BACKGROUND: Habitat models provide information about which habitat management should target to avoid species extinctions or range contractions. The willow ptarmigan inhabits alpine- and arctic tundra habitats in the northern hemisphere and is listed as near threatened (NT) in the Norwegian red list due to declining population size. Habitat alteration is one of several factors affecting willow ptarmigan populations, but there is a lack of studies quantifying and describing habitat selection in willow ptarmigan. We used data from an extensive line transect survey program from 2014 to 2017 to develop resource selection functions (RSF) for willow ptarmigan in Norway. The selection coefficients for the RSF were estimated using a mixed-effects logistic regression model fitted with random intercepts for each area. We predicted relative probability of selection across Norway and quantile-binned the predictions in 10 RSF bins ranging from low-(1) to high-(10) relative probability of selection.
RESULTS: Random cross-validation suggest that our models were highly predictive, but validation based spatial blocking revealed that the predictability was better in southern parts of Norway compared to the northernmost region. Willow ptarmigan selected for herb-rich meadows and avoided lichen rich heathlands. There was generally stronger selection for vegetation types with dense field layer and for rich bogs and avoidance of vegetation types with sparse field layer cover and for lowland forest. Further, willow ptarmigan selected for areas around the timberline and for intermediate slopes. Mapping of the RSF showed that 60% of Norway is in the lowest ranked RSF bin and only 2% in the highest ranked RSF bin.
CONCLUSIONS: Willow ptarmigan selected for vegetation types with dense field layer and bogs at intermediate slopes around the timberline. Selection coincides with previous habitat selection studies on willow ptarmigan. This is the first attempt to assess and quantify habitat selection for willow ptarmigan at a large scale using data from line transect distance sampling surveys. Spatial variation in predictability suggests that habitat selection in late summer might vary from north to south. The resource selection map can be a useful tool when planning harvest quotas and habitat interventions in alpine areas.
