Recent alarming losses of insects from agricultural landscapes in multiple countries around the world have brought into sharp focus the urgent need to identify ways to manage these landscapes to avoid further biodiversity decline. Identifying the drivers of insect declines, such as land use change, is critical to this effort. We examined ant communities at the interface between remnant vegetation patches and three adjoining farmland types (wheat crop, rested from cropping and restoration plantings) in a fragmented landscape in temperate Australia. We asked: do ant communities and occurrence of individual species differ between remnant patches and farmlands with more intensive farmland use (restoration plantings < rested farmlands < wheat crop)? We recorded 13,283 ants belonging to 102 species from 30 genera. Excluding 21 singletons, 27 species only occurred in remnant patches compared to ten species found only in farmlands. Ant community composition in wheat crop and rested farmlands significantly differed from their adjacent remnant patches and were more homogeneous. In contrast, ant communities from restoration plantings in farmland were not significantly different in composition from those in the adjacent remnant patch. The large, aggressive Australian meat ant (Iridomyrmex purpureus) showed significantly higher occurrence in the remnant patch than all farmland types, and we suggest that the absence of this strongly interacting species from farmlands may have contributed to biotic homogenisation. Our findings show that native vegetation provides crucial habitat resources for many ant species that are not provided by farmlands, and native plantings can, in some cases, ameliorate negative effects of farmland clearing over relatively short time scales (<7 years). Agricultural intensification that involves loss of remnant native vegetation or reduced revegetation will contribute to ongoing losses and changes to ant biodiversity in farming landscapes. However, replanting native vegetation can lead to rapid restoration, signifying a possible simple remedy to insect declines.
Plant diversity is threatened in many agricultural landscapes. Our understanding of patterns of plant diversity in these landscapes is mainly based on small-scale (<1000 m2) observations of species richness. However, such observations are insufficient for detecting the spatial heterogeneity of vegetation composition. In a case-study farm on the North-West Slopes of New South Wales, Australia, we observed species richness at four scales (quadrat, patch, land use and landscape) across five land uses (grazed and ungrazed woodlands, native pastures, roadsides and crops). We applied two landscape ecological models to assess the contribution of these land uses to landscape species richness: (i) additive partitioning of diversity at multiple spatial scales, and (ii) a measure of habitat specificity – the effective number of species that a patch contributes to landscape species richness. Native pastures had less variation between patches than grazed and ungrazed woodlands, and hence were less species-rich at the landscape scale, despite having similar richness to woodlands at the quadrat and patch scale. Habitat specificity was significantly higher for ungrazed woodland patches than all other land uses. Our results showed that in this landscape, ungrazed woodland patches had a higher contribution than the grazed land uses to landscape species richness. These results have implications for the conservation management of this landscape, and highlighted the need for greater consensus on the influence of different land uses on landscape patterns of plant diversity.