Modelling Environmentally-Constrained but Adaptable Society-Environment Systems

8.3.1 TEM (“Terroir”-Based Environmentally Constrained Models)

Following our classification, we describe here the “terroir” level models, where a society and its evolution is driven by their calorie and resource demand and con­strained by environmental parameters.

As an example of these models, Baum (2014) has formalized a GIS-based object model based on information from available and relevant literature and local archaeological data regarding environmental characteristics (soil, vegetation, local climate, distance to village) and cropping and livestock-keeping practices, to evaluate the environmental impact of human settlements over several village ter­ritories, along several farming scenarios (shifting, intensive garden and non-intensive cultivation) and diet assumptions (Fig. 8.1).

Further works with similar research interests include Coolen (2010), who has worked over different sites of the LBK (Linear Band Keramik) culture territory in Europe: thanks to a systematic geographical census of the archaeological sites of the LBK culture over a region, the agro-ecological characteristics of the implan­tation sites of this culture (pedology, climate, orientation, hydrology, relative dis­tance to other sites) can be statistically determined. Combined with estimates on economic and agriculture-related parameters of this agropastoral system, this methodology aims at establish a spatial discrimination of a territory along site preferences and potentialities according to a specific culture agropastoral habits.

Fig. 8.1 Three “snapshot models” illustrating the model design in (Baum 2014). Economic areas are calculated using various hypotheses and their extent modeled in a GIS.

This methodology is equivalent in Burke et al. (2008); Tipping et al. (2009); Graves (2011); Carrer (2013). For instance, Yu et al. (2012) test on the Yiluo valley a combination of demographics and agriculture-based extrapolations and social assumptions based on what archaeology but also modern information provides.

Following their work on Swiss and German archaeological sites, Ebersbach and Schade (2004) have studied livestock-keeping needs and consequences on sus­tainability of environmental resources and livestock-based farming societies: through a GIS, they have estimated the impacts on an extended village territory of the necessary livestock to both feed the population and manure the fields that are necessary for crop production for feeding a population.

The research object of all these simulations is the area or the evolving landscape; the major tool is GIS input data comprise archaeological and paleo-environmental information. However, because the subject, i.e. the research question, differs, they are representative of the “discipline-as-input” factor differentiation: Coolen (2010) used environment as input and deduced site potentialities, with the hope that further archaeological excavations may provide confirmation. Baum (2014) and Ebersbach and Schade (2004) used archaeology as input and deduce the impact on the related territory and natural resources following various scenarios. These approaches are used to test the relevance and implication of certain model parameters, such as the crop husbandry or the size of the economic territory They set up spatial hypotheses which serve for explanatory purposes and as a reference plane for future works.

8.3.2 WEM: “World” Size Environmentally Constrained

Models

Modelling prehistoric and pre-industrial society-environment systems is essential to understand the co-evolution of climate and humans over recent Millennia as well as the current state of the earth system. Such an analysis should be settled at the global scale, simply because it is the sole relevant scale for apprehending human-induced climate changes. Many of the ecosystems that are highly valued today for the services they provide to humanity are the result of long-term interactions between society and their environment. Because detailed observations of these interactions will always be limited in space and time, global human-environment models may be useful tools to bridge spatial and temporal gaps in data and to test hypotheses about the large-scale development of society and the environment.

Despite its promise, global modelling, however presents several additional chal­lenges compared to the “terroir scale” described above. The foremost among these challenges concerns data both for driving the models and for evaluating their output. Outside of Europe and parts of East Asia, critical information on subsistence life­styles, the timing of key transitions, and on paleo-demography needed to parame­terize models is not available because of a lack of investigations in these regions or a poorly preserved archaeological record. Likewise, palaeo-environmental and ethnographic information, that is highly valuable for model evaluation, is largely absent from many continents where geographic conditions and local history led to poor preservation of archives, both natural and human.

Regarding the three points described in Sect. 8.2.4, WEM-type models obviously meet the spatial genericness criterion (1) but they methodologically do not com­pletely fulfil the criterion (2) regarding time: rules of the model do apply all along simulations but the model lacks time-related adaptability. Finally, the criterion (3) is not answered as such models do belong to the ii. (b) palaeo-environmental “Disci­plines used as inputs and drivers” scheme meaning that changes are environmentally driven.

Nevertheless, there are several promising methodologies that are currently being applied to understand society-environment systems at global scale. These models may be roughly divided into two categories:

• data-driven approach where demographic and subsistence data are inputs to the model;

• an “organic” approach, where the model simulates potential human population and subsistence lifestyle as prognostic variables. Both of these approaches have advantages and disadvantages and are currently under rapid further development.

The data driven approach is typified by the ALCC scenarios KK10 (Kaplan et al. 2009, 2011) and HYDE (Klein-Goldewijk et al. 2011). These scenarios are the result of empirical models that take geographically distributed estimates of popu­lation at any time in the past and combine them with information on climate and soils in order to estimate the magnitude and spatial distribution of land use. The models used to generate these scenarios assume subsistence lifestyle implicitly, i.e., everyone on earth at a given time is presumed to have the same type of subsistence strategy. While HYDE makes a simplistic distinction between land use for crop or pasture based upon present-day geographic patterns of land use, it does not consider changes in per capita land use over time (intensification). In contrast, KK10 models intensification as a non-linear function of population density itself, so that low population densities use relatively large amounts of land.

As noted above, neither model takes directly into account the way in which dif­ferent subsistence lifestyles may use the same landscape, e.g., foragers versus shifting cultivators versus permanent agriculturalists versus pastoralists. This distinction among land use types, in particular the shift from foraging to farming, may be critical

Fig. 8.2 Comparison between global anthropogenic land cover change scenarios KK10 and HYDE for the year AD 1. The large discrepancy between the scenarios in the maps is caused by differences in the treatment of per-capita land use. HYDE fixes the spatial pattern of per-capita land use observed in AD 1961 for all time periods in the past, whereas KK10 models per-capita land use as a function of population density, with intensification occurring at higher densities (courtesy of J. Kaplan)

for understanding the pattern of land cover change and human impact on the envi­ronment during pre-industrial time. While the agricultural transition could be pre­scribed in models based on archaeological records, lack of investigations or well-preserved sites implies that in many parts of the world prescription would be based on guesswork or assumptions. An alternative approach is to use a model that explicitly simulates subsistence lifestyle changes, as in the “organic” approach mentioned above.

Currently the best example of this approach that has been developed and applied at continental to global scale is the GLUES model (Global Land Use and techno­logical Evolution Simulator (Lemmen et al. 2009). GLUES simulates human pop­ulation density, technological change and agricultural activity directly, based on the concept of gradient adaptive dynamics, where adoption of a subsistence lifestyle, e.g., Neolithic agriculture, by any given group of people at any particular time depends on endogenous environmental and social factors, e.g., potential produc­tivity, population density, and exogenous factors, including the presence of farming people in neighbouring regions.

Such models answer the criteria (1) and (2) with a formal justification of the appearance of such technical innovations (Lemmen 2012). More globally, they cannot answer the criterion (3):

• Social innovations, such as socio-anthropological family evolutions and/or political structures, are less likely to be modelled, while such social innovations may have a determining impact on the “capacity” of a society for conquering new territories, following non-Malthusian hypotheses (Boserup 1965, 1976; Lemmen 2012). For instance, Todd (2011) has suggested large variations of the family structures in Eurasia, linked with the appearance of unequal families and the consequences on the cultures’ differential “capacity” of expansion.

• The effects of “coincidences”, i.e. emergences may disappear as such conjunc­tions are smoothed while going at a broader scale, both socially and spatially.

8.4