Densities.xlsx

Predator SurveysWe conducted predator surveys in spring and summer 2018 at 33 rocky intertidal field sites from Orange County in southern California to Lincoln County in central Oregon (Fig. 1, Appendix S1 Table S1.1). The lower latitude sites represented the coastline at or nearing the southern range boundaries of the focal taxa (Frontana-Uribe et al., 2008; Marko et al., 2003; Morris et al., 1980). We surveyed five transects per site that extended from the high to low intertidal zones and spanned mussel beds (2-50 meters in length depending on site topography). We surveyed 0.5 x 0.5-meter quadrats placed at 0.5 to 2 meter intervals along each transect. Quadrat intervals were closer (every 0.5 m) near the lower limits of the mussel beds for finer spatial resolution where predators tended to be more abundant. We avoided surveying large tidepools (&gt;~50 cm wide or &gt;~20cm deep), deep cracks (&gt; ~20 cm deep) and highly rugose surfaces.In each quadrat, we counted and measured individuals of all accessible predatory sea star and snail species, including <i>Pisaster ochraceus, Leptasterias spp., Acanthinucella spp., Nucella canaliculata, N. emarginata, N. lamellosa, N. ostrina, </i>and <i>Paciocinebrina circumtexta.</i><i> </i>Size data were not collected at the two southernmost sites, Crystal Cove and Shaw’s Cove (CRCO and SHCO). Whelk size was maximum shell length (mm) from apex to the siphonal notch. Sea star size was length (mm) from the center of the central disk to the straightest arm tip (<i>Leptasterias, P. ochraceus </i>in California) or from madreporite to the opposite arm tip (<i>P. ochraceus </i>in Oregon, keeping with legacy practice). See Supporting Information Appendix S2 Fig. S2.1 and Appendix S3 section S3.2 <i>Conversion for P. ochraceus Arm Length Measurements </i>for more information.Species groupingsSince some predator taxa were rare or only occurred at some sites (<i>Acanthinucella spp., N. lamellosa, </i><i>Paciocinebrina circumtexta</i><i>)</i>, we focused only on the predator taxa that were common, including <i>Pisaster ochraceus, Leptasterias spp.</i>, <i>N. canaliculata, N. ostrina and N. emarginata</i>. Identification of <i>P. ochraceus </i>and the channeled whelk <i>N. canaliculata</i> in the field is straightforward. Field identification for species in the <i>Leptasterias </i>spp. complex (Flowers &amp; Foltz, 2001; Foltz et al., 2008; Melroy &amp; Cohen, 2021) is not possible, however. Similarly, the visually indistinguishable congeners <i>N. emarginata </i>and <i>N. ostrina</i> overlap from San Francisco Bay to Point Conception (Fig.1, Marko et al., 2003). Hence, we grouped these taxa as <i>Leptasterias </i>spp<i>. </i>and emarginate whelks, respectively.Densities of PredatorsWe calculated densities of species at the quadrat, transect, and site level only from quadrats in the ‘potential habitable zone’ of each species so that we excluded quadrats that were not within their vertical range at a given latitude (see Appendix in manuscript titled <i>Potential Habitable Zone</i> for more information). We analyzed density with latitude using a generalized linear model (<i>glm()</i> in <i>stats package </i>v4.0.0) in R (R Core Team, 2020) and specified a Poisson distribution. We tested the main and interactive effects of latitude and species on counts at each site, using the log₁₀ of total area surveyed as an offset. We analyzed the effects of species, region, and site nested within region on counts using a generalized linear mixed model (<i>glmer</i> in <i>lme4 R </i>package v 1.1-34 (Bates et al., 2015), and again used a Poisson distribution and the log₁₀ of total area surveyed as an offset. Southern California sites could not be included in this model since there were zero individuals of all species except <i>N. ostrina </i>and <i>emarginata</i>.