Incorporating host-pathogen(s)-environment axes into management and conservation planning is critical to preserving species in a warming climate. However, the role pathogens play in host stress resilience remains largely unexplored in wild animal populations. We experimentally characterized how independent and cumulative stressors (fisheries handling, high water temperature) and natural infections affected the health and longevity of released wild adult sockeye salmon (Oncorhynchus nerka) in British Columbia, Canada. Returning adults were collected before and after entering the Fraser River, yielding marine- and river-collected groups, respectively. Fish were exposed to a mild (seine) or severe (gill net) fishery treatment at collection, and then held in circulating freshwater tanks for up to four weeks at historical (14°C) or projected migration temperatures (18°C). Using weekly nonlethal gill biopsies and high-throughput qPCR, we quantified loads of up to 46 pathogens with host stress and immune gene expression. Marine-captured fish had less severe infections than river-captured fish, a short migration distance (100 km, 5-7 d) that produced profound infection differences. At 14°C, river-collected fish survived 1-2 weeks less than marine-collected fish. All fish held at 18°C died within 4 weeks unless they experienced minimal handling. Gene expression correlated with infections in river-collected fish, while marine-collected fish were more stressor-responsive. Cumulative stressors were detrimental regardless of infections or collection location, likely due to extreme physiological disturbance. Because river-derived infections correlated with single stressor responses, river entry likely decreases stressor resilience of adult salmon by altering both physiological status and pathogen burdens, which redirect host responses toward disease resistance.

Environmental DNA (eDNA) metabarcoding can rapidly characterize the composition and diversity of benthic communities. As such, it has high potential utility for routine environmental assessments of benthic impacts of marine finfish farming. In this study, 126 sediment grab samples from 42 stations were collected along an organic enrichment gradient at six salmon farms in British Columbia, Canada, and benthic biotic community changes were assessed by both eDNA metabarcoding of metazoans and macrofaunal polychaete surveys. The latter was done by analyzing 11,466 individuals using a combination of morpho-taxonomy and DNA barcoding. Study objectives were to: (1) compare biotic signals associated with benthic impacts of salmon farming in the two data types; and (2) identify potential eDNA indicators to facilitate eDNA-based monitoring in Canada. Across both data types, alpha diversity parameters were reduced in sediments near fish cage edge and were negatively correlated with pore-water sulphide concentration. Presence/absence of known indicator taxon Capitella generally agreed well between the two methods despite that they differed in both the volume of sediment sampled and the molecular marker used. In eDNA data, there was a strong negative correlation between Nematoda OTU richness and pore-water sulphide concentration, and multiple approaches were used to identify OTUs related to organic enrichment statuses. We demonstrate that eDNA metabarcoding generates biotic signals that could be leveraged for environmental assessment of benthic impacts of fish farms in multiple ways: both alpha diversity and Nematoda OTU richness could be used to assess the spatial extent of impact, and OTUs related to organic enrichment could be used to develop a local biotic index.