Towards a Better Understanding of Beta Diversity: Deconstructing Composition Patterns of Saproxylic Beetles Breeding in Recently Burnt Boreal Forest

Abstract

Compositions of ecological communities vary both in space and time, and this variation is thought to be driven by dynamical and historical ecological/biogeographical processes that act upon species distributions. Describing the spatial variation in species composition, i.e., beta diversity, and explaining underlying causes remain at the heart of ecological and conservation studies. As such, a range of indices has been devised to measure different aspects of beta diversity, which may correspond to different causal factors or explanations. In this context, one of the most fundamental and recurrent issue has been how to account for the pure spatial turnover (“species replacement”) from that caused due to variation in species richness gradients (“species loss”). A recently proposed framework appears to offer a comprehensive way to distinguish these pure “spatial” turnover and nestedness components of betadiversity. In addition, there has been suggestion of the potential utility of null models that control for richness difference, such as Raup-Crick probabilistic index, to understand processes driving beta diversity. In the present study, we adopt the betadiversity partition framework for analysing beta diversity of saproxylic beetles emerging from tree boles following forest fire. Saproxylic beetles, which depend directly or indirectly on wood of dead and dying trees, are known to colonize recently burned boreal forests. Yet, the influence of post-fire habitats on beta diversity patterns of these assemblages remains unresolved. We examined how tree species, gradients in tree size, burn severities as well as spatial gradients contributed to the turnover and nestedness components of beta-diversity. We applied null model analysis to examine how the two beta diversity components are related to betadiversity values expected under null distributions and beyond (“delta” beta diversity). We used two null models that both control for species richness, with the additional constraints on species incidence (FF null model) and not (FE null model). We extended the multivariate regression analysis of distance matrices to simultaneously model both the within- and between habitat classes variation of both betadiversity components. Our results demonstrate that the turnover component of betadiversity was linearly related to betadiversity expected beyond “random” under both, but particularly the FF null models. In contrast, the nestedness component of betadiversity was related to the null expectations under FF null model. The Raup-Crick was not related to turnover, nestedness component of biodiversity or to expectations under FF null model. We illustrate its mathematical relationship to that of FE null model. Our examination of habitat and spatial effect on betadiversity indicated that the turnover and nestedness components of beta diversity of saproxylic beetles are coherent upon different habitat attributes. The turnover component of beta diversity was primarily explained by tree species, turnover within jack pine as well as between jack pine and black spruce being higher than within black spruce stands. Among diameter at breast height (dbh) classes, turnover was higher among mid-sized trees (12-16cm and 16-20cm) than within smaller (8-12cm) or larger (20-24cm) trees as well as between them (8-12 vs. 20-24 cm). However, the nestedness component of beta diversity was primarily driven by variation in tree size, with highest scores being obtained between the smallest and largest dbh classes, and the lowest scores within and among the largest and mid-sized trees. The two beta diversity components also responded differently to burn severity; nestedness beta diversity and turnover were lowest respectively within low severity and between low and high severity. Portions of the variability of both components were explained by spatial variations within and between “forest-fire blocks”, but there was no consistent monotonic increase or decrease of with geographical distance; thus limited support for distance decay hypothesis of similarity. Our study provides explicit appraisal of null model analyses for beta diversity analysis and its link to the two phenomena of beta diversity, which we demonstrate that they were driven by different underlying causes. We will discuss the implication of identifying habitat attributes relevant to the two beta diversity components for post-fire management in the boreal forest.