Island environments present challenges to human colonization, but we have a poor understanding of how environmental difference drives heterogeneous patterns of insular settlement. In this paper, we assess which environmental and geographic variables positively or negatively affect the long-term sustainability of human settlement on islands. Using the postglacial Mediterranean basin as a case study, we assess the impact of area, isolation index, species richness, and net primary productivity (NPP) on patterns of island occupation for both hunter-gatherer and agropastoral populations. We find that models involving area most effectively accounts for sustainability in hunter-gatherer island settlement. The agropastoral data are noisier, perhaps due to culturally specific factors responsible for the distribution of the data; nonetheless, we show that area and NPP exert profound influence over sustainability of agropastoral island settlement. We conclude by suggesting that this relates to the capacity of these variables to impact demographic robusticity directly.

Satellite imagery has long been recognized as well suited for the regional and ecological questions of many archaeological surveys. One underexplored aspect of such data is their temporal resolution. It is now possible for areas to be imaged on an almost daily basis, and this resolution offers new opportunities for studying landscapes through remote sensing in parallel with ground-based survey. This article explores the applications of these data for visibility assessment and land-cover change detection in the context of the Sinis Archaeological Project, a regional archaeological survey of west-central Sardinia. We employ imagery provided by Planet, which has a spatial resolution of 3 m, in four spectral bands, and is collected daily. Using Normalized Difference Vegetation Index (NDVI) values calculated for each survey unit, we find that there is a relationship between NDVI values and field-reported visibility in general, though the strength of this correlation differs according to land-cover classes. We also find the data to be effective at tracking short-term changes in field conditions that allow us to differentiate fields of similar land cover and visibility. We consider limitations and potentials of these data and encourage further experimentation and development.

No abstract