Recovering spatial features in the ocean: performance of isopycnal vs. isobaric analysis

This study compares the properties of isopycnal and isobaric coordinate systems as a basis for spatial interpolation of oceanographic observations. The effect of 2D interpolation between observations sampled on either isobaric or isopycnal surfaces is investigated using both real and simulated vertical soundings whose locations correspond to CTD casts carried out during an intensive survey in the Alborán Sea (western Mediterranean). In the case of hydrographic data, a spatial analysis of 3D features via interpolation on regularly spaced isopycnal surfaces is clearly superior to one based on isobaric interpolation within a stably stratified ocean. The isopycnal distribution of geostrophic velocity can be derived via interpolation of a Montgomery potential function, sample values of which are easily obtained from hydrographic data. The advantage of isopycnal interpolation of current data appears to be significant only within frontal regions characterized by a pycnocline of high static stability separating water masses of relatively low static stability. In regions of more uniform stratification, isobaric spatial interpolation is equally effective in resolving baroclinic features sampled by observed current profile data.