Simulation of correlated Pareto distributed sea clutter

The memoryless nonlinear transformation method is used to simulate Pareto distributed sea clutter with a specified correlation function for the clutter power. The Pareto distribution is formed from a compound model with a negative exponential distribution for the speckle intensity and an inverse gamma distribution for the clutter power. An estimator based on the expectation value of z log z is obtained for the Pareto shape parameter, which has comparable accuracy to the maximum likelihood estimator and the advantage of also being applicable to the compound gamma distribution which arises for multiple looks.

I. Introduction

Analysis of the detection performance of maritime surveillance radars requires a good model of the sea clutter returns. The K distribution is a well established model for sea clutter which is formed by compounding a negative exponential distribution for the speckle intensity with a gamma distribution for the clutter power [1]. The K distribution often does not match the tail of the clutter distribution very well, so another component is introduced in the KA [2] and KK [3] models. Recent work has shown that the Pareto distribution is able to fit the observed clutter distribution, including the tail, without introducing another component [4], [5], [6]. The Pareto distribution can be formulated as a compound model with a negative exponential distribution for the speckle intensity and an inverse gamma distribution for the clutter power.