Strong influence of local habitat structure on mammals reveals mismatch with edge effects models

Context: What determines mammal occurrence across wildland-urban edges? A better understanding of the variables involved will help update edge effects theory and improve our ability to conserve biota in urbanizing landscapes. Objectives: For the first time, we tested whether the occurrence of mammals across urban-forest edges and forest interiors was best predicted by: (1) edge variables (i.e. edge type and distance to an urban boundary), (2) local habitat structure (e.g. proportion of understory cover), or (3) edge variables after accounting for local habitat structure. Methods: Using 77 camera stations in South-Eastern Australia, we quantified the factors influencing the occurrence of five native mammals (brown antechinus, bush rat, common brushtail possum, black wallaby and long-nosed bandicoot) and three non-native mammals (red fox, cat, and dog). Results: The occurrence of most native and non-native mammals was best predicted by local habitat structure rather than by edge variables. Although edge variables had effects on most species occurrences, local habitat structure outweighed the impacts of edge effects. Conclusions: Our findings are important for management and urban planning as they suggest that local-scale management of habitat and habitat retention at urban edges will mitigate urban impacts on fauna. Our work reveals a critical mismatch in the spatial scale of predictive variables commonly used in edge effects models (edge types and distance to a boundary) compared with the smaller scale of local habitat variables, which underlie most species occurrence. We emphasize the need to consider heterogeneity within patches in predictive frameworks of edge effects.

Overcoming solvent mismatch limitations in 2D-HPLC with temperature programming of isocratic mobile phases

This work describes a method for two-dimensional high performance liquid chromatography (2D-HPLC) that uses an isocratic mobile phase with a temperature gradient in the first dimension. Temperature programming was used to manipulate solvent elution strength in place of a mobile phase concentration gradient. This ensured that all eluent fractions transferred into the second dimension were of an identical solvent composition, i.e. the second dimension injection solvent did not increase during the course of the analysis. When applied to a complex natural product extract of coffee, the separation was completed in 35 min and had an orthogonality of 35% (calculated using the bins method) and a spreading angle of 52° as determined via a geometric approach to factor analysis. This approach, incorporating a temperature gradient in the first dimension, compared favourably to previously reported 2D-HPLC separations of coffee, with similar or shorter analysis times.