A synopticity test of a sampling pattern in the Alboran Sea

Hydrographic surveys are usually considered as synoptic, as if the corresponding data set had been sampled at the same time. However, this assumption might lead to strong biases in subsequent analyses. To obtain synoptic data at a given moment, we integrate a Lagrangian motion equation towards a 'modal sampling time', to relocate the sampling stations both in time and space, by combining objective analysis and geostrophic velocities, assumed to be either stationary or unstationary over the period of integration. The relocation algorithms are applied to a data set covering the entire Alboran Sea. The relocation of data points leads to a significant change on vertical velocities computed with the OMEGA equation. For validation purpose, the algorithm is also applied to a synthetic data set obtained from a 3D Primitive Equation (PE) model output. Raw and pseudosynoptic analysed temperature-salinity (T/S) fields and derived quasigeostrophic (QG) vertical velocities are compared. It is shown that the relocation method might correct significantly the errors arising from the synopticity hypothesis and that the sensitivity to the level of no-motion and the analysis parameters is less important than the nonsynopticity. (C) 2002 Elsevier Science B.V. All rights reserved.