Search

Narrow endemism presents challenges to species occurrence modeling particularly when the distribution of key local habitat features changes across space and time as a function of processes operating at larger scales. One need facing conservation in such settings is a better understanding of the biogeographic dynamics of the larger features that govern occurrence of critical local habitat. The Mescalero–Monahans shinnery sands region of western North America is a dynamic landscape where sand shinnery oak interacts with wind-driven sand to establish dune habitat. This ecosystem supports several narrowly endemic dune-dwelling species including the dunes sagebrush lizard. Using near-anniversary satellite and aerial imagery from 1986, 1998, and 2011, we integrated object-based image classification and statistical analysis to develop and validate a spatially explicit estimate of the sand shinnery oak ecosystem, including dynamics associated with its attrition and emergence, at high resolution throughout an 89,849-km2 study area encompassing the range of the dunes sagebrush lizard. The spatial estimate of the distribution and extent of the sand shinnery oak soil-vegetation association validated reasonably well (overall accuracy = 0.79; sensitivity = 0.49; specificity = 0.91) and showed that the association declined 10.3 % in extent during the 25-year assessment window. The presence of sand shinnery oak, patch size, and patch isolation were dynamic across space and time; a regression model showed that smaller, isolated patches on the periphery of the system were more likely to be lost over time whereas larger, less isolated, and centrally distributed patches were more likely to persist or expand. This study details broadly applicable methods to accurately delineate landforms throughout large extents, and offers mapping tools specific to issues surrounding Mescalero–Monahans shinnery sands endemics that are readily amenable to testing, refinement, and application in efforts to focus sustainable landscape management including conservation of endemic species.

Prioritizing habitat for animal conservation in heterogeneous landscapes requires an understanding of where animal occurrence coincides with human influences on demographic performance. We related broad-scale patterns of occurrence with risk of mortality among female Rocky Mountain elk (Cervus elaphus) in a human-modified landscape to develop a spatially-explicit framework for animal conservation at the landscape level. Variability in the spatial pattern of elk occurrence was driven by preference for specific habitat types as well as responses to human activity. In contrast, risk of mortality was a function of human modification of the landscape with little variation explained by habitat. Proximity to industrial development was associated with increased risk of mortality whereas proximity to residences and agricultural structures was associated with decreased risk. Individual-level results revealed added complexity, whereby risk of mortality was associated with a consistent pattern of occurrence relative to industrial development, yet the association between risk and occurrence relative to structures was highly variable and likely a function of disparate land-use priorities. Approaches to managing human-mediated risk at the landscape level are most effective when they decompose human activity into constituent parts influencing risk, and when individual variation relative to the population response is investigated. Conservation interventions need to target factors that have a consistent influence across the population rather than risk uncertainty that would arise from targeting factors that influence individuals in variable or situation-specific ways. The spatial tools developed herein provide guidance for sustainable landscape planning in the study area, while the concept of linking occurrence and demographic performance within a hierarchical modeling framework has general application for animal conservation in landscapes subject to change, human-caused or otherwise.