Geographic and environmental dispersion of clonal genotypes.
For genotypes with multiple clones collected from the wild, we calculated pairwise geographic distance between collections within each clonal genotype based on GPS coordinates of each collection with thedistHaversine function in R v3.4.1 (Hijmans, Williams, & Vennes, 2019; R Core Team, 2013). To measure pairwise environmental distance, we calculated Euclidian distance for environmental variables in three dimensions with the dist function in R, as follows: coordinates for the first and second dimension were obtained as the first and second principal components in a PCA of the 19 WorldClim BioClim variables (Fick & Hijmans, 2017) for each of our collection locations. Principal components were calculated using the prcomp function in R with scaled unit variance. The values of the first two PCs explained 50.9% and 22.7% of the variance respectively. For the third coordinate, we downloaded the average sea surface salinity dataset for the years 2005-2017 from the World Ocean Atlas (Zweng et al., 2019) and identified the salinity of the ocean water nearest to each collection site (excluding one sample collected > 40 km from the coast). All three coordinates where then centered and scaled with thescale function in R before calculating distance.