Environmental variability and the ecological effects of spawning Pacific salmon on stream biofilm

1. Variation in resource subsidies can create or reinforce heterogeneity in recipient ecosystems. Related activities of organisms delivering resource subsidies, such as ecosystem engineering by Pacific salmon spawners (Oncorhynchus spp.), also alter heterogeneity. We studied whether heterogeneity in stream environmental conditions and spawner abundances were reflected in the net ecological effects of salmon (i.e. enrichment by resource subsidies and disturbance by ecosystem engineering) on benthic biofilm. 2. We sampled seven Southeast Alaska streams over 3 years, both before and during the salmon run. In each stream and year, stream environmental characteristics and their influence on responses of benthic biofilm [mean and coefficient of variation of chlorophyll a (chl a), ash‐free dry mass (AFDM) and autotrophic index (AFDM:chl a)] to spawners were assessed. 3. Streams and periods before and during the salmon run were distinct based on their environmental characteristics. The responses of most biofilm metrics to spawners were stream‐ and year‐specific, suggesting that the ecological effect of spawners ranged from net enrichment to net disturbance depending on the stream or year studied. The environmental context, especially temperature, large wood, and sediment size, explained >50% of biofilm variability during the run, but <30% over the entire study, suggesting that salmon can alter environmental constraints. 4. Precision of biofilm estimates improved by increasing either the number of streams or the number of years sampled (i.e. spatial or temporal replication). However, combining data from different North Pacific Rim ecoregions inflated the confidence interval as compared with a single ecoregion, indicating the importance of regional environmental contexts for net salmon effects. 5. Our results suggest that biofilm responses to salmon can vary greatly, even within a single ecoregion, and that environmental conditions can modify net salmon effects. Consequently, generalisations about biofilm responses across the native range of salmon may be challenging.