Publication
Theoretical studies on oceans and large lakes have shown that submesoscale instabilities in frontal zones tend to reduce horizontal density gradients and enhance vertical density gradients, thereby re-stratifying the Surface Mixed Layer (SML). Submesoscale filament dynamics are primarily studied using numerical models and remote sensing imagery. However, in large lakes, this concept remains without substantial field validation, mainly due to the difficulty in conducting the necessary high-resolution water column measurements. Using a procedure we recently developed to predict the time and location of mesoscale and submesoscale features generated by strong wind fields, this work presents direct field evidence demonstrating the role of submesoscale cold filaments in re-stratifying the SML under weakly stratified conditions in a large lake (Lake Geneva). The dynamics of the observed filaments were further investigated with a high-resolution three-dimensional (3D) numerical model and Lagrangian particle-tracking. The numerical model accurately captured the formation of these filaments. The enhancement of thermal stratification strength, N2, reached O(10-5) s-2 in areas adjacent to cold filaments under atmospheric cooling and heating conditions. In the pelagic zone (offshore), strong vertical velocities of O(100 m d-1) were associated with secondary circulation that rapidly transports and accumulates passive particles in the thermocline and hypolimnion layers, as confirmed by Acoustic Doppler Current Profiler (ADCP) backscattering intensity data. The field observations indicate that under weak stratification, Dissolved Oxygen (DO) variability reaches 0.5 mg l-1 near cold filaments. This documentation of strong vertical motions associated with submesoscale filaments is expected to contribute to the understanding of the vertical exchange of heat, contaminants and oxygen between the atmosphere and the pelagic zone of large lakes, as well as in oceans where carrying out such field measurements is very challenging.