Relationship of AUSRIVAS-based macroinvertebrate predictive model outputs to a metal pollution gradient

Abstract

AbstractBiological assessment of river health often involves making comparisons between the biota observed at a site and the biota expected to occur in the absence of damage. Our study aimed to determine whether observed-to-expected (O/E) ratios of macroinvertebrate taxa showed a monotonic (dose–response) relationship and were correlated with pollution levels along a gradient of trace-metal contamination. The study was conducted on the Molonglo River, Australia, which exhibits a clear gradient of trace-metal contamination, resulting from previous mining activities. Macroinvertebrate samples were collected from riffle habitats at 16 test sites and 5 sites representative of reference conditions. The sites’ O/E taxa ratios, generated by a regional predictive model, ranged from 0.21 (severely impaired) to 1.13 (equivalent to reference). Physical and chemical variables measured at each site included Cd, Cu, Pb, and Zn concentrations in the water and sediment. A factor analysis reduced 19 physical and chemical ...