No abstract

Linear regression of chlorophyll concentration on total phosphorus concentration for phosphorus- limited Lake Burragorang, N.S.W., yields regression coefficients within the range reported for individual lakes in the Northern Hemisphere. Some variation in slope of published regressions is attributable to the choice of different regression subvariables (e.g. annual mean or annual maximum). The extent of this variation is quantified. Data from Lake Burragorang and other sites indicate that chlorophyll-phosphorus relationships in the Southern Hemisphere are concordant with those in the north if turbid waters are excluded from consideration. This is obviously significant in Australia, with so many turbid waters. The notion of 'growing season', as applied to Northern Hemisphere studies, is inappropriate for the warm temperate conditions of Lake Burragorang, and it was necessary instead to use the annual maximum chlorophyll concentration. Prediction of annual maximum chlorophyll concentration is of particular significance to water-quality management. Despite highly significant regressions, 95% confidence intervals and 95% prediction limits are wide, so that prediction of chlorophyll concentration from single values of total phosphorus, using double-In regressions, gives a wide arithmetic range. Use of annual mean total phosphorus concentration as the predictor variable limits the forecasting ability of the Lake Burragorang regressions but facilitates future coupling with a phosphorus loading model. This would assist in the assessment of projected management plans and the formulation of protective loading criteria.

The chemical and biological components of existing water quality models are in general described by simple first-order rate equations in which the production and uptake coefficients are fixed functions of the other variables of the model. Thus although, for example, in a photosynthesis model the specific algal growth rate may be a function of light intensity, nutrient concentration and temperature, the form of this dependence on these variables is assumed to be fixed. In this paper, the effect on the performance of a water quality model of removing this assumption for the specific case of photosynthesis is examined. An existing coupled motion-photosynthesis model has been installed in the water quality model DYRESM-WQ and the result compared with the original model, which assumes a fixed functional dependence. The resulting model has been applied to two reservoirs, without recalibration. The result shows that the removal of the assumption of a fixed functional form for the photosynthetic growth rate may have a significant effect on the magnitude and timing of predicted cyanobacterial blooms in lakes and reservoirs, which would have important implications for reservoir and lake management. In addition, the result suggests that, in general, the validity of the assumption of fixed functional form for the rate coefficients in water quality models is not assured.