SEARCH

SEARCH BY CITATION

Abstract

Benthic macroinvertebrates were quantitatively sampled from thirty sites along two river systems in southwestern Australia, and patterns in community structure related to physical and chemical parameters. Classification and ordination showed a major separation between upland and lowland sites, irrespective of river system. The change in benthic community structure reflected the rapid transition in geomorphology, stream hydraulics, and water chemistry as upland forested streams exit the Darling Escarpment to give rise to open rivers, disturbed by agriculture and urban development.

One upland site was clearly influenced by a storage reservoir immediately upstream and consistently grouped with lowland sites; evidence of recovery was apparent at sites downstream. The remaining upland sites were separated on the basis of catchment; this was most likely related to stream flow permanence than any inherent catchment difference. A seasonal pattern was also detected for upland sites. Samples taken in summer or autumn were distinct from those taken in winter or spring.

In constrast, lowland sites could not be separated into distinct groups on the basis of season or drainage basin. The presence of cosmopolitan and tolerant species with a high likelihood of dispersal, together with the homogeneous nature of the sites, may account for the high degree of similarity among benthic communities of sites along the lowland rivers.

Much of the spatial and temporal variation in benthic community structure was explained by physical characteristics of the sites. Prediction of community type using chemical data alone was poor, however, this success could be improved by combining physical and chemical data, particularly for upland sites. The poor predictive success using chemical data was likely the result of the abrupt changes in the physical nature of the streams, and the absence of large spatial differences in water quality.

The successful predictive relatioship betweenm benthic community structure and physical data will enable water management authorities to detect subsequent changes in water quality in these two river systems. The predictive power of the model could be assessed in adjacent river systems for which the patterns in benthic community structure are as yet unknown.