How did that get there? Understanding sediment transport and accumulation rates in agricultural landscapes using the ESTTraP agent-based model

Abstract

The 15th–18th century CE site of Engaruka in Tanzania is often described as primarily comprising drystone agricultural terraces, but it is now known that many of these former farming plots are not terraces per se, but are instead sediment traps. Stratigraphic excavations of these traps show that they were built by constructing low drystone walls adjacent to either natural or artificial water courses in order to capture fine alluvial sediments entrained within water flows. In the northern area of the site sediments were accumulated to a depth of up to 700 mm, while in one area in the south of the site over 2 m of deposits were accumulated over at least a 300 year period. The presence of sediment traps on archaeological sites allows investigations of the efficacy and sustainability of these structures over decadal to centennial timescales, since stratigraphic excavations can define the process of construction, and geoarchaeological analyses can explore changes within accumulated sediments over time. Although a combination of stratigraphy and absolute dating can discern the broad sequence and timing of sediment capture they cannot determine sediment-accumulation rates, and these techniques are too time consuming to be used to map the development of over 9 km2 of sediment traps. The ESTTraP agent-based model provides these data by simulating sediment accumulation under different hydrological conditions. Four scenarios were simulated for a period of 100 years: constant water availability (SIM-01), seasonal variability (SIM-02), long-term climate variability (SIM- 03), and vegetation-cover impact (SIM-04). The model results suggest that the fields can be constructed over a short period of time, approximately 1–3 months per 6 × 6 m field, and that to construct a block of 90 fields covering 3,000 m2 it would take between 8 and 13 years in periods of high water availability, and up to 27 years during prolonged dry periods. The results define the amount of time needed to construct individual fields, and suggest that farmers constructed blocks of fields concurrently rather than sequentially expanding across the landscape, and that the c. 10 km2 area of sediment traps at Engaruka could have been constructed by a number of households working independently. The ESTTraP model presents an important resource in the assessment of sediment dynamics and patterns of field development, is relevant to a range of archaeological sites worldwide that include intentional or unintentional alluvial deposition, and has applications for modern landscape management.