Termination of global warmth at the Palaeocene/Eocene boundary through productivity feedback

The onset of the Palaeocene/Eocene thermal maximum (about 55 Myr ago) was marked by global surface temperatures warming by 5-7 degrees C over approximately 30,000 yr (ref. 1), probably because of enhanced mantle outgassing and the pulsed release of approximately 1,500 gigatonnes of methane carbon from decomposing gas-hydrate reservoirs. The aftermath of this rapid, intense and global warming event may be the best example in the geological record of the response of the Earth to high atmospheric carbon dioxide concentrations and high temperatures. This response has been suggested to include an intensified flux of organic carbon from the ocean surface to the deep ocean and its subsequent burial through biogeochemical feedback mechanisms. Here we present firm evidence for this view from two ocean drilling cores, which record the largest accumulation rates of biogenic barium--indicative of export palaeoproductivity--at times of maximum global temperatures and peak excursion values of delta13C. The unusually rapid return of delta13C to values similar to those before the methane release and the apparent coupling of the accumulation rates of biogenic barium to temperature, suggests that the enhanced deposition of organic matter to the deep sea may have efficiently cooled this greenhouse climate by the rapid removal of excess carbon dioxide from the atmosphere.     

Arctic hydrology during global warming at the Palaeocene/Eocene thermal maximum

The Palaeocene/Eocene thermal maximum represents a period of rapid, extreme global warming 55 million years ago, superimposed on an already warm world. This warming is associated with a severe shoaling of the ocean calcite compensation depth and a >2.5 per mil negative carbon isotope excursion in marine and soil carbonates. Together these observations indicate a massive release of 13C-depleted carbon and greenhouse-gas-induced warming. Recently, sediments were recovered from the central Arctic Ocean, providing the first opportunity to evaluate the environmental response at the North Pole at this time. Here we present stable hydrogen and carbon isotope measurements of terrestrial-plant- and aquatic-derived n-alkanes that record changes in hydrology, including surface water salinity and precipitation, and the global carbon cycle. Hydrogen isotope records are interpreted as documenting decreased rainout during moisture transport from lower latitudes and increased moisture delivery to the Arctic at the onset of the Palaeocene/Eocene thermal maximum, consistent with predictions of poleward storm track migrations during global warming. The terrestrial-plant carbon isotope excursion (about -4.5 to -6 per mil) is substantially larger than those of marine carbonates. Previously, this offset was explained by the physiological response of plants to increases in surface humidity. But this mechanism is not an effective explanation in this wet Arctic setting, leading us to hypothesize that the true magnitude of the excursion--and associated carbon input--was greater than originally surmised. Greater carbon release and strong hydrological cycle feedbacks may help explain the maintenance of this unprecedented warmth.     

Possible methane-induced polar warming in the early Eocene

Reconstructions of early Eocene climate depict a world in which the polar environments support mammals and reptiles, deciduous forests, warm oceans and rare frost conditions. At the same time, tropical sea surface temperatures are interpreted to have been the same as or slightly cooler than present values. The question of how to warm polar regions of Earth without noticeably warming the tropics remains unresolved; increased amounts of greenhouse gases would be expected to warm all latitudes equally. Oceanic heat transport has been postulated as a mechanism for heating high latitudes, but it is difficult to explain the dynamics that would achieve this. Here we consider estimates of Eocene wetland areas and suggest that the flux of methane, an important greenhouse gas, may have been substantially greater during the Eocene than at present. Elevated methane concentrations would have enhanced early Eocene global warming, and also might specifically have prevented severe winter cooling of polar regions because of the potential of atmospheric methane to promote the formation of optically thick, polar stratospheric ice clouds.     

The Cenozoic palaeoenvironment of the Arctic Ocean

The history of the Arctic Ocean during the Cenozoic era (0-65 million years ago) is largely unknown from direct evidence. Here we present a Cenozoic palaeoceanographic record constructed from >400 m of sediment core from a recent drilling expedition to the Lomonosov ridge in the Arctic Ocean. Our record shows a palaeoenvironmental transition from a warm 'greenhouse' world, during the late Palaeocene and early Eocene epochs, to a colder 'icehouse' world influenced by sea ice and icebergs from the middle Eocene epoch to the present. For the most recent approximately 14 Myr, we find sedimentation rates of 1-2 cm per thousand years, in stark contrast to the substantially lower rates proposed in earlier studies; this record of the Neogene reveals cooling of the Arctic that was synchronous with the expansion of Greenland ice (approximately 3.2 Myr ago) and East Antarctic ice (approximately 14 Myr ago). We find evidence for the first occurrence of ice-rafted debris in the middle Eocene epoch (approximately 45 Myr ago), some 35 Myr earlier than previously thought; fresh surface waters were present at approximately 49 Myr ago, before the onset of ice-rafted debris. Also, the temperatures of surface waters during the Palaeocene/Eocene thermal maximum (approximately 55 Myr ago) appear to have been substantially warmer than previously estimated. The revised timing of the earliest Arctic cooling events coincides with those from Antarctica, supporting arguments for bipolar symmetry in climate change.     

Abrupt reversal in ocean overturning during the Palaeocene/Eocene warm period

An exceptional analogue for the study of the causes and consequences of global warming occurs at the Palaeocene/Eocene Thermal Maximum, 55 million years ago. A rapid rise of global temperatures during this event accompanied turnovers in both marine and terrestrial biota, as well as significant changes in ocean chemistry and circulation. Here we present evidence for an abrupt shift in deep-ocean circulation using carbon isotope records from fourteen sites. These records indicate that deep-ocean circulation patterns changed from Southern Hemisphere overturning to Northern Hemisphere overturning at the start of the Palaeocene/Eocene Thermal Maximum. This shift in the location of deep-water formation persisted for at least 40,000 years, but eventually recovered to original circulation patterns. These results corroborate climate model inferences that a shift in deep-ocean circulation would deliver relatively warmer waters to the deep sea, thus producing further warming. Greenhouse conditions can thus initiate abrupt deep-ocean circulation changes in less than a few thousand years, but may have lasting effects; in this case taking 100,000 years to revert to background conditions.     

Increased Arctic cloud longwave emissivity associated with pollution from mid-latitudes

There is consensus among climate models that Arctic climate is particularly sensitive to anthropogenic greenhouse gases and that, over the next century, Arctic surface temperatures are projected to rise at a rate about twice the global mean. The response of Arctic surface temperatures to greenhouse gas thermal emission is modified by Northern Hemisphere synoptic meteorology and local radiative processes. Aerosols may play a contributing factor through changes to cloud radiative properties. Here we evaluate a previously suggested contribution of anthropogenic aerosols to cloud emission and surface temperatures in the Arctic. Using four years of ground-based aerosol and radiation measurements obtained near Barrow, Alaska, we show that, where thin water clouds and pollution are coincident, there is an increase in cloud longwave emissivity resulting from elevated haze levels. This results in an estimated surface warming under cloudy skies of between 3.3 and 5.2 W m(-2) or 1 and 1.6 degrees C. Arctic climate is closely tied to cloud longwave emission, but feedback mechanisms in the system are complex and the actual climate response to the described sensitivity remains to be evaluated.     

A humid climate state during the Palaeocene/Eocene thermal maximum

An abrupt climate warming of 5 to 10 degrees C during the Palaeocene/Eocene boundary thermal maximum (PETM) 55 Myr ago is linked to the catastrophic release of approximately 1,050-2,100 Gt of carbon from sea-floor methane hydrate reservoirs. Although atmospheric methane, and the carbon dioxide derived from its oxidation, probably contributed to PETM warming, neither the magnitude nor the timing of the climate change is consistent with direct greenhouse forcing by the carbon derived from methane hydrate. Here we demonstrate significant differences between marine and terrestrial carbon isotope records spanning the PETM. We use models of key carbon cycle processes to identify the cause of these differences. Our results provide evidence for a previously unrecognized discrete shift in the state of the climate system during the PETM, characterized by large increases in mid-latitude tropospheric humidity and enhanced cycling of carbon through terrestrial ecosystems. A more humid atmosphere helps to explain PETM temperatures, but the ultimate mechanisms underlying the shift remain unknown.     

Astronomical pacing of late Palaeocene to early Eocene global warming events

At the boundary between the Palaeocene and Eocene epochs, about 55 million years ago, the Earth experienced a strong global warming event, the Palaeocene-Eocene thermal maximum. The leading hypothesis to explain the extreme greenhouse conditions prevalent during this period is the dissociation of 1,400 to 2,800 gigatonnes of methane from ocean clathrates, resulting in a large negative carbon isotope excursion and severe carbonate dissolution in marine sediments. Possible triggering mechanisms for this event include crossing a threshold temperature as the Earth warmed gradually, comet impact, explosive volcanism or ocean current reorganization and erosion at continental slopes, whereas orbital forcing has been excluded. Here we report a distinct carbonate-poor red clay layer in deep-sea cores from Walvis ridge, which we term the Elmo horizon. Using orbital tuning, we estimate deposition of the Elmo horizon at about 2 million years after the Palaeocene-Eocene thermal maximum. The Elmo horizon has similar geochemical and biotic characteristics as the Palaeocene-Eocene thermal maximum, but of smaller magnitude. It is coincident with carbon isotope depletion events in other ocean basins, suggesting that it represents a second global thermal maximum. We show that both events correspond to maxima in the approximately 405-kyr and approximately 100-kyr eccentricity cycles that post-date prolonged minima in the 2.25-Myr eccentricity cycle, implying that they are indeed astronomically paced.     

High temperatures in the Late Cretaceous Arctic Ocean

To understand the climate dynamics of the warm, equable greenhouse world of the Late Cretaceous period, it is important to determine polar palaeotemperatures. The early palaeoceanographic history of the Arctic Ocean has, however, remained largely unknown, because the sea floor and underlying deposits are usually inaccessible beneath a cover of floating ice. A shallow piston core taken from a drifting ice island in 1970 fortuitously retrieved unconsolidated Upper Cretaceous organic-rich sediment from Alpha ridge, a submarine elevated feature of probable oceanic origin. A lack of carbonate in the sediments from this core has prevented the use of traditional oxygen-isotope palaeothermometry. Here we determine Arctic palaeotemperatures from these Upper Cretaceous deposits using TEX86, a new palaeothermometer that is based on the composition of membrane lipids derived from a ubiquitous component of marine plankton, Crenarchaeota. From these analyses we infer an average sea surface temperature of approximately 15 degrees C for the Arctic Ocean about 70 million years ago. This calibration point implies an Equator-to-pole gradient in sea surface temperatures of approximately 15 degrees C during this interval and, by extrapolation, we suggest that polar waters were generally warmer than 20 degrees C during the middle Cretaceous (approximately 90 million years ago).     

Warm tropical sea surface temperatures in the Late Cretaceous and Eocene epochs

Climate models with increased levels of carbon dioxide predict that global warming causes heating in the tropics, but investigations of ancient climates based on palaeodata have generally indicated cool tropical temperatures during supposed greenhouse episodes. For example, in the Late Cretaceous and Eocene epochs there is abundant geological evidence for warm, mostly ice-free poles, but tropical sea surface temperatures are generally estimated to be only 15-23 degrees C, based on oxygen isotope palaeothermometry of surface-dwelling planktonic foraminifer shells. Here we question the validity of most such data on the grounds of poor preservation and diagenetic alteration. We present new data from exceptionally well preserved foraminifer shells extracted from impermeable clay-rich sediments, which indicate that for the intervals studied, tropical sea surface temperatures were at least 28-32 degrees C. These warm temperatures are more in line with our understanding of the geographical distributions of temperature-sensitive fossil organisms and the results of climate models with increased CO2 levels.     

Episodic fresh surface waters in the Eocene Arctic Ocean

It has been suggested, on the basis of modern hydrology and fully coupled palaeoclimate simulations, that the warm greenhouse conditions that characterized the early Palaeogene period (55-45 Myr ago) probably induced an intensified hydrological cycle with precipitation exceeding evaporation at high latitudes. Little field evidence, however, has been available to constrain oceanic conditions in the Arctic during this period. Here we analyse Palaeogene sediments obtained during the Arctic Coring Expedition, showing that large quantities of the free-floating fern Azolla grew and reproduced in the Arctic Ocean by the onset of the middle Eocene epoch (approximately 50 Myr ago). The Azolla and accompanying abundant freshwater organic and siliceous microfossils indicate an episodic freshening of Arctic surface waters during an approximately 800,000-year interval. The abundant remains of Azolla that characterize basal middle Eocene marine deposits of all Nordic seas probably represent transported assemblages resulting from freshwater spills from the Arctic Ocean that reached as far south as the North Sea. The termination of the Azolla phase in the Arctic coincides with a local sea surface temperature rise from approximately 10 degrees C to 13 degrees C, pointing to simultaneous increases in salt and heat supply owing to the influx of waters from adjacent oceans. We suggest that onset and termination of the Azolla phase depended on the degree of oceanic exchange between Arctic Ocean and adjacent seas.     

A climatologically significant aerosol longwave indirect effect in the Arctic

The warming of Arctic climate and decreases in sea ice thickness and extent observed over recent decades are believed to result from increased direct greenhouse gas forcing, changes in atmospheric dynamics having anthropogenic origin, and important positive reinforcements including ice-albedo and cloud-radiation feedbacks. The importance of cloud-radiation interactions is being investigated through advanced instrumentation deployed in the high Arctic since 1997 (refs 7, 8). These studies have established that clouds, via the dominance of longwave radiation, exert a net warming on the Arctic climate system throughout most of the year, except briefly during the summer. The Arctic region also experiences significant periodic influxes of anthropogenic aerosols, which originate from the industrial regions in lower latitudes. Here we use multisensor radiometric data to show that enhanced aerosol concentrations alter the microphysical properties of Arctic clouds, in a process known as the 'first indirect' effect. Under frequently occurring cloud types we find that this leads to an increase of an average 3.4 watts per square metre in the surface longwave fluxes. This is comparable to a warming effect from established greenhouse gases and implies that the observed longwave enhancement is climatologically significant.     

Vertical structure of recent Arctic warming

Near-surface warming in the Arctic has been almost twice as large as the global average over recent decades-a phenomenon that is known as the 'Arctic amplification'. The underlying causes of this temperature amplification remain uncertain. The reduction in snow and ice cover that has occurred over recent decades may have played a role. Climate model experiments indicate that when global temperature rises, Arctic snow and ice cover retreats, causing excessive polar warming. Reduction of the snow and ice cover causes albedo changes, and increased refreezing of sea ice during the cold season and decreases in sea-ice thickness both increase heat flux from the ocean to the atmosphere. Changes in oceanic and atmospheric circulation, as well as cloud cover, have also been proposed to cause Arctic temperature amplification. Here we examine the vertical structure of temperature change in the Arctic during the late twentieth century using reanalysis data. We find evidence for temperature amplification well above the surface. Snow and ice feedbacks cannot be the main cause of the warming aloft during the greater part of the year, because these feedbacks are expected to primarily affect temperatures in the lowermost part of the atmosphere, resulting in a pattern of warming that we only observe in spring. A significant proportion of the observed temperature amplification must therefore be explained by mechanisms that induce warming above the lowermost part of the atmosphere. We regress the Arctic temperature field on the atmospheric energy transport into the Arctic and find that, in the summer half-year, a significant proportion of the vertical structure of warming can be explained by changes in this variable. We conclude that changes in atmospheric heat transport may be an important cause of the recent Arctic temperature amplification.     

Warm tropical ocean surface and global anoxia during the mid-Cretaceous period

The middle of the Cretaceous period (about 120 to 80 Myr ago) was a time of unusually warm polar temperatures, repeated reef-drowning in the tropics and a series of oceanic anoxic events (OAEs) that promoted both the widespread deposition of organic-carbon-rich marine sediments and high biological turnover. The cause of the warm temperatures is unproven but widely attributed to high levels of atmospheric greenhouse gases such as carbon dioxide. In contrast, there is no consensus on the climatic causes and effects of the OAEs, with both high biological productivity and ocean 'stagnation' being invoked as the cause of ocean anoxia. Here we show, using stable isotope records from multiple species of well-preserved foraminifera, that the thermal structure of surface waters in the western tropical Atlantic Ocean underwent pronounced variability about 100 Myr ago, with maximum sea surface temperatures 3-5 degrees C warmer than today. This variability culminated in a collapse of upper-ocean stratification during OAE-1d (the 'Breistroffer' event), a globally significant period of organic-carbon burial that we show to have fundamental, stratigraphically valuable, geochemical similarities to the main OAEs of the Mesozoic era. Our records are consistent with greenhouse forcing being responsible for the warm temperatures, but are inconsistent both with explanations for OAEs based on ocean stagnation, and with the traditional view (reviewed in ref. 12) that past warm periods were more stable than today's climate.     

Environmental precursors to rapid light carbon injection at the Palaeocene/Eocene boundary

The start of the Palaeocene/Eocene thermal maximum--a period of exceptional global warming about 55 million years ago--is marked by a prominent negative carbon isotope excursion that reflects a massive input of 13C-depleted ('light') carbon to the ocean-atmosphere system. It is often assumed that this carbon injection initiated the rapid increase in global surface temperatures and environmental change that characterize the climate perturbation, but the exact sequence of events remains uncertain. Here we present chemical and biotic records of environmental change across the Palaeocene/Eocene boundary from two sediment sections in New Jersey that have high sediment accumulation rates. We show that the onsets of environmental change (as recorded by the abundant occurrence ('acme') of the dinoflagellate cyst Apectodinium) and of surface-ocean warming (as evidenced by the palaeothermometer TEX86) preceded the light carbon injection by several thousand years. The onset of the Apectodinium acme also precedes the carbon isotope excursion in sections from the southwest Pacific Ocean and the North Sea, indicating that the early onset of environmental change was not confined to the New Jersey shelf. The lag of approximately 3,000 years between the onset of warming in New Jersey shelf waters and the carbon isotope excursion is consistent with the hypothesis that bottom water warming caused the injection of 13C-depleted carbon by triggering the dissociation of submarine methane hydrates, but the cause of the early warming remains uncertain.     

Arctic warming aloft is data set dependent

Arctic sea ice and snow on land have retreated polewards at an alarming pace in the past few decades. Such retreat locally amplifies surface warming through a positive feedback, which causes the Arctic surface to warm faster than the rest of the globe. In contrast, ice and snow retreat causes little warming in the atmosphere above when the stable winter atmosphere inhibits vertical heat exchange. We therefore find surprising the recent report by Graversen et al. in which they claim that recent Arctic atmospheric warming extends far deeper into the atmosphere than expected, and can even exceed the surface warming during the polar night. Using a different data set, we show that there is much less warming aloft in winter, consistent with the recent retreat of ice and snow, as well as recent changes in atmospheric heat transport.     

Interdecadal variation in the extent of South Pacific tropical waters during the Younger Dryas event

During the Younger Dryas event, about 12,000 years ago, the Northern Hemisphere cooled by between 2 and 10 degrees C (refs 1, 2) whereas East Antarctica experienced warming. But the spatial signature of the event in the southern mid-latitudes and tropics is less well known, as records are sparse and inconclusive. Here we present high-resolution analyses of skeletal Sr/Ca and 18O/16O ratios for a giant fossil Diploastrea heliopora coral that was preserved in growth position on the raised reef terraces of Espiritu Santo Island, Vanuatu, in the southwestern tropical Pacific Ocean. Our data indicate that sea surface temperatures in Vanuatu were on average 4.5 +/- 1.3 degrees C cooler during the Younger Dryas event than today, with a significant interdecadal modulation. The amplified annual cycle of sea surface temperatures, relative to today, indicates that cooling was caused by the compression of tropical waters towards the Equator. The positive correlation in our record between the oxygen isotope ratios of sea water and sea surface temperatures suggests that the South Pacific convergence zone, which brings 18O-depleted precipitation to the area today, was not active during the Younger Dryas period.     

Low-latitude seasonality of Cretaceous temperatures in warm and cold episodes

The Cretaceous period is generally considered to have been a time of warm climate. Evidence for cooler episodes exists, particularly in the early Cretaceous period, but the timing and significance of these cool episodes are not well constrained. The seasonality of temperatures is important for constraining equator-to-pole temperature gradients and may indicate the presence of polar ice sheets; however, reconstructions of Cretaceous sea surface temperatures are predominantly based on the oxygen isotopic composition of planktonic foraminifera that do not provide information about such intra-annual variations. Here we present intra-shell variations in delta18O values of rudist bivalves (Hippuritoidea) from palaeolatitudes between 8 degrees and 31 degrees N, which record the evolution of the seasonality of Cretaceous sea surface temperatures in detail. We find high maximum temperatures (approximately 35 to 37 degrees C) and relatively low seasonal variability (< 12 degrees C) between 20 degrees and 30 degrees N during the warmer Cretaceous episodes. In contrast, during the cooler episodes our data show seasonal sea surface temperature variability of up to 18 degrees C near 25 degrees N, comparable to the range found today. Such a large seasonal variability is compatible with the existence of polar ice sheets.     

Eocene global warming events driven by ventilation of oceanic dissolved organic carbon

'Hyperthermals' are intervals of rapid, pronounced global warming known from six episodes within the Palaeocene and Eocene epochs (～65-34 million years (Myr) ago). The most extreme hyperthermal was the ～170 thousand year (kyr) interval of 5-7 °C global warming during the Palaeocene-Eocene Thermal Maximum (PETM, 56?Myr ago). The PETM is widely attributed to massive release of greenhouse gases from buried sedimentary carbon reservoirs, and other, comparatively modest, hyperthermals have also been linked to the release of sedimentary carbon. Here we show, using new 2.4-Myr-long Eocene deep ocean records, that the comparatively modest hyperthermals are much more numerous than previously documented, paced by the eccentricity of Earth's orbit and have shorter durations (～40?kyr) and more rapid recovery phases than the PETM. These findings point to the operation of fundamentally different forcing and feedback mechanisms than for the PETM, involving redistribution of carbon among Earth's readily exchangeable surface reservoirs rather than carbon exhumation from, and subsequent burial back into, the sedimentary reservoir. Specifically, we interpret our records to indicate repeated, large-scale releases of dissolved organic carbon (at least 1,600 gigatonnes) from the ocean by ventilation (strengthened oxidation) of the ocean interior. The rapid recovery of the carbon cycle following each Eocene hyperthermal strongly suggests that carbon was re-sequestered by the ocean, rather than the much slower process of silicate rock weathering proposed for the PETM. Our findings suggest that these pronounced climate warming events were driven not by repeated releases of carbon from buried sedimentary sources, but, rather, by patterns of surficial carbon redistribution familiar from younger intervals of Earth history.     

Greenhouse gas mitigation can reduce sea-ice loss and increase polar bear persistence

On the basis of projected losses of their essential sea-ice habitats, a United States Geological Survey research team concluded in 2007 that two-thirds of the world's polar bears (Ursus maritimus) could disappear by mid-century if business-as-usual greenhouse gas emissions continue. That projection, however, did not consider the possible benefits of greenhouse gas mitigation. A key question is whether temperature increases lead to proportional losses of sea-ice habitat, or whether sea-ice cover crosses a tipping point and irreversibly collapses when temperature reaches a critical threshold. Such a tipping point would mean future greenhouse gas mitigation would confer no conservation benefits to polar bears. Here we show, using a general circulation model, that substantially more sea-ice habitat would be retained if greenhouse gas rise is mitigated. We also show, with Bayesian network model outcomes, that increased habitat retention under greenhouse gas mitigation means that polar bears could persist throughout the century in greater numbers and more areas than in the business-as-usual case. Our general circulation model outcomes did not reveal thresholds leading to irreversible loss of ice; instead, a linear relationship between global mean surface air temperature and sea-ice habitat substantiated the hypothesis that sea-ice thermodynamics can overcome albedo feedbacks proposed to cause sea-ice tipping points. Our outcomes indicate that rapid summer ice losses in models and observations represent increased volatility of a thinning sea-ice cover, rather than tipping-point behaviour. Mitigation-driven Bayesian network outcomes show that previously predicted declines in polar bear distribution and numbers are not unavoidable. Because polar bears are sentinels of the Arctic marine ecosystem and trends in their sea-ice habitats foreshadow future global changes, mitigating greenhouse gas emissions to improve polar bear status would have conservation benefits throughout and beyond the Arctic.