Tree biodiversity in northern forests shows temporal stability over 35 years at different scales, levels and dimensions

Abstract

Abstract At global scales, species richness is declining. However, at local scales, understanding exactly how, where and why biodiversity is changing becomes challenging since researchers have assessed biodiversity trends using different indicators, data sources and methods (e.g. repeated measurements at the same site over time vs. space‐for‐time substitutions). In this study, we present a multifaceted analysis of biodiversity change by assessing how tree diversity in Québec, Canada changed between two sampling periods (1970–1977 and 2005–2016), in regards to different: levels of diversity (alpha diversity, temporal turnover and spatial beta diversity), dimensions of diversity (taxonomic, functional and phylogenetic), metrics of diversity (presence–absence and abundance based), and spatial scales of analysis (plot, 50, 100 and 200 km). We then assess how well potential drivers of biodiversity change (climate change and land cover change) explain the observed changes in alpha diversity. Since the data came from plots that remained forested over the course of the study, we used historical land cover change data and scenario analyses to test whether results from forest plots were likely to be representative of the broader landscape. Across all levels, dimensions, metrics and spatial scales of analysis, we found either increases or no net change in diversity over time, with wide distributions of values around the mean. Presence–absence metrics often indicated increases in diversity over time, while abundance‐based metrics were more likely to show no net change. Potential drivers such as climate change and land cover change explained only a small fraction of the variation in alpha diversity change (i.e. why particular sites experienced positive vs. negative change) at the plot scale (adjusted R2 ≈ 0.03), but a greater fraction at coarser spatial scales (adjusted R2 of ~0.10 to ~0.50). Results from these forest plots are likely representative of the diversity change within the study region, since estimates of alpha diversity change only became negative under scenarios with the most extreme disturbance impacts. Synthesis. None of our indicators showed evidence of declines in alpha or beta diversity of trees in temperate and boreal forests in Quebec (except for simulations with extremely high forest loss), but we did find temporal turnover in composition. Our results go beyond previous studies by analysing the many different aspects of biodiversity that might change over time, highlighting the need to shift from blanket descriptions of ‘biodiversity loss’ to more nuanced discussions around ‘biodiversity change’.