Years Ago

A.6. A "pie-chart" representation of possible contributions to the anomalous CO2 trend during the last 7,000 years from: warming of the deep ocean (resulting in decreased CO2 solubility), direct anthropogenic emissions, and maintenance of anomalous warmth in the Southern Ocean.

contact with the atmosphere, which, according to several conceptual models, would have kept CO2 concentrations from falling. The size of this additional CO2 feedback effect is not known.

New modeling results published in collaboration with atmospheric scientists John Kutzbach, Steve Vavrus, and Gwenaelle Phillipon at the University of Wisconsin support the evidence from marine and ice-core observations. We ran an experiment with an atmospheric general circulation model that included an interactive dynamical ocean in order to assess the effect of early anthropogenic greenhouse gases on the atmosphere and ocean, and the simulation produced warming effects similar to those from the oxygen-isotope and deuterium evidence. The early anthropogenic greenhouse gases would have warmed the deep ocean by an amount slightly smaller than that indicated by the oxygen-isotopic trends (fig. A.5b). They would also have made the atmosphere over Antarctica warmer by an amount that closely matched the anomalous warmth indicated by the deuterium data (fig. A.5a).

A deep ocean that warmed in the Holocene by 0.5—0.84oC would have released additional CO2 into the atmosphere and raised concentrations by about 5 to 6 parts per million. This additional CO2 release would fill much of the gap left between the direct anthropogenic emissions and the observed trend (fig. A.6), especially if the lower CO2 value in latest preindustrial time (1600-1750) is used as the "target." At this point, it is not possible to quantify the carbon exchanges in an Antarctic region that stayed warmer in the Holocene compared to previous interglaciations. As a result, we do not yet know whether additional CO2 from a warm Southern Ocean can fill the remaining "carbon gap" needed to explain the full 35-40 parts per million CO2 anomaly (fig. A.6).

Is a New GLAciAtion Overdue?

The original anthropogenic hypothesis also included the claim that new ice sheets would have begun forming by now if CO2 and CH4 levels had fallen to their natural levels rather than being driven upward by emissions from early agricultural activities. In chapter 10 of the book, I reported results from an experiment run with the GENESIS general circulation model with coauthors Steve Vavrus and John Kutzbach at the University of Wisconsin. That experiment simulated year-round snow cover in a few grid boxes along the high spine of Baffin Island. We regard the occurrence of year-round snow cover that can thicken through time as evidence of glacial inception. I commented that I had been hoping for a home run with this simulation, but the outcome felt more like beating out an infield single.

In subsequent years, Steve Vavrus, John Kutzbach, and I, along with Gwe-naelle Phillipon, have run several simulations with the National Center for Atmospheric Research's Community Climate Model (version CCSM3). In all experiments, the control-case conditions were altered by lowering the atmospheric CO2 concentration to 240 parts per million and the CH4 value to 450 parts per billion to match the values proposed in the early anthropogenic hypothesis. One simulation used the basic atmospheric version of the model, while others included vegetation/albedo feedback, ocean dynamics (noted above), and higherresolution topography.

All of these experiments simulated year-round snow cover over various high-altitude and high-latitude regions in Canada and northern Eurasia. For example, the dynamical-ocean experiment simulated twelve-month snow cover across the

A.7. Regions with year-round snow cover (incipient glaciation shown in white) from a general circulation model simulation run with atmospheric CO2 set to 240 parts per million and CH4 to 450 parts per billion, the values that the early anthropogenic hypothesis predicts would occur now in the absence of human intervention. White circles are modern-day ice caps and mountain glaciers.

A.7. Regions with year-round snow cover (incipient glaciation shown in white) from a general circulation model simulation run with atmospheric CO2 set to 240 parts per million and CH4 to 450 parts per billion, the values that the early anthropogenic hypothesis predicts would occur now in the absence of human intervention. White circles are modern-day ice caps and mountain glaciers.

northern Rockies, the Canadian archipelago, Baffin Island in northeastern Canada, eastern Siberia, and several Arctic islands north of Siberia (fig A.7). The total area (outside Greenland) with year-round snow in that experiment was 30 percent larger than the size of the modern-day Greenland ice sheet.

Many of the snow-covered grid boxes surround locations of modern-day ice caps and mountain glaciers that are now melting back from more extensive prein-dustrial (Little Ice Age) limits. Such a configuration makes sense from comparing the relative levels of the greenhouse gases. If Little Ice Age glaciers and ice caps were able to form or grow across small regions when the CO2 concentration was around 270-275 parts per million and the CH4 level was 670-690 parts per billion, then much more extensive snow cover would be expected with CO2 set to 240 parts per million and CH4 to 450 parts per billion in the experiment.

The results from these experiments felt like at least an extra base hit, but it would be premature to start a full "high-five" celebration of a successful prediction. The CCSM3 model has a moderate cold bias in summer that tends to favor "extra" snow cover, and the control-case simulation for that model with "modern" greenhouse-gas levels places year-round snow cover on a few grid boxes on islands in the Eurasian Arctic, where snow actually disappears in summer today and older ice caps are melting. To be certain that our findings are not the result of model bias, it will be necessary to run similar experiments with other models, including the new CCSM4 model soon to be made available at NCAR.

PAnDEmics, MASS MoRtAlity, AND CO2 OsciHAtions

In a corollary to the early anthropogenic hypothesis, I claimed in part 4 of the book that several dips in CO2 that lasted for several centuries were not entirely natural in origin but resulted at least in part from well-documented episodes of mass mortality during the historical era. The central idea was that mass mortality allowed forests to reoccupy abandoned land, sequester carbon from the atmosphere, and drive CO2 levels lower. At that time, only two CO2 records from Antarctica were available to test this idea: a well-dated record from Law Dome that extended back slightly less than 1,000 years and a poorly dated record from Taylor Dome that extended back through the entire historical era and beyond. The Taylor Dome ice had not been analyzed for the presence of volcanogenic layers needed to improve the dating.

In 2006, MacFarling Meure and colleagues in New Zealand published a very detailed and well-dated CO2 record from Law Dome (fig. A.8). The new analyses spanning the last 1,000 years fit seamlessly into the earlier ones, and the earlier part of the record was pushed back to nearly 2,000 years ago. Plotted below the CO2 record is a summary of the major mass-mortality events of the historical era (prior to industrialization). The correlation between the two signals is striking: every major CO2 decrease lines up with an episode of mass mortality in Europe and China or the Americas. Intervals of stable or increasing CO2 are free of mortality disasters.

As noted in the book, pandemics are the major factor in these depopulation episodes: bubonic plague and other outbreaks in the late Roman era (200-600), bubonic plague again (the "Black Death") in late medieval times (1348-1400),

CO2 (parts per million) 284 282 280 278 276 274 272

100 200 300 400 500 600 700 800 900 1000 1100 12001300 14001500 1600 1700 1800

50m Americas (11%)

Depopulation (% of global)

Population (106) 200 I-

100 i

0 100 200 300 400 500 600 700 800 900 1000 1100 12001300 14001500 1600 1700 1800


A.8. High-resolution CO2 analyses from Law Dome Antarctica show CO2 decreases that match episodes of mass human mortality caused by pandemics in Europe and the Americas and by civil strife in China.

100 i

and the "virgin soil" epidemics brought to the Americas by Europeans (1492— 1700). Subsequent work with Ann Carmichael of Purdue University, an expert on the history of diseases, made it clearer to me that the two depopulation episodes in China were probably not the result of pandemics, but of extreme civil strife caused by invasions from Mongolia, including Genghis Kahn and his descendants in the 1200s.

Correlation is not causality, but the excellent match in figure A.8 certainly supports the proposed link between mass mortality and CO2 decreases. In contrast, this more detailed CO2 record raises questions about natural explanations for the CO2 decreases, especially the drop of nearly 10 parts per million from high values in the medieval era around 1200 to the low values in the Little Ice Age between 1600 and 1700. As I noted in chapter 12 of the book, carbon-climate models suggest that each 1oC decrease in global temperature should cause a 12 parts per million drop in CO2 because of increased CO2 solubility in the cooler ocean and also in part from decreased oxidation and carbon release from vegetation litter on land. Proxy-based estimates of Northern Hemisphere temperatures now bracket the Northern Hemisphere cooling from 1200 to 1700 between 0.2o and 0.5 oC, with a mid-range estimate of 0.3—0.35 oC. According to the models, a cooling of that size could account for a CO2 drop of 3 to 4 parts per million, but this estimate would leave the other two-thirds of the observed decrease to be explained by other factors, such as anthropogenic forcing.

My attempts in recent years to simulate the effects of sequestering billions of tons of carbon on abandoned farmland suggest that reforestation could directly explain at least 4 parts per million of the observed 10 parts per million drop. This analysis left open the possibility of a larger anthropogenic amount if CO2 feedback from the ocean (especially the Southern Ocean) boosted the direct response.

Improving on these estimates would require better data on the per-capita "forest footprint" of the farmers who died in order to calculate the resulting carbon sequestration more accurately. Another uncertainty that needs further attention is the question of how much farmland that might have been abandoned when its farmers died was instead taken over by family members or neighbors and thus prevented from reverting to the wild. Population densities in the Mediterranean region during the late Roman era were probably high enough that substantial reoccupation of this kind occurred. By the time of the Black Death pandemic, much of west-central Europe was also populated at high enough levels that reoccupation would have been common, although not in less populated northeastern regions like the Baltic and Russia. This reoccupation, along with the amazingly rapid recovery to prepandemic population levels by 1500, might explain why the imprint of the European Black Death on the CO2 signal is subtle.

In contrast, the 85 to 90 percent mortality rates of indigenous people in the

Americas between 1500 and 1700 left almost no one to reoccupy abandoned farmland, and the land would have reverted to forest. This pandemic likely played an important role in the CO2 drop to low concentrations near 1600. Because European populations in the Americas remained low until 1750—1800, the areas reforested by this massive pandemic remained reforested for centuries.

In summary, the years since publication of Plows, Plagues, and Petroleum have seen numerous follow-up efforts to test and explore various aspects of the early anthropogenic hypothesis. In my (not entirely unbiased) opinion, these new contributions have had two results. First, the weaknesses of natural explanations of the Holocene greenhouse-gas increases have become clearer. The central weakness remains the continuing lack of an answer to a very simple question: If the upward trends in late-Holocene CO2 and CH4 concentrations were caused by natural factors, why do similar intervals during previous interglaciations fail to show upward trends? These multiple failures falsify the natural explanations.

Second, new research has addressed the most serious criticisms of the early anthropogenic hypothesis. Although these studies have not "clinched the case," they have provided evidence that points to plausible answers to all of the major criticisms. The net effect of this new research has been to keep the early anthropogenic hypothesis alive and well (and unfalsified).

Was this article helpful?

0 0

Post a comment