Primary plan is the combination of OMP analysis and dynamics, biogeochemical cycling studies and water age studies. The combination of water mass fractions and geostrophic velocities is one point to consider. This is possible as part of Invers Box Models or from ocean section data directly. Including potential vorticity as an additonal tracer was suggested by Matt Tomczak. at the WOCE AIMS Tracer Workshop held in February 1999 in Bremen (Germany). See also Tomczak 1999.
1. Improving the link between ocean dynamics and tracer fields.
There are generally two ways of describing the oceanic circulation, the dynamic method and the tracing of water masses. Both methods produce valid approximations of the circulation and oceanic property fields but are based on different procedures. One of the tasks of oceanography is to bridge this gap by combining the two approaches.
One first step of achieving this could be to include dynamical properties as parameters. Potential vorticity is such a parameter. It is a property determined by the physics of the flow field which acts like a tracer. Many studies have shown that large parts of the ocean offer conditions under which potential vorticity is conserved.
I suggest that future applications OMP analysis might include potential vorticity as a parameter. Its weight could be determined in the same way as weights for other conservative parameters are determined. Alternatively, its weight could be varied to produce solutions which are either strongly or weakly constrained by ocean dynamics.
2. Estimating climate variability in water mass formation regions.
As far as I can see all applications of OMP analysis to date minimise the residuals by finding the optimal fit of the data to a prescribed matrix of source water types which has to be defined a priori. This is mostly achieved by reference to historical or climatological data.
It is becoming evident that water mass properties in formation regions undergo changes on interannual, decadal and longer time scales. If OMP analysis is applied to a data set which spans many years, or to two hydrographic sections separated by several years, this climate variability translates into time changes in the residuals. This is a rather unsatisfactory state of affairs. A better approach could be as follows.
The linear system of OMP equations can be written as
where G is the matrix of source water types, d the vector of observation values, x the solution vector and r the vector of residuals. The aim of the analysis is to minimise r, which is usually done by varying x. It is possible in priciple to minimise r by varying both x and G and through that process finding the best values for the source water types which can explain the observations. If the data d are obtained as a time series, this method will allow determination of the time history of the source water types.
Implementation of this idea is not trivial. It leads to a nonlinear minimisation problem and gives away the advantage of pointwise solutions. It may therefore require more powerful computational resources than a simple PC.
3. Estimating the importance of diapycnal mixing.
This point is a reminder that OMP analysis can be used to identify regions of strong diapycnal mixing. This has been shown and exploited before (Tomczak, 1981; Tomczak et al., 1994; Klein and Tomczak, 1994) but discussion at the WOCE AIMS Tracer Workshop showed that there is renewed interest in the importance of diapynal vs. epipycnal mixing without general awareness of the capabilities of OMP analysis in this respect.
The main requirement for a reliable discrimination of diapycnal vs. epipycnal mixing through OMP analysis is a sufficiently dense data distribution along the vertical. This is not always given with normal bottle spacing. The WOCE data set should be particularly good in this regard.
Tomczak, M. (1981) An analysis of mixing in the frontal zone of South and North Atlantic Central Water off North-West Africa. Progress in Oceanography 10, 173-191. Tomczak, M., D. G. B. Large and N. Nancarrow (1994) Identification of diapycnal mixing through optimum multiparameter analysis 1: test of feasibility and sensitivity. Journal of Geophysical Research 99, 25,267-25,274.
Tomczak, M. (1999) Some historical, theoretical and applied aspects of quantitative water mass analysis. Journal of Marine Research Vol.1, 1999
creation date: 25 February 1999