Category: Climate change and palaeoclimatology

Polar ice cores have presented us with the most exquisite records of how high-latitude climate has changed in the recent past from indirect clues presented by variations in stable isotopes of oxygen and deuterium (temperature change), dust and sulfate content (aridity and volcanicity respectively) in layers of ice. That proxy record extends back to 800 ka in the Dome C core from Antarctica, showing in great detail the course of the last nine glacial-interglacial cycles, both the astronomical effect of a changeover from a 40 ka pacing to one of around 100 ka and many intricacies on a millennial time scale. The most tangible archive of information resides in the air bubbles trapped by the original snow that eventually turned into ice. That reveals how the intricate pacing of climate change has been almost perfectly tracked by the global carbon cycle as shown by changes in the concentrations of carbon dioxide and methane. This was first demonstrated by cores through the Greenland ice cap, which penetrate just the last glacial episode and the warmth before and after.

After several years of painstaking bubble analyses at many collaborating labs, the full 800 ka greenhouse-gas records from Antarctica have now appeared (Luthi, D. and 10 others 2008. High resolution carbon dioxide concentration record 650,000-800,000 years before present. Nature, v.  453, p. 379-382. Lulergue, L. and 9 others 2008. Orbital and millennial-scale features of atmospheric CH₄ over the past 800,000 years. Nature, v.  453, p. 383-386). These long records demonstrate the close connection between climate and greenhouse gases that must be maintained by complex (and not fully understood) feedback mechanisms. Different Earth processes affect the two principal gases, methane probably being controlled by effects of varying temperature and rainfall on peat-rich swamps in the tropics, whereas carbon dioxide’s main driver is capture and release of carbon by the oceans. The central feature remains that of astronomical forces, with perhaps some sign of a signal from the 413 ka component of orbital eccentricity from a shift in the range of temperatures and greenhouse gases in 100 ka cycles around 450 ka ago, and a broad change in methane concentrations. Yet, despite being a pole away from high northern latitudes where comparison of the Greenland ice record with North Atlantic sea-floor sediment data revealed a northern cause for dramatic short term shifts, much the same millennial cycles characterise the whole Antarctic record. It could be that these rapid changes are proxies for the course of northern climate vagaries – there are about 75 of them in the methane Antarctic record. So stunning are the new data that they are sure to spur attempts to go back even further by more drilling in Antarctica, probably in the eastern ice cap where current air temperature and snow fall are extremely low and a greater length of time may be preserved in a smaller thickness of ice. That is because the faster snow and ice accumulate the more rapidly flow removes the record: the reason why the thick Greenland ice, although capable of yielding time resolution of as little as individual years, cannot retain records much beyond 200 ka.

See also: Brook, E. 2008. Windows on the greenhouse. Nature, v.  453, p. 291-292.

 

The yellowing of the Sahara

As Earth emerged steadily from the last glacial maximum, around 14.8 ka when temperatures were close to those of the Holocene yet sea level still had a way to rise before reaching its current level, the Sahara became a land of wetlands, lakes and grassland. Many caves within its modern arid confines contain superb artwork depicting its fauna and the forager-hunters that preyed on it. Around the time of the earliest Pharaonic civilisation on the Nile floodplain (~3000 BCE) the humid episode ended, forcing inhabitants of the Sahara either to the Nile valley of the Mediterranean coast. Having spanned the millennium-long climatic upheaval of the Younger Dryas and the relative stability and warmth of the early Holocene, why it ended is something of a mystery. A small, amazingly beautiful lake in northern Chad seems to hold the key, as it has existed and gathered sediment for at least 6 thousand years (Kröpelin, S and 14 others 2008. Climate-driven ecosystem succession in the Sahara: the past 6000 years. Science, v. 320, p. 765-768), Lake Yoa is one of several permanent lakes fed by ancient groundwater from the vast Nubian Sandstone aquifer, yet receives negligible rainfall. The uppermost lake sediments are laminated in an annual fashion so that each layer and its contents of aquatic organisms, pollen and dust can be precisely dated.

Between 4200 and 3900 years ago the lake changed from a freshwater habitat to a salt lake when evaporation overcame recharge by rain. However, the environment as a whole did not change suddenly, but progressively. The sudden change in salinity resulted from Lake Yoa losing any outflow, which previously had removed salts accumulated by evaporation of the inflowing groundwater. The lake would then no longer have had any use for humans and their livestock, but conditions did not drive people out of the Sahara suddenly.

Almost a year ago two dozen scientists presented evidence to suggest that onset of the Younger Dryas at 12.9 ka followed upper atmosphere explosions of cometary material (Firestone, R.B. and 25 others 2007. Evidence for an extraterrestrial impact 12,900 years ago that contributed to the megafaunal extinctions and the Younger Dryas cooling. Proceedings of the National Academy of Sciences of the United States of America, v. 104, 16016-16021; see Whizz-bang view of Younger Dryas in EPN July 2007). Evidence cited included: excess iridium; tiny spherules; fullerenes containing extraterrestrial helium; nanodiamonds and evidence for huge wildfires. Not quite the Full Monty, as neither crater nor shocked mineral grains were claimed, hence the teams’ opting for a cometary airburst. In North America such signs were said to overly the last known occurrences of Clovis tools at 7 archaeological sites (see Clovis First hypothesis dumped above). It was pretty clear that the suggestion for a hitherto unnoticed event with a widespread signature – 26 sites either side of the Atlantic were cited –  was going to be challenged, and so it has (see Kerr, R.A. Experts find no evidence for a mammoth-killer impact. Science, v. 319, p. 1331-1332), perhaps not unconnected with the blaze of publicity surrounding the paper’s appearance, including several TV documentaries.

Well, say experts, sooty layers do suggest large-scale fires, but forest fires occur every year, especially when humans are around. Fullerenes or ‘buckyballs’ equally can form terrestrially, except those containing ET helium. The last is regarded by many critics as ‘inventive’; they have never been isolated since such combinations were first reported in 2001 (see Extinctions by impacts: smoking artillery in EPN March 2002). The accepted methodology for detection of tiny diamonds seems to have been ignored, and that claimed to have found them misused. The iridium ‘spike’ – crucial in identifying the global nature of the K-T event – by itself is not enough for claims of impacts. Astonishingly, the authors cited such a Younger Dryas iridium spike in a Greenland ice core, yet the originator of those data says his paper does not report abnormal iridium at 12.9 ka or anywhere during the YD. Microspherules rain down all the time with interplanetary dust, and do not constitute sound evidence either.

So, what on Earth is going on? A collaboration between 26 authors, who willingly supply other workers with materials for checking surely cannot be conspiring at a hoax. Impact experts are hinting at ‘over-enthusiasm’ by a team outside the ‘impact community’. It all sounds oddly similar to the furore that in 1980 greeted  first suggestions by the Alvarezes for the K-T impact…

Accepted geoscientific ‘wisdom’ is that the Cretaceous Period was so warm that forests reached polar latitudes and so too did cold-blooded reptiles. Planktonic foram oxygen isotopes indicate that the Cretaceous ‘hothouse’ in the Turonian (93.5-89.3 Ma) produced tropical sea-surface temperatures up to 37°C; warmer than human blood temperature. It also saw sea level reach an all time high. Both features have been attributed to the rate of ocean-floor volcanism being at its highest. It has, however, been difficult to model the warmth at high latitudes without fudging the input to general circulation models.

Measuring d¹⁸O in both planktonic and benthonic (ocean-floor) forams at centimetre spacings in Turonian ocean-floor sediments seems to have truly bamboozled specialists in the Cretaceous. They reveal a period of ~200 ka  at around 91.2 Ma where both show a sharp increase (Bornemann, A. and 8 others 2008. Isotopic evidence for glaciation during the Cretaceous supergreenhouse. Science, v. 319, p. 189-192). Respectively, the peaks reflect decreased sea-surface temperature (but only down to 32°C in the tropics) and an increase in the extraction of light ¹⁶O from the oceans; only likely when ice caps build up on land. The size of the benthonic d¹⁸O increase suggests ice caps about half the size of that now blanketing Antarctica. Other evidence includes rapid decreases in Turonian sea level in Europe, North America and Russia; only likely on such a scale as a result of glacio-eustasy. However, direct evidence in the form of tillites, striated pavements and glacio-marine sediments has yet to turn up

Until these convincing data emerged, it seemed that sufficient post-Permian frigidity for large-scale glaciation had not developed until Oligocene times. However, the paradox of high-latitude ice caps and low-latitude balmy seas is resolvable. Evaporation from the tropical sea surface would have been much greater than nowadays. Transport of moisture to cooler areas may have resulted in such immense winter snowfall at high latitudes that sufficient remained unmelted after winter darkness for its albedo to further cool the polar region. Almost certainly the site for the ice cap would have been Antarctica, which in the Cretaceous, as now, sat over the South Pole. Remove the present ice, and that continent would have had an average surface height of between 1 and 2 km that would have encouraged snow build up were sufficient to have fallen during the Turonian. Yet without the direct evidence for glaciation in sediments – much would be buried by the present Antarctic ice cap, if not eroded away – the scenario is difficult for some to believe.

Holocene cold spell and glacial lake burst

The most startling event during the gradual warming after the last glacial maximum was the millennium of icy conditions between 12.5 and 11.5 ka; the Younger Dryas. Long after Holocene warmth seemed well established and agriculture had been underway for two millennia, with perhaps increased human population, a smaller cold ‘snap’ took place, between 8.21 and 8.17 ka; i.e. for about 70 years. Its main effect was around the North Atlantic, but it was felt over the whole hemisphere. It must have been devastating for early farmers and new migrants into higher latitude lands. High-resolution records of many kinds are possible for such a young event, from both ice and marine cores, and also terrestrial pollen records. Norwegian, French and Dutch climate researchers have gleaned a great deal from a sea-floor core from between southern Greenland and Labrador (Kleiven, H.F. et al. 2008. Reduced North Atlantic deep water and the glacial Lake Agassiz outburst. Science, v. 319, p. 60-64). Their combined fossil, oxygen-isotope and mineralogical study shows anomalies from about 170 years before to 100 years after the drop in regional temperatures.  These include signs of decreased saltiness of the water in the Labrador Basin and a reduction in production of deep water in the North Atlantic. This is exactly the predicted signature for a shut-down of the Gulf Stream, similar to those implicated in Dansgaard-Oeschger events through the last Ice Age and the Younger Dryas itself.

The Younger Dryas has been linked to sudden drainage of huge glacially dammed lakes that once surrounded the ice cap of the Canadian Shield.  One scenario for that is a huge, protracted flood down the St Lawrence River into the North Atlantic, another being one down the MacKenzie River into the Arctic Ocean. Freshening of surface waters by such means would have reduced the formation of the dense cold brines that sink to form North Atlantic Deep Water today. In so doing these down-wellings drag surface waters northwards from low latitudes to form the Gulf Stream that makes the western side of the North Atlantic unusually warm. If they stop or slow significantly regional air temperatures fall, as they did again around 8.2 ka. In this case the likely cause was escape of water melted from the last dregs of the North American ice sheet that had been held in a glacial lake south of Hudson Bay: Lake Agassiz.

Following its first discovery, evidence for low-latitude glacial action at several times during the Neoproterozoic has fuelled one of the most publicised controversies in the geosciences. Was the Earth totally frozen over during these episodes, or was ice confined only to parts of the surface? Whatever, the last part of the Precambrian witnessed huge fluctuations of many kinds, and after the cold epochs the first large animals made a sudden appearance. The most dramatic geochemical ups and downs in Earth history took place, in the form of sudden extreme shifts in the relative proportions of the stable isotopes of carbon in seawater, as recorded by marine carbonate rocks. These fluctuations correlate closely with the evidence for low-latitude glaciations: large negative excursions of d¹³C with glacial epochs, and positive values developing between them. The first can be interpreted as the result of massive declines in photosynthetic fixation of organic carbon. The second suggests repeated recoveries of such biological productivity, which favours the extraction of ¹²C from seawater and an increase in the relative proportion of the heavier isotope as organic carbon becomes buried in seafloor sediments.

Since organic carbon is ultimately extracted photosynthetically from carbon dioxide in the atmosphere, a link between climate and living processes (and those that bury dead organisms) can be the basis for models attempting to explain the extraordinary events of Neoproterozoic times. If large amounts of organic carbon are buried or remain suspended in the oceans, the drawdown of atmospheric CO₂ reduces the greenhouse effect and leads to cooling. Conceivably, the effect could be to so reduce global mean surface temperature that freezing conditions grip even the lowest latitudes. Once glacial and sea ice becomes established, its high reflectivity reduces the amount of incoming solar radiation that is absorbed to warm the Earth. The two processes combined would tend to lock frigid conditions in place until such time as gradual release of volcanic CO₂ increased the atmospheric greenhouse effect. That is the theoretical essence of the Snowball Earth hypothesis in which complete ice cover sterilised surface biology for long periods. However, it leaves out two important factors: as water cools it is able to dissolve more gases from the atmosphere; organic carbon in ocean water can be transformed to dissolved CO₂ if it is oxidised, thereby reducing the amount of carbon being buried. Modelling the carbon-climate link in the Neoproterozoic requires that both factors are accounted for (Peltier et al. 2007. Snowball Earth prevention by dissolved organic carbon remineralization. Nature, v. 450, p. 813-818).

The model devised by Richard Peltier and colleagues from the University of Toronto also incorporates the distribution of land at the time. Results from it show a looping behaviour, with recovery from frigidity as increases in dissolved oxygen convert organic carbon to dissolved carbon dioxide, whose increasing concentration in turn leads to more escape of the gas to the atmosphere. The model also suggests how glacial and sea ice might have developed during such a cycle, and with the late Precambrian configuration of drifting continents it allows for low-latitude continental glaciation, but not for all-enveloping sea ice. The implication is indeed glacial events vastly greater than those of the late Palaeozoic and during the present Ice Age, but less effect on marine photosynthesis than from Snowball conditions – a ‘Slushball’ Peltier et al. explain why the cyclical processes suggested by the model stopped before the start of the Phanerozoic, from carbon-isotope evidence for a massive oxidation of suspended marine organic carbon around 550 Ma. Thereafter, abundant oxygen and large animals ensured most dead organic carbon was oxidised in the oceans.

Unsurprisingly, one of the authors of the Snowball hypothesis finds flaws in the geochemical argument for its impossibility (Kaufman, A.J. 2007. Slush find. Nature, v. 450, p. 807-808). Not only was oxygen likely to have been at far lower atmospheric concentrations than it became in the Phanerozoic, the glacial epochs provide evidence that its concentration in seawater was very low. The marine diamictites associated with each contain both ironstones and iron-oxide cements. For them to have formed demands high concentrations of dissolved iron in sea water, in the form of reduced Fe²⁺ ions; incompatible with widespread oxidizing conditions that would favour Fe³⁺ whose compounds are insoluble.

Some good news about carbon burial

The second largest ‘sink’ for atmospheric CO₂, after silicate weathering and formation of carbonate sediments, is the burial of organic carbon. Derived from photosynthesis of carbon dioxide in the air or dissolved in water, organic carbon descends from the photic zone of the oceans or is carried from the land by rivers. In the second case it is often believed that more than 70% of the carbon load of rivers is oxidised back to CO₂ before having a chance of being buried in marine sediments. To estimate the proportion that does contribute to carbon sequestration is a complicated matter, involving measurement of the carbon budgeting for an entire river basin and its offshore sediments. This has been done by a team of French geochemists for the huge Ganges-Brahmaputra system that drains the northern Indian subcontinent and much of the Himalaya (Galy, V. et al. 2007. Efficient carbon burial in the Bengal fan sustained by the Himalayan erosional system. Nature, v. 450, p. 407-410). This system carries a stupendous load of sediment, especially during the monsoon season. At 1 to 2 billion t of sediment deposited from suspension in the Bay of Bengal each year, this is the largest single flux of sediment from land to the ocean floor. Even more is delivered as bed load (rolling and bouncing sand particles) to build up the Ganges-Brahmaputra delta of Bangladesh and West Bengal, India. The authors found that recently produced organic carbon is about 4 to 5 times more abundant in the suspended sediment load than is reworked fossil carbon derived by erosion of ancient sedimentary rocks, which itself is predominant in the bed load. Fossil carbon makes no difference to the modern carbon cycle, provided it does not get oxidised, which is less likely than for recent organic carbon in a form that can be metabolised.

By comparing the recent organic carbon load suspended in the rivers’ flow with that in the fine sediments of the Bengal Fan, Galy et al. have been able to show that most of that carbon is conserved without oxidation. As a result, the Bengal Fan accounts annually for about 15% of global carbon burial. There are two reasons for this remarkable efficiency: the low oxygen availability in deep waters of the Bay of Bengal; the very high sediment load from erosion of the Himalaya that buries carbon before oxidation is possible. Orogenic belts in humid areas are therefore key factors in exerting negative feedback on climate, whereas drainages of flat areas, such as the Amazon and especially its main tributary the Rio Negro, encourage oxidation in their lower reaches and offshore and are less important.

Although the ice that makes up the upper parts of the Greenland and Antarctic ice sheets is annually layered, for times before about 70 ka the layering disappears because of plastic deformation. Earlier ages have to estimated from models of the deformation, and a second check is to match the data records from ice cores against those from sea floor sediments. Different processes contribute to those records: for instance, the marine record of oxygen isotopes in benthonic forams tracks the changing volume of ice locked on land, while the same record from ice cores depends on the air temperature above the ice cap. The correlation does seem to work, however. But not, it seems, for the very deepest ice recovered from beneath Antarctica (see Yet further back in the Antarctic ice in the December 2005 issue of EPN) which extends to around 800 ka.

French scientists involved in the EPICA Dome C ice-core project have cunningly discovered a means of checking on the otherwise undateable deep Antarctic ice (Raisbeck, G.M. et al. 2006. ¹⁰Be evidence for the Matuyama-Brunhes geomagnetic reversal in the EPICA Dome C ice core. Nature, v. 444, p. 82-84). The core penetrated to an estimated time that should include the most recent magnetic reversal, dated very precisely to 778±2 ka. Although the exact details of how the magnetic field behaved during this reversal, it is known that when its polarity flips the intensity of the field becomes very small. While the field is stable it is sufficiently strong to deflect charged particles, both in the Solar wind and in cosmic rays, so that less pass through the atmosphere. Cosmic rays are so energetic that they can perform isotopic transformations, one product being ¹⁰Be. So if the magnetic field decreased so the proportion of ¹⁰Be in the atmosphere would go up. Raisbeck and colleagues have examined the ¹⁰Be record in the EPICA core in great detail. In a 10 m thick section from a depth of almost 3.2 km the isotope rises to a peak, which they interpret as the signature of the reversal. If correct, this gives a ‘golden spike’ against which the depth to age conversion can be refined.

Balmy shores of the Precambrian

Before the appearance of fossil organisms that could give clues to past climates the only sources of information take the form of proxies. One of the best examples might seem to be the oxygen isotope composition of carbonate rocks that relate to sea-surface temperature. In fact it isn’t useful for the Precambrian because estimates of SST depend on being able to identify the shells of planktonic animals and use their d¹⁸O as a proxy. That is a pity, because limestones are common throughout the geological record and various aspects of their geochemistry have been used extensively as proxies for other crucial information, such as the relationship between their strontium isotope composition and the pace of continental weathering. Another palaeothermometer relies on the same temperature dependent fractionation of oxygen isotopes between seawater and the precipitation of dissolved silica to form cherts, whose d¹⁸O decreases with temperature. The trouble is that silica is notoriously prone to being remobilised and reprecipitated as pH changes in the fluids within sedimentary rocks. Some results from Precambrian cherts gave such low d¹⁸O that seawater temperature would have been tens of degrees higher than they were during the Phanerozoic, but they have been wisely suspected of having been affected by much later alteration by warmer fluids passing through cherty sequences. Now the approach has been given a boost by geochemists at the French National History Museum (Robert, F. & Chaussidon, M. 2006. A paleaotemperature curve for the Precambrian ocean based on silicon isotopes in cherts. Nature, v. 443, p. 969-972).

François Robert and Marc Chaussidon analysed the silicon isotopes in cherts for which oxygen isotope data are available. Since the two isotopic systems would both change, yet would behave differently during hydrothermal or metamorphic alteration, if the results correlate well both should be undisturbed. Except in samples that show the lowest d¹⁸O values (i.e. highest temperatures) there is a good correlation. That finding validates many of the O-isotope seawater temperatures, but Si isotopes fractionate during precipitation too, again in relation to temperature. So Robert and Chaussidon take Precambrian ocean temperature data to a new level with estimates based on two methods. Their results are fascinating: as well as confirming a decline from around 70°C 3.4 Ga ago to between 10 to 40°C in the Phanerozoic, the d³⁰Si data show sharp downward ‘spikes’ at about 2.5 Ga and 1.8 Ga. Between about 1.5 Ga to 600 Ma ocean temperature was steady at around 20°C, so there is no sign of continually cold oceans through the period of ‘Snowball Earth’ events – the number of samples cannot yet resolve the individual events, but the ‘Cryogenian’ is an obvious target for more work. The data are also important as they hint at all kinds of possible biological outcomes for such global warmth, and explanations are definitely needed.  Does the record suggest greater geothermal heating, or was it an outcome of the greenhouse effect? Will more details show periods of changing burial of organic carbon? Whatever, the Precambrian has become a stranger world to contemplate.

See also: de la Rocha, C.L. 2006. In hot water. Nature, v. 443, p. 920-921.

The so-called Cryogenian Period of the Neoproterozoic rests on evidence for coincident glaciation at all latitudes. It has been supposed to include at least two, maybe three and perhaps more frigid ‘snowball’ events, each with a pattern of lower diamictites and an upper carbonate cap rock. The most widely supposed glacial epochs are the Sturtian at 712 Ma, the Marinoan at 635 Ma and the Gaskiers at 580 Ma, but Precambrian sedimentary sequences are notoriously difficult to tie down in time. Only if dateable igneous events bracket evidence for glaciation is an age truly valid. Yet the global 3-fold division depends largely on correlation of stratigraphic and carbon-isotope sequences with the odd few that are dated in an absolute time-frame. The developing field of rhenium-osmium (Re-Os) radiometric dating offers a more universal check, since it provides a means of dating highly reduced black shales, that are abundant in the Neoproterozoic. The first reported results come as a blow to the ‘Snowball Earth’ community (Kendall, B. et al. 2006. Re-Os geochronology of postglacial black shales in Australia: constraints on the timing of the ‘Sturtian’ glaciation. Geology, v. 34, p. 729-732).

Bruce Kendall and colleagues from the Universities of Alberta, Canada and the Durham, UK have constrained some of the principal occurrences of the Sturtian event in Australia to between 643 and 657 Ma, by dating the shales which envelop the diamictites and cap carbonates. They are younger than even the widest range previously suggested for the Sturtian: either the glaciation was grossly diachronous, or this is yet another glaciation of ‘Sturtian’type. The best that can be concluded is that the ‘Cryogenian’ was cold but glaciation shifted from place to place – a ‘slushball’ model?

While most geoscientists use the products of processes that operate today to judge environments of the past, climatologists do the reverse: the past is the key to the present. While the climate record of the last 2.5 Ma is a key to understanding and perhaps even predicting rapid climate shifts during glacial-interglacial periods uncontaminated by human influences, such is the extent to which greenhouse emissions have affected the current climate that we have little idea what the outcomes may be. The possibility of greenhouse warming has become higher than in any previous interglacial epoch. To get even an inkling of what that might set in motion requires looking back to warmer times than the Late Pliocene and Pleistocene, at around 3 to 5 Ma. In the Early Pliocene it is very likely that CO₂ in the atmosphere was no more than nowadays. Because the Earth’s geography was little different from the way it is now and the Milankovich forcing was the same too, modelling Early Pliocene climate might seem to result in similar patterns, but it doesn’t (Fedorov, A.V. et al. 2006. The Pliocene paradox (mechanisms for a permanent El Niño). Science, v. 312, p. 1485-1489). Sea level was some 25 metres higher than it is at present and mean global temperature was an extra 3°C, and sea-surface temperatures (from the oxygen isotopes in planktonic foraminifera) were high as well. Despite much the same forcing factors as today, the Pliocene lacked large high-latitude ice caps in Arctic regions. Milankovich-related fluctuations were damped down compared with those of the Pleistocene. Both modelling and geological evidence from the Early Pliocene suggests that Earth’s climate was dominated by a perpetual El Niño in the tropical oceans, because of an inability of cold water to upwell periodically along the western tropical margins of Africa and South America. Quite probably such conditions had persisted for the previous 50 Ma, despite gradual overall cooling.

Fedorov and colleagues point to very different Early Pliocene climates in several regions: Mild winters in central and north-eastern North America; droughts in Indonesia and torrential rains in western North and South America. Overall, it was a much more humid world, and since water vapour is a powerful greenhouse gas warmth and humidity were sustained despite no higher CO₂ levels than now. At about 3 Ma, ocean surface waters began to cool, with signs that the alternations associated with El Niño and La Niña in the eastern Pacific began. An explanation for this is the gradual build up of very cold water deep in the ocean as a result of winds from continents cooling ocean surface water at high latitudes and causing it to sink. Without periodic upwellings, warm surface waters and cold deep waters could not mix, so inevitably the interface became shallower. At some critical depth, this thermocline could break surface, transforming both climate patterns and those of ocean currents, eventually to end up as the present tropical climate cyclicity which is connected with  other climate features of the Great Ice Age.

Fedorov et al  speculate that only a small descent of the ocean thermocline – a matter of a few tens of metres – could re-establish Pliocene conditions.  That might occur because of continued anthropogenic warming, and the ‘flip’ might be as quick as a few decades to centuries.

Between about 12.9 and 11.5 ka the progress of warming from the frigidity of the Last Glacial Maximum was rudely interrupted. For over a thousand years conditions returned to those of a mini ice age, with continental glaciers re-advancing on a large scale, an increase in aridity and a reversal of colonisation of high northern latitudes by both plants and humans. Pollen records become dominated by those of a diminutive alpine plant, the mountain avens (Dryas octopetala) from which the cold snap gets its name – the Younger Dryas. The pace at which cooling took place was dramatic, and glacial conditions swept in within a decade at most. The most likely scenario is failure of North Atlantic Deep Water to form, thereby shutting down the thermohaline circulation that draws the warming Gulf Stream into the Arctic Ocean off the northern cape of Norway. The reason for that was a massive and sudden freshening of surface water at high latitudes in the North Atlantic, but where the influx of fresh water came from is a puzzle. Wallace Broeker of the Lamont-Doherty Earth Observatory in New York State resurrected an earlier idea that a vast lake of meltwater in the region of the Great Lakes of North America burst down the St Lawrence Seaway, instead of quietly escaping to the Gulf of Mexico along the Missouri-Mississippi system. Broeker has recently reviewed this hypothesis (Broeker, W.S. 2006. Was the Younger Dryas triggered by a flood? Science, v. 312, p. 1146-1148).

Oxygen isotope records from sediments in the Gulf of Mexico had been recording massive influx there of water depleted in ¹⁸O; a sure sign that the Mississippi was carrying much of the water produced by melting of the Laurentian ice sheet. That signature stops abruptly at the outset of the Younger Dryas. The meltwater must have found another outlet, but so far its oxygen isotope signature has not been conclusively discovered. As well as the St Lawrence escape route there are three other possibilities: north-westwards along the MacKenzie River valley; beneath the great ice sheet and through Hudson Bay; and by massive break-up of the ice sheet to launch an ‘armada’ of icebergs that quickly melted to freshen northern Atlantic waters. One of the clearest signs that vast proglacial lakes suddenly emptied is that they carve immense channels resembling canyons, in which there is abundant evidence for extreme scouring. Examples are the ‘channelled scablands’ of the state of Washington, and the Minnesota River valley. The volume escaping at the start of the Younger Dryas would have been so immense that such overflow channels would be dominant features of northern North America’s terrain; but there are few that fit the bill, and those that do exist are poorly constrained by radiocarbon dating. The lack of accurate dates for sediments and channels associated with the demise of the Laurentian ice sheet is the main obstacle, and surely evidence for exactly how the sudden plunge into glacial conditions was triggered will emerge sooner rather than later. One thing seems certain, the Younger Dryas was a freak event. The new ice core from Antarctica (see Yet further back in the Antarctic ice in the December 2005 issue of EPN) penetrates the previous six glacial maxima and shows no sign of a similar event at their terminations.

Sedimentary evolution of the Arctic Ocean: a start is made

For the Northern Hemisphere, especially around the North Atlantic, what happens in the Arctic exerts a strong influence over climate. On the one hand, ice-cover increases the proportion of solar energy that is reflected back to space, giving a cooling effect. On the other, cooling and increasing salinity of high-latitude water at the ocean surface results in its sinking to draw in warmer waters from further south, to extend warming further north. The two are linked intricately, for sea-ice formation adds to surface waters’ salinity. How and when the delicate balances arose remained poorly known while thick sea ice prevented ships penetrating to the highest possible latitudes in the Arctic Ocean, because the key to climate evolution depends on access to long core through ocean-floor sediments. Ironically, the decrease in Arctic ice cover with global warming has created greater access by icebreakers and drilling vessels. A consortium of countries around the Arctic funded a major effort to resolve the gap in knowledge through such a marine drilling programme in 2004. Results from the polar expedition have just begun to emerge (Moran, K and 36 other 2006. The Cenozoic palaeoenvironment of the Arctic Ocean. Nature, v. 441, p. 601-605). The cores were taken almost at the North geographic pole on the Lomonosov Ridge, a sliver of continental crust separated from its connection with the northern Russian continental shelf when North Atlantic sea-floor spreading nosed into the Arctic about 57 Ma ago.

The core is from sediments deposited on the Lomonosov ridge since it became detached from Russia, and is over 400 m long. Analyses are not yet complete, and the report by the IODP Arctic Coring Expedition covers the simplest parameters to determine: sediment bulk density and lithology, and micro-organisms. Nonetheless, these preliminary results provide a major surprise. Previously it was believed that frigid conditions in northern polar regions became established long after the Antarctic developed an ice cap 43 Ma ago, which matches the Cenozoic fall in atmospheric CO₂ and other evidence for lower mean global temperatures. The first glaciation in the Arctic was thought to be at 2-3 Ma, when pebbles dropped by icebergs first appear in the cores from the North Atlantic floor. In the Arctic Ocean core, such pebbles appear at much the same time as those around the Antarctic. They become widespread by 14 Ma. At the time of the Palaeocene-Eocene global warming, in response to massive methane emissions at 55 Ma, the Arctic waters were as warm as 18°C. The record is one of transition from a greenhouse world to an ice house. Surprisingly, considering the later influence of thermohaline processes that draw in warm water from lower latitudes, the earliest period is marked by fresh or at most slightly brackish waters. That was probably a result of isolation from the Atlantic and an excess of precipitation over evaporation. The early sediments record abundant carbon, then at around 14 Ma, the percentage of buried organic carbon drops dramatically to mark the start of increasing frigidity, when icebergs dropped significantly more debris in the Arctic Ocean.

HOW THE AMAZON FORMED

The world’s largest drainage system in the Amazon basin is so huge that it might seem to be an eternal feature of South America, at least since that continent formed when opening of the South Atlantic wrenched it from Africa in the Triassic. The upper Amazon takes much of its flow from rainfall in the eastern slopes of the Andes, but that range is still in the process of formation by tectonic and volcanic forces. A review of the Amazon’s evolution in a recent issue of Scientific American (Hoorn, C. 2006. The birth of the mighty Amazon. Scientific American, v. 294 May 2006, p. 40-47) shows that the river system is much younger than you might have expected. Lots of evidence points to the major eastward flow only beginning in the late Miocene, after 15 Ma ago. Before that drainage was northwards into the Caribbean, the reason being that the north-eastern Andes of Columbia and western Venezuela had not formed. When they did begin to rise, they hindered flow to create a huge wetland in what is now eastern Columbia. Eventually a northward drainage route was definitively blocked, so that flow took the easiest remaining route to the ocean; eastwards, to create the Amazon basin.

The groundbreaking Vostok ice core from Antarctica is the deepest ever to have been drilled. It recorded 440 ka of climate and atmospheric history, but unfortunately the very depth of the ice beneath the drilling station made that the limit in time terms. Thick ice begins to deform and flow, and the lowest parts of the Vostok core were clearly scrambled by that. The European Project for Ice Coring in Antarctica (EPICA) focussed its effort on a region of the East Antarctic ice sheet (Dome Concordia) whose location may always have ensured low accumulation of snow. Hopefully that would ensure that ice thickness was not so much as to result in complex flow at depth and that a fuller record would be preserved. The idea paid off, and the Dome C core penetrates back as far as 740 ka, giving an additional 3 glacial-interglacial cycles during the early part of the 100 ka periodicity; but falling just short of the first of those major cycles that are reflected in the marine oxygen-isotope record.

Results are now starting to emerge from Dome C (Siegenthaler, U and 10 others 2005.  Stable carbon cycle-climate relationship during the Late Pleistocene. Science, v. 310, p. 1313-1317. Spahni, R. and 10 others 2005. Atmospheric methane and nitrous oxide of the Late Pleistocene from Antarctic ice cores. Science, v. 310, p. 1317-1321). The results are high-quality, and reveal some new features. The first three cycles conform to the 100 ka signal of the very weak variation in orbital eccentricity, as expected, but show lower amplitude shifts in CO₂ and methane in air trapped in bubbles than do the later four cycles.  The two ‘greenhouse’ gases vary in concert, and their earlier low levels match with less extreme shifts in temperature as shown by the changes in deuterium content of the ice itself. This is probably due to the transition from the previous dominance by the 40 ka pace of changing axial tilt. Nitrous oxide values, although patchy down the core, seem to have fluctuated but at much the same amplitude throughout the last 720 00 ka. Dome C has yet to be ‘bottomed out’ so there is a chance that the record may yet reach the 40-100 ka boundary around 900 ka ago.  What is striking – and should ring alarm bells – from the results so far is that in each of the previous 7 interglacials atmospheric neither CO₂ nor methane levels came close to those of the last century. Whatever its eventual effects, anthropogenic addition to the ‘greenhouse effect’ is an incontrovertible fact.

See also: Brook, E.J. 2005. Tiny bubbles tell all. Science, v. 210, p. 1285-7

 

A time in Earth history (~251 Ma) when life was all but snuffed out and from which the creatures most familiar to us eventually emerged is understandably revisited quite often. Causes ranging from impacts (no convincing evidence as yet), through flood-basalt emissions, catastrophic methane release, low atmospheric oxygen to ocean anoxia have all been proposed. Hesitantly, opinion is converging on a climatic crisis of some kind, and indeed the coincidence of both terrestrial and marine faunal and flora extinctions points to climate being the global transmitter of some cause or a coincidence of causes. After the waning of Southern Hemisphere glaciations, the late Permian was warm, even at high latitudes. Until recently, attempts at modelling the end-Permian climate have not been entirely convincing because of limitations in the models themselves. Jeffrey Kiehl and Christine Shields of the US National Center for Atmospheric Research in Colorado have assembled a model that couples land, atmosphere, oceans, sea-ice and palaeogeography for the period (Kiehl, J.T. & Shields, C.A. 2005. Climate simulation of the latest Permian: Implications for mass extinction. Geology, v. 33, p. 757-760).

The critical test for the model is running it with parameters for the near-present, and it performs well. Several lines of evidence point to a much higher CO₂ level in the Permian atmosphere, so this is the main input parameter. The outcome is a world with a mean surface temperature that is 8° C higher than now. Unlike today, there was no geographic hindrance to poleward heat transport, so the high mean temperature is reflected in the summer warmth and humidity of Permian high-latitude land. The sub–tropics on the other hand were scorching (around an average summer minimum of 51° C, 15° C higher than now); a clear contributor to minimising life there. Sea-surface temperatures at high latitudes are higher in the model outcomes, this warmth extending to depths of 3 km. Surprisingly, low-latitude sea temperature emerges as much the same as now. The model also suggests that seawater was saltier than now, and that results in greater uniformity of density with depth and location: a hindrance to bottomward circulation and mixing. There would probably have been no thermohaline circulation worth speaking of. The model helps confirm the likelihood of an oxygen-free lower ocean and little transfer of nutrients. The oceans too would have been inhospitable. A shutdown of biological productivity and therefore carbon burial would have accelerated warming. So, pushing the biosphere into a mass extinction would have been inevitable. The last straw may have been the additional stress of increasing acidity from sulphur dioxide emissions from the Siberian flood basalts.

Milankovich forcing and Early Jurassic methane

Periods of environmental crisis less severe than those leading to mass extinction appear throughout the fossil record. As well as minor extinction peaks they are often signified by departures of carbon-isotope records from long-lasting norms. Such a crisis appears in the d ¹³C record of the Early Jurassic, and is beautifully preserved in about 15 m of black shales on the North Yorkshire coast of England. Geoscientists from the Open University, UK and the University of Cologne, Germany have produced an extremely high-resolution time series of carbon-isotope data from the section (Kemp, D.B. et al. 2005. Astronomical pacing of methane release in the Early Jurassic period. Nature, v. 437, p. 396-399). The quality is sufficiently good to analyse the time series using Fourier analysis that yields the frequencies that contribute to the observed wave-like patterns in the data. Of course, the time in a stratigraphic time series is measured in metres, unless it is possible to calibrate the section by precise radiometric dating. The Yorkshire Jurassic contains only fossils and no dateable horizons, but the fine stratigraphic division based on ammonites is also widespread and calibration is possible from dates obtained elsewhere. The overwhelmingly dominant frequency in the carbon-isotope curve is 1.23 cycles m^-1, which represents 21 ka after the calibration of depth to time. That is the signal of precession of the equinoxes, part of the astronomical forcing bound up in Miliutin Milankovich’s theory of astronomical forcing of climate.

Astronomical pacing turns up throughout the stratigraphic column, wherever sediments are suitable for time-series analysis (steady, unbroken sedimentation), so a precessional signal is no great surprise. The important feature is the profundity of the d ¹³C excursions; a total of –7‰, largely accomplished by three abrupt shifts of –2 to –3‰. The first two coincide with bursts in extinctions. The most likely phenomenon to have produced these shifts is massive release of methane by destabilization of submarine gas hydrates. Emissions seem to have been blurting out on a regular basis as the Earth’s rotational axis precessed like a gyroscope. So, the complete time period was one in which gas hydrate was unstable, probably due to overall warming. Yet something else must have triggered vast releases three times. The Lower Jurassic extinctions link in time with massive magmatism in Southern Africa and Antarctic (the Karoo-Ferrar large igneous province). Perhaps especially large volcanic events there set the stage for large precessional methane releases. An alternative view is that volcanic emissions of CO₂ gradually produced enough widespread warming for the astronomical trigger to cause breakdown of gas hydrate simultaneously over very wide areas of the ocean floor. Other explanations have been suggested for the Lower Jurassic warming and carbon-isotope excursions, such as wildfires, impacts and connections with petroleum maturation and migration. The clear cyclicity rules them out.