Computational and Empirical Water Quality Modeling in Lakes and Ponds
MetadataShow full item record
The water quality issues facing aquatic systems have made managers and stakeholders seek help from computational and empirical models. A 3-D hydrodynamic and biogeochemical model was applied to storm-water ponds in the City of Edmonton to examine the reasons behind excessive algae growth and eutrophication. In the hydrodynamic simulation (ELCOM), the atmospheric instability corrections led to more heat loss (up to ~33%) and resulted in more accurate simulation of the temperature profiles (root mean square error (RMSE) <3.03 ℃). The calibration involved tuning albedo and light attenuation coefficients, which was optimized by applying a sensitivity analysis tool that showed the model was more sensitive to albedo. Thereafter, the biogeochemical model (CAEDYM) was coupled with ELCOM and the water quality processes in the ponds were also investigated. The results demonstrated that the model predicted nutrient (nitrogen and phosphorus) components and dissolved oxygen (DO) variability with acceptable accuracy (0.016 mg/L < RMSE <2.85 mg/L), but underestimated the chlorophyll-a (RMSE <64.3 μg/L). However, the model predicted phytoplankton community contribution to the total chlorophyll-a with marginal error (<40% for different groups). An empirical DO model, based on the measurements of hypolimnetic DO (to estimate DO depletion; dDO/dt) and the temperature profiles (to estimate hypolimnion thickness; H) in three lakes (Lake Erie, Lake Simcoe and Eagle Lake) was proposed. The model was developed by neglecting the vertical and horizontal fluxes in the DO budget during summer and considering only sediment oxygen demand (SOD) and hypolimnion oxygen demand (HOD) as the main parameters driving dDO/dt. HOD and SOD were empirically estimated by fitting to the observed dDO/dt and H. Thereafter, using computed H, estimated SOD and HOD and an initial DO, the hypolimnetic DO were simulated in the selected lakes (RMSE <2.67 mg/L). One of the main assumptions of the empirical prediction of DO, negligible vertical DO flux, was then examined by computing the flux of DO through the thermocline in Eagle Lake from microstructure casts. It is shown that the amount of vertical flux toward hypolimnion is 2 to 4 orders of magnitude lower than the other terms of the DO budget (HOD and SOD).