Empirical relationships between salinity and temperature, which vary with longitude, latitude, and depth, have been identified from past data from conductivity-temperature-pressure (CTD) probes and from Argo profiling floats. These data have been interpolated to standard pressure levels and organized spatially so that regression models can be fitted to them at each level within sub-regions of the South Atlantic. A variety of models have been examined, with all combinations of temperature, temperature squared, longitude, and latitude used as regressors. Because of the sparsity of the data, evaluating which of these models should give the most accurate estimates of salinity is difficult.
- Data used for identifying empirical relationships between salinity and temperature
- Sub-regions within which the relationship between salinity and temperature is relatively homogeneous
- Salinity in the South Atlantic has a minimum near the bottom of most XBT profiles.
- Simple estimates of salinity can be obtained using climatological mean fields.
- Regression models offer a flexible and accurate approach to estimating salinity.
- Independent verification allows the accuracy of the salinity estimates to be gauged.
- The regression models can be used for "salting" South Atlantic XBTs.