Robert N. Miller
College of Oceanic and Atmos. Sci., Oregon State Univ., Corvallis, OR 97330
Acoustic tomographic data consist of line integrals of sound speed along ray paths. The community has little experience with line integral data, and most of the methods for mapping oceanic properties and for data assimilation (i.e., combining observed data with results from dynamical models) have been designed and tested with point measurements in mind. In this work, the tools of estimation theory were used to investigate the potential for reconstructing time series of oceanic pressure and density fields by assimilation of line integral data into a simple ocean model. Under ideal conditions, with noise-free data and ocean dynamics given exactly by a simple combination of linear waves, time series of integral quantities along the edges and diagonals of a square array are sufficient to determine time series of amplitudes of a large collection of waves, most of whose wavelengths are much shorter than the length of the edge of the array. Simulation experiments were performed to investigate the consequences of imperfect, possibly nonlinear dynamical ocean models and noisy data.