Landscape disturbance models consistently explain variation in ecological integrity across large landscapes

Abstract

AbstractThe generally negative effect of anthropogenic disturbance on the quality of habitats for species viability makes it a common focus of conservation assessment and prioritization efforts. Although many available spatial models and metrics (e.g., distance to or density of disturbance) characterize impact patterns of anthropogenic disturbance on the landscape, a general evaluation of model performance against empirical measurements of ecological integrity is lacking. We tested both distance‐based and disturbance‐density models in relation to ecological indicators. The models included roads, residential and commercial development, agricultural land use, mining, energy development infrastructure, and transmission structures as disturbance sources. Model parameters were based on expert input and results from the published literature. The disturbance models were tested against two disparate and independent measures of habitat quality: a floristic quality index and measures of greater sage‐grouse population integrity. Floristic quality scores were significantly lower in vegetation plots closer to disturbances in a general distance‐based disturbance model across Colorado. Although the proportion of variation in floristic quality explained by anthropogenic disturbance was relatively low (8.5–11.8%), it appeared to represent a ubiquitous baseline negative effect of proximity to anthropogenic disturbance on the quality of vegetation communities. For both distance‐ and density‐based greater sage‐grouse models, modeled disturbance indices were significantly lower (10–12 times) near active than historic leks, and numbers of males counted at leks increased significantly (3.2–3.4 times) as modeled disturbance decreased. Our findings indicate that as a general class, geospatial models can depict effects of anthropogenic disturbance on both plant communities and individual animal species. Empirical validation of disturbance models focused on other species or regions is recommended to further evaluate the utility and reliability of these methods.