Local-landscape filtering explains insect diversity patterns across a dryland urban landscape

Context: Urbanization has a profound impact on biodiversity worldwide. The commonly applied ‘urban–rural gradient’ approach often yields contrasting patterns of insect diversity. Integrating local-scale habitat and landscape-scale attributes that act as filters of community assemblage (‘local-landscape filtering’) may help reconcile these contrasting patterns. Objectives: Our objective is to evaluate if and to what extent the urban–rural gradient or the local-landscape filtering approach can explain insect diversity patterns across a dryland city. Methods: We conducted field surveys of insects at 73 diurnal sites and 20 nocturnal sites across the urban landscape of Urumqi, China. We measured habitat variables and landscape attributes surrounding the local sites. We assessed the urban–rural gradient using distance to urban center, population density, artificial light at night, and impervious surfaces. We used generalized linear mixed models to assess the effects of these factors. Results: Insect species richness did not follow clear urban–rural gradient patterns. For diurnal and nocturnal insects, local habitat factors (including plant richness and soil properties) explained ~ 68% and ~ 77% of the variation, whereas landscape attributes explained ~ 32% and ~ 23%, respectively. Diurnal ants and nocturnal mosquitoes/flies were more strongly associated with landscape attributes, while diurnal leafhoppers, moths/butterflies, along with nocturnal thrips, were mostly associated with local habitat. Conclusions: Well-tailored local-landscape filtering approaches better explain insect diversity patterns across urban landscapes. Complementary to the urban–rural gradient, this framework facilitates understanding of how urban biodiversity patterns are shaped by particular factors at given spatial scales.