Water quality guidelines are an important tool for managing environmental pressures on freshwater streams, but guidelines are frequently set using conditions from reference sites that are assumed to be unimpacted. Using biological thresholds potentially provides a better foundation for guidelines. However, guidelines based on field observations alone may be compromised by confounding influences. This study used an outdoor stream mesocosm, an artificial substrate (rope), and six salinity concentrations to assess the veracity of a diatom–salinity threshold determined previously in natural temperate South Australian streams. In addition, shaded treatments assessed the synergistic influences of salinity and reduced sunlight. Salinity had the strongest effect on diatoms, influencing both species and functional compositions. Species diversity and richness, and functional diversity had negative correlations with salinity. Shade strongly reduced diatom concentrations and altered species composition, with no observed interaction between salinity and light. Threshold Indicator Taxa ANalysis indicated a salinity threshold of 1610 μS cm−1 for both shade treatments, lower than the upper limit of the range used in current freshwater guidelines. This study recommends a new candidate guideline of 1600 μS cm−1 for regional freshwater streams and suggests that contemporary methods for deriving water quality guidelines may not adequately protect aquatic health.