One of the most important events to happen underwater is rarely modeled. This project examines the essential ecological process of lake turnover, a chemical and physical mixing process that evenly distributes nutrients, oxygen, and other dissolved chemicals in temperate lakes each spring and autumn [1]. The inquiry focuses on the autumn turnover process of Mirror Lake, a small temperate lake in the village of Lake Placid, NY, which failed to turn over from 2016 through 2018 due to road salt pollution [2]. We created a model of a successful turnover process, which field data shows occurred in autumn 2019 [3]. This model is intended to enable further research of the lake turnover process and what environmental factors might interfere with its success. The model may also be used to design interventions that support the natural turnover process while still maintaining winter safety for cars and pedestrians around the lake.

Turnover is driven by a combination of natural convection (from temperature changes as seasons change) and wind-forced flow. At the beginning of autumn, the lake has stratified over the summer: there is a warm (less dense) layer on top, with a colder (denser) layer at the bottom, usually at around 4 °C (the temperature at which water is maximally dense). As autumn progresses, the air cools the surface water, which increases in density, sinks, and mixes with the lower layer [1]. Eventually the temperature distribution (and therefore the density distribution) becomes nearly uniform at 4 °C, allowing the wind to generate larger-scale flow patterns that mix the water more completely.

We created this model in COMSOL, representing natural convection with transient fluid flow and heat transfer driven by density variation with temperature and surface wind. A 2D cross section at the deepest part of the lake was modeled using a half-lake geometry (represented by a quarter ellipse) with a symmetry condition. Air temperature data from the nearby Adirondack Regional Airport was used for the external temperature [4]. To reduce simulation time, we created models representing half the turnover process– from summer stratification to uniformity around 4 °C. After validating the results using measured data from the lake, we modeled a full-lake geometry (represented by a half ellipse) that simulated the entire autumn 2019 Mirror Lake turnover process spanning 30 days. Our final result was a 30-day simulation and animation of autumn turnover for Mirror Lake in 2019. The model shows the streamlines and temperature profiles across the duration of the process. We learned that the lake mixes to a uniform temperature, then cools down to 4 °C homogeneously. The lake then becomes stratified again, this time with cooler water on the top and warmer water at the bottom. For future models of turnover, we recommend that users incorporate salt concentration to observe its effects, use non-isothermal physics for accurate results, and include a point constraint of zero atmospheres gauge pressure at the center of the lake’s surface.