δ15NBase isoscapes in the NP
Our isoscape shows that the eastern Bering Sea continental shelf has higher δ 15NBase (δ 15NBase > 7.0‰; Fig. 1) than any other potential habitat range of chum salmon. This highδ 15NBase is likely to be a result of coupled partial nitrification–denitrification in this region (Brown et al . 2015). High rates of sedimentary denitrification fueled by active nitrification preferentially remove 14N from the system, thereby enriching 15N in the remaining nitrate (Granger et al . 2011). Spatial patterns ofδ 15NBase in the pelagic subarctic NP are largely determined by nitrate utilization, which is the ratio of nitrate assimilation by phytoplankton to nitrate supply in the euphotic layer, and is inversely correlated with surface-water nitrate concentrations because of isotope fractionation during nitrate assimilation by phytoplankton (Fig. S3). In the western subarctic NP, which is a nutrient-rich, low-chlorophyll region, nitrate is not depleted throughout the year and phytoplankton growth is instead limited by the availability of other micronutrients, such as iron (Nishioka & Obata 2017). This results in less fractionation associated with nitrate utilization, which in turn accounts for the lowδ 15NBase in the western subarctic NP (Yoshikawa et al . 2018). Nitrate concentrations decrease with distance from the nutrient-rich, low-chlorophyll region and nitrate δ 15N increases with nitrate depletion (Yoshikawa et al . 2018), which may account for the increased δ 15NBase around the eastern coast of Japan and the Gulf of Alaska. Unfortunately, the isoscape compiled in this study suffers from a paucity of data in the Okhotsk Sea and the western Bering Sea Shelf. Although previous studies have suggested that nitrate δ 15N in the Okhotsk Sea is similar to that of the eastern coast of Japan and not as high as that of the Bering Sea Shelf (Yoshikawa et al . 2006), it remains possible that δ 15NBase in the western Bering Sea Shelf resembles that of the eastern Bering Sea Shelf. Thus, we cannot discount the possibility that the salmon analyzed in this study migrated to the western Bering Sea Shelf, although this would not affect the substance of our conclusions.
Our results also suggest thatδ 15NBase can be used as a direct indicator of the origin of marine organisms as suggested by previous studies (Hetherington et al . 2017; McMahon & Newsome 2019). This is supported byδ 15NBase values in the first vertebral section (Section 1) of two salmon (OK2: 4.4‰ and OK8: 3.6‰) similar to that of the isoscape in coastal Japan and the Okhotsk Sea (ca. 4‰; Fig. 1). A previous study also confirmed that Japanese chum salmon initially move into the Okhotsk Sea just after their seaward migration (Urawa et al . 2009). Large seasonal variations in the isotopic composition of primary producers (Bronk et al . 1994; Rolff 2000; Hannides et al . 2009) often complicate estimations of the isotopic baseline in the ocean. However, copepods effectively provide time-averaged δ 15NBase values in a given region because they have relatively long lifespans (from 1 month to 2 years) (Paffenhöfer 1993; Tsuda et al . 2001), low dispersal ability, and low intra-species TP variation (Table S3). In our study, copepod samples from the pelagic NP were mostly composed of large Neocalanus species, which have a lifespan of 1–2 years (Tsuda et al . 2001). In coastal areas, we sampled copepods which have a shorter lifespan across different seasons to take any seasonal variation into account, and to capture the annual meanδ 15NBase of each region.