Has the sex-specific structure of Finland's brown bear population changed during 21 years?

Gene flow between the two brown bear Ursus arctos populations of Fennoscandia is very weak although bear populations both in Finland and Scandinavia has been increasing substantially during the last 50 years. We examined spatial and temporal pattern in the proportion of adult females in Finland's bear population. First, we expected evidence for decreasing female proportions with increasing distance from the potential core areas in Russia and near Finnish-Russian border. Secondly, we expected increasing proportions of females during our study period of 21 years, i.e. two brown bear generations. We conducted a logistic mixed effects model that would predict a hunter-killed bear's sex based on year, the distance from the source population (Finnish-Russian border), bear management area and bear's age group (adult, sub-adult) as independent variables. Geographic coordinates were used to reduce the apparent spatial autocorrelation. Because female bears form kin clusters, in the final model we used exponential spatial autocorrelation, based on the assumption that the probability of a dead bear being a female increased when the nearest dead bear also was a female. The model demonstrated that the population's population structure did not change spatially during our study period. The only differentiating variables were bear management area and age group, the probability of females being low in reindeer husbandry area, and males predominating among subadults. Because the low proportion of females in Finland's reindeer husbandry area may weaken meta-population viability in Fennoscandia, bear management in northern Finland should save females more efficiently.