Non-parametric representation of the electromagnetic bias in radar altimetry

Abstract

Most previous analyses of the electromagnetic bias have modeled the bias as parametric function of the wave height and wind speed. Recent studies using indirect electromagnetic bias estimates from the TOPEX/Poseidon altimeter support that the bias is modeled better with a non-parametric (NP) representation. This method essentially consists of estimating the bias for each wave height and wind speed point on a two-dimensional grid and using statistical smoothing methods to generate a bias estimate contour. The authors have applied different nonparametric analysis methods to the six months of direct electromagnetic bias measurements collected from an oil platform in the Gulf of Mexico during 1990 [Arnold et al. 1995]. The differences and similarities between the oil platform measurements and the satellite estimates are shown. This study is in support of the upcoming mission of the NASA/CNES Jason-1 altimeter.