Cumulative probability noise analysis in geophysical spectral records

Abstract

The technique of rank ordering of spectral energy density samples was applied to high frequency (HF) ocean backscatter radar spectra to examine the properties of signal and noise in the record. HF ocean backscatter spectra have characteristic Bragg scatter lines and bands of energy with random phasing, and it is known that spiky noise from sferics and anthropogenic sources contribute to the observed energy density. Rank ordering of the spectral density data revealed three main features: (i) first-order Bragg scattering; (ii) continuum of randomly phased energy; and (iii) the underlying noise floor of the system. It is shown that the technique of rank ordering of spectral amplitude samples provides better insights and an improved algorithm for differentiating between bands of randomly phased sea echoes and system noise. The analysis reveals a constant offset in the spectral energy density data from the Ocean Surface Current Radar (OSCR) HF radar system, which should be removed before algorithms are applied to extract significant wave height information. The rank ordering of spectral energy density samples is a useful tool for differentiating between system noise and wide-band incoherent energy often obtained in a wide range of geophysical remote sensing.