Introduction
The old-growth cedar-hemlock forests of the Pacific Northwest of North America characterize one of the most diverse temperate rainforests in the world (Newmaster et al. 2003). This ecosystem includes disjunct coastal and inland temperate rainforest (ITR) elements, with the latter located in the Northern Rocky Mountains and separated from the larger coastal rainforest located in the Cascades and coastal ranges by approximately 200 km of xeric habitat. The entire Pacific Northwest region has been widely impacted by Pleistocene glacial/interglacial cycles (Waitt & Thorson 1983), with flora and fauna being strongly impacted by these climatic changes. The ITR has been of particular interest because of the dramatic implications of the alternative hypotheses that have been proposed to explain its history during and after the Pleistocene. One, the Recent Dispersal (RD) hypothesis, posits the recent (<5K years ago; Kya) establishment of the ITR, invoking a post-Pleistocene colonization of the inland areas from coastal populations, and implying that the ITR is a recent propagule of the coastal forest with little evolutionary novelty. The second hypothesis posits that the ITR represents an ancient disjunction between the inland and coastal forests (Brunsfeld et al. 2001) that occurred pre-Pleistocene (>2.5 million years ago; Mya). In this Ancient Vicariance (AV) hypothesis, while the onset of the glaciers caused massive contraction of the ITR, inland refugia persisted during cold periods and subsequently expanded in-situ post-Pleistocene. The AV hypothesis predicts that allopatry may have led to speciation of some taxa in the ITR, and thus, that the inland region harbors a unique endemic flora and fauna. These two hypotheses broadly encapsulate the proposed modes of the formation of the disjunction. They are critical to understanding general biogeographic processes associated with the ecosystem and the region, and they also have broad implications for conservation and management of this diversity hotspot.
The range of the PNW temperate rainforest is defined by the distributions of the two late-successional dominant species, Tsuga heterophylla Raf. (Sarg.) (western hemlock) and Thuja plicataDonn ex. D. Don. (western redcedar). Pollen records from the central and southern ITR suggest these species have only been present inland for <3 Ky (Mehringer 1996, Rosenburg et al. 2003, Chaseet al. 2008, Gavin et al. 2009), and represent the primary data point for the RD hypothesis. Suitable inland habitat for western hemlock is not completely occupied, suggesting the species range is still expanding (Gavin & Hu 2006). Similarly, Rosenburg et al. (2003) found no record of this species’ pollen in southeastern BC prior to ~ 3500 ya. Most pollen records concur that western hemlock pre-dates evidence of western redcedar (Mehringer 1996, Whitlock 1992). This coincides with microsatellite molecular evidence for western redcedar samples across the disjunction (O’Connell et al. 2008), which supports one southern coastal refugium throughout the Pleistocene, with no evidence for ancient, disjunct inland refugia, nor for northern coastal refugia (e.g., the Haida Gwaii archipelago). O’Connell et al. (2008) further suggested that, given the lack of hierarchical structure in these three clusters, the divergence between them has been recent and rapid, which is congruent with post-glacial recolonization of the northern coast and ITR (O’Connell et al. 2008). While this inference was based on microsatellite loci, recent advances with reduced representation sequencing (Peterson et al. 2012, Andrews et al. 2016) provide enhanced power to test phylogeographic hypotheses regarding formation of disjunct populations (Carstens et al.2012, Garrick et al. 2015). Indeed, a recent study used clustering analyses for population assignment based on a genome-wide panel of SNPs in western redcedar to support the presence of an ITR refugium (Fernandez et al. 2021). However, they did not attempt to estimate population parameters, such as the timing of demographic events through model comparisons, which would enable the demographic hypotheses to be evaluated.
Though much of the available evidence supports the recent, post-glacial colonization of the ITR by these dominant tree species, a number of phylogeographic investigations have been conducted to evaluate the impact of the disjunction on other species in the PNW temperate rainforest (e.g., Soltis et al. 1997, Brunsfeld et al.2001). To date, eleven other species complexes with disjunct ranges in the PNW have been investigated in a phylogeographic framework including five amphibians (Nielson et al . 2001, Carstens et al.2004, Wilke & Duncan 2004, Steele et al. 2005, Metzger et al. 2015), one mammal (Carstens et al. 2005), two plants (Brunsfeld et al. 2006; Carstens et al. 2013, Ruffleyet al. 2018), three mollusks (Smith et al. 2017, Smithet al. 2019, Rankin et al. 2019), and an arthropod (Espíndola et al. 2016). These species span the tree of life and, based on analyses of their genetic variation, they also span the possible phylogeographic histories for the PNW temperate rainforest. Some species, such as the tailed frogs (Ascaphus ; Neilson et al. 2001), salamanders (Dicamptodon, Plethodon ; Carstens et al. 2004; Steele et al. 2005), and jumping slugs (Hemphilia ; Rankin et al., 2019) show clear evidence of an ancient divergence between the ITR and coastal populations, indicating pre-Pleistocene divergence. Conversely, other species, such as Salix melanopsis(Carstens et al. 2013), Microtus richardsoni (Carstenset al. 2005), and taildropper slugs (Prophysaon andersoni ; Smith 2018) show evidence of post-glacial recolonization of the inland from the coast. Other phylogeographic models, such as pre-Pleistocene divergence with migration, have also been supported with genomic evidence (Alnus rubra; Ruffley et al. 2018). These results suggest that some ITR endemics might have been present before the ecosystem dominant species were established if these ecosystem dominants colonized the ITR more recently, as supported by pollen records and early molecular studies. Inferring the phylogeographic history of the ecosystem dominant species that establish the boundaries of the PNW temperate rainforest will provide a critical insight for the availability of suitable habitat for refugial populations in the ITR, and will be a central contribution towards the understanding of the biological history of the area.
The hypothesis that the ITR has an ancient (pre-Pleistocene) divergence from the coastal rainforest and has persisted throughout glacial cycles in refugia in the interior Northwest is compelling because its presence would support the habitat requirements of other species that show evidence of ancient vicariance. Additionally, paleontologists have questioned the plausibility of the old-growth ITR becoming so established in less than 3500 years (Mehringer 1996). However, the persistence of the ITR throughout the Pleistocene has received no support from the pollen record (e.g., Mehringer 1996, Chase et al. 2008, Gavin et al. 2009). Whether or not the ITR persisted throughout the Pleistocene also has other implications for how the PNW disjunct community as a whole has adapted to the dramatic climatic changes, either in concert or individualistically (Davis 1981, Habeck 1987, Sullivan et al. 2000, Flessa and Jackson 2005). Common insight from paleoecology suggests that modern communities of PNW forest have assembled over a long history of individual responses to climate change (David 1981, Flessa and Jackson 2005), and the hypothesis of a recently assembled, rapidly diverse ITR poses a challenge to this insight.
This question is not novel to the PNW temperate rainforest, as the idea of the species responding individualistically, rather than in-concert, in response to climatic changes is an ecological notion dating back to the early 20th century (Gleason 1926), and has been demonstrated empirically (Burbrink et al. 2016). However, Kirkwood (1922), one of the first to characterize the ecology of species in the northern Rocky Mountains in general, emphasized how the understanding of the ITR would be dramatically improved when “the individualities of the constituent species were understood”. Alternatively, there is evidence in other communities that species do, in fact, respond to climatic changes in concert (Chen et al.2014, Gehera et al. 2017). For plant communities specifically, the idea of community-wide concerted response to climatic change can be traced back to early 20th century plant ecologist Clements (1918), and his idea that communities are “super-organisms” whose interactions are interwoven and dependent on one another. Regardless of the organism or ecosystem, researchers have long been fascinated with the question of whether or not species in the same environment respond asynchronously or synchronously to climatic changes (Sullivan et al. 2000, Carstens et al. 2005, Hickersonet al. 2006).
In this study, we test predictions about the phylogeographic history of these two species, specifically with respect to whether or not they harbor cryptic diversity across the disjunction; that is, show evidence of pre-Pleistocene divergence and no subsequent migration. These predictions serve as a test of the predictive framework that was originally developed by Espíndola et al. (2016) and recently updated with life history traits by Sullivan et al. (2019). We then validate these predictions, and ultimately test whether the ITR persisted throughout the Pleistocene (Brunsfeld et al . 2001) by generating genomic data for individuals from these species throughout their ranges. For this, we rely on coalescent simulations, the joint site frequency spectrum, and machine learning inference procedures to develop and test our phylogeographic hypotheses. We also validate the power of predictive phylogeography in detecting the presence and absence of cryptic diversity. Finally, we evaluate the potential role of genomic data in uncovering the history of the past and explore how our inferences can be influenced by various datatypes and perspectives in genomics and paleontology.