Fish farming in lakes and acceptable total phosphorus loads: Calibrations, simulations and predictions using the LEEDS model in Lake Southern Bullaren, Sweden

Publisher Website
Google Scholar
Abstract
This work presents a dynamic model, Lake Eutrophication, Effect-Dose-Sensitivity (LEEDS), designed to predict how large a fish farm a lake could sustain without incurring eutrophication problems. Since the model included various time-dependent phosphorus cycling processes, the second goal was to estimate the relative importance of those processes. We studied the dimictic and mesotrophic Lake Southern Bullaren, Sweden for several reasons. The lake had a fish farm which produced between 70 and 500 tons of rainbow trout per year from 1980 to 1992 (the permit was for a maximum of 70 tons y−1). Rather severe blooms of Cyanophyceae occurred twice during the last six years of production. Based on comprehensive water chemistry and biology data sets, the LEEDS model was calibrated at 14 specific points in order to yield reliable predictions for this lake. The dominating phosphorus loads were the fluxes from tributaries and the fish farm (assuming a production of 500 tons y−1), which accounted for about 70% and 25% of the total phosphorus load, respectively. The simulations suggested that about 50% of the annual deposition of phosphorus was resuspended, of which about 60% reached the productive surface waters. Further, the lake should have been able to sustain a fish farm producing about 500 tons of rainbow trout per year without sustaining marked ecosystem effects such as increased algal volume. LEEDS includes all major fluxes for evaluations of fish farm emissions in temperate lakes. The relative importance of the various processes vary from lake to lake. Among internal fluxes, the largest uncertainty lies in the rate of total phosphorus sedimentation. All other internal fluxes (resuspension, diffusion, etc.) depend on this rate. By identifying the major fluxes in a given lake, one could also identify the major uncertainties of model predictions of target variables, e.g., lake total phosphorus concentration and maximum algal volume. A future improvement of this model would be the development of a general, validated sub-model for the rate of total phosphorus sedimentation as well as the development of a generic sub-model for the distribution coefficient (Kd), which regulates the amount of phosphorus in dissolved and suspended phases and hence the rate of phosphorus sedimentation.