Short-term effects of retention felling at mire sites on boreal spiders and carabid beetles

Green-tree retention is an integral part of forest management in the boreal zone. Retention of small spruce mires, proposed as ‘key habitats’ for many forest organisms, is recommended while logging, but the efficiency of such practices for the maintenance of forest species is poorly understood. Hence, we studied boreal spiders and carabid beetles at 11 retained mire patches (up to 0.55 ha) in Eastern Finland during 1998–2001. The adjacent surrounding drier forests of the focal mires were clear-cut during the winter of 1998/1999. We evaluated the importance of micro-habitat type (retention, edge or clear-cut plot), time since logging, and retention-patch size in determining the assemblages of spiders and carabids. Following logging, species associated with forests and mires generally decreased both in the retained mire patches and in their adjacent clear-cuts. In a GLMM, the number of standing trees in a retention patch – a surrogate for retention-patch size – had no significant effect on most of the tested species, but plot type (retention patch, its edge or clear-cut) was significant for many species. Semi-open-habitat species and open-habitat specialists increased following logging, especially in the clear-cut plots but even within the retention patches. In contrast, mature-forest and moist-habitat specialists became significantly less abundant in clear-cuts than in retention patches following logging. Spider assemblages showed pre-harvest differences among the mires, their edges and adjacent drier forests, but the fauna of logged plots rapidly diverged from that of mire and edge plots. However, after a 1–2 summers time lag, the spider fauna of mires and edges changed toward clear-cuts. For spiders, the post-harvest sample heterogeneity was significantly lower in clear-cuts than in retention patches, including their edges. Carabid responses were generally ambiguous. Multivariate regression trees showed that the number of trees in a forest patch better determined the spider assemblage structure than study area, study year or micro-site type (retention patch, its edge or clear-cut), indicating a strong impact of logging. For carabids, however, the study area better determined the assemblage structure; the other factors were of minor importance. Our results suggest that, as the spider and carabid faunas of the retention patches had drastically changed following logging, (i) retention patches should be considerably larger than the studied size range to efficiently maintain a ‘mire core’ spider and carabid assemblage; (ii) the effect of logging may take years to appear; (iii) spiders were more sensitive to habitat change than carabids; and (iv) harvesting not only changes the relative abundances of forest- and open-habitat associated species but it may also locally decrease the faunal variation.