Spatiotemporal complexity in stream food web responses to salmon subsidies

Nutrient mitigation to enhance fish production in stream ecosystems assumes strong bottom-up effects, yet many stream salmonids obtain food resources from freshwater, marine, and terrestrial sources. Subsidized trout may also influence lower trophic levels through top-down control, but the effects may be lagged, and detection of those effects may require multi-year experiments. From summer 2008 through 2011 we conducted an experiment across nine tributaries of the N. Fork Boise River, Idaho, consisting of 500-m stream reaches treated with salmon carcasses (n=3), salmon analog (pasteurized and pelletized salmon tissue) (n=3), and un-treated reference reaches (n=3). We observed 2-8 fold increases in streambed biofilms in the 2-6 weeks following additions of either salmon subsidy treatment. Adding salmon carcasses caused a 1.5 fold increase in biomass of aquatic insect larvae (especially Chironomidae midges), but only over the first year of the experiment. Analysis of diets of resident trout indicated direct consumption of salmon carcass and analog contributed to 17% and 6% of annual production, respectively. In streams treated with carcasses, 7% of annual trout production was fueled by consumption of maggots that had colonized carcasses exposed to air or transported to riparian areas by animals. Both treatments caused 2-3 fold increases in annual production of resident trout. However, this did not translate into differences in trout population biomass or density, perhaps owing to shifts in fish size distribution or emigration of trout from treated reaches. Over the four years of the treatments, responses across trophic levels suggest that by year 2, trout in treated reaches cropped any additional benthic invertebrate biomass, reducing them further in year 3. Our findings indicate that nutrient mitigation programs must incorporate stream-riparian linkages into existing frameworks to better account for the multiple pathways that salmon subsidies can directly and indirectly influence trout production. Our results further indicate that bottom-up food web patterns are not predictable through time, as opposing top-down forces may mask these effects