9.2. Natural processes caused the atmospheric CO2 peak nearly 10,500 years ago and the subsequent decrease until 8,000 years ago, but humans have caused the anomalous CO2 increase since that time.
was the case with methane, this rise may not be the full anomaly. It is also necessary to try to take into account the amount by which CO2 would have dropped if natural processes had continued in control during the last 8,000 years. We can do this by examining the previous interglaciations to find the times most analogous to today, checking what the CO2 values were at those times, and in this way predicting what the natural CO2 values should be today.
The CO2 concentrations at those earlier times averaged about 242 parts per million. This value lies along a reasonable projection of the natural CO2 decrease that occurred between 10,500 and 8,000 years ago before the anomalous reversal in trend (fig. 9.2). If this analysis is correct, then the full CO2 anomaly by the industrial era would be 40 parts per million, the difference between the 280-285 level reached just before the industrial era and the projected 240-245 natural value. This offset amounts to almost half the range of the natural glacial-interglacial variations (fig. 9.1) and would require the addition of at least 300 billion tons of carbon to the atmosphere during the last 8,000 years. But where could so large an amount of CO2 have come from?
Two early attempts to explain this anomalous-looking CO2 increase invoked natural causes. One idea was that the extra CO2 came from a natural release of carbon from the continents in response to changes in solar radiation caused by
Earth-orbital changes. One part of this idea sounds plausible: we know that the weakening summer monsoons over the last 5,000 years allowed tropical deserts to spread into areas previously occupied by grassland (chapter 5). The net loss of tropical biomass from these changes would have sent CO2 into the atmosphere. But other evidence refutes this explanation. Global-scale models show that natural carbon losses in some areas over the last few thousand years would have been offset by gains in others. Overall, the loss of biomass carbon is far too small to provide the amount needed to account for the anomalous CO2 trend.
I concluded that this or any other attempt to explain the rising CO2 trend since 8,000 years ago based on the natural behavior of the climate system is doomed to fail for a very simple reason. All of the major factors in the climate system (changes in solar radiation, rates of retreat of the ice sheets, rises in sea level, changes in vegetation, and so on) behaved in a similar way throughout the last four intervals of ice melting and then for several thousand years afterward, yet only the current interglaciation shows a CO2 rise during the early interglacial interval. All of the three prior intervals showed steady drops in CO2 (fig. 9.1).
As a result, any explanation of the recent CO2 increase based on natural factors is doomed to fail when tested as an explanation of the CO2 drops during the three previous interglaciations. With natural explanations apparently eliminated, once again only one explanation seemed to be available: humans must have been the cause of the anomalous CO2 rise. Somehow, humans had seemingly added 300 billion tons or more of carbon to the atmosphere between 8,000 years ago and the start of the industrial era.
But this conclusion collided head-on with the conventional wisdom. The initial reaction of most climate scientists I spoke to was that the small number of humans on Earth 8,000 years ago could not possibly have taken control of the atmospheric CO2 trend, especially considering the primitive technologies then at their disposal. This skepticism seemed well justified: How, indeed, could so few humans have had so large an effect? The primary way humans would have released CO2 to the atmosphere during this interval was by cutting forests. A carbon release of 300 billion tons between 8,000 and 250 years ago would require more than twice as much forest clearance before the Industrial Revolution as has occurred during the 200 years of the industrial era. Modern rates are very high because of rapid clearance of tropical rain forests in South America and Asia. By comparison, the estimated annual clearance rate just two centuries ago was almost ten times smaller than today, and the rates faded away back in time toward trivially small amounts for earlier intervals. From this perspective, it seems preposterous to propose more than twice as much total forest clearance prior to 1750 as afterwards.
But the conventional-wisdom view failed to take into account one key factor— time. The average rate of carbon emissions from forest (and other) clearance during the last 200 years has been about 0.75 billion tons per year. At that rate, 200 years of forest clearance would emit a total of 150 billion tons of carbon. But the preceding era of slowly rising CO2 stretches over 7,750 years, an interval 40 times longer. To match the 300-billion-ton carbon emissions target during this earlier era, the average rate of emission would have to have been only about 0.04 billion tons per year, or just 5% of the industrial-era average. With one-twentieth of the rate but an interval 40 times longer, you end up with twice as much in total emissions (divide 40 by 20). As in Aesop's fable, the tortoise (which moves slowly, but starts very early) challenges the hare (faster moving, but late getting started). Look back at figure 1.1. With this simple calculation, the total of 300 billion tons of early carbon emissions began to look more plausible.
Yet this calculation falls well short of a convincing demonstration that humans are responsible for the anomalous CO2 rise. To strengthen the case, it would help to find evidence consistent with two obvious characteristics of the CO2 rise in figure 9.2. First, we need evidence that Stone Age humans began clearing forests at a substantial rate as far back as 8,000 years ago, when the CO2 curve began its rise. Second, we need evidence that the cumulative effects of forest clearance by humans can account for the very large rise in CO2 well before the industrial era (in this case, as early as 2,000 years ago).
Definitive evidence exists that cutting of forests began nearly 8,000 years ago. As agriculture slowly spread from the eastern Mediterranean Fertile Crescent, it eventually reached the forests of southern Europe. As shown in figure 7.2, the distinctive package of cereal grains and vegetable seeds first entered the forested Hungarian Plains nearly 8,000 years ago and then moved north and west into other forested areas. For agriculture to be practiced in forested regions, trees have to be cut to allow sunlight to reach the soil. Much of this clearance was done by slash-and-burn: high-quality flint axes first developed in the Near East nearly 9,000 years ago were used to girdle and kill trees. A year or two later, the fallen debris and the dead trees were burned during the dry season. Crops were then planted in the ash-enriched soil between the dead stumps. In some places, when the soil began to lose its nutrient content a few years later, people packed up and moved elsewhere. In others, they remained and built permanent settlements near their fields.
A wide range of other evidence from Europe points to forest clearance. Along with the distinctive crop grains and seeds, lake sediments contain higher percentages of herb and grass pollen that indicate openings in forests or permanent fields near lakes. Also present are remains of distinctive types of European weeds associated with clearance: plantain, nettles, docks, and sorrels. Increased charcoal in the soils also points to prevalent burning. This evidence from Europe matched the first requirement of the hypothesis perfectly—humans had begun to cut forests in small but significant amounts 8,000 years ago, even though they had only flint axes at their disposal.
Elsewhere, the evidence is less firm, but still suggestive. The first evidence of agriculture in China dates to 9,500 years ago, and by 6,000 years ago pollen sequences in lakes show a noticeable decrease of tree pollen. Although some of this change could reflect moisture loss caused by the weakening summer monsoon, many scientists consider it to be related to increasing human impacts in the valleys of the Yellow and Yangtze rivers. In India, the first evidence of agriculture from the Indus River Valley dates to 9,000 or 8,500 years ago, and expansion of agriculture again required cutting of trees. More studies are needed in both regions to pinpoint the earliest human impacts.
The second requirement—cutting of forests on a much larger scale well before the industrial era—can be tested by examining the start of the historical era, 2,000 years ago. By that time, the world was a very different place compared to 8,000 years ago. In the intervening 6,000 years, human ingenuity had completely altered the practice of agriculture (chapter 7). In place of Stone Age axes and wooden digging sticks were Bronze Age and then Iron Age axes and plows, and domesticated oxen and horses to pull the plows. As a result, evidence of extensive clearance and land disturbance appeared during the Bronze Age in most regions and intensified during the Iron Age.
From the combined efforts of scientists in many disciplines comes a snapshot of world agriculture 2,000 years ago (fig. 9.3). By this time, sophisticated agricultural practices (diverse crops, multiple plantings, and tending of livestock) were in use in China, India, southern and western Europe, and Mediterranean North Africa, as well as the lowlands of Central America, the Amazon Basin, and the highlands surrounding the Peruvian Andes. These regions are all naturally forested, and large-scale agriculture required that those forests be cut.
In such regions, the evidence for increasing human influences on the landscape is plentiful long before the industrial era, but most of this information is not easily converted to specific percentages of forest clearance. By chance, however, the historical literature provided me with one way to make such estimates. In AD 1089 William the Conqueror ordered up the Domesday Survey, a comprehensive analysis of his new domain (England). Included in that survey were two particularly useful numbers: (1) the percentage of land still in forest (rather than in crops or pastures), and (2) the total population. The first number was astonishing: 85 percent of the countryside was deforested, as well as 90 percent of the arable land (elevations below 1,000 m). A full 700 years before the industrial era, Britain was almost entirely deforested, with many of the remaining "forests" protected as hunting preserves for English royalty and nobility. Other evidence indicates that most of this clearance had already occurred by 2,000 years ago. For example, lake
9.3. By 2,000 years ago, sophisticated forms of agriculture were practiced in naturally forested areas of China, India, southern Europe, and Mediterranean North Africa.
■ Complex agriculture ■ Simple "peasant" agriculture □ Other
9.3. By 2,000 years ago, sophisticated forms of agriculture were practiced in naturally forested areas of China, India, southern Europe, and Mediterranean North Africa.
and river sediments recorded large influxes of silt and mud that resulted from destabilization of steeper terrain by deforestation. Here was a firm benchmark that proved massive deforestation far in the past at one location.
The Domesday Survey provided me with an important quantitative link between human population density and the total amount of deforestation. If 1.5 million people required near-total deforestation of a known amount of arable land to make their living as farmers, then every human living in England in 1089 required an average of 0.09 km2 (9 hectares) of cleared land. Both the clearance and the growth of population had occurred gradually over many millennia, and in effect each new person added to the population base had cleared an average of 0.09 km2 of "new" land. I think of this number as the "human forest footprint" for Iron Age England.
The Domesday Survey had profound implications for similar changes in other regions. All of the arable land in any region with a population density equal to or larger than that of England in 1089 should have been completely deforested. The earliest reliable census counts in China date from the Han dynasty nearly 2,000 years ago. By that time, China had 57 million citizens at a density some three times that of Domesday-era England. If England was deforested at a much lower density of people, why not China at this higher density? The same conclusion applied to other highly populated areas such as India and Indonesia. This approach suggested that most of the arable land in southern, central, and western Europe should have been deforested by AD 1000.
Gradually I became aware of an enormously rich literature on early deforestation lying scattered among different disciplines. I also discovered a newly published (2003) book by Michael Williams called Deforesting the Earth, a superb compilation of the wide-ranging evidence for massive preindustrial deforestation. I learned that much of China was deforested more than 3,000 years ago and that many countries in western Europe began passing laws to protect their remaining wooded areas after AD 700, evidence that the extent of forests had shrunk alarmingly.
Those of us working in the field of climate science tend to think of our research as unusually broad in scope, as indeed it is compared with the kinds of science that focus on smaller-scale processes. To do our job, we need expertise in climatic records from ocean sediments, ice cores, lakes, trees, and corals, and we need to know about the history of changes in the surface and deep ocean, sea ice and ice sheets, the land, vegetation, and air. We also need to be able to fit all these pieces together in order to figure out key cause-and-effect relationships. We can justifiably take some pride in our breadth of knowledge. This also makes studying climate history fun.
Yet here was a case in which at least initially I found myself almost completely ignorant of an entire field of research—the early history of humans and our possible effects on the climate system. I am still filling that gap. Just recently I was made aware of a paper published in 2002 by a group of French scientists who had pointed out that sediments from several regions show increasing amounts of charcoal during the last few millennia that generally match the rise in atmospheric CO2 observed in the ice cores. Those data also argue for an early human impact on the CO2 trend.
Although this wide range of evidence argued for large early human releases of CO2, it did so only in a qualitative way. A more quantitative assessment of the hypothesis was still needed. That assessment started with the Domesday "forest footprint" of 0.9 km2/person for Iron Age England. I then found in the historical ecology literature a similar estimate of the footprint for late Stone Age people in north-central Europe 6,000 years ago. This estimate was derived by calculating the amount of cleared land required for a Stone Age village of 30 people (6 families): the amounts needed for homes and crops, for hayfields and pastures, and for a woodlot cut on a rotating basis. The estimated Stone Age footprint came to 0.03 km2 (3 hectares) per person. The tripling of the footprint from 0.03 km2/person 6,000 years ago to 0.09 km2/person in AD 1089 seemed reasonable, given the advent of iron axes and plows and the use of draught animals in the intervening millennia.
Now I had a way to attempt a quantitative estimate of the cumulative amount of carbon emissions by 2,000 years ago. This calculation required three numbers: the human populations (on a regional basis), the amount of forest (in km2)
cleared per person, and the amount of carbon emitted by clearing each km2 of forest. Multiplying these numbers would produce an estimate of the cumulative amount of carbon emitted by total forest clearance as of 2,000 years ago:
per-capita forest tons of carbon emitted total carbon
clearance by clearance emitted
The number of people alive 2,000 years ago was available from historical population estimates. Census data were available from Europe and China, where head counts were made primarily for tax collection purposes (yes, even 2,000 years ago!). Populations were less well known in areas like India, Indochina, and the Americas. In all, some 200 million people were alive 2,000 years ago.
I grouped these 200 million people according to whether they lived in naturally forested regions or nonforested areas (deserts and steppes). Roughly 10 percent lived in deserts and steppes and probably had no significant effect on carbon emissions. The other 90 percent lived in naturally forested regions, where practicing agriculture meant cutting forests. The people living in naturally forested regions were then divided into Iron Age cultures (all of Eurasia and Mediterranean North Africa) with a forest footprint of 0.09 km2/person and Stone Age cultures (the Americas and sub-Saharan Africa) with the smaller 0.03 km2/person footprint. At this time, by far the most people (130 million) lived in the Iron Age cultures of southern Eurasia (China, India, and Europe).
The final number needed for the calculation—the amount of carbon emitted by clearing 1 km2 of forest—came from the field of ecology. This number ranged from 1,000 to almost 3,000 tons of carbon per km2, depending on the type of forest (more from dense tropical rain forests, less from forests in areas with seasonally dry or cold climates). Each region was partitioned into the type of natural forest that ecologists estimated would exist in the absence of human influences.
The total figure I arrived at was large: 200 billion tons of carbon or more could have been released to the atmosphere from clearance by 2,000 years ago. This value was well along the way toward the more than 300 billion tons estimated for all preindustrial carbon releases. It seemed likely that the second requirement of the hypothesis might be met: forest clearance might well be able to explain the growing size of the CO2 anomaly over the last 8,000 years.
I also discovered other kinds of human activities that would have released carbon long ago. As forest fuels began to disappear from some regions in northern Europe, people began unearthing and burning deposits of peat for cooking and for heat. A crude attempt can be made to calculate how much peat carbon might have been added to the atmosphere in preindustrial times: If an average of 5 million households burned 10 half-kilogram (~1-pound) bricks of peat every day of the year over a time interval of 2,000 years, some 10 billion tons of carbon could have been added to the atmosphere during that time. In addition, the Chinese have been mining and burning coal for the 3,000 years or more since they deforested the arable lands in the north-central parts of their country. A calculation similar to that for the burning of peat suggests the possibility of another 20 billion tons of carbon emissions from China over that interval.
Although these findings seemed promising, they certainly did not settle the question of the size of the human impact on the atmospheric CO2 trend over the last 8,000 years. I suspect the debate over this issue will go on for years and maybe decades. But I keep coming back to that one central fact I noticed at the start of my investigations: the CO2 concentration has risen during recent millennia, yet it had always fallen during the most similar intervals of previous interglaciations. It went up when it should have gone down. Humans are the most obvious explanation for this wrong-way trend.
In a way, this hypothesis challenges many of our basic instincts about what is or is not natural. Think of a beautiful Mediterranean Sea coast with blue-green water and grassy hillsides grazed by goats tended by herders. As beautiful as that scene is, it is anything but "natural." If those goats were kept out and nature were allowed to reassert control for a century or more, some of these regions would once again be covered with rich Mediterranean forests. Other hillsides have lost so much of their soil cover and have so few nutrients left in the soils that remain that they cannot revert to their naturally forested state.
As for those charming Mediterranean seaports tucked into the coastline between the steep hills, many of the ports near river mouths were located farther inland 2,000 years ago. As early deforestation stripped the terrain of its natural cover, the steeper slopes were no longer able to retain the soil when heavy rains fell. Eroded silts and muds began to clog the ports, forcing relocation of coastal towns seaward to catch up with the advancing shorelines. Similar stories apply across southern Asia. By 2,000 years ago, that part of the world had been heavily transformed by humans, and nature was no longer in control of the landscape. Nor was it in control of the atmospheric CO2 trend. Humans had taken control.
Was this article helpful?