Recovery of temperature, salinity, and potential density from ocean reflectivity

This work explores a method to recover temperature, salinity, and potential density of the ocean using acoustic reflectivity data and time and space coincident expendable bathythermographs (XBT). The acoustically derived (vertical frequency >10 Hz) and the XBT-derived (vertical frequency <10 Hz) impedances are summed in the time domain to form impedance profiles. Temperature (T) and salinity (S) are then calculated from impedance using the international thermodynamics equations of seawater (GSW TEOS-10) and an empirical T-S relation derived with neural networks; and finally potential density () is calculated from T and S. The main difference between this method and previous inversion works done from real multichannel seismic reflection (MCS) data recorded in the ocean, is that it inverts density and it does not consider this magnitude constant along the profile, either in vertical or lateral dimension. We successfully test this method on MCS data collected in the Gulf of Cadiz (NE Atlantic Ocean). T, S, and are inverted with accuracies of Tsd=0.1 degrees C, Ssd=0.09, and sd=0.02kg/m3. Inverted temperature anomalies reveal baroclinic thermohaline fronts with intrusions. The observations support a mix of thermohaline features created by both double-diffusive and isopycnal stirring mechanisms. Our results show that reflectivity is primarily caused by thermal gradients but acoustic reflectors are not isopycnal in all domains.