‘Scalescape’: an R package for estimating distance-weighted landscape effects on an environmental response

• Published in: Landscape ecology 2022-07, Vol.37 (7), p.1771-1785
• Main Authors: Lowe, Erin B., Iuliano, Ben, Gratton, Claudio, Ives, Anthony R.
• Format: Article
• Language: English
• Subjects: Biomedical and Life Sciences; Decision analysis; Ecology; Environmental Management; Landscape Ecology; Landscape studies; Landscape/Regional and Urban Planning; Life Sciences; Methods; Nature Conservation; Organisms; Research Article; Statistical analysis; Sustainable Development; Variables
• ISSN: 0921-2973; 1572-9761
• DOI: 10.1007/s10980-022-01437-5

Context Landscape studies often focus on determining how the landscape around a discrete set of field sites affects an abiotic or biotic response measured at those sites. To run this type of analysis, a decision must be made about the spatial extent over which the landscape affects the response. Many authors have acknowledged the limitations of common approaches to estimating this spatial extent of a landscape effect and some have proposed alternatives. However, many alternative methods have thus far been difficult to implement. Objectives The R package, ‘scalescape’ which we introduce in this paper, builds on and improves the usability of these alternative methods for estimating the spatial scale of a landscape effect on an abiotic or biotic response. Methods and results The package uses established approaches that weight landscape variables based on their distance from where a response is measured. It integrates well-used functions for performing regression in R, such as lm(), glm(), lmer(), glmer(), and gls(), with landscape weightings of spatial predictor variables. Here, we provide an introduction to ‘scalescape’ including a user guide detailing its functionality and step-by-step instructions. We also conduct simulations to illustrate and validate the ‘scalescape’ approach. Conclusions We conclude that ‘scalescape’ builds on previously proposed methods for landscape analyses by making it easier to implement distance-weighted approaches and by improving the statistical validity of these analyses through the use of GLS models and bootstrapping.