Test and evaluation of objective mapping applied to oceanographic data from GIN Sea'86

Abstract

This study is an investigation to produce an objective analysis

procedure to assimilate oceanographic data from the GIN Sea program of SACLANTCEN into numerical model initialization. More specifically, this study suggests a preprocessing analysis to define the most appropriate interpolation scheme. Here, interpolation is considered accurate if the resulting field distribution preserves the dominant spatial features contained in the data set. The method develops as follows. A known continuous function is given and mapped on an irregular grid. This artificial distribution is evaluated on a regular grid using different interpolation algorithms. The interpolated solutions are compared with the values of the analytical function at the same mesh points, by computing standard statistical distributions. Application of interpolation routine to oceanographic data illustrates differences and similarities between interpolated distributions preserving synoptic and mesoscale features, and indicates the limits of the solutions for a valid scientific analysis of the observed field.