Final Report Summary - AAREA (The Archaeology of Agricultural Resilience in Eastern Africa) The twin concepts of sustainability and conservation that are so pivotal within current debates regarding economic development and biodiversity protection both contain an inherent temporal dimension, since both refer to the need to balance short-term gains with long-term resource maintenance. Proponents of resilience theory and of development based on ‘indigenous knowledge’ have thus argued for the necessity of including archaeological, historical and palaeoenvironmental components within development project design. Indeed, some have argued that archaeology should lead these interdisciplinary projects on the grounds that it provides the necessary time depth and bridges the social and natural sciences. The AAREA project (Archaeology of Agricultural Resilience in Eastern Africa) accepted this logic from the outset because to ask whether a system is sustainable is to simultaneously ask: sustainable for how long, sustainable for what size of population, and sustainable under what economic and climatic conditions? However, archaeologists have tended to downplay weaknesses in archaeological data sources that are highly relevant to addressing questions of sustainability, including difficulties in distinguishing active conservation from lack of exploitation, and the fact that significant agricultural conservation practices such as fallowing regimes and irrigation schedules may be impossible to detect archaeologically. The project thus aimed to use a combination of archaeological excavation, geoarchaeology, archaeobotany and agent-based computer modelling (ABM) to assess the sustainability of two agricultural systems employing terracing and irrigation in eastern Africa, and in doing so to present a frank and realistic appraisal of the role archaeology can play in sustainability debates worldwide. The project’s first case-study was Engaruka in Tanzania, archaeologically famous as the largest abandoned irrigated and terraced landscape in eastern Africa. The site has been cited as an example of economic and ecological collapse on the assumption that its abandonment is evidence of unsustainable management practices, and it has long been assumed that Engaruka was irrigated out of necessity: the argument being that agriculture would always have been near impossible without irrigation in what is now a semi-arid environment. Our geoarchaeological research refutes this second assumption, demonstrating that parts of the site flooded with sufficient regularity to allow the construction of more than 1000 ha of alluvial sediment traps, in places more than 2m deep. This accumulation of sediments initially enhanced sustainability through several important mechanisms: levelling the topography, burying rocky topsoils below deep deposits of fertile, fine sediments that are easily to till and irrigate, and avoiding salinization that is a common problem in irrigated soils in arid locations by repeatedly adding more sediments to irrigated surfaces. Computer simulations of sediment accumulation show that incremental construction of these sediment-trap fields could be managed at a household level, and thus did not require highly bureaucratic or centralised control, both of which are likely to limit flexibility and reduce resilience. Simulations of household decision making through an ABM show that even relatively minor changes to ecological conditions could prompt rational subsistence farmers to abandon the system, although instructing households to follow pre-set cultural preferences or to be influenced by the decisions of neighbours could reduce the rate and scale of abandonment. There is thus no need to assume that abandonment would coincide with a marked change in ecological or climatological conditions. The question of why Engaruka was abandoned is thus unresolved. However, the accumulation of sediments in the lowlands required soil erosion within the highland river catchments to the immediate west of the site; a trade-off that might ultimately have led to the agronomy becoming unsustainable if erosion rates exceeded rates of soil formation, and/or if erosion within the highlands reduced the water storage capacity of the river catchments, thereby reducing the river flows that supplied the ancient irrigation system. A trade-off of this sort also occurred at the project’s second case-study at the World Heritage Site of Konso, Ethiopia: a landscape consisting of 40 walled towns surrounded by over 200 km2 of dry-stone agricultural terracing. In contrast to Engaruka, Konso has been described as a model of sustainable soil and water conservation, and by the UN FAO as a ‘lesson from the past’; an assessment that equates the apparent longevity of these agricultural practices with long-term sustainability. Our geoarchaeological research confirms earlier ethnographic evidence that parts of the system are at least 500 years old, but also shows that previous work by ethnographers, geographers, agronomists and researchers interested in the developmental potential of ‘indigenous knowledge’ had misread the landscape by assuming that the agricultural terraces preserved the original topsoil, and that forests conserved locally as sacred groves represent fragments of largely undisturbed woodland. Neither is true, since clearance of hillside vegetation by fire – presumably for slash and burn agriculture – led to widespread and severe soil erosion on hillslopes, as evidenced by the loss of all of the topsoil and most of the subsoil; at least in the river valley that was the focus of our research. This eroded material was captured in riverside sediment traps very similar to those at Engaruka. These sediment traps were not recognised by earlier researchers, and it was thus not realised that these irrigable fields were formerly the mainstay of the economy, and that they remain highly economically valuable during wet years and are a significant risk-spreading measure in dry periods. Indeed, hillside terraces were first constructed to prevent exposed weathered bedrock from eroding on to these valuable plots, with terrace construction preceding upslope from the sediment traps to the hilltop settlements, in the process creating and containing colluvial topsoils. The terraces are thus erosion mitigation measures rather than erosion avoidance features, and contain soils of stony, heavy clays that are considerably harder to work than the fine material within the sediment traps. Archaeological data from both Engaruka and Konso thus have clear policy implications for landscape managers and for those who seek to learn from, encourage, or export traditional practices elsewhere. In the case of Konso, these data provided highly pertinent information that research techniques based on interviews and observation did not detect, despite a considerable number of studies. The project thus fulfilled its aim of assessing whether archaeology has a role to play in sustainability assessments, but it is evident too that archaeological data alone will have limited impact. For example, archaeological research on the plants grown and consumed at both sites provided evidence of a wide range of crops, but interviews and observations of cultivation practices at Konso were necessary to note that planting multiple crops in a single plot is a risk mitigation measure: if rains are poor, thirsty plants can be weeded out in favour of drought tolerant crops, but if rains are good, high-yielding crops can be prioritised. The same is true of the field system: whilst we are accustomed to viewing irrigated and terraced landscapes as highly specialised interventions designed to intensify and maximise production, they can also be used to spread risk by creating a range of agricultural niches. Indeed, periods of depopulation at Konso recorded by anthropological expeditions undertaken by the Frobenius Institute in the 1950s strongly suggest that these niches included periodic farming and grazing of animals up to 20 km away from the Konso heartlands. The project thus recommended that cost-benefit analyses should be undertaken of the agricultural landscape at Konso (AAREA Policy Brief No 1), with specific recommendations to quantify the economic and ecological trade-offs created by soil erosion and capture; note that the terraced landscape exists within a broader web of resources; and recognise that valuing economic benefits purely in terms of maximising yields might underestimate the importance of risk-spreading behaviours. Archaeologists are ill qualified to undertake these analyses, but they can and should form partnerships with the individuals and organisations that are, and in doing so can both contribute to sustainable management plans and refine their archaeological interpretations in the process.
