Predictive Phylogeography
Before assessing whether these species harbor cryptic diversity, we made
phylogeographic predictions of cryptic and non-cryptic for both species
following the procedure introduced by EspĂndola et al (2016) using
Random Forest with bioclimatic variables associated with sample
localities and taxonomic ranks. Following Sullivan et al. (2019), we
also included trait values along with the bioclimatic and taxonomic
variables. The error rates we recovered were congruent with those found
by Sullivan et al. (2019) and thus these classifiers were used to make
predictions about Thuja plicata and Tsuga heterophylla .
Each classifier predicted neither species to harbor cryptic diversity
(Table 2), with the only variation in the prediction being the less
accurate classifier (bioclimatic data only).