Beoordeling van bosgebonden grondstoffen en milieu-impact van menselijke activiteiten in het Sagalassos-gebied van het verleden

Sagalassos, located in what is now southwestern Türkiye, approximately 100 km north of Antalya, was an antique city that also lent its name to its adjoining territory. The territory is mountainous and highly diverse in terms of geology, elevation and climate, which gives rise to different types of forest. Hunter-gatherers first entered the area during the Middle Palaeolithic (120.000-45.000 BCE) after which a long history of changing climates, transitions in power, fluctuating populations, different economic activities and associated land use decisions left an indelible impact on the landscape. The Roman Imperial period (25 BCE-250 CE) forms a focal point in the research on the area because several socioeconomic processes intensify or are otherwise altered: the city gained considerable regional power during a period of overall peace and prosperity, a monumental city centre was constructed, a production centre for fine tableware was established to the east of the city and agriculture in the countryside intensified and became more complex. The city's population swelled to approximately 3000 people, while the countryside was home to about 16.000 more. In the face of this considerable expansion in several dimensions, questions on the sustainability of these combined dimensions arise. Indeed, concepts like 'sustainability' or 'resilience' have entered the archaeological lexicon and societies of the past are now presented as systems embedded in a dynamic environment that provides inputs to and receives outputs from these societies. However, estimating the magnitude of energy and material flows within these past coupled systems remains challenging. Therefore, the impacts of societies of the past on their environment, which, in turn, may cause new impacts on those same societies further down the line, can remain hidden in time in the absence of clear physical evidence. This dissertation sheds light on one of the most important and understudied flows at work in societal systems of the past: that of fuelwood. Forests, woodlands and the wood they provided were essential to many societies, but archaeology so far has only taken simplistic approaches to their study. This is due to (1) difficulties in quantifying fuel use in the past and (2) a failure to understand forests as complex, adaptive systems. Both of these hurdles are overcome through a set of entirely new computational modelling methodologies, which are applied to the case study of Sagalassos in the Roman Imperial period. In a first step, a general framework is proposed that translates information on material culture, practices and cognitive norms to estimates of biofuel (wood, charcoal, animal dung and crop processing residues) consumption by households. It does so by quantifying the effects of three drivers of biofuel consumption: household size (increases in household size drive down energy consumption per capita), outside temperature (lower temperatures require more expenditure of heating energy) and land use (biofuel preferences can be associated with different landscape compositions). This framework is then applied to the city of Sagalassos and the adjacent Ağlasun Valley for the Roman Imperial period. The local diet is reconstructed from a variety of sources, from which the required cooking energy is inferred. The combination of this estimate with the three drivers from the framework yields the following results: inhabitants of the countryside would have approximately consumed 1,3 kg of wood per day while inhabitants of Sagalassos would have consumed 1 kg of fuelwood and 0,3 kg of charcoal. For the sake of contrast, the same methodology was repeated for the preceding late Achaemenid to early Hellenistic period (late 5th to late 2nd century BCE), prior to the start of urbanisation at Sagalassos. Model results for this period suggest a daily consumption of approximately 1,6 kg of wood per capita. Model estimates of dung and crop processing residue consumption were negligible in both cases. These estimates are only averages, as settlements at higher altitudes expend more energy to compensate for the colder temperatures. To account for the individual resource requirements and ensuing gathering behaviours of different settlements, a new agent-based model, SAGAscape, was constructed. This model simulates the gathering of food, wood and clay by individual settlements in a reactive landscape, characterised by heterogeneous soil fertility, forest growth speed and clay availability. The Iron Age to Hellenistic period (1000 BCE-25 BCE) was used as an example case to illustrate the workings of this first version of SAGAscape. The results indicate that every individual settlement during these periods could have met its own resource requirements without the need for territorial enforcement. Additionally, the hilltop settlement of Düzen Tepe, neighbouring Sagalassos to the southwest, seems to have been in a slightly more optimal position than Sagalassos with regards to access to resources. This is somewhat surprising, as habitation at Düzen Tepe seems to disappear during the Hellenistic period, at the same time when Sagalassos starts off on a process of population growth and urbanisation. At the same time, however, this feature matches existing explanations of the complex origins of urbanisation and polis formation at Sagalassos. Afterwards, a second version of SAGAscape is applied to the Roman Imperial period. The new version includes improved forest growth estimates, allows individual settlements to specialise in gathering a particular resource and also allows them to exchange surpluses. The wood consumption estimates used in this model were the results of the aforementioned framework. The results of this second version of SAGAscape indicate that the landscape could sustain the extraction of the required quantities of wood without suffering extensive deforestation, even when factoring in the large requirements of the Bath-Gymnasium and the tableware production activities of the potters' quarter. Furthermore, the model predicts a land use pattern for agriculture and specific roles in the exchange network for individual settlements that aligns with the existing archaeological knowledge base. The simulated exchange of clay throughout the study area, however, does not match the evidence. The spatially explicit wood harvested estimates generated by SAGAscape served as inputs for the forest gap succession model GREFOS. A new version of this model was run for a transect of 10 km, starting in the city centre and running north. Results show that the projected harvesting intensities would likely lead to only limited changes in total forest standing stock, but that small to moderate shifts in species compositions could result from succession dynamics following a harvest. This overall stability in standing stock could however only be reached by means of knowledgeable management, including, for instance, the retainment of enough young trees. The results furthermore indicate that a small increase in average temperatures over the course of the Roman Imperial period improved the growth and expanded the altitudinal range of Pinus brutia, the local lowland pine species. One of the few primary sources of high-resolution, long-term environmental data for the territory consists of pollen material. The pollen data have, over the years, become an underlying fixture of the many narratives concerning the evolution of the study area. However, it is now often these narratives, not the data themselves, that are being used as a source. The final part of this dissertation concerns itself with the deconstruction of these narratives and provides two new data visualisation methods that allow laypersons to engage with the complex pollen datasets. In line with SAGAscape and GREFOS results, the pollen data show that overall forest cover did fluctuate over time and across the territory, but never fell below 40%, not even during the Roman Imperial period. In conclusion, the available model results and data suggest that wood extraction in the territory of Sagalassos during the Roman Imperial period, a time of high resource requirements, was sustainable and that, in fact, the landscape could have provided more wood if required. Many of the approaches that were developed in this dissertation can be readily applied to other contexts, and will hopefully lead to a better understanding of coupled social-environmental systems in the past.