Wave shadowing effects on phase resolved wave spectra estimated from RADAR backscatter: Part 1: An investigation of the errors due to wave shadowing in the spectral averaging approach

Abstract

This is the first part of a short two part series concerned which is acknowledged as the primary source of error involved in the estimation of phase resolved sea wave spectra from RADAR backscatter measurement. A mathematical model is produced of what is termed here as spectral averaging which is the most commonly used technique for such estimation purposes. It is found that the spectral averaging technique can be described by a wave-number domain convolution integral. The kernel of the convolution must be a good approximation to a delta function for good estimates to be obtained. Using simulations of the wave shadowing process it is shown that good approximations can be obtained to the delta function requirement using approximately one tenth of the data employed in typical implementations of spectral averaging. This large data requirement is due to the need to impose the wave dispersion relationship as a constraint during the estimation process. Two new estimation methods are briefly introduced to resolve this situation and can operate on small data sets, one of which, termed conflation, is developed in detail in Part 2.