In this study, we first show that it is difficult to reconstruct the vertical structure of vortices using only surface observations. In particular we show that the recent SQG and ISQG methods systematically lead to surface intensified vortices and those subsurface intensified vortices are thus not correctly modelled. We then investigate the possibility to distinguish between surface and subsurface intensified eddies from surface data only, using the sea surface height and the sea surface temperature available from satellite observations. A simple index, based on the ratio of the sea surface temperature anomaly and the sea level anomaly, is proposed. While the index is expected to give perfect results for isolated vortices, we show that in a complex environment, errors can be expected, in particular when strong currents exist in the vicinity of the vortex. The validity of the index is then analysed using results from a realistic regional circulation model of the Peru-Chile upwelling system, where both surface and subsurface eddies coexist. We find that errors are mostly associated with double core eddies (aligned surface and subsurface cores) and that the index can be useful to determine the nature of mesoscale eddies (surface or subsurface- intensified) from surface (satellite) observations. The errors however reach 24% and some possible improvements of the index calculations are discussed.