In Northern Europe, human activities have caused a substantial decrease in the number of old deciduous trees over the last two centuries, leading to a decline in species populations associated with this habitat. One way to mitigate this trend is to increase the abundance of mature and old deciduous trees in commercial forests, such as by tree retention at final harvest. We analysed the biodiversity value of retained mature oaks in the production forests of Norway spruce in southern Sweden, using oaks in pastures as reference. The forest oaks were grown in two different levels of shade. We analysed two categories of saproxylic (i.e. dead wood-dependent) beetles: those utilizing oaks (Group I) and those utilizing oak but not spruce (Group II, which was, therefore, a subcategory of Group I). We found that forest oaks sustained high beetle diversity, in particular, Group I beetles, which were significantly more abundant in forest oaks in heavily thinned patches, as compared with pasture oaks and oaks in moderately thinned patches. For both beetle groups, the composition differed between the forest oaks and pasture oaks, indicating that the forest oaks can be a complementary habitat to that of pasture oaks. There was a positive relationship between oak dead branch diameter and beetle biodiversity, but only for older oaks (?200?years old). We conclude that retaining oaks in production spruce forests can increase the diversity of oak-associated beetles at the landscape scale. Since many oak associated species depend on relatively high levels of insolation, management of retained oaks in production forests should include periodic removal of encroaching trees.

Stumps comprise up to 80% of the residual deadwood following clear cutting and are a significant source of biomass for bioenergetic applications. However, stump harvesting may pose significant conservation risks for saproxylic organisms that occur in residual deadwood. To define retention targets for stump harvesting operations, we compared abundance and species richness of saproxylic beetles within individual stumps as well as species accumulation curves in replicated pairs of clear cuts with and without stump harvesting in northern Sweden. Using 20 stands, we sampled 1049 stumps using eclector traps and collected 9821 beetles representing 253 species with known saproxylic biology. Nineteen of these species were red-listed in Sweden. We hypothesized that individual stumps left following stump harvesting would contain higher densities and species richness than in clear cuts without stump removal due to crowding of beetles into increasingly limited habitats. However, we found no difference in density or richness within individual stumps between control clear cuts and stumped stands. We also compared species richness between control and stumped treatments using rarefaction within individual stands and across all stands and found no difference. As with density and richness, beetle composition at the stand-level did not differ between control and stumped stands. Thus, the density of surrounding stumps within a stand had very little effect on beetle assemblages in residual stumps. We estimated the effect of stump harvest on species richness at the stand level by combining all samples and extrapolating a rarefaction curve derived from the landscape-level species pool to an accumulated sample volume of 48 m3 which corresponds to the total volume of stumps on average-sized clear cuts in Northern Sweden. Using this curve, we compared differences in species richness in average-sized clear cuts with 100% (48 m3) and 25% (12 m3) stump retention and found that stump harvest resulted in a 26% (95% C.I. 7–41%) loss of species. While the absolute scaling of the landscape-derived rarefaction does not reflect species loss at the stand-level because the combined curve reflects all rare species in the landscape, the relative species loss derived from this curve may serve as credible benchmark for species loss at the stand level following current stump harvesting practices. This benchmark may be further calibrated with additional information on number of singleton species and estimates of maximum species richness.