Fig 1.
An example of the hierarchically nested used-available design for one individual (Pronghorn Ear Tag 171) to predict pronghorn (Antilocapra americana) migratory route habitat selection during spring at two scales across the Northern Sagebrush Steppe.
Pronghorn second-order selection compares migratory neighborhoods to the entire study area whereas third-order selection compares migratory pathway relocations to the migratory neighborhood. Scales of migration habitat selection are nested as a result of third-order available locations also operating as used locations at the broader, second-order scale. This original map was created using ArcGIS Desktop version 10.4.1 (ESRI, Inc., Redlands, California, USA) and includes vector data from Natural Earth (“Admin 1 –States, Provinces”) which complies with CC BY 4.0 license; available from https://protect-us.mimecast.com/s/pxvaCxkVBDILmvE2uAfAL2?domain=naturalearthdata.com.
Fig 2.
Resource selection spatial predictions for female pronghorn across the Northern Sagebrush Steppe from 2004–2010 during spring and fall migrations.
Second-order selection during spring and fall are identified in the first row, third-order selection during seasonal migrations are identified in the second row, and scale-integrated step selection function (ISSF) map during seasonal migrations are identified in the third row. ISSF was calculated by multiplying the second- and third-order selection values across the Northern Sagebrush Steppe into a scale-integrated spatial prediction. This figure presents original map predictions that predicts the probability of use by pronghorn where green pixels indicate high probability areas for migration habitat, while red pixels indicate low probability areas.
Fig 3.
Pronghorn predicted relative probabilities of occurrence within ten ordinal habitat categories binned by resource selection probability functions during spring (A) and fall (B) for second-, third-, and scale-integrated step selection function (ISSF) spatial maps. Using k-fold cross validation, maps indicate that second-order models and ISSF scale maps during both seasons were more effective in discriminating between low versus high quality (bins 8–10) migration habitat than third-order models.
Table 1.
Top second-order resource selection function model standardized coefficients for female pronghorn migratory neighborhood selection during spring and fall migration across the Northern Sagebrush Steppe, 2004–2010.
Table 2.
Spearman rank-based model validation results for second-order, third-order, and scale-integrated step selection function (ISSF) models of female pronghorn migrations across the Northern Sagebrush Steppe, 2004–2010.
Table 3.
Top third-order step selection function conditional logistic regression model coefficients for pronghorn migratory pathway selection during spring and fall migration across the Northern Sagebrush Steppe, 2004–2010.