Mapping assemblage-level tip-based metrics across space
We considered the assemblage-level tip-based metrics derived from all 10,000 estimates (100 ancestral character simulations for each of the 100 phylogenies) to build maps showing spatial variation on average and uncertainty (standard deviation) of aTR, aST, and aLT.
Relationship between assemblage-level tip-based metrics and richness
We used a Generalized Least Squares (GLS) regression (Pinheiro and Bates 2000) to analyze the effect of assemblage richness on aTR, aST, and aLT, because tip-based metrics can be high for assemblages with many species. We used GLS to account for spatial correlation in the relationship between richness and tip-based metrics. Correlation structure was exponential with nugget effect —the same we used in LMMs. Spatial and statistical analyses were conducted using packages ‘raster’ (Hijmans 2020), ‘sp’ (Bivand et al. 2013), and ‘nlme’ (Pinheiro et al. 2020) in R 4.0.2 (R Core Team 2020).
Relationship between assemblage-level tip-based metrics and phylogenetic diversity
We used the same GLS regression just described to test the influence of phylogenetic diversity on aTR, aST, and aLT, because tip-based metrics can be high for assemblages composed by phylogenetically distinct species. Furthermore, we also tested whether phylogenetic diversity varies between ecotones and cores. Ecotone assemblages can consist of species from different neighboring regions, thus resulting in higher phylogenetic diversity and likely higher values of tip-based metrics in ecotones than cores.