| Summary: | Approved for public release, distribution is unlimited === In a computer simulation experiment, acoustic tomography is assessed as a means of measuring the seasonal flux of heat advected by the Norwegian Atlantic Current. Oceanic heat flux has traditionally been measured by various direct or indirect techniques that are prone to error or large uncertainty. The tomographic technique offers distinct advantages over conventional methods in that temperature and current fields, that combine to yield heat flux in the ocean, can be determined at various spatial and temporal scales. The adequacy of the tomographic technique thus hinges on the question of how well can the temperature and current by resolved spatially? The spatial resolution of tomography varies with array size, number of transceivers used and the characteristics of the sound channel. In the assessment we use the General Digital Environmental Model (GDEM), a climatological data base, to simulate an ocean area 550 x 550 km squared off the Norwegian Coast. Resolution and variance analysis are performed on two circular arrays consisting of six transceivers. An important finding is that the horizontal resolution lengths of the current and temperature fields differ. For a six element array the horizontal resolution length is approximately one fifth the array diameter for the current field, whereas for the temperature field it is one sixth the array diameter. We then generate synthetic travel time data that have embedded within them temperature and current signals as well as random noise. We invert the synthetic travel time data to form estimates of the original fields using a linear optimal estimator based on the Gauss-Markoff theorem. We relate the sound speed perturbation field to potential temperature and compare these estimates to the original values. Finally we use the estimated fields to compute heat flux across a transect located within the array. We compare the actual to the estimated heat flux to asses the quality of the tomographically derived value. We have found that the quality of the heat flux estimates depends critically on how well the flow field is resolved. A six element array can adequately resolve the current in the Norwegian Sea, provided that its diameter is shorter than 250 km. Such an array is able to measure net heat flux through a transect at the center of the array with only a 10% error.
|
|---|
