Implications of spatial variability of fish assemblages for monitoring of Australia’s tropical estuaries

1. Although often seen as ideal targets for monitoring environmental change because of their high public profile, fish assemblages rarely appear as indicators in monitoring and reporting. 2. Published data were evaluated to develop a simple approach to quantifying the temporal consistency in assemblage structure suitable for routine monitoring and assessment. Data were examined as catch per unit effort (CPUE) and probability of encounter (PoE), and compared using three approaches with the potential to produce simple indices quantifying the patterns of similarity within an estuary over time: species richness, Bray-Curtis Similarities and a new approach, DeltaPoE. Indices derived from published data were then tested against a time series of data from two estuarine lakes with a history of fish kills. 3. Multidimensional scaling based on PoE emphasizes the temporal consistency of fish assemblages within estuaries at least as well as one based on mean CPUE while providing operational advantages. Similarities based on PoE were more sensitive to change from ‘natural’ assemblage structure than the simpler indices and showed comparable results with Similarities from (log) CPUE data. The one drawback to Similarities is that their complex statistical formulation often makes them less effective vehicles for reporting and communication. Where this is the case the conceptual simplicity of DeltaPoE and its performance relative to Similarities suggests it is a good candidate from which to develop monitoring indices suitable for routine reporting. 4. The performance of each of the indices were considered against the known fish kills. Species richness tracked the observed changes, an expected outcome in this case because a major impact directly removed species. Both indices based on Similarities also tracked the changes faithfully. Of the two, Similarities based on PoE seemed to react more strongly to assemblage changes