Assessing the Spatial Scale Effect of Anthropogenic Factors on Species Distribution

Abstract

Patch context is a way to describe the effect that the surroundings exert on a landscape patch. Despite anthropogenic context alteration may affect species distributions by reducing the accessibility to suitable patches, species distribution modelling have rarely accounted for its effects explicitly. We propose a general framework to statistically detect the occurrence and the extent of such a factor, by combining presence-only data, spatial distribution models and information-theoretic model selection procedures. After having established the spatial resolution of the analysis on the basis of the species characteristics, a measure of anthropogenic alteration that can be quantified at increasing distance from each patch has to be defined. Then the distribution of the species is modelled under competing hypotheses: H0, assumes that the distribution is uninfluenced by the anthropogenic variables; H1, assumes the effect of alteration at the species scale (resolution); and H2, H3 … Hn add the effect of context alteration at increasing radii. Models are compared using the Akaike Information Criterion to establish the best hypothesis, and consequently the occurrence (if any) and the spatial scale of the anthropogenic effect. As a study case we analysed the distribution data of two insular lizards (one endemic and one naturalised) using four alternative hypotheses: no alteration (H0), alteration at the species scale (H1), alteration at two context scales (H2 and H3). H2 and H3 performed better than H0 and H1, highlighting the importance of context alteration. H2 performed better than H3, setting the spatial scale of the context at 1 km. The two species respond differently to context alteration, the introduced lizard being more tolerant than the endemic one. The proposed approach supplies reliably and interpretable results, uses easily available data on species distribution, and allows the assessing of the spatial scale at which human disturbance produces the heaviest effects.

Context-dependent expression of sexual dimorphism in island populations of the common wall lizard (Podarcis muralis)

“…Even though no direct measures of habitat features that may potentially represent environmental stressors for the expression of head shape in lizards has been carried out to date, our data suggested that small island size, low habitat diversity, no human presence and therefore lack of human related habitats, drier and warmer climate regimes might promote reduced intensity of head's shape sexual dimorphism (Figs , ). This pattern is consistent with the hypothesis that unfavourable environmental conditions may actually represent a stress which limits the expression of SD, since common wall lizards are well known to be well adapted to human made habitats (such as buildings and drywalls or olive groves and vineyards) and are generally less thermophilus than other Italian lizards like, e.g., Podarcis siculus (Biaggini et al ., ; Mangiacotti et al ., ). Thus, the environmental conditions of small islets or islands kept at low habitat diversity may reliably impose more severe limits to lizard survival, and consequently to the allocation of resources in the expression of secondary sexual characters.…”

Differential effects of habitat loss on occupancy patterns of the eastern green lizard Lacerta viridis at the core and periphery of its distribution range

“…Variables related to landscape composition were calculated at different buffer-distances (hereafter, "scales") around each patch in each region. The different scales were selected based on reported dispersal distances for L. viridis [37][38][39]. Scales selected were: 50m, 150m, 250m,…”

Testing for taxonomic bias in the future diversity of Australian Odonata

“…It is important to understand that the SDM outputs assume a species is present throughout all suitable habitats and overlaying them to estimate richness and composition is predisposed to include errors of commission and overestimate richness (Guisan & Rahbek, 2011;Pineda & Lobo, 2012). Species could be absent or uncommon at sites that are predicted to be environmentally suitable due to dispersal constraints, biotic interactions, unsuitable micro-habitats and stochastic effects (Heikkinen et al, 2006) or human habitat modification (Mangiacotti et al, 2013). For example, SDMs predicted 26 species could occur at Middle Creek in Victoria where Hawking and New (2002) sampled odonates (larvae and adults) intensively on 20 visits over three years.…”