Using an "inverse dynamic method" to determine temperature and salinity fields from ADCP measurements

In heavily fished areas, upward looking acoustic Doppler current profilers (ADCPs), moored at depth, may
be the only option for long-term current measurements. Arrays of ADCP moorings that cross a current can thus
be the optimal strategy for monitoring the volume flux. These instruments only measure water properties at the
instrument, not through the water column, however. By itself, an ADCP array, therefore, does not give flux
estimates of specific water masses unless temperature and salinity profiles can be derived from the velocity
profiles. This is the opposite of the classical problem of determining currents from temperature and salinity
observations, and in principle it should be possible to solve it by inverting the classical dynamic method. As
for the classical method, this problem requires additional reference information. Using observations from the
Faroe Current between Iceland and the Faroe Islands, it is demonstrated that this procedure can indeed be used
by applying empirical orthogonal function (EOF) analysis to CTD and ADCP data from a section that crosses
this current. It is found that one of the empirical velocity modes is highly correlated to the dominant temperature
and salinity modes. Employing this relationship, ADCP measurements are used to reconstruct temperature and
salinity fields with the same temporal resolution as the velocity field. For the Atlantic inflow of the Faroe Current,
the reconstructed fields are found to explain 60% of the temperature and 44% of the observed salinity variances