Conserving natural heterogeneity is crucial for designing effective ecological networks

Abstract

Large-scale ecological networks (ENs) are an important mitigation measure in agriculturally transformed landscapes. However, understanding the multitude of pressures influencing the presence/absence of species, and subsequent degree of species spatial heterogeneity, is important when planning effective ENs. We aim here to measure these pressures and determine species heterogeneity in ENs against natural reference sites. We use arthropods, as they are effective bioindicators for measuring these pressures and heterogeneity, as many are habitat sensitive. Here we use many arthropod taxa to determine how a suite of variables influences the spatially sensitive grassland interior species of both EN corridors and protected areas (PAs). At each of 48 selected sites, nine stations were sampled for arthropods, with six stations in plantation block (i.e. transformed grassland) or natural indigenous forest, as well as associated edge zone and three stations in EN corridor or PA interior. Eleven variables were measured and classed into environmental, design, and current and historical management variables. Data were split into: overall data, recording all species found in interior zones, and datasets containing only species that had >50 or >75 % of their abundance sampled in the interior zone. These datasets were split into total species and singleton-removed datasets. Overall, the richness of non-singleton species, i.e. those frequently sampled in grassland interiors, were most responsive to natural background environmental variables, while design and management variables were most important for datasets with singletons retained. This means that when planning ENs, we first need to conserve the natural range of environmental heterogeneity to conserve a range of interior specialists. This natural spatial heterogeneity then needs to be incorporated into design and management planning to conserve the full range of biodiversity in ENs, as if in PAs.