Author(s): Jorge C. Toledo; Cintia L. Ramon; John C. Little; Francisco J. Rueda

Linked Author(s):

Keywords: Bubble plume; Parametric uncertainty; Structural uncertainty; Plume-lake interaction

Abstract: Bubble plume oxygenation devices are commonly used in lakes and reservoirs to prevent problems associated with low dissolved oxygen concentration during summer stratification in the hypolimnion. However, the bubble plume operation introduces vertical mixing, which has the potential to alter the density structure of the lake or reservoir and induce large-scale mixing. For multi-basin lakes, if the depth of the maximum plume rise is above the sill elevation, the plume transport could even modify interbasin exchanges. To understand the dynamics of the oxygen plume and how it affects lake circulation, plume models are often coupled to 3D hydrodynamic models. The results of the plume models are strongly sensitive to the assumptions in the model structure (structural uncertainty) and the defined value of the input parameters (parametric uncertainty). Both types of uncertainty are, however, rarely assessed. Analyzing the uncertainty in the plume model output is especially relevant when the plume-induced circulation is hypothesized to explain the large-scale mixing and particularly the interbasin transport in multi-basin lakes. That is the case of Amisk Lake, a two-basin lake in Canada, where the bubble-plume-induced currents have been proposed as a mechanism responsible for oxygen exchanges between the untreated and treated basin. To evaluate the structural and parametric uncertainty of the plume-model output we conducted a Monte-Carlo analysis and propagate the plume-model uncertainty to the large-scale transport and mixing in Amisk Lake. We identify the model structures and parameters that contribute the most to the variability in the depth of the maximum plume rise and plume detrainment depth and flow. Our results show that, in Amisk Lake, only an extreme combination of parameters can produce detrainment depths above the sill depth, leading to exchange profiles in the sill that are similar in shape (three-layered exchanges) although lower in magnitude to those measured in the field. Either way, the wind-induced oxygen transport between basins always prevails over the transport induced by the plume for any of the parameter combinations of the model plume.

DOI: https://doi.org/10.3850/IAHR-39WC2521711920221511

Year: 2022