Väli G., Zhurbas V. M., Laanemets J., Lips U.
Velocity field simulated by a circulation model with extremely high horizontal resolution (the grid bin is 232×232 m) in the Gulf of Finland during a period of summer upwelling events is used to calculate advection of floating Lagrangian particles that are uniformly distributed on the sea surface initially. For a relatively short time of advection τ (of the order of one day), the particles are found to gather within narrow, elongated stripes characterized by extremely high, positive values of vorticity, Finite-Time Lyapunov Exponent, and lateral thermohaline gradients module (fronts). The clustering rate, defined as the time derivative of the standard deviation of normalized particle concentration, tends asymptotically at small τ to the standard deviation of flow divergence. The standard deviation of flow divergence, in its turn, displays a considerable growth with the refinement of the model grid, confirming the paramount role of submesoscale dynamics inclustering of floating stuff. At large τ, the probability density function of floating particle concentration is shown to tend to lognormality. Based on the backward-time integration of the Lagrangian velocity convergence, a criterion for finite-time clustering is introduced.