Abstract
The spatial distribution of biomes has changed considerably over deep time, so the geographical opportunity for an evolutionary lineage to shift into a new biome may depend on how the availability and connectivity of biomes has varied temporally. To better understand how lineages shift between biomes in space and time, we developed a phylogenetic biome shift model in which each lineage shifts between biomes and disperses between regions at rates that depend on the lineage's biome affinity and location relative to the spatial distribution of biomes at any given time. To study the behavior of the biome shift model in an empirical setting, we developed a literature-based representation of paleobiome structure for three mesic forest biomes, six regions, and eight time strata, ranging from the Late Cretaceous (100 Ma) through the present. We then fitted the model to a time-calibrated phylogeny of 119 Viburnum species to compare how the results responded to various realistic or unrealistic assumptions about paleobiome structure. Ancestral biome estimates that account for paleobiome dynamics reconstructed a warm temperate (or tropical) origin of Viburnum, which is consistent with previous fossil-based estimates of ancestral biomes. Imposing unrealistic paleobiome distributions led to ancestral biome estimates that eliminated support for tropical origins, and instead inflated support for cold temperate ancestry throughout the warmer Paleocene and Eocene. The biome shift model we describe is applicable to the study of evolutionary systems beyond Viburnum, and the core mechanisms of our model are extensible to the design of richer phylogenetic models of historical biogeography and/or lineage diversification. We conclude that biome shift models that account for dynamic geographical opportunities are important for inferring ancestral biomes that are compatible with our understanding of Earth history.[Ancestral states; biome shifts; historical biogeography; niche conservatism; phylogenetics].
Highlights
Biomes are ecologically and climatically distinct species assemblages that vary in size, shape, and continuity across geographical regions, in large part due to regional differences in temperature, precipitation, seasonality, altitude, soil types, and continentality (Whittaker 1970; Wolfe 1985; Olson et al 2001; Mucina 2019)
It is accepted that clade-wide variation in regional biome occupancy was generated and is maintained by evolutionary forces including speciation, extinction, dispersal, and adaptation to new biomes, it remains difficult to estimate exactly when, where, and under what conditions phylogenetic lineages first shifted into the biomes that their descendants inhabit today
We aim to address the aforementioned challenges facing current phylogenetic models of biome shifting by incorporating four key properties: (1) that biome shifts and dispersal events share a common state space over biomes and regions, (2) that any discrete region may contain a number of different biomes, (3) that the geographical structure of biomes within and between regions can vary over time, and (4) that lineages adapted to different biomes and located in different regions will experience different dispersal rates between regions and different shift rates into new biomes
Summary
Biomes are ecologically and climatically distinct species assemblages that vary in size, shape, and continuity across geographical regions, in large part due to regional differences in temperature, precipitation, seasonality, altitude, soil types, and continentality (Whittaker 1970; Wolfe 1985; Olson et al 2001; Mucina 2019). Ancestral regions and biome affinities are often estimated independently of one another, and relationships between regions and biomes are compared post hoc (e.g., Crisp et al 2009; Weeks et al 2014). Such studies yield important evolutionary insights, the estimates themselves do not account for how lineages might move between regions or adapt to newly encountered biomes given the temporally variable spatial configuration of biomes across regions. How a biome is geographically distributed should influence how a lineage might disperse into a new region or shift into a new biome, an effect Donoghue and Edwards (2014) termed geographical opportunity. Cardillo et al (2017) carried out such an analysis in studying the biogeography of the Australian plant clade, Hakea (Protaceae), using method features developed by Matzke (2014), where total regional area and shared perimeter lengths tuned dispersal rates between bioregions
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
