GHER Contribution


Gher Contribution

Gher Contribution

3D variational inverse analysis of in situ temperature and salinity data

From observational data sets collected during the two synoptic surveys, the variational inverse method will be used to reconstruct the 3D temperature and salinity to obtain a regular 3D density matrix suitable for its use in a numerical model. This will be done in an independent way for the three consecutive samplings in the same area of the Western Alboran Basin, in such a way that three consecutive analysed density fields will be achieved.

Diagnosis of the ageostrophic motion

through the density dynamical assimilation using the Digital Filter Initialization and the numerical model. Three independent and consecutive adjusted states (density and 3D velocity) will be obtained. Boundary conditions of the primitive equation model will be fixed according to the particular hydrodynamic regimes prevailing at the time of the cruises, in order to allow the model to reproduce the dominant features of the general circulation; those boundary conditions will be determined from repeated surveys and from nesting in a general circulation model.

Data assimilation algorithm determination and optimisation

A number of parameters in the data assimilation technique will be tuned to optimise the prognoses of the PE model. These include the noise/signal ratio used in the T,S data inversion, the strength of the assimilation constraint and the type of assimilation (nudging or sequential, assimilation of only T,S or assimilation with ADCP data). The repeat survey observations will be used to constrain the assimilation and model parameters.

Diagnoses of the vertical velocity

will be performed through density dynamical assimilation of the observed data sets in the primitive equation model. The numerical model initialized with the first adjusted state will be integrated forward in time in order to forecast two successive states coincident in time with the second and third samplings. The assimilated fields will be analysed and their sensitivity with respect to the model and the assimilation parameters will be investigated. In addition we will investigate the sensitivity to the spatial resolution of the PE model (horizontal and vertical), etc. Once again the repeat survey result will be used to constrain the solutions from the model integrations.

Correlate remotely sensed SST, surface roughness, elevation and ocean colour

A methodology will be developed to correlate SAR detected structures with SST, that until now has been very useful in locating the Alboran gyres at low resolution. This correlation will be related to meteorological measurements taken during the cruise. Also, surface mesoscale structures identified in SAR images will be compared with ocean colour mapping and with in situ physical and biological measurements of surface and sub-surface structures Athorough understanding of the correlations will be obtained using coupled biochemical/physics models and data assimilation techniques.

Comparison of vertical velocity estimates

Solutions to the QG omega equation will be sensitive to smoothing of the combined data sets and the boundary conditions used to constrain the solutions (Tintore et al. 1991, Allen and Smeed 1995. More realistic boundary conditions resulting from the assimilation of the data sets into a primitive equation numerical model will also be available to constrain the omega equation. We will use the new density dynamical assimilation high temporal resolution modelling techniques described by Viúdez et al. (1995) to provide an independent diagnosis of the vertical motion present in the data sets. The comparison of these diagnostic estimates will establish the level of uncertainty inherent in omega equation analyses. We expect that qualitatively the derived vertical velocity maps will be similar and this will be reflected in rms values, but that peak magnitudes will be noticeably sensitive to the length scales used in the smoothing algorithms. In-situ direct measurement of vertical velocities will be recorded by SOFARGOS floats. The data recorded by these floats will be directly compared with the mapping of our diagnosed parameters.

Examine the short term temporal variability in the observations

will be determined by looking at repeated surveys. Vertical flow can be considered to be made up of two components (Strass 1994), flow along sloping isopycnal surfaces and the vertical disturbance of isopycnal surfaces themselves. Both cruise proposals incorporate repeated finescale SeaSoar and ADCP surveys. Repeated surveys enable observations to be made of the advection of salt, temperature and biomass along isopycnal surfaces. Thus qualitative and quantitative estimates will be made of the relative importance of the two components of vertical motion.

Estimate and map vertical transports

We will make these estimates by considering the relative importance of the components of vertical motion. The large vertical velocities associated with the disturbance of isopycnal surfaces following the propagation of meanders and vortices relative to the underlying water are temporary with timescales of the order of the period of the instabilities. Significant vertical mixing near the surface, however, will result in the permanent vertical transport of salt and heat. The vertical excursions are considerable, O(100 m), and their role in bringing nutrients and zooplankton up into the euphotic zone and drawing phytoplankton down out of the euphotic zone will be identified. Advection of properties along isopycnal surfaces is proportional to the cross-frontal component of the total velocity field (Lindstrom and Watts 1994) and may generally represent a more permanent vertical transport of water properties.

Develop software to calculate vertical velocities routinely from CTD, ADCP and other data sources

The comparison of vertical velocity estimates will culminate in the development of standardised scientific software package to calculate vertical velocities routinely from SeaSoar/CTD, ADCP and remotely sensed data. Programs will be written in FORTRAN 77 and will include routines for smoothing data and combining density and horizontal velocity measurements.

Error analyses for the exploitation of the published methodologies

An accompanying report will include guidance on implementation of the software and guidance for the error analysis of the derived fields. This and the software will be generally available by Email, ftp and publication on the OMEGA WWW server.


[GHER Contribution] [Objectives] [In Situ Data Extraction] [T/S and W Analysis] [ Vertical velocity field computation package] [Publications]
[Main page]