Project Focus
Benthic boundary layer turbulence and its role in differential food supply to benthic organisms in Lake Ontario

Background and Motivation
Understanding turbulent mixing at the bottom boundary in the nearshore of the Great Lakes is of critical importance due to the role turbulence plays in mediating near-shore nutrient cycling in the context of changes to the food web caused by invasive dreissenid (zebra and quagga) mussels. The nearshore zone of the Great Lakes extends roughly 10 km from shore and is the region where the summer thermocline typically intersects the lakebed at depths between 15-30 m. There are always significant vertical movements of the thermocline in all of the Great Lakes and the movements are dominated by the high amplitude near-inertial internal Poincaré waves. For an  internal Poincaré wave, the maximum amplitude of the thermocline movements occurs at the edges of the lake, whilst the maximum velocity signal associated with Poincaré waves occur in the centre of the lakes. The ubiquitous presence of large-scale thermocline motion means that it will impact benthic stratification and mixing dynamics everywhere the moving thermoclien intersects with the lake bed.

The main objective of the study was to examine how internal Poincaré wave-driven thermocline movements influence nearshore benthic stratification and mixing dynamics at the depth of the seasonal thermocline in Lake Ontario. A secondary goal was to analyze the implications of these processes on the context of food supply to dreissenid mussels, nutrient supply to nuisance benthic algae (cladophora), and biotic responses at the lake bed.     

Methods

• The project was mainly based on field observations made on the northern side of Lake Ontario between 2 and 3 km from shore. Two sets of summer observations were made, the first lasted for 20 days in 2012 and the second lasted for 55 days in 2013. In both cases, detailed structure of the near bed stratification was acquired  at high temporal (= 2s) and spatial resolution (=0.1m) with three benthic thermistor chains. Water currents and surface wave data were measured using ADCP (acoustic Doppler current profiler) units. We  also used long-term data from the nearby water quality monitoring station – the Land/Ocean Biogeochemical Observatory (LOBOviz, http://ontario.loboviz.com/), operated by the Ontario Ministry of Environment and Climate Change.
• The acquired data by the deployed instruments were processed to analyze the offshore thermocline movement and the associated near-bottom flow, stratification and mixing dynamics at the depth of the thermocline by calculating a variety of parameters, including the buoyancy frequency, depth of the thermocline, and the inferred benthic turbulent dissipation and diffusivity caused by internal Poincaré wave-driven cross-shore flows as well as surface waves. 
• Diapycnal mixing rates (K_z) caused by the thermocline movements were estimates applying the Thorpe-scale method on the measured temperature profiles by benthic temperature chains.
• surface-wave driven benthic dissipation rates was calculated from measured significant wave heights and peak wave periods.

Figure 1. Thermal stratification in water column in (a) and  near-bottom water thermal stratifications in (b-c). Variation in near-bottom temperature in four spatial locations is shown in (d). Positions of the thermocline and 13oC isotherm are marked by a black and a white line respectively in (a).	Figure 2. Variability of near-bottom temperature grdient, depth-averaged buoyancy frequency and currents.
Figure 3.  Estimated inferred intensity of turbulence based on the Thorpe-scale method	Figure 4. Possible impact of benthic turbulence on the food supply to mussels from deep chlorophyll maximum at the thermocline.

Key Findings

• The thermocline in Lake Ontario undergoes significant vertical excursions at near-inertial periods, indicative of strong and persistent internal Poincaré waves [Fig. 1]. The thermocline oscillation at the near-inertial internal Poincaré wave period induces a striking asymmetry in near-bed stratification and benthic turbulence along the sloping lakebed at the depth of the seasonal thermocline between its rise and fall [Fig.2].
• During the falling phase of the thermocline, the warmer down-slope flows raise the nehe near-bottom water temperature, leaving strong positively stratified near-bed waters with temperature gradients. In contrast during the colder up-slope flow phase, the mean near-bottom water temperature was reduced and there was unstable stratification at the lakebed with temperature gradients.
• Due to the presence of unstable stratification, the near bottom water of the up-slope phase was characterized by high incidence of temperature inversions and relatively higher values of turbulent mixing rates. The mechanism is identified as shear driven near-bottom convective mixing driven by large amplitude internal Poincaré waves.  The opposite occurs during the down-slope phase [Fig. 3]. 
• We also found that, at least during the observation period of 2012 and 2013, benthic mixing due to thermocline movement was much more important than that caused by the orbital motions due to surface waves.
• We hypothesized that the observed dynamics of benthic turbulence may control a differential food supply mechanism of mussels [Fig. 4]. During the rising phase of the thermocline, the isotherms slow down at boundary and high turbulent mixing associated with convective overturning occurs. Thus, during this phase mussels will potentially able to filter feed more effectively experiencing possibly elevated level of phytoplankton under mixing conditions which replenish the supply of food near the lakebed (Fig. 4a). In contrast, during the falling phase, the isotherms get very close together and the boundary layer is strongly stratified and turbulence is greatly diminished. In this phase it is more likely that the mussels’ filtration is only acting on a small distance from boundary due to the weak mixing. Consequently, the immediate boundary becomes depleted of plankton and the mussels are not in optimal growth conditions (Fig. 4b).