Incorporating species losses and gains into a fish-based index for stream bioassessment increases the detection of anthropogenic disturbances

摘要: The taxonomic completeness index (ratio of observed to expected species; O/E) is widely used in stream bioassessment programs infer ecological impairment. However, its sensitivity detecting anthropogenic disturbances may be reduced by (1) the modelling procedure determine species at a site (2) inability account for assemblage shifts through gains as well losses; and (3) frequent use threshold that only allows assessment absence prevalent species. We version BC biotic (an adaptation Bray-Curtis distance) incorporated alien translocated into component, generated native probabilities using single ensemble models (‘BCA’). Sensitivity analysis, bivariate correlations multiple linear regression analyses were test whether BCA better detected than standard (i.e. without species) O/E50 derived from same models. also tested three additional fish indices currently Ecosystem Health Monitoring Program Southeast Queensland, Australia. Of tested, explained greatest amount variance disturbance variables, followed proportional sample abundance was 18% more sensitive non-reference conditions, 20% responsive an gradient, had twice number significant with variables underlying predictive model. suggest improved performance relative lies ability detect addition alien, translocated, some whose traits allow them persist or thrive degraded inclusion low prevalence taxa mild levels disturbance. Given generation requires no further information already provided traditional multivariate models, we recommend based indices.