A General Model of Discretized Rarity

We introduce a general model of discretized rarity (GDR) that incorporates geographic, functional, and phylogenetic dimensions at regional and local scales. This model can be simplified based on research question, management goal, spatial scale, and data availability to create workable definitions and link existing approaches.

The GDR includes 64 novel and existing definitions of rarity

The general model of discretized rarity encompasses many different definitions of rarity and can be visually interpreted similarly to a codon table. Interestingly, because these different definitions are all simplifications of the same general model, they can be compared directly using the R package GDRarity in their ability to explain biology.

Rarity explains variation in distribution change and flowering phenology

When applied to 1,011 British plant species, the most comprehensive sub-models (i.e., those characterized by high multidimensionality in rarity) of the GDR performed best at explaining biological processes, such as flowering phenology and distribution change over 32 years. While rarity is often used to classify species, these results demonstrate that rarity can also be used to understand broader ecological processes. This raises an important question: should rarity be defined by multidimensional completeness or by its ability to explain important biological patterns?

The role of rarity in communities and ecosystems is likely complex because the interactions between rarity dimensions are equally complex

Rarity dimensions interact, such that the effect of any single dimension on biological and biogeographic variables depends on whether other dimensions are also present. Therefore, how we understand the impacts of rare species on community structure and ecosystem function will vary by community and across the landscape. Addressing this complexity will require novel experimental designs that better capture the contribution of multidimensional rarity. The GDR and its novel simplifications offer a new set of tools to integrate rarity across diverse research and conservation efforts to address new questions and generate novel hypotheses tailored to specific goals, scales, and data availability.

The Evolution of Rarity

There are many ecological and anthropogenic drivers of rarity, but there are also key evolutionary processes that have contributed to the existence of multidimensional rarity through time.

Using the regional and local geographic, functional, and phylogenetic rarity of 1,011 British plant species, we applied univariate and bivariate hidden state speciation and extinction models to investigate the origins of rarity, rarity at evolutionary equilibrium, and associations between rarity and differential diversification.

Contemporary rarity is reflective of non-equilibrium dynamics

Rarity represented the most probable ancestral state for many rarity dimensions, but many dimensions were also predicted to occur in very low frequency at equilibrium. The current prevalence of rare species is likely driven by a combination of ecological factors and non-equilibrium dynamics. The low turnover rates of rare lineages may enable these non-equilibrium dynamics to persist for millions of years.

Rarity is generated by repeated and rapid transitions from commonness

The non-zero or increased frequency of rarity dimensions at equalibrium may be explained by state changes through time. Across all dimensions, lineages lost rarity 1.8 to 5.1 times faster than they gained it. However, all transitions to and from rarity occurred relatively rapidly. When comparing the average speciation, extinction, and transition rates for rare versus common lineages, the next probable step at any given time for a rare lineage is to transition out of rarity, while common lineages are most likely to speciate or go extinct.

The current prevalence of rarity reflects non-equilibrium dynamics: at equilibrium, rare lineages are expected to exist at low frequency but will be generated by evolutionary state changes rather than by diversification processes.