Work

‘envar’: An R package to retrieve and process environmental variables for macroecology and biogeography

Year: 2026

Type: article

Abstract: The selection of environmental variables is a cornerstone of biogeographical analyses. However, most studies use standard sets of predictors, without fully leveraging the variety of available environmental data. Here, we present ‘envar’, an R package to retrieve and process 372 global variables at ~1 km resolution from 23 different sources, describing climate and land cover (past, current, and future), landscape heterogeneity, habitats, soil characteristics, topography, freshwater/marine environments, human impacts, and protected areas. Through ‘envar’, the user defines a study area, a buffer in km, a coordinate reference system, the resolution, the variables to download, and the correlation and extrapolation checks. The output is one or multiple raster(s) with the downloaded and processed variables, or a table with the extracted variables over the specified points, and optionally information on correlation and extrapolation. We also illustrate the package's potential to streamline the development of Species Distribution Models (SDMs), using as a case study the butterfly Parnassius apollo in the European Alps. We use ‘envar’ to: 1) create a raster template with a buffer around occurrences where bias-corrected background points are selected, 2) extract predictors for occurrences and background points across Europe checking for collinearity, and 3) obtain predictors for the European Alps checking for extrapolation. ‘envar’ can foster the selection of a set of uncorrelated proximal variables tailored to the specific research question, across all of biogeography and macroecology, while limiting niche truncation, collinearity, and extrapolation.

Source: Ecological Informatics

DOI: 10.1016/j.ecoinf.2026.104033