The influence of fine-scale habitat heterogeneity on invertebrate assemblage structure in upland semi-natural grassland

The Scottish uplands consist of a variety of semi-natural habitats of high nature value that are typically managed by extensive livestock grazing. However, as a result of the decoupling of financial subsidies from production in the recent Common Agricultural Policy reform, sheep numbers in the uplands have significantly declined and there is concern that the conservation value of these habitats is at risk. A large-scale experiment that manipulated sheep grazing management provided an excellent opportunity to examine the influence of fine-scale habitat characteristics on invertebrate assemblages. A Geographical Information System was used to extract habitat data at four spatial scales ranging from 0.25m to 5.0m in ellipses around pitfall transects. Multivariate analyses were applied to determine the most appropriate spatial scale and the primary habitat characteristics influencing three categories of invertebrates (i.e., mobile arthropods, immobile invertebrates and carabid beetles). Sites that were grazed year-round, and thus more intensively, were associated with smaller mobile arthropods (e.g., money spiders), tipulid larvae and earthworms, while sites that were grazed during summer months only were associated with larger mobile arthropods (e.g., wolf spiders and Carabus beetles) and sawfly and lepidopteran larvae. Grazing regime and the area of fine and broad-leaved grasses were found to influence all three groups of invertebrates at a range of spatial scales, thus indicating the importance of these variables in predicting assemblage structure. The optimum spatial scale at which to collect associated habitat information was from within 0.25m for immobile invertebrates, and from within 3m for mobile arthropods and carabids. This supports the hypothesis that immobile invertebrates react to their surrounding habitat at a finer spatial scale than mobile invertebrates and draws into question the effectiveness of collecting habitat information from directly adjacent to the pitfall traps when studying mobile arthropods.