Differentiating drainage based multiple refugia from nunatak refugia
Although our results suggest multiple ancestral populations and high levels of genetic diversity in some high elevation populations, which could fit the predictions of glacial survival within the ice sheet (nunatak survival hypothesis), the pattern of several isolated glacial refugia of N. ingens complex is fundamentally different from nunataks. Conventionally, nunatak refugia refer to the small, ice-free geographical areas within the glacial ice sheets. In such habitats, populations would be restricted to small size and exposed to extremely challenging environmental conditions (cold and windy, and thus xeric), before expanding in postglacial periods (Dahl 1987; Holderegger & Thiel‐Egenter 2009). In the case of the N. ingens complex, high elevation sites with elevated genetic diversity reflect recent admixture among ancestral population clusters (conStruct results, seeTable 1 ). Ancestral populations are not associated with ice-free areas in the north and central part of the Sierra Nevada (Figure 1 ). Similar contact zones are found in a large set of alpine plants from the European Alps (Thiel‐Egenter et al., 2011), and a recent synthesis of alpine studies suggests this might be a common phenomenon across long linear mountain ranges (Wallis et al., 2016). Many alpine or sub-alpine species that have patterns consistent with multiple glacial refugia are certainly not related to persistence in nunataks (Kubow, Robinson, Shama, & Jokela, 2010; Homburg et al., 2013), suggesting that the deep divergence of population structure or multiple ancestral populations might not be sufficient to be directly linked to nunatak refugia.
The other common character of populations recolonized from the nunataks is the vestige of a bottleneck during glacial periods (Segarra‐Moragues, Palop‐Esteban, González‐Candelas, & Catalán, 2007; Kosiński et al., 2019). Our data suggest the opposite has happened, as the stairway plots show drastic population decline initiating at the last glacial maximum and accelerating into the present (Figure 7 ). Similarly, estimates of ancestral population size (reference population, Nref) in δaδi were several fold larger than the descendant populations (Table S2 ). Although other studies report glacial bottlenecks or postglacial population increases (Bidegaray-Batista et al., 2016; Sim, Hall, Jex, Hegel, & Coltman, 2016; Starcová, Vohralík, Kryštufek, Bolfíková, & Hulva, 2016; Weng, Yang, & Yeh, 2016; Huang et al., 2017), it seems logical that alpine species would undergo founder events as they recolonize high elevation areas following glacier retreat (Schoville and Roderick, 2009).
We acknowledge that the continuous spatial structure of population could possibly lead to a false signature of recent population decline for stairway plot analysis and possibly other sfs-based analyses including δaδi (Battey, Ralph, & Kern, 2020). However, there are several reasons we interpret the larger ancestral population sizes estimated by the stairway plot and δaδi to represent a true demographic decline. First, the census population size of populations within the N. ingenscomplex is mostly very small and populations are isolated from each other by complex topographic terrain (e.g . suitable habitat occurs within the first few hundred meters from the headwaters of streams or small seep areas within glacier cirques). Thus, estimates of a contemporary effective population size numbering in the hundreds or several thousand is realistic in our opinion. Second, all eight sample sites show a consistent trend of population decline, even though they differ in their neighborhood size and geographical position in the geographical cline, which are major factors that bias the stairway plot estimates in continuously structured populations (Battey et al., 2020). Lastly, the cool and wet glacial period preceding the glacial maximum would likely be optimal for a cold-specialized, riparian beetle, and thus explains the larger population size at this time. A few other studies have shown population expansion during glaciation followed by postglacial population declines in alpine and subalpine species (Galbreath, Hafner, & Zamudio, 2009; Huang et al., 2016; Zaman, Hubert, & Schoville, 2019).