A resource selection function (RSF) is any model that yields values proportional to the probability of use of a resource unit. RSF models often are fitted using generalized linear models (GLMs) although a variety of statistical models might be used. Information criteria such as the Akaike Information Criteria (AIC) or Bayesian Information Criteria (BIC) are tools that can be useful for selecting a model from a set of biologically plausible candidates. Statistical inference procedures, such as the likelihood-ratio test, can be used to assess whether models deviate from random null models. But for most applications of RSF models, usefulness is evaluated by how well the model predicts the location of organisms on a landscape. Predictions from RSF models constructed using presence/absence (used/ unused) data can be evaluated using procedures developed for logistic regression, such as confusion matrices, Kappa statistics, and Receiver Operating Characteristic (ROC) curves. However, RSF models estimated from presence/ available data create unique problems for evaluating model predictions. For presence/available models we propose a form of k -fold cross validation for evaluating prediction success. This involves calculating the correlation between RSF ranks and area-adjusted frequencies for a withheld sub-sample of data. A similar approach can be applied to evaluate predictive success for out-of-sample data. Not all RSF models are robust for application in different times or different places due to ecological and behavioral variation of the target organisms. #

We investigated grizzly bear (Ursus arctos) selection of three road types in the northern United States and southern British Columbia from 1986 to 1991. We hypothesized that grizzly bears select against open (public use allowed), restricted (forestry use only), and closed roads (no public use allowed) in that order. We analyzed use of roads for 11 bears (five females and six males) in an area containing open and closed roads and 11 bears (seven females and four males) in an adjacent area containing restricted roads. We used χ2 and log–linear models to test for selection of habitat type and distance to road categories. Ten of 12 females and 5 of 10 males (15 of 22 bears) selected against (P < 0.05) low-elevation interior cedar-hemlock and for (P < 0.05) high-elevation Englemann spruce (Picea engelmannii Parry ex Engelm.) – subalpine fir (Abies lasiocarpa (Hook.) Nutt.). After accounting for habitat, 4 of 5 females and 3 of 6 males (7 of 11 bears) selected against open roads and 3 of 5 females and 0 of 6 males (3 of 11 bears) selected against closed roads. No females (n = 7) or males (n = 4) (0 of 11 bears) selected against restricted roads. Our results are inconsistent with the hypothesis that bears select against open, restricted, and closed roads in that order. Most females and males selected against open roads, most females selected against closed roads, and no bears selected against restricted roads. The type of human activity along roads plays a role in bear responses to roads, and this aspect should be incorporated into future bear-road studies.

No abstract

We tested the commonly held hypotheses that grizzly bears (Ursus arctos) select against clearcuts and young forests and select for natural openings and old forests in the Selkirk Mountains from 1986 to 1991. We compared use versus availability using χ2 goodness-of-fit for 11 bears (five females, six males) in a south study area containing both open roads (public use allowed) and closed roads (no public use allowed) and 11 bears (seven females, four males) in a north study area containing restricted roads (forestry use only). Zero of 11 females and 1 of 11 males (1/22 bears) selected against (P < 0.05) clearcuts. Five of 11 females and 2 of 11 males (7/22 bears) selected against (P < 0.05) young forests. The apparent selection against young forests appeared to be due to selection against associated open roads, not against young forests themselves. Forestry activities alone (managed forests and restricted roads) appeared to have no negative impact on grizzly bear habitat use. Because of small sample sizes, pooling of seasonal data, and lack of experimental replication, our results should not be extrapolated until similar studies are conducted elsewhere.

The association between fetal development and neonatal survival and preceding pregnancies of the mother would be compatible with the action of male and female specific antigens capable of affecting selective implantation of blastocysts, which commands subsequent fetal development as well as early neonatal survival.