in the Precambrian, and what they were compared with later times.  Whether the Earth was warm or cold has a bearing on this.  Cosmological theory suggest that stars similar to the Sun progressively grow more energetic with time.  Without some kind of greenhouse effect, the Earth would have been condemned to frigidity from its outset.  Even today, with a more radiant Sun, only atmospheric retention of solar heat keeps overall temperature from being well below freezing.  The further back in time, the greater the “greenhouse” effect would have to have been to stave off complete ice cover and a runaway “icehouse”.  Methane almost certainly played a part in this once methane generating organisms evolved, up to about 2200 Ma, when there are signs (continental redbeds and soils rich in iron oxides) that atmospheric oxygen was appreciable.  However, warmth prevailed for about 1.5 billion years thereafter, until the plunges into frigid conditions of the so-called “Snowball Earth” period from about 700 to 550 Ma.  Somehow, the greenhouse effect lingered.

Alexander Pavlov of the University of Colorado, and colleagues from Pennsylvania State University have addressed the implications of this continued warmth in terms of maximum oxygen levels needed to avoid complete oxidation of methane releases (Pavlov, A.A. et al. 2003.  Methane-rich Proterozoic atmosphere?  Geology, v. 31, p. 87-90).  Today, more than 90% of all methane production beneath the ocean floor is consumed by bacteria, depending on the amount of dissolved oxygen and sulphate ions (for aerobic and anaerobic methanotrophs).  There is plenty of evidence that deep Precambrian ocean water was anoxic, so a great deal more methane would have emerged from them.  That it was also poor in sulphate ions is shown by their low levels in solid solution with carbonates and Proterozoic sulphur isotopes in marine sediments.  The authors argue that this signifies low atmospheric oxygen levels, around 5 to 18 percent of modern concentrations.  The scene may have been set for an excess of methane production over its oxidation, thereby keeping the “greenhouse” warming above the levels when glaciation would have been widespread..  If so, something completely upset this balancing act in the Neoproterozoic, to drive down temperatures several times – the “Snowball Earth” events.  The trigger may have been a boost in oxygen production and retention in the atmosphere.

El Niño in the Eocene

The oceanographic-climatic phenomenon in the equatorial Pacific, known as the El Niño-Southern Oscillation (ENSO), now seems to be major force in driving climate shifts far afield, such as the current drought in the Horn of Africa.  Its cyclicity relieves the suffering brought by El Niño events, yet the processes may well be highly unstable.  Some believe that it is only a matter of time before ENSO reverts to a permanent El Niño condition, with disastrous consequences.  Such a stabilisation in the past may have resulted in warming at high latitudes that permitted lush vegetation in near-polar regions, during the Cretaceous and the Eocene.  The Eocene was much warmer than now, as a result of a massive release of methane from seafloor sediments around 55 Ma.  So it makes sense to look at its climate record to check for a permanent El Niño.  Matthew Huber and Rodrigo Caballero of the University of Copenhagen have compared climate records from annually layered lake sediments from the Eocene of Germany and Wyoming in the western USA with climate models to test the hypothesis (Huber, M. & Caballero, R. 2003.  Eocene El Niño: Evidence for robust tropical dynamics in the “hothouse”.  Science, v. 299, p. 877-881).  The climate data from the lake sediments (thickness variations in annual layers) show clear signs of a roughly 5-year cycle of climate change, attributed to an Eocene ENSO.  This tallies nicely with simulations for the Eocene continent-ocean set-up.  Although the authors claim that their findings refute the hypothesis that global warming tends to shut down ENSO, which is a comforting thought, Eocene ocean and air circulation was not the same as now by any means.  There have been interglacial periods during the Pliocene to present climate system in which temperatures exceeded those of the Holocene.  Surely, annually layered sediments from those times will provide a better test.

Records of seawater oxygen isotopes and its Ca/Mg ratio shows that a substantial permanent ice sheet first formed in Antarctica in the Oligocene Epoch, about 34 Ma ago.  The favoured explanation, until this month, was that the South polar continent became thermally isolated from the rest of the planet when circumpolar currents were able to flow around it, once South America and Australia had separated from Antarctica and opened the “gateways” of the Drake and Tasmanian Passages.  But what if atmospheric CO₂ played a role?  A drop in the “greenhouse” effect and global cooling could have driven polar temperatures low enough for ice formation without an oceanographic influence.  Once established, the albedo effect of a large ice sheet would seal Antarctica into permanent freeze-up.  Factoring all the likely components in a general circulation model leads to a surprise (DeConto, R.M. & Pollard, D. 2003.  Rapid Cenozoic glaciation of Antarctica by declining atmospheric CO₂, Nature, v. 421, p. 245-249).  The opening of the Drake and Tasmanian Passages was not accompanied by a sufficient depth of water to support massive current reorganisation until several million years after the ice cap left its clear imprint on the marine record.  DeConto and Pollard’s model shows that even with closed Passages an ice cap would have formed, if CO₂ levels had fallen below three times those that prevailed in the Holocene, before industrial emissions began.  Global cooling had begun somewhat earlier than Antarctic freeze-up, following the high around the Palaeocene/Eocene boundary (~55 Ma), falling to a plateau about 40 Ma ago.  Undoubtedly CO₂ concentrations had fallen globally for this to have happened.  Of course, there is no Oligocene ice, from which glaciologists might extract trapped bubbles and samples of ancient air with which to refute or confirm the model.  However, a decrease in carbon dioxide would also cause the acidity of rainfall to decrease as well as the amount of rainfall globally, and that might show up in changed weathering processes, especially in the tropics of the time. 

How patterned ground forms

Visiting flat areas of permanently frozen ground brings you face to face with truly bizarre patterns at the ground surface.  Some are perfect hexagons of stones around finer soils, others doughnut-like circles and then a perplexing range of other features that look for all the world as though they were built by humans.  Undoubtedly, they result from the forces at work when the top soil layer freezes and thaws annually, together with soil creep down extremely shallow slopes, repeated over millennia.  However, exactly how the patterned ground develops has eluded geomorphologists for more than a century.  Rejecting the reductionist approach that any landform’s evolution can be deduced from basic principles of physics seems to be the key (Kessler, M.A. & Werner, B.T. 2003.  Self-organization of sorted patterned ground. Science, v. 299, p. 380-383).  Kessler and Werner of the University of California modelled the two likely processes of ice lensing that sorts stones and finer soil, and the transport of individual stones along the lines of accumulated stones as freezing fines expand, building in elements of spatial and time scales plus other parameters such as surface slope.  Their model is self-organising, and proceeds to mimic many of the intricacies of patterned ground, even the most labyrinthine.  It might seem a little heavy handed to crunch numbers to help explain what are really quite minor features.  But having demonstrated the power of non-linear modelling here, the authors open up a novel approach to landscape evolution of every scale and antiquity.

The curious mix of water ice and methane, known as gas hydrate or clathrate, which is stable at ocean depths greater than 300 m, is one of the largest potential components of the active carbon cycle (~10¹³ t).  Its methane content stems from bacterial breakdown of organic matter buried in anaerobic sea-floor sediments.  As well as being pressure sensitive, gas hydrate also has a narrow stability “window” as regards temperature.  Geothermal heat therefore limits the depth of gas-hydrate accumulations to a few tens to hundreds of metres below the seabed.  Its vast methane content is clearly something on which energy transnationals have an eye.  However, methane is almost four times more powerful as a “greenhouse gas” than CO₂ emissions.  Carbon-isotope studies from sedimentary rocks show signs that several times in the distant past methane was released catastrophically to the atmosphere, the timing coinciding with signs of rapid global warming.  The last major event of this kind was around 55 Ma ago, when the end of the Palaeocene Epoch witnessed an 8°C global temperature rise in a matter of a few thousand years (Thomas, D. et al. 2002. Warming the fuel for the fire: Evidence for the thermal dissociation of methane hydrate during the Paleocene-Eocene thermal maximum.  Geology, v. 30, p.1067-1070).  The warming “spike” eases because methane is quickly oxidised to water and CO₂ in the atmosphere, but that still allows abnormally warm conditions to linger.

Sonar surveys of the seabed, including that of the North Sea, reveal pits and funnels that probably mark sites of past methane releases from destabilised gas hydrates.  In theory, two general processes lead to their instability: falling global sea level that reduces the pressure on gas hydrates formed at shallow water depths; a rise in the temperature of ocean-bottom water.  The second could produce more widespread methane release than the first.  Refining these crude prognoses needs detail about the structure of gas-hydrate zones beneath the seabed.  Conventional seismic surveys conducted at the sea surface show the clathrate-rich zones just beneath the sea floor, but no detail.  Towing sources and receivers just above the seabed reveals intricate structures (Wood, W.T. et al. 2002.  Decreased stability of methane hydrates in marine sediments owing to phase-boundary roughness.  Nature, v. 420, p. 656-660).  Wood and co-workers from the US Naval Research Laboratory, the University of Victoria and the Pacific Geoscience Centre in British Columbia, Canada surveyed the Pacific floor off Vancouver Island.  Their most striking observation is of many vertical, chimney-like structures that puncture the gas-hydrate zone in the upper sediment layer.  They reckon that these structures are where methane and warm fluids find their way to the seabed; they are probably the expression in cross section of the surface pitting formed by past degassing.  They also may supply gas to the zone where it becomes locked in metastable water ice.  The sheer number of the “chimneys” indicates that the surface area of gas-hydrate stability is many times larger than previously supposed, as a result of their “roughening” effect.  Since the base of the gas-hydrate stability zone is most prone to the effect of warming of sea-bottom water, which shifts the geotherm slightly, an increase in its surface area, together with its closer approach to the seabed around the “chimneys”,  could further increase its sensitivity to small changes.  Up to now, many specialists have suggested that major methane releases resulted from sudden collapses of sea-floor sediments in tectonically unstable areas, such as the Storegga Slide off western Norway.  They may instead have been due to more widespread instability resulting from environmental change.  Since the largest pressure decreases due to sea-level falls accompanied glacial epochs, some clues to whether the “chimney” effect has had an influence may come from a fresh look at methane contents of trapped air bubbles in Antarctic and Greenlandic ice cores.  The extent to which methane releases might effect climate depends on how much is oxidised to CO₂ in sea water, before it can enter the atmosphere to enhance the “greenhouse” effect.  Little is know about such processes.

See also:  Pecher, I.A. 2002.  Gas hydrates on the brink.  Nature, v. 420, p, 622-623.

Palaeomagnetic data from localities famed for their Neoproterozoic glaciogenic rocks point persuasively to several epochs between 750 and 550 Ma when widespread continental glaciation took place at low latitudes.  It is this evidence, along with theoretical consideration of drastic changes in the Earth’s albedo that would result from tropical land ice, that encouraged the idea of pole to pole ice cover.  Only a build-up of volcanogenic CO₂ in the atmosphere could prevent such a “Snowball Earth” lasting indefinitely, and even with such relief it would have endured for millions of years.  Much of the geological evidence cited by those who support and promote this neo-catastrophic idea comes from excellent, but geographically quite limited occurrences of tillites or glaciomarine sediments, such as those of Namibia.  Some occurrences have never been seriously analysed, except as examples that superficially support the hypothesis.  One such sequence is that of Arabia, easily accessed in northern Oman and described by a British-Swiss team (Leather, J. et al. 2002.  Neoproterozoic snowball Earth under scrutiny: Evidence from the Fiq glaciation of Oman.  Geology, v. 30, p. 891-894).

Isotopic studies of carbonates from glaciogenic sediments (see Meltdown for Snowball Earth? in Earth Pages News for February 2002) seriously undermined several arguments by “Snowball Earth” supporters, but are open to various interpretations.  Hard geological evidence is less easy to rationalize.  A growing number of  Neoproterozoic glaciogenic sequences, such as the Port Askaig Tillite of the Scottish Dalradian Supergroup and others from the Congo and Kalahari cratons, and Laurentia, show dropstone-rich diamictites interbedded with sediments that show little if any sign of a glacial influence (Condon, D.J. et al. 2002.  Neoproterozoic glacial-rainout intervals: Observations and implications.  Geology, v. 30, p. 35-38).  Such evidence can be explained by climatic change and a fully functioning hydrological cycle.  The report on the Omani example by Leather and colleagues highlights splendid examples of sediments that mark cycles of glacial advance and retreat, reminiscent of those of the Pleistocene glacial epoch and more or less the same as in many Neoproterozoic occurrences.  It can only be a matter of time before Australian geologists enter the fray decisively, for glaciogenic sediments comprise up to 30% of the many-kilometres thick Umberatana Group in the Neoproterozoic of the Flinders Range in South Australia, and there are several other stratigraphically distinct diamictite sequences.

It seems likely that the “Snowball Earth” hypothesis is waning; an embarrassment for those geologists who have promoted it so assiduously over the last several years.  However, the enigma of low-latitude glaciation on a vast scale is likely to remain, unless, that is, all the diamictites can be shown to have non-glacial origins, which is not as unlikely as it might seem.  The Fiq sequence of Oman, like the Dalradian example in Scotland, formed in an actively extending basin.  Repeated seismicity on rift-bounding faults could have launched debris flows to deposit diamictites (a purely descriptive term for sediments containing a wide variety of clast sizes).  The most spectacular diamictite in the Dalradian Supergroup, and perhaps anywhere, is the Great Breccia of the Garvellachs.  Recent work suggests strongly that it is not glaciogenic, but the product of such a debris flow (Arnaud, E. & Eyles, C.H. 2002.  Catastrophic mass failure of a Neoproterozoic glacially influenced continental margin, the Great Breccia, Port Askaig Formation, Scotland.  Sedimentary Geology, v. 151, p. 313-333).  The supposedly clinching evidence for diamictites’ origin from iceberg armadas is the way in which some clasts (“dropstones”) puncture underlying stratification.  All that is required is a means of puncturing, and sediment compaction around large, resistant clasts in a water saturated matrix is quite capable of doing that.  Even the long-held belief that glaciation is uniquely signified by polished and striated surfaces beneath diamictites containing similarly scratched clasts is coming into question.  Sites of large impacts, such as the Ries crater in Germany, include exactly similar features caused by ejecta blasted from the crater, cited by Vern Oberbeck, formerly of NASA, in a little-cited paper that proposed an impact origin for diamictites (Oberbeck, V.R. et al. 1993.  Impacts, tillites and the breakup of Gondwanaland.  Journal of Geology, v. 101, p. 1-19).

Post-apocalypse weathering in the Early Triassic

Environmental crises do not come bigger than that at the end of the Permian, when marine ecosystems virtually collapsed, and similar extinctions of terrestrial flora and fauna are becoming clear.  Whereas the Siberian Traps may indeed have been a triggering mechanism, there are carbon-isotope indicators that vast amounts of methane entered the atmosphere shortly afterwards, rapidly being oxidised to CO₂.  The density of respiratory openings (stomata) in fossil leaves from the lowest Triassic is unusually low, indicating an abundance of CO₂ in the atmosphere and probably enhanced “greenhouse” conditions.  Hot and humid conditions encourage weathering of the continental surface, and there are many Early Triassic palaeosols, some which mimic those in the tropics being found at unusually high palaeolatitudes.  Such soils harbour crucial evidence for surface conditions, and the high-latitude ones present a surprise (Sheldon, N.D. & Retallack, G.J. 2002.  Low oxygen levels in earliest Triassic soils. Geology, v. 30, p. 919-922).  Unlike tropical laterites, which are rich in kaolinite, high-latitude soils are dominated by illitic clays that signify incomplete breakdown of silicates.  The surprise comes in the form of an unusual mineral, berthierine; a green, serpentine-like mineral that is easily confused with chlorites in hand specimen.  It can form by reaction between clays and ferric oxy-hydroxides, but only under highly reducing conditions.  Because most soils since about 2000 Ma ago have formed in contact with an increasingly oxygen-rich atmosphere, achieving suitably reducing conditions demands input of a reductant to the soil “atmosphere”.  The most likely candidate is methane, whose oxidation would consume oxygen.  However, methane’s residence time in the air is around 10 years, because it is quickly oxidised to CO₂, so methane release following the P-Tr boundary event seems as if it was sufficiently prolonged to influence considerably longer term soil formation.

Melting of low-latitude glaciers in Africa is so rapid that, unless they are cored soon, their content of long-term climate data may soon be gone forever.  So the first detailed isotopic record from Africa’s highest glacier on Kilimanjaro is cause for some relief.  Intrepid glaciologist Lonnie Thompson welded a large team together for this important task (Thompson, L. 2002. Kilimanjaro ice core records: evidence of Holocene climate change in tropical Africa.  Science, v. 298, p. 589-593).  The annually layered ice goes back only about 12 ka, but nonetheless gives a precious account of climate change at the heart of the continent, far more detailed than sparse lake-bed cores from various places.

The core confirms a broad pattern of warm, wet conditions from 11 to 4 ka, before the long-term cooling and drying of historical times.  These reflect likely weakening of monsoonal conditions in the late Holocene.  However, assigning precise ages to depth in the cores is not as easy as in those from high-latitude ice sheets, because of a lack of good layering (presumably) and dateable carbon.  At about 5200 years ago, the record shows an abrupt fall in d¹⁸O, a sign of drying and cooling that took place over perhaps a matter of decades.  This correlates with disruption of early civilisations in India, Egypt and the Middle East, and probably stemmed from cooling in the North Atlantic.  However, an equally rapid deterioration occurred around 6300 years bp, although not so extreme, to presage a millennium of arid conditions at the heart of Africa.  Important as these data are, the team’s estimates of current retreat rates of the Kilimanjaro glaciers are alarming.  Quite probably, the white cap of Africa’s highest mountain will have disappeared within the next 20 years.

Lonnie Thompson is obviously both keyed- and clued up about extracting climatic data from ice at high elevations.  So much so, that Science has printed a lengthy account of his exploits, mainly on low-latitude glaciers (Krajick, K. 2002.  Ice man: Lonnie Thompson scales the peaks for science.  Science, v. 298, p. 518-522

Earth Pages News  has commented several times on developments in the connection between ocean currents and climate, over the last 3 years.  The subject has many aspects, and these have been bundled and brought up to date in one of a series of review articles on the relationship between climate and the hydrological cycle in Nature’s occasional Insight series (Rahmstorf, S. 2002.  Ocean circulation and climate during the last 120,000 years.  Nature, v.  419, p. 207-214).  Stefan Rahmsdorf covers the evidence to date that implicates changes in deep circulation in rapid and dramatic climate shifts, such as changed air temperatures over the Greenland ice cap and iceberg armadas in the North Atlantic.  Another review outlines the longer-term perspective of links between atmosphere, oceans, ice sheets, solid-Earth processes and astronomical forcing in shifts of climate and sea level over the last 3 Ma.  Central to this linked system is the transfer of tens of millions of cubic kilometres of water from tropics to poles, and from ice sheets to sea levels (Lambeck, K. et al. 2002.  Links between climate and sea levels for the past three million years. Nature, v.  419, p. 199-206).

Alaskan source proposed for end-Palaeocene warming

Between 58 and 52 Ma, around the Palaeocene-Eocene boundary, Earth’s climate bucked the long-term cooling trend during the Cenozoic, by warming considerably.  Since the warming lasted for so long, it seems likely to have been caused by an enhanced atmospheric “greenhouse” gases rather than by either astronomical or oceanic causes.  Carbon isotope data around the P-E boundary can be interpreted in terms of massive releases of biogenic methane, perhaps from gas hydrates on the sea floor.  However, such releases are likely to have been sudden, and a more continual release of “greenhouse” gases fits the record better; but that begs the questions where and how?  Catastrophic methane release has been invoked for the dramatic rise in deep-ocean and high-latitude temperatures within 10 thousand years exactly at the P-E boundary.

Lengthy climatic warming can stem from increased volcanism and sea-floor spreading, but there is scanty evidence for either during this period.  Another possibility is production of gases as a result of tectonic activity, either by involvement of carbonate sediments in metamorphism, which releases CO₂, or “stewing” organic matter in thick sedimentary sequences.  Candidates for the last are the thick accretionary prisms at Pacific destructive margins, an especially appropriate example being that of the Gulf of Alaska which grew rapidly during this period (Hudson, T.I. & Magoon, I.B. 2002. Tectonic controls on greenhouse gas flux to the Paleogene atmosphere from the Gulf of Alaska accretionary prism.  Geology, v. 30, p. 547-550).  Oceanic and continental margin sediments scraped off descending oceanic lithosphere contain buried organic matter.  Increased heat flow, perhaps associated with rising magmas, can cause organic debris to break down to hydrocarbons.  Over-maturation results in the formation of methane, potentially in vast volumes, that can leak continually to the atmosphere.  Methane rapidly oxidizes to CO₂, decreasing the warming effect, but able to linger for considerable periods.  Hudson and Magoo calculate such enormous releases, that even disputes over the amount of accreted sediment in the Gulf of Alaska do little to rule out its being a major source for climatically implicated gases.  This first suggestion of a role for accretionary prisms in climate change may spur studies of such processes elsewhere, in an attempt to remove much of the load from the BLAG hypothesis that involves metamorphic release of CO_2­ in a difficult to verify process of lithospheric flatus.

See also:  Clift, P. & Bice, K. 2002.  Baked Alaska.  Science, v.  419, p.129-130

Provided the Milankovich theory of astronomical influences on insolation is indeed behind the pacing of glacial-interglacial episodes of the near past, it should be easier to forecast future change in overall climate than that of weather.  It turns out that the fluctuation of Earth’s orbital eccentricity (behind the roughly 100 ka periodicity of climate change for the past 1 Ma) is entering an historic low, due to the 400 ka period of one of its two cycles.  Modelling future insolation at high northern latitudes results in a damping of its fluctuations over the next 100 ka (Berger, A. and Loutre, M.F. 2002.  An exceptionally long interglacial ahead?  Science, v. 297, p. 1287-1288).  Left to climates own devices, the small changes in insolation may prolong the Holocene interglacial for as much as another 50 ka, instead of being now on the cusp of a descent into more frigid conditions.  Until recently, many climatologists looked to the last, Eemian interglacial as the model for the current one, and that lasted only 10 ka.

Of course, climate is no longer at the whim of astronomical forces and the Earth’s own circulation of energy, principally by the flow of energy in North Atlantic water, driven by deep water formed by sea-ice around Iceland.  Atmospheric CO₂ stands about 30% higher than during previous interglacials, because of anthropogenic emissions.  Berger and Loutre factor in the “greenhouse” influence of the additional CO₂, to find an ominous possibility that the Greenland ice sheet might well melt, with the climate entering an irreversible warming.  The climate, however, is not a model, and there is really no inkling of what surprises are in store from counter-intuitive behaviour of the many forces at work in it, under conditions that have no analogue during the whole of human evolutionary history.

Analogue of Archaean carbon cycle in Black Sea reefs

The Archaean world almost certainly had an atmosphere and oceans that were more or less free of oxygen.  Under such conditions the fate of dead organisms in the ocean, perhaps the remains of photosynthesizing cyanobacteria, would have been bacterial fermentation and the production of massive amounts of methane.  Along with volcanic emissions of carbon dioxide, methane in the atmosphere would have helped warm the planet at a time when the Sun emitted considerably less energy than it does now.  Methane is more strongly depleted in ¹³C than any organic or inorganic carbon compound.  So large falls in the d¹³C composition of organic carbon in Archaean rocks, around 2700 Ma have been taken by some palaeobiologists to signify methane metabolism.  Most methane-consuming bacteria today produce oxygen as a biproduct, so the negative excursions might indicate an early build up of more than a trace of oxygen in the Archaean atmosphere.  Discovery of bacterial communities on the floor of the Black Sea, which consume methane without oxygen production (Michaelis, W. and 16 others 2002.  Microbial reefs in the Black Sea fuelled by anaerobic oxidation of methane.  Science, v. 297, p. 1013-1015), suggest strongly that there may be little reason to suppose that Archaean conditions did involve free oxygen.

Off the coast of Crimea there are numerous sea-bed methane seeps in shallow water.  Surprisingly they are well-colonized by primitive bacteria, which produce thick mats held together by carbonate precipitates in completely anoxic conditions.  Laboratory cultures of the communities reveal that the consist of archaea and bacteria that respectively consume methane and reduce sulphate ions to sulphide.  The net result is that methane is oxidized by sulphate to produce calcium and magnesium carbonates, and lots of hydrogen sulphide (methane donates electrons for sulphate reduction, thereby becoming a source of carbon for cell metabolism).  Since much of the methane’s carbon ends up in stable carbonate – perhaps ten times more than in organic matter, such a process in the Archaean would have helped stabilize the “greenhouse effect” then.

One of the fundamental discoveries about climate change during the Plio-Pleistocene ice ages is how many climate fluctuations with periods too short to be ascribed to astronomical forcing link to shifts in deep-ocean circulation.  In the case of the North Atlantic Ocean, if high-latitude seas become diluted by fresh water cold dense brines are less able to form.  It is their sinking as a residue from the formation of sea ice that helps drive the “ocean conveyor” and draws warmer water into the Arctic from the tropics.  If they do not form, then the conveyor shuts down and high-latitudes cool.  The most spectacular of these ocean-driven events was the Younger Dryas cooling from about 12.9 to 11.6 ka, and it may well have occurred because of the sudden drainage of a giant lake of glacial meltwater down the St Lawrence Seaway to dilute the North Atlantic.  The waning of every major ice sheet covering North America would have generated vast amounts of freshwater, and because repeated glaciation created basins by erosion and sagging of the low-relief surface, drainage of such lakes would have been characteristic of every transition to interglacial warmth.  Steven Colman of the US Geological Survey reviews recent attempts to model how flooding may have escaped from the ice-sheet margins (Colman, S.M. 2002.  A fresh look at glacial floods.  Science, v. 296, p. 1251-1252).

The Hadean was cool

James Hutton’s observation that the geological history of Scotland had “no vestige of a beginning” applies everywhere, for no-one has dated rocks that are older than about 4.0 billion years (Ga) old, despite a great deal of effort.  It seems that continental crust only became capable of remaining at the surface in large volumes almost 600 Ma after the Earth formed from the Solar nebula.  Indirect isotopic evidence and dating of meteorites do indicate that the Earth accreted from dust and planetesimals about 4.56 Ga ago.  There are terrestrial materials that break the 4 Ga barrier, but they are so few and so tiny that they could be lost with one powerful sneeze.  These are crystals of the highly resistant mineral zircon, found as detrital grains in mid-Archaean sandstones in Western Australia.  The oldest of these is a single grain dated at 4.404 Ga.  All of them formed in igneous rocks produced by partial melting of the mantle, which concentrates zirconium in magma.  Following their liberation to sedimentary processes by weathering, the zircons have probably been through several sedimentary cycles since the formed.  So the pre-Archaean history of our world has left relics, but they are minuscule.  Because of the absence of pre-4Ga crust, that period was probably turbulent, partly through rapid convective turnover of the mantle and higher degrees of melting because of higher heat production, and partly due to far more large impacts that the lunar surface shows during those times.  Dating of lunar cratering and impact glasses suggests that bombardment reached a crescendo around 4.0 to 3.9 Ga.  It is now fairly certain that the Moon formed from incandescent material ejected from the Earth when it collided with a Mars-sized planet around 4.45 Ga.  Earth and its companion would, in that likely scenario, have begun their geological evolution completely molten in the case of the Moon and with a deep magma ocean on Earth.  “Hellish” is a barely adequate adjective for such conditions, and the period before 4 Ga has been termed the Hadean.  A vital question concerns when such extreme conditions waned to become potentially supportive of biochemistry and the origin of life.

Minute as they are, the pre-4.0 Ga zircons provide useful oxygen-isotope data, and their d¹⁸O is no different from that of more common zircons throughout the Archaean Aeon.  The explanation for this is that the mantle and the magmas produced from it contained an H₂O phase.  Either the mantle has always had a water content – no surprise as it still does – or the magmas from which the zircons crystallized encountered near-surface water vapour, possibly as a result of hydrothermal exchange with a hydrosphere.  Reviewing these data, John Valley and colleagues from the University of Wisconsin USA and Curtin University Australia pursue the second conjecture (Valley, J.W. et al. 2002.  A cool early Earth.  Geology, v. 30, p. 351-354), and argue for a surface temperature below the boiling point of water since 4.4 Ga, only 50 Ma years after geochemical “year zero”.  The crux of their argument is that the high d¹⁸O values of four Hadean zircons indicate their equilibration with water vapour at temperatures below water’s critical point (374°C).  If crystallization at depth was below that temperature, then the Earth would have had surface oceans.  But is this such a surprising conclusion?  Loss of heat by radiation being proportional to the fourth power of absolute temperature, an incandescent Earth’s surface at the time of Moon formation would have cooled below 100°C well within 50 Ma, unless it was blanketed by an opaque atmosphere.  Impacts of the size of those which produced the lunar maria around 4.0-3.9 Ga could have boiled away any surface water from time to time, only for the surface to cool quickly once again.  Conditions for bio-geochemistry could well have been present throughout the Hadean.  The significance of that for the origin of life is hard to judge, because large impacts and ocean boiling would have extinguished any progress, so that the process may have had to restart again and again.

Hothouse conditions were forced by massive emission of CO₂ during the mid-Cretaceous superplume event that created huge submarine basalt plateaux and began the development of many island chains that litter the floor of the central Pacific.  It was at this time that dinosaur-infested forests cloaked high latitudes, almost to both poles.  Terrestrial evidence suggests that conditions cooled somewhat in the later Cretaceous, and sequence stratigraphy indicates cyclic sea-level fluctuations, ascribed by some to the development of Antarctic ice sheets.  Resolving later Cretaceous global mean temperatures, and the ice-sheet question relies on oxygen isotopes from sea-floor sediments.  These are now available with sufficient precision and resolution to show that hothouse conditions lasted a great deal longer than suspected (Huber, B.T. et al. 2002.  Deep-sea paleotemperature record of extreme warmth during the Cretaceous.  Geology, v. 30, p. 123-126).

The Pacific superplume’s maximum activity was over a period of 15 Ma from 125 to 110 Ma (Barremian and Aptian), although it lasted until the early Campanian (80 Ma).  Contrary to the supposed magnitude of CO₂ release by volcanism, heating reached a maximum from 94 to 80 Ma.  Even at high southern latitudes, deep-ocean water remained at 14 to 19°C for these 14 Ma.  Until the end of the Cretaceous it rarely fell below 10°C.  The data rule out any circulation of cold, dense brines into the deep ocean basins from the formation of boreal sea ice, and consequently any influence by polar ice sheets.  Sea level reached its highest during this period, almost certainly because the volume of the ocean basins shrank, being floored by young, warm, low-density crust formed by the superplume.  Mid to late-Cretaceous flooding of the continental margins created uniquely favourable conditions for an explosive development of carbonate-secreting organisms of many kinds.  Despite the burial of vast carbonate platforms, as well as thick boreal coal seams, these limestone “factories” seem incapable of having kept pace with greenhouse warming.  Was CO₂ the only means then of global warming?