U33B-01 INVITED
Chasing An Analogue For The Holocene : The Astronomical Forcing
Astronomical theories of paleoclimate, e.g. the Milankovitch theory, explain the long-term variations of climate by the changes in the Earth orbit and position against the Sun, and consequently by the change in the distribution of the solar energy reaching the Earth. On the other hand, anthropogenic activities from early agricultural practices to more recent fossil fuel burning also impact climate change. Models, from conceptual ones to the most sophisticated general circulation models can be used to try to disentangle the contribution from human activity and the natural contribution in the record of climate change. Alternatively, a comparison of the climate records during past interglacials similar to the one we are living in can give some insight into the natural behaviour of the climate system during an interglacial. As far as the long-term climate change is mostly orbitally driven, we will search orbital and insolation time series for past analogues of the non-human perturbed Holocene. When doing so, some key questions must be answered. The first one is related to the time interval to be used as target. A previous study (Loutre and Berger, 2000) focused on the interglacial itself, assuming that the preceding deglaciation had a negligible impact on the interglacial. Rather we decided here to choose a target time interval that includes the deglaciation. Another question is related to the choice of the variable that will be used for correlation. It can be the orbital parameters themselves. It is also possible to use top-of-the-atmosphere insolation. If daily insolation is chosen, the latitude and time in the year for which it is computed are crucial; if a more time- integrated insolation is used (e.g. seasonal insolation), the time interval for the integration is an essential feature. Loutre and Berger (2000) used mid-June insolation at 65N and identified MIS11 as the most recent potential analogue for the future climate. A higher correlation of the insolation was even obtained between the future and MIS19. Here, we focus on the Holocene, starting from the last glacial maximum up to the next millennia (21 ka BP to 9 ka AP). There is a good correlation between the insolation changes during most of the recent interglacials and this target time interval, but the correlation can be poor for the orbital parameters taken separately, due to different leads and lags between the orbital parameters among the different interglacials. Namely, the maximum of eccentricity of the Holocene (14 ka BP) leads the minimum in climatic precession by 2 ka and the maximum of obliquity by 5 ka. At MIS5, the maximum of obliquity occurs 4 ka before the minimum in climatic precession and several thousands of years before the maximum of eccentricity. This rapid comparison underlines that certain interglacials are very poor analogues for the Holocene by reference to the astronomical parameters. A more systematic investigation confirms that no interglacial is displaying high correlations with the Holocene for all the astronomical parameters. For example, MIS11 shows as strong correlation with the present and future insolation (Loutre and Berger, 2000) but the correlation is much smaller for the orbital parameters. On the point of view of the orbital parameters, MIS15 might be the best compromise of the Late Quaternary. Loutre M.F. and Berger A. 2000. Future climatic changes: are we entering an exceptionally long interglacial? Climatic Change 46: 61–90.
U33B-02 INVITED
Ice core records of the evolution of atmospheric methane in the Holocene
Atmospheric methane mixing ratios declined from peak values of ~680-730 ppb in the early Holocene, to a broad minimum of ~560-610 ppb in the mid-Holocene, then rose toward values of 690-725 ppb in the late pre-Industrial Holocene. The mid-Holocene minimum is unusual relative to the long-term ice core record, leading to suggestions that the post 5 ka increase may have been due to early human sources. We generated two new high-precision Holocene ice core methane records, from Greenland and Antarctica, analyzing over 380 new samples in duplicate. This allows an accurate comparison of absolute methane levels in the high latitude northern and southern hemispheres, providing clues to changes in source locations. The late Holocene rise is associated with a decreasing inter-polar methane gradient, indicating a dominating increase in low latitude sources. Could the rise be due to natural changes in emissions from wetlands? Northern hemisphere summer isolation decreased over this time period, which should have decreased temperature and precipitation in many methane-producing regions. One important observation, though, is that over 60% of the late pre-industrial Holocene rise happened after 2,000 years ago. Asian monsoon proxies from Chinese cave deposits, and the isotopic composition of atmospheric oxygen (Severinghaus et al., this meeting), suggest increased rainfall in the tropics after 2,000 years B.P. In addition, southern hemisphere wetland methane sources must have increased as southern summer insolation increased through the Holocene, though the contribution to the global budget is unclear. Another puzzling aspect of the data is that mid-Holocene results seem to require a source 'flip-flop'. The mid- Holocene minimum is associated with a high gradient, indicating relatively larger northern sources and smaller tropical sources. This may reflect the drying of a tropical region, perhaps in Africa (which is also consistent with the atmospheric oxygen isotope data), and increases in boreal emissions as the Laurentide ice sheet disappeared. In the early Holocene, a slow decline of methane concentrations and inter-polar gradient from 11,000 years B.P. to ~6,000 years B.P. reflects decreasing sources globally, but a relatively larger decline in the high latitude north. This trend is plausibly related to declining northern hemisphere insolation. We also produced a record with decadal resolution for the last 1000 years from the WAIS Divide ice core in Antarctica. The data show obvious methane maxima at about 1150, 1300, 1550, and 1700 AD, consistent with the one other record (Law Dome) that covers this time period in detail (MacFarling Meure et al., 2006; Geophysical Research Letters, 33). Preliminary high-resolution data from Greenland match these patterns extremely well. A direct relationship with population trends or climate patterns is not yet obvious.
U33B-03
Speleothem records of changes in tropical hydrology during the Holocene
The largest natural source of atmospheric methane today is wetlands, and approximately two-thirds of wetland methane emissions come from the tropics. Tropical methane emissions are primarily controlled by temperature and water-table variations. During the Holocene, when tropical temperatures varied little, changes in tropical methane emissions should dominantly reflect changes in tropical hydrology, with a particularly strong response to changes in the monsoons. One method of estimating changes in the tropical hydrological cycle and, therefore, a potential measure of the magnitude of tropical methane sources, is oxygen isotope time series from speleothems. High-resolution oxygen isotopic time series from several complete, well-dated Holocene speleothems exist for the Indian and East-Asian monsoon regions. Although often from caves thousands of kilometers distant from one another, these records show remarkable coherence, suggesting that they are excellent recorders of regional-scale changes in the Indian-East Asian monsoons. For the Holocene, the first order trend in these records is a monotonic increase in oxygen isotope values that closely follows the precessional cycle in summer solar insolation. The trend in oxygen isotopes is interpreted as reflecting an insolation-driven, steady southward migration of the mean location of the ITCZ, and an associated steady decrease in Indian - East Asian monsoon summer monsoon rainfall throughout this period. Other records of Northern Hemisphere tropical moisture support this interpretation. If interpreted as indicating decreased tropical methane emissions, these records suggest that Northern Hemisphere wetlands are unlikely to be the source of the observed late Holocene increase in atmospheric methane. As the ITCZ migrates southward, however, a complimentary increase in southern hemisphere tropical rainfall should be expected. And, indeed, such an increase is observed in southern hemisphere speleothem and other records. Thus, whether tropical methane sources can or cannot account for the late Holocene rise in methane would seem to depend on the relative changes in methane emissions from southern vs northern hemisphere tropical wetlands.
U33B-04 INVITED
Dynamics of CO2 and CH4 changes: MIS 11 versus Holocene
The behaviour of CO2 and CH4 during MIS 11 has been of particular interest, in particular in the light of Ruddiman's suggestion (2003) that the rise of these greenhouse trace gases during a large part of the Holocene has been of anthropogenic origin, in contrast with the traditional view of a limited anthropogenic influence starting with the industrial era. One observation of ice cores supporting Ruddiman's hypothesis was the fact that the Holocene long-term increases of CO2 and CH4 are unusual compared with a corresponding decrease prevailing during MIS 5, 7 and 9. We present here the high-resolution EDC ice core records of CO2 and CH4 during MIS 11, which enable to discuss in details the dynamics of these greenhouse gases. In particular they show a long-term increase like during the Holocene. It has been argued (Ruddiman 2007) that one should look at the trends in greenhouse gases in comparison to particular alignements of orbital characteristics. We also compare this new high- resolution record with the record of Antarctic temperature and different insolation parameters and conclude that the ice core record may also question this interpretation. The MIS 11 record will also be compared with other high resolution interglacial records of greenhouse gases, highlighting the diversity of past interglacial periods.
U33B-05
Holocene Changes in Land Cover and Greenhouse-gas Concentrations: Rethinking Natural vs Anthropogenic Causation
The Holocene has witnessed a switch from a nature-dominated to a human-dominated Earth system. Although globally-significant human impacts (wildfire, megafaunal extinctions) occurred during the late Pleistocene, it was the advent of agriculture that led to the progressive transformation of land cover, and which distinguishes the Holocene from previous interglacial periods. A wide array of data provide clear evidence of local-to-regional human disturbance from ~5 ka BP, in some cases earlier. There is more uncertainty about when the anthropogenic "footprint" became detectable at a global scale, and there has consequently been debate about how much of the pre-industrial increase in atmospheric greenhouse gas concentrations is attributable to human causation, linked to processes such as deforestation (CO2) and wet rice cultivation (CH4). Although there has been recent progress in developing quantitative methods for translating pollen data into palaeo-land cover, such as the REVEALS model of Sugita (Holocene 2007) coupled to GIS, this has yet to be widely applied to existing data bases, and most pollen-based land-use reconstructions remain qualitative or semi-quantitative. Lake trophic status, sediment flux / soil erosion, and microcharcoal records of biomass burning provide alternative proxies that integrate regional-scale landscape disturbance. These proxy data along with documentary sources imply that globally-significant changes in land cover occurred prior to ~250 BP which must have altered atmospheric greenhouse gas concentrations by this time. The polarised debate for and against early anthropogenic impact on global carbon cycling mirrors our industrial-era division between nature and society, both conceptually (e.g. Cartesian dualism) and on the ground (e.g. demarcating land between monoculture agriculture and wilderness). However, for the period before ~1750 AD, this likely represents a false dichotomy, because pre-industrial societies more often formed part of the natural world, while at the same time modifying and transforming it. Attempts to partition carbon emissions between natural and anthropogenic sources during the Holocene may therefore be misplaced. Many landscapes, such as savannas, are the result of synergistic – and in some cases contingent – relationships between people, other animals, plants and other components of nature. The issue is thus not whether early humans altered carbon cycling (they did…), but rather at what point it became detectable at a global scale, and what form it took.
U33B-06
Simulations of carbon cycle dynamics during the Holocene
Effects of different biogeochemical processes of natural origin on atmospheric CO2 during the course of the Holocene are tested using the coupled climate-carbon cycle model of intermediate complexity CLIMBER-2. The model includes dynamic atmospheric, oceanic, sea ice, and land surface modules. A coarse geographical resolution of the physical climate system allows a high computational efficiency of the model. The oceanic biogeochemistry module includes the inorganic carbon cycle and marine biology. Carbonate compensation is simulated by a model of deep sea sediment diagenesis. A model of coral reef growth is added to account for changes in carbonate sedimentation in shallow waters. The dynamical global vegetation model LPJ is used to simulate the land carbon dynamics. The model operates on a high spatial resolution of 0.5 degree (longitude and latitude) and is coupled to CLIMBER-2 via the climate anomalies approach. Changes in land carbon storage in response to altered climate and atmospheric CO2 are accounted for interactively in the global carbon cycle. A weathering module calculates changes in the riverine bicarbonate flux. A scenario of peat accumulation in boreal regions is taken as an external forcing. The coupled model is driven by changes in the orbital forcing from 8,000 yr BP to pre-industrial. We compare contributions of different natural factors: changes in land carbon storage including peat, changes in sea surface temperatures and circulation, carbonate sedimentation in shallow waters and deep sea. We found that coral reef growth is the most essential factor explaining the CO2 increase. The land biomass dynamics is essential, but overall the land is a sink for carbon due to significant peat accumulation. The contribution of changes in sea surface temperatures and ocean dynamics is quite small. A delayed response to the reorganization of oceanic circulation and the regrowth of the terrestrial biosphere before 8,000 yr BP might be significant but its quantification requires a long-term transient simulation of the model through the compete glacial cycle. The CO2 analysis is complemented by a comparison of simulated d13CO2 changes with ice core data.
U33B-07 INVITED
Climate Model Tests of the Early Anthropogenic Hypothesis
We test the hypothesis that greenhouse gas emissions produced by the combination of early and recent human activities, augmented by additional rises in greenhouse gases through ocean feedbacks, have kept the climate warmer than its natural level and offset an incipient glaciation. We use four different configurations of NCAR's Community Climate System Model to investigate the natural climate that should exist today if CO2 and CH4 concentrations had fallen to their average levels reached during previous interglaciations. The model simulations consist of three using a coupled atmosphere-slab ocean configuration---fixed land cover at moderate (T42) and high (T85) model resolution and interactive vegetation composition at T42 resolution—--and one employing a coupled atmosphere-dynamical ocean configuration and fixed land cover at T42 resolution. With greenhouse gas concentrations lowered to their estimated natural levels, global mean temperature falls by 2.5-3.0 K in all four experiments. Of the total global cooling with fixed land cover and moderate model resolution, 38% (62%) is attributable to early agricultural activities (industrialization), while early agriculture accounts for approximately half of the expanded permanent snow cover area. The greenhouse cooling triggers widespread glacial inception in the Northern Hemisphere, where permanent snow cover expands by at least 80% and even more with the addition of enhanced model processes: 130% with the dynamical ocean, 150% with high (T85) model resolution, and 200% with vegetation feedbacks included. The regional pattern of incipient glaciation is strongly influenced by atmospheric and circulation changes, sea ice feedbacks, and model resolution. The simulation with a dynamical ocean produces a decrease in vertically integrated global ocean temperature of 1.25 K, a 20% weakening of the Atlantic meridional overturning cell, and an expansion of sea ice and reduced upwelling in the Southern Ocean. Viewed from the perspective of explaining the unusual late-Holocene increases of CO2 that occurred prior to the Industrial Revolution, these simulated changes in ocean temperature, sea ice cover, and circulation (with sign reversed) support the hypothesis that early agriculture played a role in initiating anomalous warming that thwarted incipient glaciation beginning several thousand years ago. Decreased ocean solubility globally and positive ocean/sea-ice feedbacks in the Southern Hemisphere probably augmented the initial CO2 increase and caused additional warming.
U33B-08
Falsification of Natural Explanations for Late Holocene Greenhouse-Gas Increases
The cause of the slow increases in atmospheric CO2 and CH4 concentrations during the late Holocene is vigorously debated. Natural explanations predict that similar increases should have occurred early in previous interglaciations when insolation forcing was similar, but six comparable interglaciations (isotopic stages 5, 7, 9, 11, 17 and 19) show gas decreases rather than increases during the most similar intervals. These dozen failures falsify natural explanations for the Holocene gas increases. In addition, simulations with carbon-climate models have not yet been able to reproduce both the gas increases during the late Holocene and the decreases early in previous interglaciations. Falsification of natural explanations leaves only the anthropogenic one. The start of the Holocene CO2 rise coincides with extensive forest clearance in Europe ~7500 years ago. The start of the CH4 increase coincides with a rapid spread of rice irrigation across China ~5000 years ago.