Datasets and script for niche comparisons and GLMMs
Species distribution models are often used to predict the potential distributions of invasive species outside their native ranges and rely on the assumption of realized niche conservatism. Analyses observed that freshwater invasive species often show high degrees of niche expansion, suggesting limited reliability of species distribution models. However, observed niche shifts can arise because of both actual niche shifts, determined by biological factors, and apparent shifts, due to methodological issues. We compared metrics of niche dynamics calculated using different sets of variables to identify factors that could influence the rate of niche shifts. We collected presence-only data for 40 freshwater invasive animal species, then measured niche shift dynamics using 14 different combinations of environmental variables. Shifts were assessed measuring niche overlap, expansion, and unfilling, and testing for niche conservatism. We then built generalized linear mixed models relating niche shifts to methodological choices and biological features. Our results showed that methodological choices strongly affected all the considered niche dynamics metrics, while the effects of biological features were less prominent. Moreover, different niche dynamic measures sometimes provided contradictory assessments of niche conservatism. Niche analyses are powerful tools to predict areas at risk of invasion, but inappropriate methodological choices can lead to apparent niche shifts, questioning niche models reliability and biological interpretation. The high rate of niche expansion observed in freshwater invasive species highlights the importance of delineating objective criteria to determine the set of variables to be used in niche dynamic assessments.
We provide dataset and scripts used to perform niche comparisons between native and exotic distribution for 40 freshwater invasive species and generalised linear mixed models (GLMMs).
Species occurrences ("observations.zip") were downloaded from iNaturalist (https://www.inaturalist.org/).
The areas used as native and exotic backgrounds ("backgrounds.zip") were created using IUCN Red List distribution data (https://www.iucnredlist.org), literature data, and the Freshwater Ecoregions of the World (Abell et al. 2008, available at: https://www.feow.org/).
Niche comparisons were performed using the ecospat package (Di Cola et al. 2017). The file neded to perform niche comparisons are found in the "spbg.zip" folder, and they are needed to run the "niche_comparison_c.R" script.
GLMMs were performed to test the effects of methodological choices (bioclimatic variables, chemical variables, distance-weighted variables, limited background selection, number of parameters) and of biological features (taxonomical group, native range extent, exotic range extent, residence time) on nine measures resulting from niche comparisons. These models were performed using the ‘lme4’ (Bates et al., 2015) and the ‘GLMMtmb’ packages (Brooks et al., 2017). The script used to perform GLMMs is provided as "glmm.R" and it needs the "data_table.RData" file.
