Category: Climate change and palaeoclimatology

The central factor in abrupt climatic shifts during the last glacial period was change in thermohaline circulation (THC), particularly in the Atlantic Ocean.  Two general processes underpin THC: differences in solar heating from low to high latitudes drive polewards flow of surface water; formation beneath sea-ice of dense brine that sinks to form an equatorwards flow of North Atlantic Deep Water (NADW).  Freshwater influx at high latitudes suppresses the formation of NADW, which, together with enhanced low-latitude evaporation, slows polewards surface flow..  Currently, the thermal influence and NADW formation dominates heat transport northwards in the North Atlantic, by carrying about a petaWatt at mid latitudes.  THC is of little consequence in the North Pacific, partly because its fresher surface water hinders dense-brine formation, and partly because any deep water formed beneath sea ice in the Arctic cannot flow through the very shallow Bering Straits.

Clearly THC is a sensitive mechanism, inseparable from other factors in climate forcing.  Having such a vast influence on heat transport, if it changes there are likely to be dramatic outcomes for climate, particularly along the eastern flank of the North Atlantic where much of the transported heat arrives.  Sea-ice formation around Iceland is decreasing, so a review article on THC and rapid climate change is essential reading (Clark, P.U. et al. 2002.  The role of thermohaline circulation in abrupt climate change.  Nature, v. 415, p. 863-869).  It is now known that the last glacial period was punctuated by short-period (~ 1-2 ka) warming-cooling episodes, known as Dansgaard-Oeschger events, one aspect of which was the launching of “armadas” of icebergs to latitudes as far south as Portugal (known as Heinrich events), which left their mark as occasional gravel layers in the otherwise muddy sediments on the deep Atlantic floor.  These episodes involved temperature changes over the Greenland icecap of as much as 15°C.  They began with warming on this scale within a matter of decades followed by slow cooling to minimal temperatures, before the next turn-over.  The deep cooling seems to have accompanied slowing and shut-down of THC.  Current global warming is likely to do three things:  increasing low-latitude evaporation, increasing freshwater influx to high-latitude Atlantic surface water and a decrease in sea-ice formation at the site of NADW formation.  Because all three drive down polewards heat flux, anthropogenic warming may well result in contrary climate shift in Western Europe and Scandinavia – freeze rather than thaw.  If it happens, chances are that it will be upon us with little warning.

Following on from their linking carbon-isotope excursions associated with Neoproterozoic diamictite-cap carbonate sequences to methane release (see Methane and Snowball Earth in Earth Pages, December 2001), Martin Kennedy, Nicholas Christie-Blick and Anthony Prave turned to the d¹³C values in the diamictites for which a glaciogenic interpretation forms the main plank of the Snowball Earth hypothesis (Kennedy, M.J. et al. 2001.  Carbon isotope composition of Neoproterozoic glacial carbonates as a test of paleoceanographic models for snowball Earth phenomena.  Geology, v. 29, p. 1135-1138).  Complete ice cover of the oceans would chemically isolate ocean water from the atmosphere, and would effectively shut down the organic sinks for atmospheric carbon dioxide.  While they operate, the exclusion of ¹³C relative to lighter carbon by organisms drives up d¹³C in sea water, to be preserved in carbonate sediments.  The Snowball Earth model predicts negative d¹³C, approaching the -5‰ of the mantle, in carbonates deposited during all-enveloping glacial epochs.  However, few researchers have made the measurements needed to test that part of the hypothesis.

Kennedy and co-workers show from three such diamictite sequences that the carbonate precipitated as cement in them has consistently positive d¹³C.  Although that does not disprove the existence of glaciation at tropical latitudes, it is not consistent with the dreadful scenario of totally ice-bound oceans devoid of life.  Nor, for that matter, is there any evidence from strontium isotope variations in carbonate cap rocks for the massive continental weathering that the Snowball Earth devotees propose as a means of escape from the eventual hot-house that build up of volcanic CO₂ emissions to release Earth from the mothers of all cold snaps would create.  Expect interesting news in later Earth Pages of how the greatest Earth science debate of the 21^st century develops.

Much of the information about glacial and interglacial climate change has come from cores drilled either from ocean-floor sediments or ice caps.  However, both suffer from limits to time resolution of the order of more than 100 years, although ice younger than about 5 thousand years clearly shows annual layers.  While groundwater is able to flow, speleothem (flowstone)  grows continuously in caves, under conditions of extremely stable temperature and humidity.  Depending on how they are analysed and how thick the deposits are, stalagmites and stalactites should give fine time resolution.  A half-metre long stalagmite from an Irish cave has grown since the start of the Holocene.  Using high-precision uranium-series dating, its length has been calibrated in calendar years before present.  A laser probe that releases oxygen from the speleothem calcite has provided oxygen isotope data whose resolution (between 7 and 18 years) is an order of magnitude better than sea-floor sediments and between 5 to 20 times better than from pre-5 ka ice cores (McDermott, F., Mattey, D.P. and Hawkesworth, C.  2001.  Centennial-scale Holocene climate variabilty revealed by a high-resolution speleothem d¹⁸O record from SW Ireland.  Science, v. 294, p. 1328-1331).

Until recently, the best documented climate variations that are more rapid than can be explained by the Milankovich effect are the Dansgaard-Oeschger cycles in the Greenland ice cap.  They are of the order of 1 ka, but somewhat variable in their periodicity.  The Irish stalagmite shows that there were climate shifts throughout the once supposedly stable Holocene, with frequencies equivalent to periods of 625, 169 and 78 years, the latest of which coincide with warm and cool periods since Roman times.  One caution is that the oxygen isotope variations cannot be ascribed directly to variations in air temperature, because they would have been affected by differences in the surface seawater from which water vapour evaporated to fall as rain in SW Ireland.  Before about 4.5 ka 8 clear peaks and troughs occur at the same times in both the Irish stalagmite and the Greenland ice core; clear signs of regional changes.  These probably reflect releases of glacial meltwater to freshen surface waters of the North Atlantic.  Over Greenland they resulted in atmospheric cooling, in response to weakening of the effects of the Gulf Stream by reduced thermohaline circulation.  The correlation breaks down for the last 4 ka, and the fluctuations in the Irish data do not show features that coincide with ice-rafting events known from sea-floor sediment cores.  That suggests that ice-rafting was no longer able to cap the North Atlantic with fresher water.  Nonetheless, something was going on to impart isotopic changes to rain falling on Ireland, and that did coincide with the widespread climate changes of the recent past.  What the driving processes were is not known, but it seems inescapable that underlying the drive to global warming through industrial CO₂ emissions is a more fundamental process.  Should anthropogenic warming reinforce it, as seems to be happening, their combined effects could flush fresh water into the North Atlantic’s surface layers, thereby slowing thermohaline circulation and the warming effect of the Gulf Stream.

One process connected to long-term climate change is the way that weakly acid rainwater (containing dissolved CO₂) weathers silicates in continental rocks, one product being carbonate in soils.  The process should draw CO₂ from the atmosphere, thereby reducing its “greenhouse” effect.  The idea is by no means new, but received a boost in the mid 1990’s from Maureen Raymo’s suggestion that fluctuations in the strontium-isotope composition of the oceans through geological time should be a proxy for changes in the rate of continental weathering.  The ⁸⁷Sr/⁸⁶Sr of marine carbonates does show clear correlation with long-term climate shifts during the Phanerozoic..

Continental weathering should increase as topographic relief becomes greater through mountain building episodes.  The Himalaya’s rise through the late-Tertiary has been suggested as a major influence over climatic deterioration, partly by its effect on the Asian monsoon and partly as a huge site for the sequestration of atmospheric CO₂ by chemical weathering.  Himalayan rivers have enormous flows and equally large sediment and dissolved element loads.  In particular they carry far more strontium than other rivers, and it has a highly radiogenic content of ⁸⁷Sr.  There are three means of attaining these levels: from average continental crust which has a higher ⁸⁷Sr/⁸⁶Sr ratio than oceanic crust (the other main source of seawater strontium); from strontium rich limestones that acquired their isotopic signatures from the ocean when they were deposited; or from sources with unusually high ⁸⁷Sr/⁸⁶Sr ratios.  The Himalaya are well known for carbonate sediments, and for granites formed by melting of deeper, older continental material that gives them very high proportions of radiogenic strontium.  Recent work now shows that a significant contribution of highly radiogenic strontium to Himalayan rivers is hydrothermal activity (Evans, M.J. et al. 2001.  Hydrothermal source of radiogenic strontium to Himalayan rivers.  Geology, v. 29, p. 803-806).  Hot springs feeding a major tributary of the Ganges contribute up to 30% of its strontium load, and incidentally a great deal of CO₂.  Both result from hydrothermal alteration of deeper rocks, and are unrelated to weathering if the water involved emanates from the deep crust.  It seems that these waters are recycled rainwater, so this is a case of a high-temperature chemical weathering.  Whatever, it further complicates the original notion of linkage between mountain building and climate.

Methane and Snowball Earth

The well-publicized “Snowball Earth “ model for Neoproterozoic glaciogenic rocks that occur at tropical palaeolatitudes has to involve an escape mechanism from global frigidity.  Without some means of warming, the high albedo of widespread ice would have locked the Earth into perpetual glaciation, which of course did not happen.

The main proponents of the model, Paul Hoffman and Dan Schragg of Harvard University suggested a gradual build up of volcanogenic CO₂ during “Snowball” conditions, when a dry atmosphere would have retained the “greenhouse” gas instead of its being sequestered to the oceans and carbonate rocks by acid rain and continental weathering.  Gradually, atmospheric temperatures would have risen due to trapping of outgoing, long-wave radiation by CO₂.  This simple aspect of the model leads to scenarios where warming overruns once ice sheets disappeared, to give extremely high-temperature conditions.  Using carbon-isotope data from marine carbonates is a means of supporting or refuting this escape mechanism, and also of detecting the influences of other components of the carbon cycle.  Carbonates take up carbon dissolved in seawater without fractionating its different isotopes, and provide measures of the degree to which organic processes did contributed to fractionation.  Cell processes preferentially take up ¹²C, and if large masses of undecayed organic matter ends up in seafloor sediments, the proportion of “heavier” ¹³C (indicated by the standardized ratio of the two main isotopes d¹³C) increases in seawater and the atmosphere.  Carbon of mantle origin, that emerges as volcanic CO₂, has a constant d¹³C of about -5‰.  So these two processes contribute to an isotopic balance, which for most of the Mesozoic and Cenozoic Eras established a d¹³C of between 0 and +4 ‰ in sea water and limestones.  This is interpreted as a sign that the recent carbon cycle achieved a balance between volcanic additions and organic carbon burial weighted towards trapping of undecayed carbohydrate in sea-floor sediments.  Explanations for broad climate changes since 250 Ma therefore rely more on other mechanisms than on the carbon cycle

The most comprehensive study of Neoproterozoic carbon (Walter, M.R. et al. 2000.  Dating the 840-544 Ma Neoproterozoic interval by isotopes of strontium, carbon and sulfur in seawater, and some interpretative models.  Precambrian Research, v. 100, p. 371-433) does indeed show dramatic see-sawing of d¹³C through supposed “Snowball” events, from highly positive values (<+10‰) before glaciogenic sedimentation to highly negative (>-10‰) in the immediate aftermath.  However, few data were available from within glaciogenic sediments, and resolution is insufficient to detect tell-tale trends.  The key approach needs detailed carbon isotopes through a single event, and such data appeared recently for the famous Neoproterozoic glaciogenic-cap carbonate sequence of Namibia (Kennedy, M.J. et al. 2001.  Are Proterozoic cap carbonates and isotopic excursions a record of gas hydrate destabilization following Earth’s coldest intervals.  Geology, v. 29, p. 443-446)

Kennedy et al. measure d¹³C in carbonate cements in the glaciogenic diamictites, in overlying cap carbonates and in cement to later clastic rocks.  Interestingly, there is little sign of a gradual decrease in ¹³C through the glaciogenic rocks.  Constant oceanic carbon composition would be expected if no volcanic CO₂ entered seawater during frigid, dry conditions, and living processes were minimal.  In the cap carbonates d¹³C plummets from +3‰ to -4‰.  One simple explanation would be massive “rain-out” of volcanic CO₂ (d¹³C of -5‰) that had built up in the air during the “Snowball” episode.

The view that all manner of processes connected with climate – volumes of land ice, ocean temperature and flow, aspects of atmospheric composition and its motion, and the expansion and contraction of biological communities – are locked into the cycles of changing solar input steadily evolves into something less mechanical, as new data flows in.  The first serious doubts about Milutin Milankovic’s theory of astronomical forcing of the world’s climate, since oxygen isotope fluctuations in sea-floor sediments began to reveal the periodicities predicted by him, stemmed from a very different kind of deposit.  Devil’s Hole near Las Vegas, a fissure being slowly filled by calcite flow stone that precipitates from groundwater, presented a detailed record of oxygen-isotope variations over the last 600 thousand years.  Though showing the same patterns as ocean cores available at the time, Devil’s Hole revealed changes in continental climate that differed from those in land-ice volume by thousands of years.  Ice-core time series of Antarctic air temperatures also show that warming began up to 9000 years before the last four terminations of glaciation.  As more proxies for climate are devised, the more complex global climate shifts appear to have been.

The latest measure stems from fluctuations in the structure of compounds produced by marine algae as a result of shifts in sea-surface temperature.  Applied to sea-floor sediments deposited off California, an area influenced today by the southward, cold California Current, they reveal regional warming of the sea that began 10 to 15 thousand years earlier than the last five deglaciations of the northern hemisphere (Herbert, T.D. and 8 others  2001.  Collapse of the California Current during glacial maxima linked to climate change on land.  Science, v. 293, p. 71-76).  In cores south of the modern cold current, no such large discrepancies emerged.  In fact they accompanied the maximum extents of land ice.  It seems that, like the Gulf Stream, the California Current is prone to shutting down, but as a result of changed Pacific wind patterns in response to the North American ice sheets rather than to thermohaline deep circulation.  Here is an explanation for the vexing record from Devil’s Hole – regional climate shifts that do not “knock” Milankovic.

There is no doubt that changes in ice volume on the northern continents are the main characteristic of environmental change going back more than 2 Ma.  However, the mechanistic view that lots of ice means a cold, dry world and a great deal less points to warmth and more moist conditions is dead in the water as a useful paradigm.  Yet all models of climate are little more than Heath Robinson tangles of such reductionism, despite claims for their increasing incorporation of ideas that stem from measured realities.  As always, the devil lies in the detail, and Herbert et al.’s paper also shows from pollen records in the marine cores that dense warm-climate forests cloaked the Pacific seaboard during the last 5 glacial maxima.  For a vast area of western North America to be warm while ice sheets elsewhere were at their maximum should be a warning of unpredictable future climate shifts.

Growing concern about unpredictable and contrary change was amply expressed by a meeting of 1800 climate specialists in Amsterdam in early July.  They endorsed the distinct possibility of sudden shifts in regional climates that may stem from increased global warming, such as return of vegetation to the Sahara, aridity in the Amazon basin, and Europe’s plunging into a frigid climate as the Gulf Stream slows because of reduced thermohaline circulation (Pearce, F.  2001.  Violent future.  New Scientist, 13 July 2001, p. 4-5).

The kingdom of the eukaryotes rests on a very simple environmental economy.  Plants are producers of carbohydrate through photosynthesis, thereby generating excess oxygen from the photo- and molecular chemistry involved.  Animal consumers use up oxygen in their metabolism and return carbon dioxide, the ultimate source of carbohydrate, to the air.  A simple view is that animals contribute to global warming, whereas plants help cool the world.  Perhaps because of that “common sense” view, most environmental scientists take a very different line, linking it with volcanic exhalation of CO₂, “capture of carbon through rock weathering and the burial of dead organic matter  in the global carbon cycle.  Greg Retallack of the University of Oregon is about to publish a reappraisal of the animal versus plant part of the C-cycle (in press, Journal of Geology) that is based on observed imbalances between the two opposed kinds of respiration.  Specialists in the C-cycle hold that there is a an overall balance, taking all components into account, whose inevitable result is the build up of oxygen in the atmosphere of an inhabited world.  Yet oxygen is extremely reactive and should quickly combine in mineral oxides and hydroxides – after all, the iron in an untended car reverts to its oxide ore in the space of a few decades at most.

Partly following James Lovelock’s Gaia hypothesis, Retallack focuses on the major fluctuations in atmospheric chemistry evidenced in the geochemical record, the most immediate being the see-saw fluctuation of modern levels of CO₂ in the atmosphere – a 2% annual variation controlled by the waxing and waning of vegetation in the northern hemisphere (where plant cover is greatest) according to season.  One of the largest shifts in atmospheric CO₂ concentration followed the evolution of land plants from about 450 Ma ago.  To thrive, they had to develop hard cellular material (lignin) that formed stems and trunks, which animals of the Palaeozoic were unable to oxidise efficiently.  Both living biomass and burial of undigested lignin drew down CO₂ and boosted oxygen levels.  Animal evolution eventually exploited this “free lunch” through the humble termite and reptilian and then mammalian megafauns.  Retallack believes that heavy breathing that resulted from lignin digestion reversed the declining CO₂ trend for the 200 Ma following the Carboniferous to Permian glacial epoch in Gondwana.  Though displaying some ups and downs, the Mesozoic saw a “greenhouse” world.  Removal of the mighty and extremely abundant herbivorous dinosaurs by the K-T mass extinction provided and opportunity for plant diversification.  Many Mesozoic plants evolved armour against browsing dinosaurs, exemplified by the surviving Andean “monkey puzzle” tree Araucaria.  Their demise removed the need, and the plant Kingdom’s evolutionary response was the appearance of grasses.  Reatallack points out that grass itself is not as good as lignin-rich plants in holding CO₂, but grasslands encourage the development of thick carbon-rich soils that hold more than the soils of the forest floor.  It is this development that Retallack believes lay at the base of the decline in average global temperature through the Cainozoic, to culminate in the present Ice Age.  Unsurprisingly, proponents of the complexity and diversity of the C-cycle, particularly in the oceans, are disinclined to have truck with the hypothesis.

Source:  Pearce, F.  The Kingdoms of Gaia.  New Scientist, 16 June 2001, p. 30-33.

Carbonates and biofilms

Above the low level that is essential for their role in molecular “information” transfer, calcium ions pose a fatal threat to cell processes.  That is simply because excess calcium combines with carbonate ions to form minute calcium carbonate crystals within the cell when the solubility product of calcite is exceeded.  The solubility product is the concentration of calcium ions multiplied by that of carbonate ions, so that increase in one or the other can lead to supersaturation of calcium carbonate and imminent precipitation.  Because CO₂ is an essential need for photosynthesis and a product of animal metabolism, this risk is always present.  In the most common photosynthesising bacteria, the cyanobacteria that have been around for at least 3.6 billion years, the drawing in of CO₂ in the form of carbonate (CO₃^2-) or bicarbonate (HCO₃^–) ions in water can result in supersaturation immediately around the cell.  When it occurs, the “blue-green” bacterial biofilms induce precipitation of calcium carbonate.  That is why such micro-organisms can act as reef builders, as they did to great effect during the early Precambrian (stromatolites), and also from Cambrian to Cretaceous times.

Calcite mineralization by biofilms is, however, a complicated process.  It is connected with highly reactive substances that cyanobacteria exude outside their cell walls.  Depending on their degree of ordering and the supply of calcium ions, these substances control the manner in which calcium carbonate precipitates.  The detailed biochemistry and the form of calcite biofilms obtained by study of modern cyanobacteria in different watery environments has allowed Gernot Arp and co-workers at the University of Göttingen to evaluate varying calcium and CO₂ concentrations in ocean water since 540 Ma, and suggest differences in Precambrian oceans (Arp, G. et al. 2001.  Photosynthesis-induced biofilm calcification and calcium concentrations in Phanerozoic oceans.  Science, v. 292, p. 1701-1704).

Their studies suggest that up to the Cretaceous, the Phanerozoic oceans must have had higher calcium contents than they do today.  Microbial reefs formed in that period preserve details of the “blue-green’s” cell structure, suggesting that calcite was nucleated directly by the extracellular substances.   Vast burial of the calcite shells of planktonic metazoan organisms to form the Chalk deposits of Cretaceous age reduced very high levels to give the calcium-depleted oceans that prevailed during the Cainozoic.  Microbial carbonates of these younger ages show no structure.  The stromatolites that are so characteristic of Precambrian limestones are stuctureless too, although they show evidence of progressive build-up from myriads of thin layers.  Irrespective of the Precambrian oceans’ calcium content, this lack of structure can be explained by more dissolved CO₂ that resulted from its higher concentration in the atmosphere.  About 700-750 Ma ago, stromatolites that contain calcified cyanobacterial cells appear, and that may signify the massive drawdown of CO₂ from the atmosphere that is implicated in creating icehouse conditions on a global scale during the late Proterozoic Aeon.

Because climate depends partly on the retention of solar heat by carbon dioxide in the atmosphere, a record of past CO₂ fluctuations is important in linking evidence for shifting climate and environments to models.  Conversely, models that seek to mimic climates of the past depend heavily on the assumption that the “greenhouse” effect and the carbon cycle underpin global temperature and precipitation.  Current theorists consider that shifts in CO₂ content of the atmosphere reflect a balance between its release through volcanism (itself a reflection of the rate of plate tectonics) and  its removal by weathering of silicate minerals and burial of dead biomass. 

The GEOCARB III model predicts rising atmospheric CO₂ following the ice-house condition of the late-Precambrian, when rapid sea-floor spreading broke up and began to reassemble supercontinents during the Lower Palaeozoic.  In the early Cambrian CO₂ levels come out at 25 times the modern amount.  Colonization of the land by plants through the Upper Palaeozoic, and the burial of a proportion of the increased amount of carbon fixed by them, allows the model to predict a massive fall in CO₂.  That tallies very well with the long period of glaciation in southern Pangaea during the Carboniferous and Permian.  GEOCARB III suggests a recovery in levels through the Mesozoic, punctuated by extraordinary releases from plume activity, such as that implicated in the formation of ocean plateaux beneath the Pacific about 120 Ma ago.

From GEOCARB modelling stem predictions of the overall forcing of global temperatures.  However, only the last 100 Ma can be assessed as regards temperatures, by using accurate proxies provided by oxygen isotopes and the Ca:Mg ratio of marine carbonates.  Two of the leading climatic theorists, Thomas Crowley and Robert Berner of Texas A&M and Yale universities usefully summarise the range of other proxies that help validate their kind of modelling (Crowley, T.J. and Berner, R.A. 2001.  CO₂ and climate change.  Science, v. 292, p. 870-872).  These include estimates from fossil soils, carbon isotopes in sediments, the pores in plant leaves (see Plant respiration and climate below) and how much boron is taken up in the shells of fossil animals.  There are considerable discrepancies with modelling, albeit encompassed by the high uncertainties in the calculations.  Crowley and Berner acknowledge the complexity of other factors that affect the global redistribution of heat, such as continental configurations in terms of area, geographic position, their effects on ocean circulation and even on the pace of the carbon cycle.  They see the need to expand climate models, taking other factors on board, in an attempt to quantify the discrepancies.

Methane and escape from Snowball Earth

Palaeomagnetic pole positions determined from areas characterized by thick glacigenic deposits around 750 Ma old leave little doubt that large volumes of ice covered the Earth to tropical latitudes.  Such evidence suggests an ice-bound world from which escape would have been very difficult because much of the Sun’s energy would have been reflected back to space.  Extreme and prolonged frigidity, from which Earth’s climate did escape is seen by a growing number of palaeobiologists as the most profound influence over later evolution and diversification of life.  The first fossil metazoans appear in the record shortly after a “Snowball Earth” event at 650 Ma, and the Cambrian explosion of animals with hard parts followed close on the heels of the last.  Carbon isotope studies from marine carbonates suggest that each global glaciation witnessed massive extinctions of single-celled organisms, and surviving life was presented with a virtual tabula rasa of niches to fill.  Such survivors, possessing characters that had ensured their survival – at which we can only guess – exploited them to the full.  It is reasonable to speculate that without such climatic upheavals life would not be as it is now, and that our eventual appearance depended on them.

That Earth’s climate broke out of runaway ice-house conditions is obvious, the question being how was that possible.  Volcanic emissions of carbon dioxide, which neither the Neoproterozoic biosphere nor silicate weathering were able to draw down into ocean water and sediments, would have accumulated in the atmosphere, to create “greenhouse” conditions.  That simple scenario, envisaging a spectacular shift from frigid to hot conditions, has its problems.  In order for climate to stabilize, without rushing into runaway heating along the path followed by Venus, demands implausibly high rates of silicate weathering to draw down CO₂ in the period following the end of each “Snowball” event, and strontium isotopes that record the rate of continental weathering shwo no sign of anything so dramatic.  It also poses the question of how global ice cover could remain while CO₂ slowly built up.  The key seems to lie in carbonates that everywhere cap the glacigenic deposits of this age.  The cap carbonates record rapid falls in the ¹³C proportion of the carbon in carbonate.  ¹³C shows a rise in the glacial epochs that signifies massive burial of dead organic matter (enriched in lighter ¹²C), probably through mass extinction.  In a review of the geochemical basis for changes in oceanic carbon isotopes, and high-resolution data from cap carbonates, scientists from the University of California and the Lamont-Doherty Earth Observatory, suggest that the isotopic excursions could reflect massive release of methane from gas-hydrate layers in sediments that were frigid during the Snowball event (Kennedy, M.J. et al. 2001.  Are Proterozoic cap carbonates and isotopic excursions a record of gas hydrate destabilization following Earth’s coldest intervals?  Geology, v. 29, p. 443-446).  Backing up this hypothesis are examples of structures in cap carbonates that are identical to those formed in modern sediments affected by break down of gas hydrates and release of methane from the sea floor.

Plant respiration and climate

Leaf surfaces are pockmarked by pores (stomata), through which cell metabolism draws in the carbon dioxide involved in photosynthesis and transpires its products, including oxygen.  When CO₂ levels are low, more pores are needed, and vice versa.  Surprisingly, museum specimens of leaves collected since the start of the Industrial Revolution do show a decrease in the density of such pores that matches the documented rise in atmospheric CO₂ levels.  Were it possible to find fossils of the same plant species, pore density would be an excellent proxy for the “greenhouse” effect.  That is not possible, because of evolution.  However, plants related to the Ginkgo have a pedigree that goes back about 300 Ma.  Morphologically, the four genera of Ginkgo-like leaves are very similar, so using them potentially gives an independent record of the “greenhouse” effect.

Gregory Retallack of the University of Oregon has measured the stomatal index of sufficient Ginkgo and related leaves to assess CO₂ levels in a broad-brush sense for the period since the early Permian (Retallack, G.J. 2001.  A 300-million-year record of atmospheric carbon dioxide from fossil plant cuticles.  Nature, v. 411, p, 287-290).  His results tally broadly with oxygen-isotope and other proxies for palaeotemperature variations, and to some extent with CO₂ modelling (see Phanerozoic CO₂ levels above).  However, the stomatal record shows changes up to 10 Ma in advance of shifts in temperature.  That might be due to coarse resolution in Retallack’s data, but could signify other forces at work other than the “greenhouse” effect.  The most significant advance provided by leaf studies is that they help account for mismatches between evidence for cooling and predictions of highCO₂ by modelling, for the Jurassic and Cretaceous, that have been a thorn in the side of the modellers.  Given fossil leaves more closely spaced in time, and using other plant groups, Retallack’s method potentially could revolutionize climate analyses and extend them back as far as 400 Ma ago.

See also:  Kürschner, W.M.  2001.  Leaf sensor for CO₂ in deep time.  Nature, v. 411, p. 247-248.

The most favoured means whereby the weak fluctuation in solar radiation due to the Milankovich-Croll Effect become amplified to affect climate’s ups and downs is the switching on and off of thermohaline circulation in the North Atlantic Ocean.  The key to such ocean circulation is formation today of dense, cold brine through sea-ice formation around Iceland.  To set circulation in motion, however, depends on these brines being able to move southwards, which they do now in a sea-floor channel between Shetland and the Faeroe Islands.  When the North Atlantic began to open, this route was blocked by a ridge between Greenland and Shetland, buoyed up by residual warmth in the lithosphere from volcanic activity at the Iceland plume.

It is important to assess when the Shetland-Faeroe “gateway” formed, so that the effects of thermohaline circulation on pre-glacial climate can be assessed.  Petroleum exploration using high-resolution seismic reflection profiles and drilling has resolved this particular issue.  Geologists and geophysicists from Exxon and Cardiff University have found signs that sediment drift dragged by such a deep flow began in the early Oligocene (about 35 Ma ago) (Davies, R.  et al. 2001.  Early Oligocene initiation of North Atlantic Deep Water formation.  Nature, v. 410, p. 917-920).  The evidence takes the form of multiple, moat-like erosion surfaces down to the base of sediment fill between the Faeroes and Shetland, shown superbly by the seismic data.  Drilling shows that these signs of deep-water flow stop abruptly in Early Oligocene sediments.

Astrology and ice

The early Oligocene marked the onset of serious ice cover on Antarctica, and it shows as a dramatic increase in d¹⁸O values in the ocean-floor record of benthic forms – lighter ¹⁶O had been trapped in land ice.  That may or may not be a coincidence with the finding about the start of  North Atlantic thermohaline flow in the previous item.  A lesser, but still dramatic increase marks the Oligocene-Miocene boundary, suggesting further growth of the Antarctic ice sheet, which is not so readily matched empirically.  Detailed study of the isotopic  “blip” at this time by a team from the Universities of California, Cambridge and South Florida (Zachos, J.C. et al. 2001.  Climate response to orbital forcing across the Oligocene-Miocene boundary.  Science, V.  292, p. 274-278) suggests that it related to a remarkable coincidence in the astronomical record of solar heating.

Round 23 Ma ago, the orbital eccentricity dropped almost to zero – Earth’s orbit would have been circular – at the same time as its axial tilt became very stable, the one reinforcing the climatic effect of the other.  The isotopic “blip” coincides exactly with the coincidence.  The detailed record also shows very clearly that minor fluctuations in climate at that time were in step with the 400 and 100 ka periods in the eccentricity variations, and with those of 41 ka that relate to changes in axial tilt.  If nothing else, these results confirm that it is unnecessary to turn to extraterrestrial influences over climate other than those which are predictable from Milankovich’s theory (see Impacts and human evolution, above).

Additional source:  Kerr, R.A. 2001.  An orbital confluence leaves its mark.  Science, v. 292, p. 191.

Start of Pleistocene environmental change in tropical Africa

Pollen records from an ODP core drilled off the Congo estuary provide a record of the fluctuation in the monsoon of western tropical Africa (Dupont, L.M. et al.  2001.  Mid-Pleistocene environmental change in tropical Africa began as early as 1.05 Ma.  Geology, v.  29, p. 1195-198).  Before 1.05 Ma there is little sign of a glacial-interglacial pulse in the fluctuation of vegetation in the Congo Basin.  Thereafter, ups and downs in pollen from various vegetation groups correlate well with the benthic foram oxygen-isotope time series.  However there are a few surprises.

Conventional wisdom is that Africa experienced drying during glacial epochs, rain forest expanding during interglacials.  In the Congo basin, grasses and savannah trees increased during interglacials while mountain trees fell in their influence, up to 600 ka.  This suggests the opposite trend of  warm, dry interglacials and cool, humid conditions during glacial periods, similar to the record for tropical South America.  In the later Pleistocene, the fluctuation switched to that indicated by fluctuating lake levels throughout the continent.  The pollen variations are backed up by variations in dinoflagellate cysts, which show that discharge from the Congo dropped during interglacials.  The other surprise is that the onset of astronomically paced environmental change in west Africa predated the change to a 100 ka domination of global climate, and the increase in amplitude of changes in land-ice volume at 900 ka by a hundred thousand years.  Dupont et al. suggest that the changes in albedo in tropical West Africa in response to vegetation changes could have had an influence on global climate when the fluctuations began.

As well as being interesting in terms of climate change, the new data throw doubt on the hypothesized link between climate in Africa and pulses of migration of early human species, such as H. ergaster and H. erectus.  There were fluctuations in humidity in the earlier Pleistocene, but they show no link to global climate change.  So, it seems unwise simply to look to the Milankovich forcing as a pacemaker in early human affairs.

A Late-Jurassic methane “gun”

Massive releases of methane from gas hydrate layers beneath the ocean floor, and its subsequent oxidation to carbon dioxide have been implicated in major climatic and oceanographic changes in the mid-Jurassic, Cretaceous and Palaeocene.  They can be detected by drops in the ¹³C content of marine carbonates, caused by the “light” carbon trapped in biogenic methane.  All those known also correlate with evidence for climatic warming.

The Swiss Jura mountains are a repository of great thicknesses of Jurassic carbonates, whose ammonite faunas allow fine stratigraphic division.  Between 157 and 156 Ma (late Middle Oxfordian) there is a major negative excursion in d¹³C whose duration was as short as 180 ka (Padden, M. et al. 2001.  Evidence for Late Jurassic release of methane from gas hydrate.  Geology, v.  29, p. 223-226).  The Swiss-French geochemists who discovered the anomaly believe that the release may have linked to opening of the ocean gateway that connected currents between Tethys and the easter Pacific oceans through what is now the Atlantic.

The presence of abundant oxygen in Earth’s atmosphere defies Le Chatelier’s Principle – it should react rapidly with the rest of the environment through oxidation.  That it does not is sufficient evidence for an alien observer to conclude that our planet is dominated by photosynthetic life at its surface and the burial of carbohydrate by geological processes.  So, Le Chatelier is not defied on the long term, because the CO₂ + H₂O = carbohydrate + oxygen equilibrium does not reach a balance because of continual removal of organic material from the right-hand side!  That Mars has no atmospheric oxygen bears witness to its lifelessness in that respect, as concluded decades back by James Lovelock.

Before 2.5 Ga ago, in the Archaean, atmospheric oxygen was a trace gas.  Preservation of detrital grains of sulphides and uranium oxides in Archaean clastic sequences, that would have broken down in an oxidizing environment, is the main evidence for that.  The other side of the coin is that oxygen-producing photosynthesizers – the cyanobacteria – were abundant throughout the Archaean, leaving their trace as common stromatolitic carbonates and signs of the crucial enzyme rubisco in kerogens and the carbon-isotope record.

If cyanobacteria generated oxygen, then why did it not build up in the atmosphere throughout the Archaean, instead of from about 2.2 Ga ago?  The most likely explanation is that Archaean magmatism released vast amounts of Fe-II or ferrous iron to sea water, which then reacted with available oxygen to form the ferric oxide of banded iron formations (BIFs), with the biproduct of hydrogen gas that further drove Archaean environmental chemistry into a reducing condition.  Seawater circulating through Archaean ocean crust would also have enriched basalts in ferric iron by the same oxidizing reaction.  Such a chemical model still leaves unexplained the shift to an oxygenated atmosphere after the Archaean.

Norman Sleep of Stanford University, reviews an article by Kump et al. in  Geochemistry, Geophysics, Geosystems (2001) that deals with this dilemma (Sleep, N.H.  2001.  Oxygenating the atmosphere.  Nature, v. 410, p. 317-319).  Kump and his co-workers suggest that, rather than relating to a change in palaeoecology, the shift arose from subduction of dense ferric oxide-rich lithosphere to settle at the core-mantle boundary.  By the end of the Archaean oxidized material filled the lower mantle.  Heating reduced its density so that it became buoyant.  If that deep oxidized layer rose to displace more primitive, reducing mantle, later magmatism would have released less Fe-II, thereby allowing biologically generated oxygen to build up.  The converse effect would have been to bring down levels of reducing atmospheric gases, such as hydrogen, methane and carbon monoxide, to trace levels.

Except to its primitive producer – cyanobacteria – oxygen would have been anathema to the dominant anaerobic Bacteria and Archaea that constituted Archaean life.  An end-Archaean mantle overturn, implicated by the tectonic pandemonium from 2.7 Ga, could well have triggered accelerated extinction and evolution that encouraged the rise of the eukaryote cell that requires oxygen for its basic metabolism.  Nonetheless, such an upheaval would have been directly connected with earlier living processes.  That is something which will delight followers of the Gaia hypothesis.

Climate researchers at MIT in Cambridge, Massachusetts have analysed Northern Hemisphere climate data from 1972 to 1999, in the search for correlations that might help improve long-term weather forecasting.  The most striking match to emerge is that of winter climate with the extent of autumn snow cover in Siberia.  Snow reflects back to space a far greater proportion of incoming solar energy than any other kind of surface, with the exception of salt.  More snow results in less warming in the area.  Although Siberia is at the heart of the Asian continent, and therefore pretty dry, it has cold winters, so that when snow falls it covers large areas and tends to remain.  It is the focus for an enormous mid-continent high-pressure area in winter, appropriately named the Siberian High, which is one of three systems that dominate northern climate.

High-pressure areas do two things: air spills from them into surrounding areas; they isolate the area beneath them from warming, moist winds blowing from the oceans.  In winter the second creates cooling so intense that temperatures can steadily drop to -50°C or below , further building pressure because of the increase in air density.  Siberia sheds cold air westwards into Europe and over the North Pole into North America.  The MIT study bears out the obvious prediction based on this tendency.  However, it may also add the Siberian High to the range of large-scale terrestrial processes – shifts in air pressure over oceans, such as the El-Niño of the tropical Pacific and the North Atlantic Oscillation, and thermohaline controls over Atlantic surface currents – that make ice-age climate patterns so complex.

Cooling of northern Europe and the Canadian Shield does not have to be very extreme to lower the topographic elevation at which snow remains permanently, the glaciation limit – at present that level is only a couple of hundred metres above the tops of Britain’s highest mountains.  Should permanent snow cover return to the highest areas around the North Atlantic, that would amplify the present effect of Siberian autumnal snow and expand the high-pressure area.  That is a positive feedback driving climate towards increased frigidity, and larger winter highs would hold back maritime warming influences.

Computer modelling of the air-flow patterns over Asia shows that the primary influence is the Himalaya and Tibetan Plateau.  In particular, they dry out air passing over them during the South Asian Monsoon, and hinder its influence further into central Asia.  The two huge massifs seem to have risen rapidly and recently, beginning about 8 million years ago, despite the fact that India collided with Asia about 50 million years ago.  Together with other roughly E-W high mountain ranges in central Asia, they also channel Siberian cold air to spill westwards and eastwards, and over the pole.  Behaviour of the Siberian High almost certainly dates from the uplift of the Himalaya and Tibetan Plateau.

Adding another controlling factor to long-term northern climate has an intrinsic potential in refining academic studies of Pleistocene climate.  However, there is an immediacy to the observations.  For snow to cause cooling by reflecting away solar heat it does not have to be thick; a few centimetres will suffice.  The critical factor is the area covered by it.  Siberia is so cold in autumn and winter that it will snow there, provided moist air can enter.  Should more get in then more snow will cover a greater area, to feed the positive feedback to cooling.  Perversely, the more the climate warms globally, the more moisture evaporates from tropical and mid-latitude oceans to move polewards and towards continental interiors……

Mismatches from north to south proven

Whether or not climate changes, especially those of shorter duration than the full glacial-interglacial cycle, occur at the same time everywhere is something that vexes all climatologists.  It encapsulates all the problems of causation: orbital forcing, thermohaline circulation, shifts in the Polar Front and Intertropical Convergence Zone, etcetera.  The problem mainly stems from uncertainties in the correlation of  time series that show proxies for climate change.  This is particularly bad for ocean-floor sediment cores, which depend upon radiometric dates for calibration from depth to time sequences and an assumption of constant rates of sedimentation between dated samples.  Imprecision often means that correlations are not believable, except at a very general level.  Many analyses end up by correlating the patterns shown by the proxies, which defeats the object of assessing the degree of global synchronicity of climate changes.

Cores taken through ice sheets offer a way out, for annual layers of ice are there to be counted, but only in the upper parts.  For deeper parts, converting depth to time relies on models of how ice compacts and how it thins by glacial flow.  Another seeming advantage of ice-core records is that a great deal more ice accumulates than does ocean-floor sediment over a particular time.  That means that the resolution of ice core records can be finer – potentially at the level of decades compared with hundreds of years for sediment cores.  A seeming key to correlation between ice cores lies in the way that ice traps air.  Being rapidly mixed, the atmosphere should have the same composition everywhere.  This is particularly so for methane, partly because it soon becomes oxidised to carbon dioxide, and partly because its level is highly variable from emissions by rotting vegetation and unstable gas hydrate on the shallow ocean floor.  Thomas Blunier and Edward Brook of Princeton University and the University of Berne used the methane records of Greenland and Antarctic ice to correlate the other proxies therein over the last 90 thousand years (Blunier, T. and Brook, E.J. 2001.  Timing of millennial-scale climate change in Antarctica and Greenland during the last glacial period. Science, v. 291, p. 109-112).  They show a consistent mismatch between rapid warmings of the air over the two polar ice sheets, where Antarctic changes precede those over Greenland by 1500 to 3000 years.  Interestingly, when frigidity gave way to comparative warmth in a matter of a few decades over Greenland, the Antarctic was shifting from warm to cool conditions.

Commenting on the paper in Science’s Compass, Nicholas Shackleton of Cambridge University shows yet more emerging oddities (Shackleton, N. 2001.  Climate change across the hemispheres.  Science, v. 291, p. 58-59).  In the North Atlantic Ocean, surface water temperatures apparently changed according to Greenland’s pace, while those for deep water match that of the Antarctic.  To add to the complexity of climate change through the last glacial period – until a few years ago it was all supposed to link to the astronomical forcing of solar heating at high northern latitudes – the oxygen isotope changes in the same deep water of the North Atlantic match those of ice volume around the north pole.

Whereas Blunier and Brook have proved that air-temperature changes above ice sheets at high northern and southern latitudes are not synchronous, this still leaves problems in correlating between ice and sediment cores, and between the oceanic record at the many sites world-wide, especially those at low latitudes.  With a growing number of hypotheses for climate changes of the order of a few thousand years – driven by changes associated with northern ice sheets, Antarctica and the tropics – onlookers await with interest the development of a means of precise correlation among all the time series.

The Cariaco Basin off Venezuela lies in an area that is sensitive to climate change and has been for at least the last 90 thousand years.  As the trade winds change with the seasonal migration of the Intertropical Convergence Zone (ITCZ), cold, nutrient-rich waters well up along the coast of northern Venezuela.  Biological productivity waxes and wanes on an annual rhythm, as too do sediments transported into the basin by the great rivers of this part of South America – shifts of the ITCZ also impose annual wet and dry seasons over land.  This cyclicity seems to have functioned since at least 90 ka ago, and drill cores from the Cariaco trench are dateable at the annual level because of the colour banding of seasonal sediments (Peterson, L.C. et al.,. 2000.  Rapid changes in the hydrological cycle of the tropical Atlantic during the last glacial.  Science, v. 290, p. 1947-1951).

Matching the varying thicknesses of colour bands beneath the Cariaco Basin to the high-latitude climate record preserved in Greenlandic ice-cores shows a remarkable correlation.  Warming (interstadials) over Greenland correspond to periods of increased rainfall and ocean bio-productivity (the layers are thicker) off Venezuela.  Peterson and his co-workers believe that this could signify periods of greater transport of water vapour from the Atlantic to the Pacific.  That would increase the salinity of the Atlantic.  Working through to high latitudes, saltier surface water would more easily become dense cold brine once sea ice had been frozen from it.  That would enhance the thermohaline deep circulation of the North Atlantic, so that warm, tropical waters might be dragged further to the north during interstadials, in the manner of today’s Gulf Stream.  It is hard to see how just melting ice sheets during interstadials could do that; in fact that would encourage a further reduction of deep circulation.  So, a tropical connection seems plausible.  However, interstadials stopped extremely rapidly, repeatedly plunging high latitudes into full glacial, or stadial conditions.  That may well have been an outcome of all the fresh water from melting glaciers acting to dilute surface waters’ saltiness, and thereby shutting down thermohaline processes.

The annual precision of sediment cores from the Cariaco Basin carries a bonus, by helping better to calibrate ¹⁴C dating.  Radiocarbon dates have long been known not to correspond predictably to calendar years.  For instance, dates from around 11 ka ago, the time of the last major glacial advance (the Younger Dryas) show a mismatch of about a thousand years between dates based on counting tree rings and annual ice layers (exact calendar years), and those provided by ¹⁴C dating of carbon-rich samples.  The reason for this is partly fluctuations in the production of ¹⁴C by bombardment of nitrogen atoms in the stratosphere by cosmic radiation and the solar wind.  The Cariaco Basin layering extends calendar dating at least 5 000 years further back, into the period when deglaciation accelerated as the Earth’s climate emerged from the last glacial maximum (Hughen, K.A. et al., 2000.  Synchronous radiocarbon and climate shifts during the last glaciation. Science, v. 290, p. 1951-1954).  That helps to evaluate shifting rates of ¹⁴C production over this part of the core (maybe related to varying solar output because they match shifts in ¹⁰Be, also produced by upper atmosphere processes), and to add meaning to radiocarbon dates from it.  However, not all the shifts in ¹⁴C can be due to solar fluctuations, and it is clear that the largest, during the Younger Dryas event, stemmed from increased carbon preservation on the ocean floor, that removes all isotopes of such carbon from the atmosphere and upper ocean.  This supports the notion that the Younger Dryas, and perhaps all the stadial-interstadial events of the last 90 ka stem from changes in ocean processes.

Several gases have the property of absorbing radiation in the wavelength range emitted by the Earth because of its surface temperature, including methane as well as carbon dioxide, the usual culprit.  By doing so, they delay the escape of thermal energy through the atmosphere to outer space and give the Earth a higher surface temperature than it would have if they were not present.  Because methane oxidizes to CO₂ more rapidly than the latter gas’s recycling time, a record of atmospheric carbon dioxide is the best guide to fluctuations in the “greenhouse” effect through the past glacial-interglacial cycles.  Bubbles in cores through the ice sheets of Greenland and Antarctica trapped air at the time when snow converted to ice within a few decades after it fell in polar regions.

The publication of data of all kinds from the ice-core drilled beneath the Vostok camp in Antarctica (see Earth Pages archive – Milankovic forcing flawed? July 2000) opened up 420 000 years worth of atmospheric composition shifts.  Daniel Sigman and Edward Boyle, of Princeton University and MIT, Massachusetts, review all the bio-geochemical factors that might have contributed to the CO₂ time series for the last 4 major climate cycles (Sigman, D. M. & Boyle, E.A.  2000.  Glacial/interglacial variations in atmospheric carbon dioxide.  Nature, v. 407, p. 859-869).

While work continues to fully grasp this climate forcing function, Sigman and Boyle argue convincingly that the overwhelmingly dominant influence on it is the combined biological and physical carbon “pump” of the ocean around Antarctica.

News from the South

Increasingly, evidence of many kinds points to a dominant influence on climatic ups and down through the last 2.5 Ma by processes in the northern hemisphere.  Empirically, at least, the global-climate time series seems to show patterns that closely resemble Milankovic’s predictions of varying insolation at high northern latitudes.  For millennial-scale fluctuations, such as Heinrich events and the Dansgaard-Oeschger cycles in ocean and ice-sheet cores respectively, the focus is on changes in deep-water formation in the North Atlantic.  The South cannot be set aside, however, and there are two important issues that crop up in October’s publications.  One is the extent to which climatic events in the southern hemisphere tracked those in the North, and the other is the role of the southern oceans in the global carbon cycle that underpins the climate-related fluctuations in atmospheric CO₂ concentrations.

Both the Greenlandic and Antarctic ice cores show synchronicity of CO₂ trapped in air bubbles with the records of local air temperature and global land-ice volume, going back over 400 ka in Antarctica.  With more or less constant additions from volcanism, the ups and downs of the primary “greenhouse” gas have to be mediated by removal of carbon in one form or another from the ocean-atmosphere system through the agency of biological processes.  Just what process, where it is most active and the controls underlying it form a topic of continual discussion and research.  One possibility is variation in the biological productivity of the open oceans, coupled with removal of carbon from the ocean-atmosphere interface.

In terms of size and potential, the Southern Ocean is overwhelmingly the most likely candidate for a control.  It is today the largest repository of unused nutrients in surface waters (by comparison with its potential for supporting phytoplankton it is a “wet” desert), but also a major source of deep-water formation that could sequester carbon from the surface environment.  The late John Martin suggested that the main control over ocean productivity is soluble iron, currently at low concentrations far from land.  The first realistic experiment to verify this involved “seeding” a small area of the equatorial Pacific with iron sulphate in 1995.  Sure enough, that provoked a short-lived bloom of microscopic marine plants and local changes in dissolved CO₂, but a boost in productivity at low latitudes is unlikely to lead to carbon removal from the surface part of the C-cycle.

Eighteen months ago, a multinational team of 35 ocean scientists conducted a similar experiment off Antarctica at 60°S (Boyd, P.W. et al. 2000.  A mesoscale phytoplankton bloom in the polar Southern Ocean stimulated by iron fertilization.  Nature,  407, 695-702.  See also: Chisholm, S.W. 2000.  Stirring times in the Southern Ocean. Nature,  407, 685-687).  Once again bio-productivity soared by three times, and an input of 9 t of ferrous sulphate into about 50 km² of ocean triggered an estimated 600 to 3000 t of extra algal carbon production.  The “bloom” lasted for at least 6 weeks, being transformed into a swirling ribbon 150 km long.  But it did not seem to be absorbed into deep water, merely mixing at the surface.  In principle, iron dissolved from dust blown far from land during cold, dry episodes might have drawn down CO₂ levels, but it is still uncertain.  Yet the dust records trapped in ice cores do show a pronounced negative correlation with both CO₂ and climate proxies.

Millennial-scale climate shifts are best known from the area around the North Atlantic.  The most recent of these, and the most dramatic, was a sudden reversal from the warming trend out of the last glacial maximum around 13 ka ago, which lasted around 1800 years.  This is recorded in many ways everywhere around the North Atlantic, and takes the name Younger Dryas (YD) from the associated increase in sediment cores of pollen of the cold-resistant mountain avens (Dryas octopetala).  For some years there have been reports of a YD signal in climate records from the southern hemisphere, and some suggesting it was not felt there at all, the most detailed counter-evidence being the lack of the YD signal in Antarctic ice cores (ascribed by some to climatic inertia of the ice-bound continent).

The YD interrupted warming and wetting in the lead-up to the Holocene interglacial, so its signal ought to be easy to verify or rule out, simply because no later glacial advances have obliterated suitable investigation sites and many lakes at high altitudes and latitudes formed about that time.  The problem for southern-hemisphere work has been a lack of precise dates.  Southern Chile proves to be an excellent place to check, because lakes there go back further and contain evidence for many glacial advances and retreats (Bennett, K.D. et al. 2000.  The last glacial-Holocene transition in Southern Chile.  Science, 290, 325-328).  Moreover, the sediment cores provide sufficient high-precision dates to construct a believably detailed time scale.  Bennett and co-workers show that during the YD Chilean glaciers were retreating rather than advancing.  That seems to knock the idea of “teleconnections” spanning both hemispheres for this particularly dramatic event, although its signal extends to the north Pacific.  Like the mountain avens, however, disputing palaeoclimatologists are a hardy lot.  It could be that the site of Bennett and colleagues work was far from a boundary between pollen-shedding species that was sensitive to climate change, despite the excellence of their record (see also Rodbell, D.T. 2000.  The Younger Dryas: cold, cold everywhere?  Science, 290, 285-286).

No escape from global warming?

Palaeoclimatology is well-funded because it is believed to shed light on the likely consequences of anthropogenic warming caused by CO₂ emissions, and perhaps even technical solutions that allow us to continue burning fossil fuels.  There is no doubt that throwing money at the range of associated phenomena and data has produced many astonishing findings and connections for the last 2.5 Ma.  There is now sufficient high-quality data to reviewing them in their proper context; that of the climatic aspect of the “human condition”.  That is the task that yet another multinational group of scientists set themselves at an International Biosphere-Geosphere Programme (IBGP) workshop at the Royal Swedish Academy of Science in November 1999 (Falkowski, P. and 16 others 2000.  The global carbon cycle: a test of our knowledge of earth as a system.  Science, 290, 291-296).

The workshop used two generalized outcomes of many years of work on Antarctic ice cores: the variation over more than 400 ka of CO₂ in trapped air bubbles with temperature shifts; the frequencies and amplitudes of changes in atmospheric CO₂.  They compare these with human effects over the last 200 years.  A great deal of discussion and qualification surrounds the workshop’s conclusions, but they are stark and simple.  Anthropogenic change falls way outside that induced by natural processes (whatever they are), and its period bears no relationship to those involved in short- to long term processes.  Despite the seeming attraction of technical fixes, such as boosting ocean productivity and the deep-water carbon sink (above), and intervention in terrestrial plant processes to increase CO₂ sequestration from the atmosphere, both face the likelihood of weakening natural feedbacks due to the massive change that has taken place.  Indeed, the consequences of strategies of these kinds aimed at mitigating climate change cannot be known in advance.  This grim conclusion stems from the fact that no matter how well we get to know the climate system of the past, it is no longer what it was.  Even a complete halt to all anthropogenic emissions now cannot reverse the trend in the short to medium term.

The group suggests Earth’s entry into a new Epoch (the Anthropocene) of uncertainty, but brimming with growing knowledge.  To them, this seeming paradox must not be “used as an excuse to postpone prudent policy decisions based on the best information available at the time”.  They also highlight the disciplinary compartmentalization of research that hinders a “proper” understanding of the Earth system. I suppose what they are getting at is the continuing ethos of Descartes’ 400-year old reductionism in science, yet surely their call for a “systems approach” is merely dressing up reductionist empiricism in a more complicated guise; hurling yet more intricate maths at the problem.  That is indeed the goal of climate modelling and has been for well over a decade.  Perhaps the solution lies not in descriptive retrospection by scientists and in “policy”, but with society as a whole that now begins to confront the mismatch between several thousand years of divided human activity with the rest of the world.

Daniel Sigman and Edward Boyle, of Princeton University and MIT, USA, usefully review the whole issue of varying CO₂ through the 420 ka Antarctic ice-core record, together with its environmental buffering (Sigman, D.M. and Boyle, E.A.  2000.  Glacial/interglacial variations in atmospheric carbon dioxide.  Nature,  407, 859-869).  Their article helps see the views of Falkowski et al. from a broad and detailed context, and links to News from the South (above), because Sigman and Boyle conclude that while the pacing of climate change tracks the combined effects of orbital processes on solar energy input at high northern latitudes the “greenhouse” effect  changes because of biological and physical processes in the Southern Ocean that surrounds Antarctica.

The nudge of noise

The emergence of a signal in the climate shifts through the last ice age and the Holocene with a roughly 1 000 to 1 500 year period (see Earth Pages Archive, A new regular pulse in recent climate, September 2000) finds no link with processes linked to Earth’s orbital behaviour.  It must be generated within the Earth system itself.  That being said, there is a lot of debate over what precisely is involved.  It’s safe to say that debate will continue.

However, another factor might well be involved; one that is as much to do with statistics as with phenomena with sufficient power to flip climate patterns.  Random noise is everywhere in nature.  If strong enough at a critical time, such stochastic noise might resonate with an otherwise weak, periodic phenomenon to give it sufficient push that it shows up in a climate change.  Let’s say that there is some weak pulsation that bears on climate – not really known with certainty, but having a 1 500 year period.  If resonance with noise was involved, we might expect to see 1 500, 3 000, 4 500 year periods in the climate record (1-, 2- and 3-cycle shifts), with the first more common than the last two – that is how the statistics should work.  The fact that short-term climate pulses (the stadial-interstadial events) cluster around 1 000 to 1 500 years might indicate that random noise is implicated.  However, only the last 120 000 years of climate data have sufficient precision for such statistical analyses, so it might be fortuitous.

The same nudge of randon climatic noise has also been called on to explain the jump from a roughly 41 000 year cyclicity to the present one of 100 000 years about 700 000 years ago.  The first correlates with the period of changes in the Earth’s axial tilt, and the second with changes in the eccentricity of its elliptical orbit. The effect of orbital variations on the energy received from the Sun is so very small that it cannot have much of an effect on climate by itself, but changes related to axial tilt are ten times bigger.  The change in behaviour seven ice ages ago is therefore hard to explain, without the nudge of noise.

Source:  Kerr, R.A.  2000.  Does a climate clock get a noisy boost.  Science, v. 290, p. 697-698.

Various geochemical signals show that the Palaeocene-Eocene boundary (at 55 Ma) was a time of global warming superimposed on the general Cainozoic cooling from the ‘hothouse’ of the Cretaceous Period.  Some also point to an enhanced ‘greenhouse’ effect driven by massive methane release from gas hydrates on the sea floor.  Methane, a ‘greenhouse’ gas in its own right, oxidizes to CO₂ in the atmosphere, transferring its carbon that eventually ends up in the shells of marine organisms.  It is the carbon-isotope blip at the P-E boundary that points to methane as a source of the warming.  Not only does it appear in the marine C-isotope record from foraminifera shells in cores, but also in the teeth of terrestrial mammals, which means that the carbon reservoirs of both atmosphere and seawater were globally changed.  Using the magnitude of that signal allowed palaeoclimatologists to estimate the amount of methane released – about 1 500 billion tonnes.  On a millennial scale, that is comparable to a rate of warming similar to that currently induced by human activities.

The P-E boundary marks the most dramatic biological changes since the mass extinction 10 million years before at the Cretaceous-Tertiary boundary.  But its underlying control is sufficiently close to what is happening to climate now to form both an object lesson and a means of modelling what may happen if current emissions continue.  One of the important aspects needing scrutiny is how such warming events come to an end.  British and American oceanographers have taken a look at the P-E record in ocean sediment cores, and believe they have come up with an answer, at least in part (Bains, S., 2000.  Termination of global warmth at the Palaeocene/Eocene boundary through productivity feedback.  Nature, v. 407 14 September 2000, p. 171-174).

Most such studies focus on oxygen- and carbon-isotope records in the carbonate of foraminifera shells, revealing ups and downs in seawater temperature and volume of land ice, and of biological productivity and releases of ‘greenhouse’ gases.  Unfortunately, neither isotopic record properly resolves the alternative contributions to variation.  Santo Bains and colleagues add another parameter that helps resolve the influence of biological productivity in the oceans.  Marine organisms, especially plankton, either precipitate barium sulphate (barite) in tiny crystals within their cells or induce its precipitation once they die and decay.  Because barite is not prone to much change by later events on the sea floor, counting its crystals in marine cores is a reliable proxy for the varying abundance of plankton through time.

One strong possibility during major warming events is that ocean circulation becomes sluggish, perhaps stopping altogether.  That slows the re-supply of nutrients to sunlit upper layers, and works to reduce photosynthetic life in the oceans.  The barite record produced by Bains et al. shows the opposite for the P-E events.  For about 40 000 years after the P-E event biogenic barite rose to more than twice its normal abundance.  The ocean biosphere responded to the methane blurt by blooming.  Why it did so is not yet clear, but such a spurt in drawing CO₂ into living and dead and buried tissue would work to reverse the warming event.  The barite peak coincides exactly with the oxygen- and carbon-isotope records’ features that signify temperature and the influence of isotopically light carbon from methane released by gas-hydrate breakdown.  It might seem as if life did regulate climate in a geologically rapid manner following the P-E event, to the delight of Gaians.  However, the control over biological productivity is ultimately nutrients, and life has little influence over their supply to the oceans.  Among the possibilities for an essential nutrient bonanza, and increased circulation of the oceans is definitely ruled out during major warmings, are hugely increased rainfall to wash terrestrial sediments and dissolved matter into the oceans, and increased volcanism that would supply fine ash to the distant ocean surface.

Converging on an explanation for the end of a period of global warming is far from showing how this might be achieved for a warming induced by human activities.  That might well prove eventually to be a life-or-death necessity for our species, bearing in mind that the P-E warming was a fatal crisis for many land mammals of the time.

See also: Schmitz, B., 2000.  Plankton cooled a greenhouse.  Nature, v. 407 14 September 2000, p. 143-144.

A new regular pulse in recent climate

Gerard Bond of the Lamont-Doherty Earth Observatory at Columbia University, Palisades, New York has taken his analysis of high-frequency climatic shifts in the last glaciation into the Holocene record.  Previously, Bond had tried to make sense of the sharp fluctuations of the order of a few thousand years that are seen as gravel layers in the uppermost levels of sea-floor cores and in the oxygen isotope records of cores through the Greenlandic and Antarctic ice sheets.  The first signs of short ups and downs in climate were the coarse layers first found by Hartmut Heinrich in the glacial part of the sea-floor record.  Heinrich ascribed them to periodic releases of iceberg armadas as the ice sheets of the last glaciation became unstable.  Bond’s latest work also focuses on Heinrich events, but he has used specific lithologies as markers rather than merely grain-size variations.  In particular, hematite-stained quartzo-feldspathic materials seem likely to have come from altered rocks in east Greenland and Svalbard, far distant from the drill sites whose cores he has examined.  The proportion of reddish grains varies systematically in the cores, some layers coinciding with Heinrich events, but there are many more.  The layers appear roughly every 1500 years.  This periodicity coincides with cycles of dust blown from the Sahara to form layers in cores from the west African coast, so whatever the pulses represent, they are global signals.

Interestingly, the cycles show little sign of change in the period after the melt back that signified the beginning of the Holocene interglacial.  Behind the long-term climatic shifts in glacials and interglacials, that coincide with the 100, 41, 23 and 19 thousand year fluctuations in solar warming of the northern hemisphere, some other process must be put-putting in the background.  The 1500 year cycles may stem from processes that shift heat in the oceans and atmosphere.  A likely candidate is the production of deep currents by sea-ice formation in the northern North Atlantic.  However, detailed calculations of tides suggest a similar pacing that might change the mixing of surface and deep water in the ocean conveyor system.

Whatever the driving force, this periodicity strikes a chord with emerging details of Holocene climate changes from lake-sediments studies and the historic record.  One such recent cooling pulse that might have delivered icebergs to mid-latitudes in the North Atlantic was the Little Ice Age that peaked in the 17^th century that saw prolonged stresses on the population of Europe, and major political changes that resulted from such events as the Peasants’ Revolt and repeated famines.

Source:  Pearce, F.  2000.  Feel the pulse.  New Scientist, 2 September 2000, p. 30-33.

The 30- or more year long debate about anthropogenic climate change resulting from the ‘greenhouse’ effect of carbon dioxide releases by fossil-fuel burning has grown sharper in recent years.  Some specialists have cast doubt on climatologists’ ability to unravel human effects on the undoubted rise in mean surface temperature over the last 150 years from underlying fluctuations that stem from natural processes.  Considering the number of forcing factors, both large and small and with different periodicities, such doubts are valid, even though they may well be overemphasised in order to support continued and rising use of petroleum and coal.

The climate record for the last millennium is known to have been one of considerable change, involving a mediaeval warm period during its first half and the so-called ‘Little Ice Age’  from around 1600 to the mid 19^th century.  Even within the recent warming trend there have been climatic ups and downs in the northern hemisphere, such as the mid-20^th century warm period and several documented examples of cooling associated with major volcanism than punched aerosols into the stratosphere.

Thomas Crowley of Texas A&M University (Crowley, T.J., 2000.  Causes of climate change over the past 1000 years.  Science, v. 289, p 270-277) has attempted to isolate the known natural forcing functions and model their individual effects on mean surface temperatures in the northern hemisphere, to isolate meaningful signs of human effects from the various records temperature changes.  Even charting the changes is no simple task, as most records are local to regional, rather than valid for the whole hemisphere.  The most comprehensive climate data model uses proxy indicators from ice cores, tree-ring studies and coral time series, scaled to instrumental temperature records for the century from 1860-1965.  Uncertainty increases backwards with time, to around + 0.3°C in year 1000 AD.  This is somewhat greater than the fluctuations recorded in the temperature reconstructions, that Crowley and others have derived by a complex statistical method that fits a smoothed trend to the temperature fluctuations modelled from proxy data.

The approach used in Crowley’s analysis is to identify each likely, natural factor that influences energy balance in the northern hemisphere, and then to model the trends that they produced, and should continue to be producing.  There are two factors that are significant in millennial to shorter timescales: influences of volcanic aerosols, as timed by ash layers in ice-sheet cores; and variations in solar output based on ¹⁰Be and ¹⁴C variations in ice cores and tree rings (solar radiation generates both at the top of the atmosphere).  Possible anthropogenic forcing factors are numerous, including industrial aerosols and emitted gases other than the usual suspect, carbon dioxide.  The end product is a temperature time series which removes the effects of all known forcing factors, except those connected with ‘greenhouse’ gases, from the northern-hemisphere temperature model.  Until the early 19^th century, this hovers close to zero variation, and then starts an upward rise to a value of about 0.75°C by now.

Crowley’s modelling adds significantly to the case for a detectable economic influence on climatic warming, by showing that known natural forcings cannot account for the rising trend since the start of the Industrial Revolution.  It does not prove the case, and leaves several features in recent climatic change unexplained, notably the cooling in the late 19^th century.  Also, this approach does not exhaust all possibilities involved in climate shifts, such as linked fluctuations in energy movement by ocean-atmosphere circulation that work to make some regions experience cooling while others warm.  Such processes might respond to variations in solar-energy input by a kind of complex resonance, which remains to be looked at.

See also 20 percent more oil in the ground

Methane hydrate – more evidence for the ‘greenhouse’ time bomb

Where ocean water is more than 400 metres deep and bottom temperatures fall below 1 to 2°C methane and water can freeze to form crystals of methane hydrate.  These efficiently absorb more methane in a gas-solid solution, known as a clathrate.  Being lighter than seawater, methane clathrates do not carpet the ocean floor, but occupy pore spaces in sediments.  Under the anaerobic conditions of marine sediments, bacteria break down buried organic matter to release methane.  Build up of the gas in clathrates forms distinct reflecting horizons seen on many seismic sections of marine basins.  Estimates suggest that methane clathrates contain around the same amount of buried carbon as all fossil fuels lumped together.  Since methane is a powerful ‘greenhouse’ gas when released into the atmosphere, breakdown of the clathrates is a potential mechanism for global warming.

In 1995, evidence began to emerge that 55 million years ago, at the Palaeocene-Eocene boundary, a pulse of global warming probably stemmed from catastrophic release of methane from ocean-floor clathrates.  The signs lay in the proportions of ¹²C to ¹³C in organic matter within marine sediments of that age.  Since organisms selectively take up the light isotope of carbon, when organic matter becomes buried, seawater becomes enriched in ¹³C.  Buried carbon, both in organic molecules and in marine carbonates, takes on the isotopic signature of seawater at the time.  This means that the ups and downs of carbon burial leave an imprint in the carbon-isotope record of marine sediments.  When warming of deep-ocean water or a pressure release caused by a lowering of sea-level releases biogenic methane from clathrates, its high ¹²C content quickly appears in the carbon in seawater as a whole, by oxidation to CO₂ and solution.  This reduces the proportion of heavy to light carbon in both buried organic matter and marine carbonates, forming a downward ‘spike’ in the carbon-isotope record.  Other processes can produce the same effect, such as increased release of volcanic CO₂, which is also isotopically light, or a collapse of the marine biosphere.  So ¹²C ‘spikes’ need to be matched with other evidence.

Co-workers from Britain and Denmark have just reported in detail on just such an excursion that took place about 183 Ma ago, during a period of massive burial of carbon in the Jurassic Period (Hesselbo, S.P. et al., 2000.  Massive dissociation of gas hydrate during a Jurassic ocean anoxic event.  Nature, v. 406, p. 392-395).  The Early Toarcian was a period where circulation in the deep ocean stopped, to give anoxic conditions, ideal for burial of dead organisms.  While that lasted, important hydrocarbon-rich source rocks for petroleum reserves were laid down, and the carbon isotopes in sea water became unusually ‘heavy’.  Several stratigraphic sections show evidence for a ¹²C ‘spike’ in the middle of this period of general ¹³C enrichment of the oceans.  Hesselbo and co-workers isotopically analysed Toarcian mudstones exposed on the Yorkshire coast in England, which contain both marine matter and fragments of wood formed in terrestrial ecosystems.  Both show ¹²C enrichment, and that means that the entire carbon cycle at the time was somehow perturbed.  With no sign of massive extinction, the signature pointed either to a methane release or to hugely increased volcanism.

Although there was strong volcanic activity on Earth during the Toarcian, it came nowhere near being able to generate the anomaly.  Also the ‘spike’ occupies at most a period of a few tens of thousand years.  Only catastrophic release of methane from clathrates, equivalent to 20% of those estimated to be present today, is able to account for the anomaly.  Why it happened is nor certain; it may have been a result of either increased temperature of deep-ocean water by general global warming, to which it added, or perhaps too great a release of methane by decay in organic-rich sediments to be taken in by clathrates.  Another trigger, for which evidence is lacking at present, is through a comet impact in an ocean basin.

Methane hydrate layers in the oceans pose an ever-present threat today, because of their extreme sensitivity to temperature and pressure.  Some scientists believe that small releases may lie behind inexplicable disappearances of ships due to the drop in bulk density of seawater frothed by bubbles.  Also many areas of shallow seas are pockmarked by vents marking methane release when sea level stood lower during glacial epochs, and at least one methane spike in ice-core records can be correlated with a massive submarine landslide off western Norway.

Silica as a control over atmospheric CO₂ levels

Today the oceans far from land are the equivalent of deserts, having very low biological productivity.  This is not due to a lack of the main nutrients, potassium, nitrogen and phosphorus, or to too little sunlight for photosynthesis.  For some time, marine specialists have suggested that the culprit is too little soluble iron – a micronutrient at the core of pigments and the enzyme RuBisCO, on which photosynthesis and the fundamental Calvin cycle depend.  The halving of atmospheric CO₂ levels during glacial maxima is widely believed to reflect more efficient ocean bioproductivity and thus burial of dead organic matter.  The idea that general dryness and windiness during glacial epochs delivered soluble iron to the remote ocean surface is one means of explaining this.  However, it took CO₂ 8 000 years to rise to pre-industrial levels after the last dusty period when ice sheets reacehd their maximum extent, whereas iron lingers in seawater for only a few tens of years at most.  Dust carries far more silica than soluble iron, and SiO₂ resides for 15 000 years or so.  This encourages the blooming of silica secreting diatoms in competition with calcium-carbonate secreting plankton.  Carbonate production by cells actually generates CO₂, so less carbonate secreters relative to those producing silica shells means that tendency is offset by a greater contribution to buried carbon from dead silica secreters (Source:  Tréguer, P and Pndaven, P., 2000.  Silica control of carbon dioxide.  Nature, v. 406, p. 358-359)

Milutin Milankovic built on James Croll’s notion that perturbation of Earth’s astronomical behaviour is likely to cause variations in solar heating that might lie behind repeated glacial epochs.  One of the most fertile discoveries in Earth science since World War 2 is that the periodicities that Milankovic calculated do seem to dominate the time-record of climate change over the last 2.5 million years, when 50 glacial-interglacial cycles forced changes in the oxygen isotope content of fossils from deep-ocean cores.  These data record the variations in long-term storage of water in continental ice sheets, and are a near-incontrovertible ‘proxy’ for both varying extents of glaciation and sea levels    Much the same kinds of signals also appear to turn up in time series of other kinds from sediments of a much wider range of ages.  Milankovic theory now has as much popular support as Alfred Wegener’s idea of drifting continents.  But problematic aspects refuse to go away.  Not the least of these is the conversion of depth to time in oceanic sediments from which the longest and most detailed records have emerged.  An analysis of the frequencies involved in past climate change rests or falls on the accurate conversion of depths in sediment cores to time.  In oceanic sediments this is by no means an easy job, because of a lack of material that can be dated precisely.

The first attempt to unscramble the complex variation in ocean cores used a few calibration points in climate time series onshore, whose shapes seemed to match those of ocean records.  The most crucial of these was that for the rise in sea level at the end of the ice age before last (called Termination II) recorded in coral reefs in the Caribbean.  The most widely used date for Termination II in the Caribbean is 127+6 thousand years (ka).  It was from using this date as a global time correlation that the Milankovic signals of 100, 41, 23 and 19 ka periodicities popped out of the mathematical analysis.  One surprise was that the match with the prediction of varying solar heating referred to the Northern rather than the Southern Hemisphere, or the planet as a whole.  A great deal of later work hangs on that, and much of it has simply assumed the Northern-Hemisphere pacemaker, such as the widely used SPECMAP time scale.

Daniel Karner and Richard Muller of the University of California in Berkeley summarise the most contrary pieces of evidence for the timing of climatic change in a recent issue of Science (Karner, D.B. and Muller, R.A., 2000.  A causality problem for Milankovitch. Science, 288, p.2143-2144).    Using a different dating approach the Caribbean timing of Termination II comes out at 132 ka, while for a series of coral reefs in Papua New Guinea it appears as early as 142 ka.  Detailed climate changes recorded in stalactitic material from a cave in Nevada (Devils Hole) also show an ‘early’ Termination II.  All these ages are as precise or better than the accepted 127ka date for Termination II, so Karner and Muller see a big problem.  Whereas the end of the last glaciation (Termination I) is pretty well tied down to about 12 ka, and corresponds to increased solar heating in the Northern Hemisphere from the Milankovic predictions, Termination II bucks that by 5 to 10 thousand years.  A theory that is only 50% believable needs a serious seeing to!

For the last four terminations, the most varied and informative data come from cores through the Antarctic ice sheet.  Though that too has its problems regarding calibration of depth to time, a recent evaluation of the climate variations over the last 420 ka (Petit, J.R. et al., 2000.  Climate and atmospheric history of the past 420,000 years from the Vostok ice core, Antarctica.  Nature,  399, p. 429-436) shows significant differences between the last four terminations.  Karner and Muller suggest that each glacial-interglacial cycle may have different controls, and encourage a new look at the wealth of data, unbiased by earlier ideas centring on pacing by a single, astronomical pacemaker and accepting that climate controls are multidimensional.