Effects of post-fire salvage logging on boreal mixed-wood ground beetle assemblages (Coleoptera, Carabidae)

The frequency and intensity of salvage logging has recently increased in burned forests of the Canadian boreal so that post-fire areas make up a significant annual share of all harvested forest land in some years. However, little is known about how this practice affects re-establishment of animal and plant communities that already have been strongly altered by the fire. We pitfall-trapped carabid beetles (Coleoptera, Carabidae) for two consecutive years following a large-scale boreal wildfire (276,000 ha) in burned mixed-wood forests of 12 landscape units (2.5 km x 2.5 km each). These units included four salvage treatments, each replicated three times: control (no salvaging), low (ca. 20-30% of merchantable mixed-wood removed), moderate (40-50%) and high salvage intensity (60-70%). We established 16 sampling sites in each unit: on the control landscapes, all 16 sites were in un-logged stands; in low-salvage units, 4 out of 16 sites were in salvaged areas; in moderate- and high-salvage units, 8 and 12 out of 16 sites, respectively, were in salvaged areas. Salvage logging positively affected carabid species richness. However, there was an overall salvage-caused decrease in the abundance of many common forest-dwelling carabids, and an increase in the abundance of disturbance or open-habitat specialists. Interestingly, the effects of salvage logging on the total catch and several abundant forest species (e.g., Calosoma frigidum, Harpalus laevipes, Pterostichus punctatissimus, P adstrictus and Platynus decentis) appeared to be more important at the landscape level (the four landscape treatments) than at the level of sampling sites (site logged versus not). We suggest that this observation resulted from ambiguous site-level responses to salvage that collectively contributed to the (mostly negative) significant responses at the landscape level. Effects of fire severity (estimates based on tree survival) on carabids were species-specific; however, the impact of this measure was most often significant at the site, and not at the landscape, level. A Multivariate Regression Tree revealed that fire severity and the overall (pre-fire) amount of mixed-wood forest on the landscape were significant determinants of assemblage structure, with the local effects of fire severity being strongest. (c) 2006 Elsevier B.V. All rights reserved.