Burgers dynamics for Poisson point process initial conditions

Abstract

We investigate the statistical properties of one-dimensional Burgers dynamics evolving from stochastic initial conditions defined by a Poisson point process for the velocity potential, with a power-law intensity. Thanks to the geometrical interpretation of the solution in the inviscid limit, in terms of first-contact parabolas, we obtain explicit results for the multiplicity functions of shocks and voids, and for velocity and density one- and two-point correlation functions and power spectra. These initial conditions gives rise to self-similar dynamics with probability distributions that display power-law tails. In the limit where the exponent $\alpha$ of the Poisson process that defines the initial conditions goes to infinity, the power-law tails steepen to Gaussian falloffs and we recover the spatial distributions obtained in the classical study by Kida (1979) of Gaussian initial conditions with vanishing large-scale power.