As human impacts in estuaries are often pervasive (estuary-wide) and/or pre-existing, the identification of suitable reference points, from which to assess the extent of impacts, is problematic. One solution is to compare potentially degraded estuaries with estuaries deemed to be largely unmodified by human activities. However, there is a perception that individual estuaries are too spatially and temporally dynamic to allow valid comparisons to be made using such an approach. We tested this idea for a commonly used indicator, benthic macroinvertebrates, using a factorial design incorporating both temporal and spatial scales between and within three adjacent meso-tidal river estuaries in northern Tasmania. Variation in macroinvertebrate assemblage structure was analysed using permutational multivariate analysis of variance. Most variance occurred within estuaries (68–82% variance) relative to spatial differences between estuaries (24–14%) corresponding with a strong upstream estuarine gradient and small-scale spatial patchiness. Seasonal variation accounted for 9–4% of total variance indicating that temporal differences were relatively insignificant when contrasted against spatial variability within and between estuaries. We suggest that with sufficient spatial replication at the within estuary-scale, entire estuaries may act as whole reference systems, allowing studies to examine potential impacts within estuaries with spatially diffuse, pre-existing human impacts.