Distance-Dependent Landscape Effects in Terrestrial Systems: a Review and a Proposed Spatio-Temporal Framework

Purpose of Review We review recent methodological advancements in estimation of distance-dependent landscape effects on terrestrial species. These methods address key theoretical elements from landscape and metapopulation ecology that were ignored in previous approaches. Models that treat landscapes as circles within which all land features are equally important to a focal population ignore distant-dependent population processes, such as dispersal, resource selection, and social interactions. Realistic models that estimate variation in landscape-scale effects over space and time are necessary to understand the complex processes that influence population dynamics. Recent Findings The addition of kernel smoothers to generalized linear models has potential to increase the biological realism of landscape-species models. These models include estimation of parameters that dictate the relationship between distance and importance of landscape features to focal populations. There are examples of implementing these models in both maximum likelihood and Bayesian frameworks, as well as examples using model selection to determine appropriate smoothing kernel shape. One key limitation of these models is computational effort, although we provide some guidance for reducing model runtime. Summary Models allowing for inference on explicit ecological processes are critical to advancing knowledge of the basic landscape ecology of species and will benefit efforts to prioritize conservation and evaluate species recovery efforts. We describe how distance-dependent landscape-scale effect models can be used for these purposes in a variety of scenarios. We conclude by proposing a process-based, spatio-temporal framework for understanding the mechanisms behind the spatial scale at which landscapes influence species.