Spatial distribution patterns and predictive models: archaeological sites on the lower courses of the Nhamundá and Trombetas rivers

Abstract

The methodological focus of this article is spatial data modeling, utilizing the concept of predictive archaeological models. This method is based on software using species distribution models with a maximum entropy algorithm. The model was constructed using n = 76 archaeological point locations from the region between the lower courses of the Nhamundá and Trombetas Rivers in the state of Pará, Brazil. The entire analysis involved open source software (Qgis, Grass-Gis, R and MaxEnt) and free satellite and radar images (Landsat 5 and SRTM). The results obtained indicate the presence of various ecological niches with high archaeological potential, ranging from the banks of the major rivers and lakes through the hinterland areas surrounding the plateaus in the region. The discussions about the results and analysis indicate that the use and construction of predictive archaeological models in the Amazon region should consider both the archaeological complexes and the region in question, observing the diversity of established indigenous groups along the rivers in Amazonia as well as the various ways of selecting and the occupying the area.

Keywords
Amazon archaeology; Predictive model; Maximum entropy algorithm