Modelling Maize Agriculture by the Pre-Columbian Casarabe Culture of Amazonian Bolivia: An Agent-Based Approach

Scholars have long debated the extent to which pre-Columbian (pre-1492 CE) indigenous cultures modified and ‘domesticated’ the landscapes of Greater Amazonia. Compelling evidence to support large-scale pre-Columbian landscape modification can be found in the forest-savanna mosaic environments of northern Bolivia, where the Casarabe Culture constructed hundreds of earthen settlement mounds, integrated by a vast causeway-canal network. Operating between 400 and 1400 CE, recent research suggests this culture practiced a form of low-density agrarian urbanism. However, as just two mounds have been excavated in any detail and few palaeoecological data are available, little is known about the extent to which this culture modified the surrounding forest and savanna ecosystems. Here, we present the results of experiments conducted with an exploratory agent-based model of the Casarabe Culture, which we developed to generate hypotheses regarding how they utilised this landscape under a range of different scenarios. Based on our model outputs, we hypothesise that the Casarabe Culture only modified localised areas of the landscape, driven by their desire to maximise cultivation on land with ‘desirable’ environmental characteristics. For this same reason, land that possessed characteristics desirable to the Casarabe Culture is likely to have been intensively modified. More than sufficient forest and savanna was locally available to most settlement mounds to facilitate cultivation without the Casarabe Culture needing to encroach into undesirable areas, but their close spacing also suggests that a level of inter-settlement cooperation may have been necessary. The outputs of our model will play an important role in guiding future research on the Casarabe Culture, identifying viable sites of interest for archaeological and palaeoecological fieldwork. They can also be compared with future empirical research as it becomes available, improving our understanding of past underlying human-environment interactions on these landscapes.