To provide capability for accurately estimating salinity from observations of temperature.
While the number of accurate salinity measurements is increasing with the advent of the Argo Program's autonomous profiling floats, temperature profiles from expendable bathythermograph (XBT) probes without salinity measuring capability are still more numerous, and the archived XBT data are the richest source of information about the ocean's past state. Providing accurate estimates of the salinity to complement the XBT data allows for better estimates of the ocean's density field, which is crucial for computing currents and the transport of heat that is central to understanding the earth's climate and to predicting its changes. In particular, the salinity estimates are needed when XBT data are assimilated into numerical models of the oceanic circulation so that corrections to the model's temperature field maintain the correct water-mass properties.
Empirical relationships between salinity and temperature can be identified from past data, mostly from conductivity-temperature-pressure (CTD) probes but also from the more recent data from Argo profiling floats. These empirical relationships vary from region to region and from depth to depth, so a great amount of work is needed to provide such estimates for the entire ocean. Consequently, the approach is to treat the world ocean one region at a time, with the choice of which regions to consider first based on the amount of data available for the region and on need for such estimates by other on-going research projects. For a particular region, temperature and salinity profiles are interpolated to regularly spaced levels spanning the depth ranges of XBT probes, and at each level regression models are fitted to the data. Typically, these models describe salinity as linear or quadratic functions of temperature and possibly of latitude and/or longitude. Near the ocean's surface where temperature might provide scant information about salinity, if there are sufficient data, estimates might be improved by accounting for the time of year. In other regions, other information might be exploited: e.g. the nino 3 index can be used as a proxy for longitudinal variability of the western-central tropical Pacific that accompanies El Nino events.
described by Hansen and Thacker (1999). It was subsequently applied to the Gulf of Mexico (Thacker 2005) and to a large region in the northwestern Atlantic containing the Gulf Stream and its recirculation (Thacker and Sindlinger 2005). Currently salinity-estimation capability is being developed for the South Atlantic between 25°S and 45°S.
Hansen, D. V. and W. C. Thacker (1999) Estimation of salinity profiles in the upper ocean. Journal of Geophysical Research, 104, 7921--7933.
Thacker, W. C. (2005) Estimating salinity to complement observed temperature: 1. Gulf of Mexico. To appear in Journal of Marine Systems, 58.
Thacker, W. C. and L. Sindlinger (2005) Estimating salinity to complement observed temperature: 2. Gulf of Mexico. To appear in Journal of Marine Systems, 58.