Bioassessment without reference sites: use of environmental filters to predict natural assemblages of river macroinvertebrates

Sets of reference sites often are used to create a benchmark for biological assessment of river condition. Unfortunately, human modification of river systems is now so widespread and complex that this practice can result in comparisons against a reference condition that embodies an appreciable and poorly defined degree of human impact. Our paper describes an alternative means of deriving reference conditions that does not require the use of reference sites. Instead, the tolerances and preferences of each taxon in the regional taxon pool are established for various environmental factors. These factors can be regarded as filters that exclude members of the regional pool from sites where environmental conditions are incompatible with their tolerances and preferences. The suite of taxa that can occupy an assessment site under natural conditions can be estimated if the natural values of the environmental filters at that site can be determined. A comparison between this suite of potential taxa and the taxa actually collected at an assessment site provides an indication of human impact on taxonomic richness. This approach is illustrated for aquatic macroinvertebrate families in rivers in northeastern New South Wales, Australia, and the environmental filters of annual water temperature range, flow regime, and river bed type. An assessment based on this approach showed substantial correlation with several independent measures of human influence, in contrast to very little correlation for a previous macroinvertebrate bioassessment based on reference sites. The performance of the filters approach probably could be improved with further development of the underlying techniques and data.