INTRODUCTION

Speciation is a key evolutionary process that results from the independent evolution and adaptation of populations and ultimately acts as a major driver responsible for the generation of species-level biodiversity (Kopp, 2010; Schluter & Pennell, 2017). Species richness is unevenly distributed across the Tree of Life, and its current patterns of distribution result from biotic and abiotic processes that operate over space and time (Benton, 2009; Simpson, 1953; Vargas & Zardoya, 2014). Evidence for the mechanisms that promote population differentiation and speciation are currently better understood in terrestrial than in marine environments (Butlin et al., 2012; Coyne & Orr, 2004; Nosil, 2012), where the lack of obvious physical barriers would suggest that neutral processes of panmixia, or isolation-by-distance, will prevail, especially in highly mobile species (Moura et al., 2013). However, counterintuitive evidence of fine-scale differentiation among populations and species in a number of marine taxa has been described as the “marine species paradox” (Bierne, Bonhomme, & David, 2003; Palumbi, 1994). Thus, there is a need for explicit evaluations of the role of selective processes in driving patterns of differentiation in marine systems.
In species complexes that are geographically widespread, the gradual evolution of reproductive isolation in allopatry can make species delimitation challenging, especially in young radiations (Carstens, Pelletier, Reid, & Satler, 2013; Cutter, 2013). Many allospecies first tend to differ from their close relatives at traits subjected to sexual and other forms of social selection (Price, 2008; Seddon et al., 2013). When this occurs, our ability to delimit species may be further hampered by morphological stasis, especially when changes in ecological niche in allopatry are minimal (Fišer, Robinson, & Malard, 2018). In cases of morphological stasis and limited behavioural information, genomic data can provide informed hypotheses on species limits of allopatric taxa and can be conclusive in parapatric or sympatric taxa. Despite the extent of disagreement about how genomic data should be applied to species delimitation (Leaché, Zhu, Rannala, & Yang, 2018; Sukumaran & Knowles, 2017), agreement exists that genomic data can provide additional perspective on species limits when used together with other data types such as phenotypic and ecological information.
Seabirds of the order Procellariiformes present some of the most extreme examples of the marine speciation paradox. Procellariiformes are highly mobile pelagic seabirds with a high dispersal ability and perform some of the longest animal migrations on Earth (covering more than 120,000 km a year) (Shaffer et al., 2006; González-Solís, Croxall, Oro, & Ruiz, 2007; Weimerskirch, Delord, Guitteaud, Phillips, & Pinet, 2015). However, Procellariiformes also show high philopatry to their breeding grounds (Coulson, 2002), which is expected to limit gene flow and therefore reinforce genetic differentiation (Friesen, Burg, & McCoy, 2007).
Shearwaters are a monophyletic group in the family Procellariidae, and they offer an excellent case study for examining the mechanisms of population differentiation and speciation in marine environments. First, shearwaters are globally distributed and breed mostly in allopatry. Second, the current taxonomy recognises three genera and 30 species with a well-resolved phylogeny showing clear periods of rapid diversification (Ferrer-Obiol et al. under review). Third, the three recognised genera exhibit different ecologies and degrees of species richness. Fourth, their high mobility makes them an ideal model to evaluate the roles of founder events and vicariance using biogeographic analyses. Fifth, abiotic and biotic factors are known to promote speciation in the shearwaters and related Procellariiformes; for instance, paleoceanographic changes such as the Pleistocene climatic oscillations can act as historical drivers of speciation (Gómez-Díaz, González-Solís, Peinado, & Page, 2006; Silva et al., 2015) and intrinsic biotic factors such as different foraging strategies and allochrony can also promote speciation (Friesen, Smith, et al., 2007; Lombal, Wenner, Lavers, & Austin, 2018; Rayner et al., 2011). Sixth, species limits are controversial, mostly due to high morphological stasis (Austin, Bretagnolle, & Pasquet, 2004; Austin, 1996); indeed, only a few phenotypic traits, such as vocalisation characteristics, slight plumage colour differences and in particular, body size, may differ between closely related species. A comprehensive study using genomic data will assist in resolving species delimitation within the context of the factors that promote diversification and speciation.
To accurately relate historical environmental and oceanographic changes to the timing of speciation events, it is necessary to estimate accurate divergence times. Analyses based on concatenation can lead to biases in branch lengths and misleading age estimates, particularly at recent timescales (Angelis & Dos Reis, 2015; McCormack, Heled, Delaney, Peterson, & Knowles, 2011; Mendes & Hahn, 2016). For such events, the multispecies coalescent model (MSC) offers a more accurate solution by incorporating the effects of incomplete lineage sorting (ILS), which is likely the most common source of phylogenetic incongruence in rapid diversification events (Edwards et al., 2016; Maddison, 1997; Suh, Smeds, & Ellegren, 2015).
The reconstruction of ancestral ranges and evaluation of alternative biogeographic models are critical to our understanding of shearwater diversification throughout the world in light of environmental and oceanographic events. Of particular interest is the importance of founder events during the evolution of shearwaters. The foundation of colonies is believed to be a rare event in most seabird species despite their great potential for long-range dispersal (Milot, Weimerskirch, & Bernatchez, 2008). However, in several shearwater species, contemporary colony foundation events have been reported (Munilla, Genovart, Paiva, & Velando, 2016; Storey & Lien, 1985). Understanding the relative importance of founder and vicariant events during the evolution of shearwaters can have important implications for the conservation of these endangered pelagic seabirds.
Here, we use paired-end double-digest restriction site-associated DNA sequencing (PE-ddRAD-Seq) for almost all extant shearwater taxa to explore the drivers of diversification and speciation in this group of pelagic seabirds. Specifically, we produce the first time-calibrated shearwater species tree using the MSC to account for the high levels of ILS affecting the shearwater phylogeny. We then infer the biogeographic history of the group by estimating ancestral ranges and evaluating the roles of founder events, vicariance and surface ocean currents in driving their diversification. Furthermore, we explore the ecological forces responsible for the variability in a key phenotypic trait, body size. Finally, we assess the validity of the current taxonomy of the group by analysing genomic patterns of recent shared ancestry and differentiation among shearwater taxa.