Earlier IGsEarlier IGs

A. 2. A thought experiment. If the small-amplitude insolation changes during the Holocene drove the observed CO2 increase (top), then the larger-amplitude insolation changes during previous interglaciations should have driven larger CO2 increases (bottom). Yet ice-core records show CO2 decreases.

Land Use and Population

Another general criticism of the early anthropogenic hypothesis has centered on the idea that there could not have been enough people alive so many thousands of years ago to have taken control of greenhouse-gas trends and driven them upward during subsequent millennia. For the start of the historical era 2,000 years ago, the best estimates point to a global population of about 200 million people.

CO2 (parts per million)

Ha per Capita

Total Ha (108)

7000

CO2 (parts per million)

Ha per Capita

Total Ha (108)

800 600

6000

5000

2000

1000

4000 3000 Years ago

A.3. Trends in atmospheric CO2 concentration and global population (A), estimated per-capita land use (B), and total (global) forest clearance (C) during the middle and late Holocene.

6000

5000

2000

1000

800 600

4000 3000 Years ago

A.3. Trends in atmospheric CO2 concentration and global population (A), estimated per-capita land use (B), and total (global) forest clearance (C) during the middle and late Holocene.

Prior populations are not known, but hindcasts based on an assumed population doubling time of 1,000 years suggest that perhaps 10 to 20 million people lived 6,000 years ago (fig. A.3a).

In chapter 7 (fig. 7.2) of the book, I showed that the beginning of the CO2 increase (then dated to 8,000 years ago in Vostok ice, now at 7,000 years ago in Dome C ice) coincided with the first spread of agriculture and land clearance across the forests of Europe. But the early rise of the CO2 signal that followed, and the leveling off of that early CO2 increase after 2,000 years ago, bears little resem-

blance to the late-rising population curve (fig. A.3a). This mismatch is an obvious argument against the hypothesis that humans were the major cause of the CO2 increase. As further support of this criticism, several model-based estimates of preindustrial land use that assume a close (i.e., nearly linear) relationship between population size and total land clearance have concluded that most global deforestation occurred within the last 300 to 400 years, during the time of the most rapid prein-dustrial population increase, and much later than the early CO2 rise.

But this assumption that land use and population are linked in a close relationship is not supported by results from several field-oriented disciplines. Anthropological studies across a range of contemporary cultures that still practice forms of shifting cultivation, such as slash-and-burn farming, provide insights into farming practices used millennia ago, when all agriculture was likely of this form. Studies in land-use archeology, paleoecology, paleobotany, and sedimentology provide constraints on past changes in type and extent of agriculture, in gradual replacement of natural vegetation by domesticated crops, and in enhanced erosion of slopes bared by deforestation and tilling of fields. The common message from these field disciplines is that per-capita land use during the last 7,000 years has not remained constant (that is, tied to population in a linear way), but instead has decreased by a large amount.

Decades ago, Ester Boserup proposed a sequence summarizing how land use has changed with increasing population (table A.1). In the earliest and least populated phase of agricultural development (the long-fallow phase), early farmers set fire to patches of forest and planted seeds in ash-enriched soil between charred stumps. When soil nutrients became depleted after a few years, people simply moved on to another plot, and then another, returning to the original plot only after twenty to twenty-five years or longer. This kind of agriculture required little per-capita labor, but the continuing rotation among plots used a large amount of land.

Through time, as increases in local population densities reduced the amount of available land, farmers were forced to shorten the fallow period. At some point, with still-ongoing population increases, farming became restricted to the same plot of land every year (annual cropping). They began to make use of technological innovations that increased yields per acre, including improved plows, livestock traction, irrigation, and fertilizers. Ultimately, many farmers began rotating two or more crops per year in the same field and developing sophisticated and extensive irrigation systems. Despite the benefit of iron tools and other new technology, this later phase of intensive farming required very large amounts of labor per person—collecting and spreading manure and compost, tending the livestock that supplied most of the manure, eliminating weeds and insects, maintaining irrigation canals, and other such efforts.

Table A.1

The Boserup Sequence of Changing Holocene Land Use

Table A.1

The Boserup Sequence of Changing Holocene Land Use

Shifting Agriculture

Intensive Agriculture

Type

Long-fallow Short-fallow

Annual cropping Multicropping

Population

Low >

> High

Hectares/person

2-6 1-2

0.3-0.6 0.05-0.3

Figure A.3b summarizes one estimate of changes in per-capita land use along the "Boserup Sequence" proposed by myself and Erie Ellis, an expert on land use. During the early part of the sequence, when all those who farmed used the long-fallow method, per-capita land use was several hectares per person (1 hectare = 2.4 acres), with a best estimate of 4 hectares. At the later (preindustrial) end of the sequence, land use had fallen to an estimated 0.4 hectares per person, prior to reaching even lower average values (0.2—0.3 hectares) during the 1800s and 1900s. Because the intervening trend is not known, several possible trajectories are shown. We infer that the actual trend is likely to have some kind of convex shape, based on the abundance of evidence of widespread technological innovation during the historical era.

The total area of land cleared over time on a global basis can be estimated as the product of global population (fig. A.3a) and the trends in per-capita land use (fig. A.3b). All three trends (fig. A.3c) show global clearance leveling out nearly 2,000 years ago, and the convex case even shows a small reduction in total land use since 1,500 years ago. The most critical factor in the leveling out of these trends since 2,000 years ago is the decrease in per-capita land use, which cancels out part or all of the rapid rise in population. With allowance for the Boserup land-use trend through the late Holocene, the long-term trend in total land use (fig. A.3c) now looks more like the CO2 signal (fig. A.3a).

Atmospheric CO2 concentrations depend, however, on the rate of clearance rather than the total amount. In addition, other factors would have to be considered to make a full comparison between changes in land use and atmospheric CO2 concentrations. Because early farming was concentrated in well-watered valleys with rich soils, forests cleared from those areas are likely to have been more carbon-dense than those on hillsides and slopes cleared in more recent times. Allowance for this gradual change in carbon density would further steepen the rate of the earlier Holocene CO2 emissions compared to the more recent ones. Finally, realistic comparisons will have to take into account the very long residence time of CO2 in the atmosphere. Some 15 to 20 percent of the total amount of CO2 injected into the atmosphere remains there for countless millennia.

A.4. The spread of rice cultivation in China prior to 4,000 years ago based on several hundred archeological sites. The gray background pattern shows modern-day rice paddy areas.

A second area of doubt concerning the early anthropogenic hypothesis has been whether early agriculture can account for the reversal in trend of the methane signal nearly 5,000 years ago. In a joint effort with Zhengtang Guo and his colleagues in Beijing, I found new evidence, based on a compilation of several hundred archeological sites from China, that shows a large increase in the number of sites with rice remains at the time of the methane reversal (fig. A.4). Prior to 7,000 years ago, scattered sites contain remains of what are thought to be strains of dry-adapted natural rice that may have been replanted in slightly moister soils. Between 7,000 and 6,000 years ago, as rice remains begin to become somewhat more numerous in the archeological record, direct evidence of irrigation. such as radiocarbon-dated wooden sluiceways. first appears. Between 6,000 and 4,000 years ago, the technique of growing wet-adapted rice strains in irrigated paddy fields spread rapidly across the area in south-central China where irrigated rice is grown today.

The number of sites dating to the interval between 5,000 and 4,000 years ago exceeds those dating to between 8,000 and 7,000 years ago by a factor of 10, an increase rapid enough to suggest a causal link with the reversal of the methane trend nearly 5,000 years ago. In addition, archeologists have concluded that rapidly expanding rice farming in south-central China and dry-land agriculture in north-central China caused a major surge in population during these millennia. These population increases would have produced still more methane from additional factors such as the growing numbers of livestock tended, greater biomass burning, and more generation of human waste.

Although these archeological compilations support the idea that humans played a role in the methane reversal nearly 5,000 years ago, this link cannot easily be quantified. The number of individual rice paddies in China (and in all of southern Asia) today is incalculable—probably at least in the tens of millions— whereas the archeological database consists of just a few hundred ancient sites located near paddies whose actual sizes are, in most cases, unknown.

Carbon Budgets and CO2 Feedback

The most telling criticism of the early anthropogenic hypothesis has been the conclusion that even deforestation of much of southern Eurasia, along with substantial parts of the Americas and Africa, could not have driven CO2 concentrations upward by about 20 parts per million or more prior to the industrial era (fig. A.1b), much less explain the full 35-40 parts per million size of the CO2 anomaly I proposed. In chapter 11 of the book, I conceded that my critics (most notably Fortunat Joos and colleagues) were right. And in a paper published in 2007, I concluded that direct anthropogenic emissions from deforestation, and lesser contributions from early coal burning in China and severe degradation of soil profiles in parts of Eurasia, could not account for more than about 9 to 10 parts per million of the 35-40 parts per million CO2 rise, or about 25 percent of the total.

Although some scientists inferred that this concession invalidated my hypothesis, they missed the fact that the CO2 trend in the last 7,000 years still departed from those in previous interglaciations by the same 35-40 parts per million. As a result, the anomaly had not disappeared. These two observations presented a major enigma: a 35-40 parts per million anomaly that appears to have been anthropogenic, but a direct contribution from deforestation limited to only 9-10 parts per million.

In chapter 11 (fig. 11.4), I proposed that the answer must lie in feedbacks operating within the climate system that boosted the direct CO2 contributions from deforestation and early coal burning by releasing additional CO2. My reasoning was that the warming of the atmosphere caused by the direct anthropogenic emis-

A.5. Evidence that the current interglaciation has remained warmer than the trend toward cooler climates during previous ones: (A) oxygen-isotopic measurements (S18O) of deep-sea benthic foraminifera, with lighter (more negative) values indicating warmer temperatures; (B) deuterium (SD) measurement from Dome C ice cores, with lighter values indicating colder Antarctic air temperatures.

sions of greenhouse gases (both CO2 and CH4) would also have warmed the ocean, or at least have kept it from cooling. In this way, an anomalously warm ocean could then provide positive CO2 feedback to the atmosphere via an indirect, but still anthropogenic, effect.

At first glance, this explanation may seem unlikely. How could a 9—10 parts per million initial CO2 "push" from direct anthropogenic emissions produce another 26-30 parts per million of CO2 feedbacks? But framing the problem this way leaves out methane, which also plays a prominent role. The estimated anthropogenic CH4 anomaly from farming was 250 parts per billion, which would have had an effect on climate equivalent to roughly 12 parts per million of carbon dioxide. With this additional help from methane, direct greenhouse-gas emissions equivalent to approximately 21 to 22 parts per million of CO2 would have given the climate-carbon system an initial push that resulted in the CO2 feedbacks totaling 26 to 30 parts per million.

In chapter 11, I also speculated that the most likely source of CO2 feedback is the ocean, particularly the deep ocean and the Southern Ocean, both of which are known to have played a significant role in natural CO2 changes during glacial-interglacial cycles. New evidence now supports this suggestion.

Oxygen-isotope (S18O) ratios from bottom-dwelling foraminifera trend toward heavier values during all six previous interglaciations (fig. A.5a). This trend indicates either that new ice sheets were growing or that deep water was cooling, or (more likely) both. In contrast, the S18O trend during the Holocene since 7,000 years ago has been toward lighter values. Because ice sheets were neither growing nor melting in significant amounts during this time, this trend toward lighter values suggests that deep-ocean temperatures warmed during the middle and late Holocene, before cooling slightly in late preindustrial time. The approximately 0.2o/oo negative trend would translate to a deep-ocean warming of about 0.84oC. Because CO2 is less soluble in warmer water, an ocean warming would have released CO2 to the atmosphere and driven concentrations higher.

Deuterium (SD) ratios from Antarctic ice cores show a similar "warm anomaly" during the Holocene. In the early parts of all six previous interglaciations, the ratios drifted toward more negative values, indicating that air temperatures directly above the Antarctic ice sheet were cooling (fig. A.5b). In the Holocene, however, the ratios varied only slightly and ended up at nearly the same value as 7,000 years ago. In this case, the Antarctic registered a "warm anomaly" in the sense that it failed to register the "normal" cooling observed in previous interglaciations.

As a result, ocean temperatures near Antarctica would also have remained warmer, and sea ice would have failed to advance as it did early in previous interglaciations. These changes would have kept Southern Ocean waters in greater

Was this article helpful?

0 0

Post a comment