multilandr: an R package for multi-scale landscape analysis

Abstract

Context Typical landscape-scale studies comprise the delimitation of landscapes followed by the calculation of one or more landscape metrics. Performing an analysis at multiple spatial scales is often required, occasionally followed by the selection of a particular scale according to the response variable of interest. More complex research goals might require a thorough inspection of landscapes, plus a selection of landscapes that would fulfill certain conditions regarding their landscape metrics. These tasks can usually be programmatically challenging, especially if multiple spatial scales are being analyzed. Objectives The R package multilandr builds on several spatial-oriented R packages to provide a toolbox to develop and inspect multi-scale landscapes based on simple spatial inputs. Methods and results The package delivers functions to calculate metrics within a multi-scale framework. Also, it provides several utility functions to visualize correlations between metrics, filter landscapes that fulfill certain predefined conditions or select a wide-range gradient of landscapes for a given metric, among other useful tasks. This paper introduces the functionality of multilandr through a step-by-step instruction guide and case studies. Conclusions The R package multilandr provides a set of functions to facilitate typical methodological workflow of landscape-scale studies in the R environment, for both beginner and expert R users. It provides the functionality to perform a systematic filtering and selection of landscapes according to a given experimental design. The package is especially programmed to develop multi-scale designs but is also useful for the calculation of metrics of a set of landscapes from any GIS-related project.