The study of natural gradients in nutrient subsidies between ecosystems allows for predictions of how changes in one system can affect biodiversity in another. We performed a large-scale empirical test of the role of Pacific salmon (Oncorhynchus spp.) in structuring riparian plant communities. A comparison of 50 watersheds in the remote Great Bear Rainforest of British Columbia's central coast in Canada shows that salmon influence nutrient loading to plants,shifting plant communities toward nutrient-rich species, which in turn decreases plant diversity.These effects are mediated by interactions between salmon density and the physical characteristics of watersheds. Predicting how salmon affect terrestrial ecosystems is central to conservation plans that aim to better integrate ecosystem values into resource management.

Anadromous Pacific salmon (Oncorhynchus spp.) subsidize terrestrial food webs with their nutrients and carcasses, a process driven largely by selective foraging by bears (Ursus spp.). We quantify wildlife transfer of salmon carcasses to riparian zones on two watersheds in coastal British Columbia and estimate total terrestrial fly production from remnant carcasses. Large-bodied chum salmon (Oncorhynchus keta) were transferred into the forest at a greater rate than were pink salmon (Oncorhynchus gorbuscha) (chum salmon mass = 6089–11 031 kg, 16%–48% of salmon run; pink salmon mass = 2266–2808 kg, 4%–6% of salmon run). Blow flies (genus Calliphora) and other Diptera dominated colonization (>90% of salmon carcasses). Between the two watersheds, 196 and 265 g of Calliphora larvae per metre of spawning length (4 and 7 million larvae for whole watersheds) were generated from salmon carcass transfer. Stable isotope analysis of δ15N and δ13C of spring-emerging adult Calliphora revealed that >80% of individuals had salmon-based signatures. Flies are a dominant consumer and vector of salmon nutrients in terrestrial habitats and supplement the diet of at least 16 vertebrate and 22 invertebrate species. Anticipated further declines of salmon in the North Pacific can be expected to further erode the complex associations coupling marine and terrestrial ecosystems.

Resource subsidies across habitat boundaries can structure recipient communities and food webs. In the northern Pacific region, bears Ursus spp. foraging on anadromous salmon Oncorhynchus spp. provide a key link between marine and terrestrial ecosystems, with salmon density, fish size and watershed size as potential predictors of the magnitude of marine subsidy to terrestrial habitats. We use nitrogen and carbon stable isotopes to provide an assessment of the patterns of marine‐enrichment in riparian plants (11 species, 4 guilds) and litter invertebrates (4 guilds) sampled from 27 watersheds in coastal British Columbia, Canada. Watersheds occurred in three geographical regions (Vancouver Island, mainland midcoast and Haida Gwaii) and varied in size, and in biomass (kg m−1 of spawning length) and species of salmon (chum O. keta, pink O. gorbuscha and coho O. kisutch). δ15N values in all plant species and invertebrate guilds were positively predicted by total salmon biomass (kg m−1) and negatively predicted by watershed size. We observed replicated parallel slopes among plant species and invertebrate guilds across the gradient in salmon biomass, with differences in means hypothesized to be due to plant fractionation and animal trophic position. As such, we derived a watershed δ15N‐index averaged across guilds, and using an information theoretic approach we find that the biomass of chum salmon is a much stronger predictor of the δ15N‐index than either pink or coho salmon, or the sum biomass of all species. The top linear model contained chum biomass and watershed size. Chum salmon biomass independently predicted δ15N‐index variation in all three regions of British Columbia. Chum salmon are larger than pink or coho and provide an energetic reward for bears that facilitates carcass transfer, tissue selective foraging, and nutrient distribution by insect scavengers. Analyses of biodiversity and habitat data across many watersheds moves towards a long‐term goal in fisheries ecology to better integrate ecosystem values in salmon conservation.

Although managers safeguard protected areas for migratory species, little consideration has been given to how migratory species might benefit parks. Additionally, whereas land-sea connections are considered in management of protected areas, most effort has focused on reducing negative "downstream" processes. Here, we offer a proposal to promote positive "upstream" processes by safeguarding the seasonal pulse of marine nutrients imported into freshwater and riparian ecosystems by spawning migrations of Pacific salmon. Currently, high rates of fishing limit this important contribution to species and processes that terrestrial parks were designed to protect. Accordingly, we propose limiting exploitation in areas and periods through which salmon runs bound for terrestrial protected areas can migrate. Best suited for less commercially valuable but relatively abundant and widespread pink and chum salmon (O. gorbuscha and keta), our proposal thus considers ecosystem and societal needs for salmon. We conclude by outlining strategies to overcome socio-economic barriers to implementation.

Bears ( Ursus spp.) and other predators can capture and transport large numbers of Pacific salmon ( Oncorhynchus spp.) to riparian areas beside small coastal streams, a process that may affect site productivity and local plant communities. We used a novel experimental manipulation of salmon carcasses to analyze understory plant uptake of salmon-derived N. A hierarchical before–after, control–impact design was used with the addition of chum salmon ( Oncorhynchus keta (Walbaum in Artedi, 1792)) carcasses to forest sites along 11 streams on the central coast of British Columbia, Canada. Eight months after carcass placement, the foliar %N and δ15N in three herbaceous and one moss species had increased by an average of 14%–60% (%N) and 0.5–3.3‰ (δ15N) at treatment carcass sites versus control sites. Treatment effects for %N were typically greater than for δ15N. Nitrogen isotope signatures at carcass sites were highly variable (δ15N range = 30.1‰) and were mediated by plant species, stream salmon density, carcass mass, and individual plant foliar %N. These results show that understory plants use N from salmon during an important period of plant growth many months after carcasses were deposited in riparian areas. However, they also indicate that habitat variation across spatial scales from individual plants to streams needs to be considered when estimating the contribution of salmon to plant nutrition.