Bird-community responses to habitat creation in a long-term, large-scale natural experiment

Ecosystem function and resilience are compromised when habitats become fragmented due to land-use change. This has led to national and international conservation strategies aimed at restoring habitat extent and improving functional connectivity (i.e. maintaining dispersal processes). However, biodiversity responses to landscape-scale habitat creation and the relative importance of spatial and temporal scales is poorly understood, and there is disagreement over which conservation strategies should be prioritised. Addressing these knowledge gaps has been challenging because (1) there can be a significant time lag between habitat creation and biodiversity responses, and (2) many taxa respond to landscape characteristics over large spatial scales. These conditions can be difficult to replicate in a controlled setting but can be simulated using ‘natural’ experiments. Here, we used 160 years of historic post-agricultural woodland creation as a natural experiment to evaluate biodiversity responses to landscape-scale habitat creation. Specifically, we disentangle the direct and indirect relationships between bird abundance and diversity, ecological continuity, patch characteristics and landscape structure, and quantify the relative importance of local and landscape scales. Results suggest that ecological continuity has an indirect effect on total bird species richness through its direct effects on stand structure. However, for functional groups most closely associated with woodland habitats, ecological continuity had little influence. This was probably because woodlands were rapidly colonised by woodland generalists in < 10 years (the minimum patch age), but were on average too young (median 50 years) to be colonised by woodland specialists. Local, patch characteristics were relatively more important than landscape characteristics. We conclude that biodiversity responses to habitat creation are dependent on local and landscape-scale factors that interact across time and space. We also suggest that knowledge gained from studies of habitat fragmentation/loss should be used to inform habitat creation with caution, since the two are not necessarily reciprocal.This article is protected by copyright. All rights reserved