Interactions among stressors may be weak: Implications for management of freshwater macroinvertebrate communities

Aim: Ecological models that do not account for interactions among stressors, if interactions are important, could be inaccurate and lead to inefficient conservation strategies. Conversely, if interactions are not important (i.e., stressors operate largely independently), then actions concentrating on a stressor-by-stressor basis would be warranted. Here, we investigated whether interactions among multiple stressors affected widely used indices of freshwater macroinvertebrate biodiversity, which are sensitive to environmental change at management-relevant scales (i.e., reaches and catchments). Location: State of Victoria, south-eastern Australia. Methods: We used a 7,418-sample dataset for stream macroinvertebrates from 2,165 sites distributed over 237,630 km(2) for 20 years. We calculated the interactive effects on stream macroinvertebrates of stressors operating at different scales, namely vegetation loss at the catchment and reach scales and hydrological change and salinization at the local scale. The importance of interactions among multiple stressors was assessed by comparing the cross-validated predictive performance of models with and without multiple stressor interaction terms. Results: Cross-validated models explained 31%-63% of the variation in the macroinvertebrate responses. The most important stressors were catchment vegetation loss (the proportion of remaining native vegetation cover) and salinity. The inclusion of interaction terms did not increase cross-validated predictive performance, which indicates that there was little evidence that interactions among stressors were important for explaining variation in commonly used freshwater macroinvertebrate condition indices. Main conclusions: Interactions among vegetation, salinity and hydrological change stressors may not always be of importance for determining patterns of stream macroinvertebrate biodiversity, so that such interactions may not necessarily be critical considerations for catchment and reach scale management, at least if based on these or comparable condition indices. The mitigation of the impacts of vegetation loss, salinization and hydrological change stressors one-by-one probably is sufficient to guide conservation activities and might be advantageous if socio-political contexts make it difficult to address interactions among stressors.