Predictions from AUSRIVAS were used to compare macroinvertebrate data from samples collected over several years from two nearby river sites in the Hunter Valley Region in NSW, Australia. One of the sites was in agricultural land and the other one was in a National Park. Both sites had good water quality and a diverse range of macroinvertebrate habitats.
The AUSRIVAS output most commonly used for assessing river condition is the Observed/Expected (O/E50) value. Only relatively common taxa, ie those that have greater than 50% probability of occurrence are included in the calculation of O/E50 values. T-tests used to test differences between the mean O/E50 values for the individual samples showed no significant differences between the two sites. When similar analyses were performed on outputs calculated using rare taxa, significant differences were found between the two sites for the riffle habitat but not for the edge habitat.
Including the rare taxa in calculating AUSRIVAS outputs and utilising multiple-year outputs for assessing site condition may allow detection of impacts not detected by examining the commonly used outputs.