Clearcutting and the breeding productivity of golden eagles (Aquila chrysaetos) in Västerbotten County

Sammanfattning: The golden eagle (Aquila chrysaetos) is red-listed as near threatened in Sweden, but there are many basic ecological and demographical parameters important for successful management and conservation that are unknown for the Swedish golden eagle population. Forestry, and especially clearcutting (or final cutting/felling or regeneration cutting/felling), the most common harvesting system in Sweden since the 1950’s, are likely important factors affecting the Swedish golden eagle population, but exactly how is largely unknown. Effects of both clearcutting and forestry in general on golden eagles and their breeding productivity may be positive, negative, direct and indirect. In this thesis I have investigated the effects of clearcutting on the breeding productivity of golden eagles between 2002 to 2016, 15 years in total, in 143 territories in Västerbotten County (each territory inventoried in at least 8 of the years of the study period). Golden eagle breeding data was provided by Kungsörn Sverige Västerbotten, and clear-cut data was provided by the Swedish Forest Agency. Analyses were made in three groups: mountain territories, lowland territories, and all territories together (lowland + mountains). Using simple linear and 2ⁿᵈ order polynomial regressions in the program R, I made territory-level regressions of the average number of nestlings born per year (in three breeding productivity variables: year with territory visited by surveyors, occupied by golden eagles, and with breeding golden eagles, respectively) on the cumulatively summed area of clear-cuts weighted by territory area (in three clear-cut variables: ordinary cumulative sum, a.k.a. including clear-cuts made over the entire study period except for in 2016, thus including clear-cuts up to at least 15 years of age, and successively excluding clear-cuts when they reach 5 and 10 years, respectively). Clear-cut area was extracted from two modelled territory sizes, ca 5 km2 and ca 30 km2, which correspond to the minimum and maximum core area sizes of a golden eagle territory reported in a previous study. The clear-cut and breeding productivity variables were calculated over the whole study period, thus the regressions were only made on a spatial scale. Golden eagle breeding productivity fluctuated with peak and low years and appeared to be declining in Västerbotten overall, and while the decline was not extremely steep it was also not inconsiderable. The regression results were not particularly conclusive, with many of the regressions being far from statistically significant (with statistical significance being p ≤ 0.05) and explained very little of the variation in the breeding productivity. The only statistically significant relationships were for the breeding productivity variable nestlings per visited year in the group with all territories together. Here, significant results were found for all clear-cut variables and both territory sizes, and the relationship appeared to be positive. These models also had the highest adjusted R² values, but at best only 8.73% of the breeding productivity variation was explained. However, low R² is not uncommon in ecological studies and even noisy, high-variability data can have a significant trend which indicates that there is something going on, which I did not consider unlikely for my significant results. Coupled with e.g. the results of previous studies and observations from golden eagle surveyors, I would say that my results are still relevant, implying that there is a relationship worth investigating here, and that it could, at least so far, be positive. That clearcutting, and forestry in general, affect golden eagles and their breeding is still considered likely, and the possibility of a significant correlation between them, possibly also to clear-cut area alone, should not be dismissed based on my results, as there are several things which should be considered when interpreting my results. For example, the relationship between golden eagle breeding productivity and clearcutting, and forestry in general, is likely very complex. There are many potential factors which could be affecting the relationship and serving as important sources of error and variance, and my methods and models were too simple and could only account for a few of these, partly due to there being restrictive gaps in both the breeding and clear-cut data. Further study is therefore highly recommended in order to more thoroughly investigate the relationship and hopefully acquire more conclusive results.