| Popis: |
Turbulence close beneath a free surface leaves recognisable imprints on the surface itself. The ability to identify and quantify long-lived coherent turbulent features from their surface manifestations only could open up possibilities for remote sensing of the near-surface turbulent environment, e.g., for assimilation into ocean models. Our work concerns automatic detection of one type of surface feature – “dimples” in the surface due to surface-attached “bathtub” vortices – based solely on the surface elevation as a function of time and space. Two-dimensional continuous wavelet transformations are used together with criteria for eccentricity and persistence in time, to identify candidate surface-attached vortices and track their motion. We develop and test the method from direct numerical simulation (DNS) data of turbulence influenced – and influencing – a fully nonlinear, deformable free surface. Comparison with the vertical vorticity in a plane close beneath the surface reveals that the method is able to identify long-lived vortical structures with a high degree of accuracy. Further tests of success rate included the vortex core identification method of Jeong and Hussain (1995). Different mother wavelets were tested, showing that the simplest option – the Mexican hat – outperforms more advanced options. Jeong, J., & Hussain, F. (1995). On the identification of a vortex. Journal of fluid mechanics, 285 69-94. |
