Passive scalar intermittency in random flows

This thesis concentrates on reconstructing the complete probability density function (PDF) for a passive scalar governed by a random advection-diffusion equation using a variety of mathematical tools, primarily from partial differential equations, perturbation theory, numerical analysis and statistics. First we present a one-dimensional model which is essentially a random translation of pure heat equation. For some deterministic initial data, the ensuing scalar PDF and its statistical moments can be explicitly calculated. We use this model as a testbed for validating a numerical reconstruction procedure for the PDF via orthogonal polynomial expansion. In this model, the Péclet number is shown to be decisive in establishing the transition in the singularity structure of the PDF which affects the effectiveness of the series expansion, from only one algebraic singularity at unit scalar values (small Péclet), to two algebraic singularities at both unit and zero scalar values (large Péclet). Next, we study the more complicated, two-dimensional model in which the underlying flow is a random linear shear in one dimension. For planar, Gaussian random initial data, we identify the scalar PDF as an integral representing a conditional mixing of Gaussian probability measures averaged over all realizations of a single random variable, namely, the renormalized L2-norm of standard Wiener process. Rigorous asymptotic analyses and solid numerical simulation are performed to the integral formulation to study the evolution and the parametric dependence of the scalar PDF. During these analyses, we discover a transient, nonmonotonic "breathing" phenomena that is related to the multiple spatial scales in the initial random field. Lastly, some preliminary analytical and numerical results are presented to explore the potential of applying the reconstruction methodology to more general, physically relevent models, such as a rotating, viscous, wind-driven shallow water equation