Fig 1.
Location of Tanana Valley study area in Alaska.
The data used to produce the background layers in this image are available under the Open Database License from OpenStreetMap.org (https://www.openstreetmap.org/copyright).
Fig 2.
Map of archaeological potential, randomly selected sites.
Archaeological potential was modeled on 90 randomly selected sites (white) and tested with 92 other sites (blue). The five classes of archaeological potential are coded from 1 (lowest potential) to 5 (highest potential).
Fig 3.
Relationship between LAMAP class and the number of sites found on land of each class, randomly selected sites.
The grey vertical bars indicate the number of sites located in cells with the given potential estimate. The solid black trend-lines indicate the expected number of sites for a given LAMAP class according to the Poisson regression model. The 95% and 99% posterior predictive intervals for the regression models are indicated by the darker and lighter grey shaded ribbons, respectively.
Table 1.
Number of sites per LAMAP class in the two Tanana Valley analyses.
‘Random’ shows the results obtained when the model was built on 90 randomly selected sites and tested with a different set of 92 randomly selected sites. ‘Pre/Post’ shows the results obtained when the model was built on pre-10,000 cal BP sites and tested with post-10,000 cal BP sites.
Fig 4.
Map of archaeological potential, pre/post-10,000 cal BP sites.
Archaeological potential was modeled on pre-10, 000 cal BP sites (white) and tested with post-10,000 cal BP sites (green). The five classes of archaeological potential are coded from 1 (lowest potential) to 5 (highest potential).
Fig 5.
Relationship between LAMAP class and the number of sites found on land of each class, pre/post-10, 000 cal BP sites.
See Fig 3 for key.