Geographical expansion rate of a brown bear population in Fennoscandia and the factors explaining the directional variations

Abstract: The brown bears, Ursus arctos L., in the Scandinavian peninsula were distributed in almost all counties before aimed reduction during the 1700-1900s (Swenson et al. 1995). From 1981-2013 the population increased more than five times (Chapron et al. 2014) to about 3000 individuals (Kindberg et al. 2014). The aim of the master thesis was to find the geographical expansion rate in this bear population in the period 1981-2019 and identify the factors influencing the expansion in this period. The study area is within latitude 58-70 degrees North and longitude 11-27 degrees East. By two methods in ArcGIS, I found the expansion rates in eight directions in four subpopulations. The source data of bears is shot female bears in Sweden, and adjacent areas in Norway and Finland. When using linear regression, the expansion varied from 1.19 km/year (0.23-2.15 km/year 95% confidence interval) in direction 270-315 in the second Northern subpopulation to 5.16 km/year (4.05-6.27 km/year 95% confidence interval) in direction 90-135 degrees in the second-Southern subpopulation. The expansion rate was significant in 18 of 32 directions. It was significant positive in all directions in the Southern subpopulation and in direction 90-135 degrees in all subpopulations. By using Minimum Convex Polygon, MCP, the estimated average expansion from 1981-2019 in the different directions varied in the three Southern subpopulations from 1.02-5.08 km/year but in the Northern subpopulation the expansion was negative in direction 135-315 degrees and about 0 in direction 315-45 degrees. The average expansion for each subpopulation from South to North was estimated to 3.20, 2.63, 2.55 and 0.67 km/year. Linear estimation by MCP give in general higher expansion rate than linear regression due to methodical reasons. The expansion is generally highest towards East and South-East and lowest to the West and partly to the North and South-West. The fit of seven models was estimated and validated in R by using Akaike’s Information Criterion defined by ∆AICc and AICcWt. Forest has the highest positive impact on all four targets. Higher density of roads has some positive impact. Percent of calving areas and mountain are the most negative single factors to expansion. The model with highest coefficient of determination, R2=0.3556, include the factors forest, mountain, percent calving areas, spring pastures in mountain and density of roads and railways for the target expansion rate by MCP. The results suggest that barriers in West and partly North and Southwest of the subpopulations are highly related to the less suitable bear habitat mountain and calving and spring pastures in reindeer husbandry