A model for inferring past conductivity in low salinity waters derived from Murray River (Australia) diatom plankton

Detecting human-induced salinisation in rivers and wetlands of the Murray-Darling Basin has proved problematic. A diatom-based model that permits the estimation of past electrical conductivity (EC) from sedimentary diatom sequences has been developed from Murray River planktonic diatoms. Canonical Correspondence Analysis indicates that EC explains the greatest amount of variance in Murray River planktonic diatoms and that its influence is partially independent of that associated with velocity, turbidity, pH and nutrients. A weighted-averaging based model for inferring past EC was therefore derived from the relationship between diatom composition and EC in Murray River plankton samples. The model works well when comparisons are made between measured and diatom-inferred EC determined by jackknifing based leave-one-out computer resampling (r(jack)(2) = 0.71, root-mean-square-error of prediction = 115 muS cm(-1)). Application of the model will enhance understanding of the nature of pre-European variability in electrical conductivity and permit detection of changes in conductivity through the period of European occupation at key sites. Such reconstructions will provide a firm empirical basis for assessing European impact on aquatic ecosystems and a means by which to assess restoration efforts.