A null biogeographical test for assessing ecological niche evolution

Abstract

Aims Quantification of the degree to which ecological niches change over evolutionary time‐scales is important for deepening our understanding of evolutionary and ecological processes. Phylogenetic niche conservatism (PNC) is when closely related species differ less ecologically than expected by chance, whereas Phylogenetic Niche Divergence (PND) is when closely related species differ more ecologically than expected by chance. We present a new null model to test for PNC and PND (the random translocation and rotation (RTR) significance test), which we combine with a novel metric for quantifying niche overlap. Location Europe, North America and Madagascar. Methods The RTR null model comprises many thousands of replicates generated by randomly translocating and rotating the set of occurrence records for two populations (e.g. sister species) while maintaining the spatial configuration between all occurrences within each replicate. For each replicate, we calculate niche overlap as the proportion of the combined niche breadth that is shared by the two species, averaged over n environmental dimensions. This approach enables us to test whether the observed niche overlap is more or less than expected by chance given the environmental conditions present in the study area. We test the performance of our approach in comparison to other methods using both simulated and real case scenarios, including crested newts in Europe, pocket gophers in North America, and lemurs in Madagascar. Results We find that our measure of niche overlap performs better than other metrics in an artificial simulation scenario, and we find evidence for both PNC and PND in our case studies for Europe, North America and Madagascar. Our results demonstrate that both the RTR significance test and the novel metric of niche overlap are consistent with evolutionary theory and are suitable methods to test for PNC and PND. Main conclusions We make available scripts to implement the RTR test and metric of niche overlap, and expect that the methods will prove useful for addressing a broad set of questions relating to ecological niche evolution and speciation, particularly for restricted‐range species for which few known occurrence records are available.

The relationship between environmental niche breadth and geographic range size across plant species

“…Our RDTR model is similar to the RTR null model presented by Nunes and Pearson (), but here, we additionally allow the random translocation of the centroid to be weighted by the empirical density of all species range centroids within each continent. Density of species range centroids was estimated using two‐dimensional kernel density estimation with the ‘bkde2D’ function in the ‘KernSmooth’ package for R. Centroid translocations were then done by randomly sampling from the estimated density distribution.…”

“…Our RDTR model is similar to the RTR null model presented by Nunes and Pearson (2017), but here, we additionally allow the ran- Each null model was run 500 times. For each run, we quantified niche breadth and tested the association between range size and niche breadth using PGLS, in the same way as for the observed data.…”

Predicting Future Seed Sourcing of Platycladus orientalis (L.) for Future Climates Using Climate Niche Models

“…In cases of occurrence records missing geo-coordinates, Google Map (http://www.google.cn/maps/) was used to determine their latitude and longitude. Incorrect species occurrence information (e.g., presence in extreme environments or unsuitable habitat) may inflate the measure of niche overlap [46]. To reduce this error, outliers were identified by a multi-dimensional-based approach using Mahalanobis distances [47] and were eliminated if they occurred outside the 5-95th range of the niche space using the raster package [48] in R [49].…”

Local niche differences predict genotype associations in sister taxa of desert tortoise

“…For example, under conditions of allopatric speciation, SDM can be used to develop niche models for sister species in order to quantify niche differences using ecological overlap metrics (e.g., Broennimann et al, ; Godsoe, ; Rödder & Engler, ) and statistical tests (e.g., Nunes & Pearson, ; Warren et al, ; Warren, Glor, & Turelli, ). Substantial similarity between niches may suggest niche conservation between the two allopatric species (Warren et al, ; Wiens & Graham, ), such that the species–environment relationships are maintained through time and across taxa, even in the presence of environmental change or speciation (Wiens & Graham, ).…”