Tuesday 7 May 2013

Used Planet

Last week saw the publication of "Used planet: a global history" in PNAS [pdf] in which I teamed up with some land use and palaeoenvironmental modellers to offer an alternative synthesis of the evidence for human environmental modification. Much of the time in studies of land use and environmental change it is assumed that major environmental change and destruction is quite new, something brought on by the industrial era in the last few hundred years. Land use models like the HYDE model illustrate this, with something that approximate the projection of modern population: land use ratios backwards to periods when very few people were around. The key authors of the original HYDE model were part of this paper, and are quick to point out the many uncertainties and flaws in the HYDE dataset and its application (as in this recent paper). This is illustrated by the graph on the left (above) with most of the transformed land (in light purple) only dating to the past few centuries). By contrast in a model that assumes landuse intensification fewer people used land more extensively and only more intensively as rising population forced them to (the right hand charts above). Comparing maps of these different models, makes it look like "two different planets"  (to quote the blog of lead author Erle Ellis). Whereas archaeologists and anthropologists have debating and modifying Boserup's intensification concept or the notion of "agricultural involution" of Geertz for decades, this seems to have came rather late to global climate and land use modelling studies. Intensification in the use of all sorts of things has been a hallmark of human prehistory, from the broad spectrum revolution, the development of post-harvest intensification, in the use of grinding stone and other cook techniques (explored elegantly by Wollstonecroft 2011), the development of pre-harvest intensification (i.e. cultivation), improved yields through the evolution of domestication traits (with its own new labour demands, a kind of intensification too, see, e.g. my "domestication as innovation" paper), and agricultural intensification as it is normally defined (on which, see Morrison 1994). That our review of these two planets, which fairly clearly comes out in favour of one in which transformations by people are old and intensification processes are long-term, catches some sort of zeitgeist, is suggested both by the press coverage of this paper in New Scientist, Scientific American and The Breakthrough.org, and a session devoted to the same theme at last months SAA (summarized by Michael Balter). Of interest is that the latter two take the younger "impact" date of 3000 years, while the former takes 5000 years (I guess on my suggestion). One of the things that came up in discussions that lead to this coverage is what kind of impacts were there and when should we put the start of the "Anthropocene." 

From Early Holocene impacts to the Anthropocene

Should we replace "Holocene" with "Anthropocene"? Some might tend towards this view, since the start of Holocene is afterall when cultivation began, i.e. human "niche construction" intensified. My own view was that while that was a watershed in human behaviour, it was a long time before cultivation got underway in most parts of the world, or even in those early centres, before it developed into true agriculture, cultivation at the scale that impacted landscapes and made economies that were basically dependent on domesticates, leaving little subsistence space or time for wild foods. Early impacts were not early Holocene but notably middle Holocene from around 6000 BC-1000 BC, and with increasing intensity.

To the put the early impacts into chronological perspective, I would make the following points. 8000 years ago represents the approximate point at which agriculture was established and agriculturally-supported permanent villages appeared in several parts of the world. This is the period of the first villages in China, focused on millet cultivation and with domesticated pigs on the North Chinese loess plateau and the Yellow river valley. This appear in several places from Gansu to Hebei to inner Mongolia. By this period there is clear evidence fo established cultivation of domesticated maize and squash in southern Mexico, and various crops being farmed on the northwest coast of Peru, peanuts among them, and these had there origins in Amzonia east of the Andes. In West Asia this is the period of well-integrated agroo-pastoral systems with a whole suite of crops (wheat, barley, flax, pea, lentil) and animals (cow, goat, sheep horse). This period (9000-8000 years ago) see the rise of not just permanent villages around the Near East but large "mega-sites", for example Catal Hoyuk in central Turkey, which supported permanent population in the 1000s (~6000 is one estimate of site's size). This may not sound big in modern terms, but when you consider that hunter-gather bands are on the scale of 30-40, with large seasonal gatherings at maybe twice or 3 times that at most (i.e. there were unlikely to ever be more than 100 people in one place at one time for even for a seasonal festival for the 150,000 or so years of Homo sapiens history prior to that), then a few 1000 people living together in one place, supported by local agriculture, is a big difference. 9000-8000 years ago is when farming started to spread from the Near East reaching Turkmentistan, Pakistan in the east by 8000 (which big permanent villages established in those areas) and reaching in SW Europe (Greece, and Balkans. The establishment of agriculture on several continents means sustained transformations of local landscapes and ecosystems. Of interest is that this roughly the time point at which some, such as Bill Ruddiman, have inferred that global carbon-dioxide level just start to divert from the expected interglacial trend (see his recent Earth & Planetary Sciences review paper and his Real Climate blog post). This lends extra importance to the 8000 BP landmark. Our paper is not about the greenhouse gas story, which involves lots of complicated factors of the carbon cycle, but we are certainly with Ruddiman in as much as one needs to factor human activities into the equation at that time when considering global carbon issues since 8000 years ago.

Another landmark for me is in the 5000 BP sort of timeframe. This is the period which sees the beginning of tropical savannah farmng, in subsaharan  Africa, inner India, also the central and eastern USA, mainland SE Asia. This sees the major spread of rice out of China and into SE Asia (see my previous blog), and the land area in the Old World (Africa and Asia) that support pastoralism (sheep, goat, cattle) doubling. This period sees the spread of agriculture with llama-keeping to high elevantions in the Andes, and by about 4500 BP the establishment of maize-beans-squash farming in the American southwest. In these more marginal environments the impact of early farming on erosion and local flora may have been more severe than in naturally better-watered regions. In the established agricultural centres (Mesopotamia, Egypt, the Yellow River, southern Mexico) this is the period in which the urbanism get establish (slightly later in Mexico perhaps) with all sorts of new demands on agricultural intensification and specialized production to support urban populations that do not carry out their own subsistence. The rise of major textile and metal industries at this time, especially well documented in Mesopotamia and Egypt, means larger herd of sheep that were not for eating, land for flax that was not for eating, and increase wood fuel demands for bronze furnaces (and new materials like Faience or in China proto-porcelains). Bronze is a bit later in China and SE Asia (after 4000-3500 years ago), but still falls into this time horizon generally. There is a new scale in deforestation to go along with the establishment of agriculture on all continents and subcontinent (except Australia or the polar regions). In Ruddiman's Greenhouse gas story this period is when global methane is meant to diverge from expected interglacial trends. In other words after 5000 BP both the greenhouse gases have started to look unnatural. He has hypothesized that this was to do with the spread of wet rice, and elsewhere (in a Holocene paper) I have argued that one also must count the cattle which spread like wildfire through the tropical savannahs of Africa, India and perhaps SE Asia at this time (see previous blog). This is also roughly the dairying revolution in Europe, not when people first used a bit of milk, but when people began to herd cattle to specialize in milk production (and European people evolved adult lactase enzymes): this may have also been a upturn in animal herd density. How human activities translate into greenhouse gases is complicated, because of interacting carbon sinks and sources, and not the point of the "Used Planet" paper, but the implication of out paper is that human activities should not be ruled as a contributing to global processes

By 3000 BP or so it is hard to deny human impacts, although I would put the significant shift earlier. In archaeological terms 3000 BP roughly corresponds to the Iron Age. The thing about Iron is that is much more egalitarian than copper alloys, in as much as Iron ores are really widespread and found in all regions. Mesopotamia had import copper from Oman or distant parts of Turkey, and time from Afghanistan. Africa south of Egypt and Nubia never really had a bronze age, nor did southern India, but they all took to Iron because iron ores were available. Everybody had iron ore on their doorstep. But iron smelting requires twice as much wood fuel (at least) as copper. One set of wood fuel just to turn wood into charcoal, and then charcoal represents a second set. So a upturn in wood demands and deforestation. There is no Iron Age in the Americas but Peruvian and Mexican metallurgy (copper, gold) are full swing by this time. The other thing about the centuries after 3000 BP is the upscaling and cities and the first empires. Think Assyrians, followed by Persians. The Qin emperor of China may not unify China until the 3rd c. BC but he comes at the end of process of expansionary attempts by various rulers ("The Warring States", etc). This is the era of Maya pre-classic (in southern Guatemala) and Olmecs (on the Gulf coast of southern Mexico), which represent first urban-like communities and kingdoms in Mesoamerica. In South America this period sees the arrival of farming on the shore of Lake Titicaca (which was to become a focus of intensive farming and urbanism of Tiwanaku about 1000 years ago), and first big ceremonial centres, foci of dispersed but farming growing populations, at sites like Chavin du Huantar. This period sees the founding of sites that were to development into cities in the Niger river in west Africa, the earliest phases at place like Jenne-Jeno and Dia. In short, all over the work human population densities increase, proportions of the populations not engaged in their own subsistence increase (but are still a miniscule minority bu modern western standards) and technologies requiring heavy wood fuel use increase in scale and global frequency. 

Time for Big Archaeology

When I was asked by the journalist from New Scientist where I put the start of the "Anthropocene" I suggested 5000 BP. By 3000 BP there is certainly no doubt (the date chose by the headlines in other coverage). Erle I know does not want to commit to any one date for this transition, and I think we would both emphasize that there have been several step changes, often different in timing in different regions, in the relationship between human population and landuse and land transformation. These changes and landuse relationships need to be better-studied and better quantified. Archaeology, which I represent among the contributors on the PNAS paper, has a lot to contribute here. Archaeologists in a sense have been gathering such data without realizing its relevance to issues of long-term global change, and archaeologists are generally very region and period focused-- archaeologists, we might say, have their heads down holes-- with less awareness of the macro-scale and really long-term comparisons. I think that it is time for that to change, and I hope that more archaeologists will start see how their evidence can contribute to better models of long-term landuse and ecosystem management. It is time for archaeologists to become a bigger part of these discussions. 

We need to get some BIG ARCHAEOLOGY going. Then we can return to the issue of when the Anthropocene began.