Sea surface salinity (SSS) is a good indicator of net surface water exchange [ Wüst, 1936; Jacobs, 1948]. SSS is controlled by E-P, advection and mixing with underlying water. With a sufficiently well resolved SSS, surface velocities and mixed layer depth, one could infer E-P from ocean data alone. P. Niiler [ Personal communication, 1994] has reported progress with this technique using climatological data. With data of sufficient spatial resolution, the detailed patterns of the mean water flux should be discernible. However, temporal variability will be difficult to establish, given the paucity of SSS data. Features such as the climatically significant Great Salinity Anomaly [ Dickson, et al., 1988] are at present detected only by serendipity. Thus, an important priority for future ocean monitoring is the improvement of surface and upper-ocean salinity measurements. Ultimately, the incorporation of upper ocean salinity data into assimilative ocean models will provide an additional method of constraining the surface water fluxes.
Fortunately, the measurement of salinity has become more routine due to the development of compact, stable conductivity sensors. It is now possible to obtain reliable salinity records from shipboard thermosalinographs and long term records from moorings or drifting buoys. In addition, expendable and autonomous profilers are now available. Finally, remote sensing techniques have been proposed. These approaches are discussed below.
Moored salinity measurements. McPhaden et al. [1990] have successfully deployed conductivity- temperature recorders on upper ocean moorings in the Equatorial Pacific for periods of 6--7 months.
Drifting salinity measurements. Swenson et al. [1991] have explored the feasibility of deploying C/T sensors on surface drifters, with good results.
Thermosalinographs. Somewhat trouble-prone in the past, there is reason to believe that new instrumentation and a systematic approach to deployment on the Volunteer Observing Ships (VOS) will yield a valuable enhancement of surface salinity data. Several pilot projects are currently underway.
XCTD. While long in development, it appears that some success has been achieved in using the eXpendable Conductivity-Temperature-Depth probe (XCTD) as an operational instrument [ Roemmich, personal communication 1992]. If deployed in quantity from the VOS, it would greatly enhance the subsurface data sets.
Retrievable CTD. Small conductivity-temperature-
depth (CTD)
packages with internal recording could be easily deployed on light line
and retrieved, much like the mechanical bathythermograph. This would allow
inexpensive CTD profiles to be obtained from a moving ship.
Autonomous Profilers. R. Davis (Scripps Institution of Oceanography) has deployed profiling floats (Autonomous Lagrangian Circulation Explorer, ALACE) that can be fitted with conductivity sensors. Promising results have been obtained to date. Other autonomous profilers under development, such as the self powered Slocum [ Stommel, 1989], will also provide salinity profiles.
Remote Sensing. The Electonically Scanned Thinned Array Radiometer (ESTAR) is a potential remote sensing technique for monitoring the large scale distribution of surface salinity [ Swift, 1993]. It depends on the influence of salinity on microwave emissions; this effect is strongest at 1.4 GHz. However, temperature has a larger effect and high accuracy temperature measurements (to .02 K) must also be made. This can be done using another frequency (2.65 or 5.0 GHz) or temperature data from another source. This is a very demanding requirement, and yields a salinity accuracy of 0.05 parts per thousand (ppt). It can only be achieved by long time and space averaging, i.e. 30 days by 100 km. This could be useful for climate scale monitoring. It is estimated that the Great Salinity Anomaly could have been detected by an ESTAR, though the sensitivity is reduced to 0.1 ppt at low temperatures. Ideally, long term service by a series of satellites would be continued for decades to be of utility for ocean climate. If improved resolution of 10 km is desired in coastal areas then the accuracy degrades to 2 ppt. Another motivation for ESTAR is its primary application over land: soil moisture measurements. However, at the present time there are no plans for deployment of an ESTAR.