Inversion of multimegameter-range acoustic data for ocean temperature

Ocean sound speed (a surrogate for temperature) derived from the ray travel times obtained from acoustic transmissions may be inaccurate when the refe rence ocean state is inadequate for linearized inversion. When the referenc e (e.g., the Levitus ocean atlas) is significantly different from the "true " ocean, the reference ray paths inaccurately represent the true sampling. In addition, natural oceanic variations, such as the evolution of a summer mixed layer, can significantly change the ray sampling over time. The guidi ng principle for inversion is that ray travel times associated with the inv erse solution must match the measured travel times, A time-dependent refere nce ocean can reduce both the nonlinearities and the solution uncertainties since the model variances mag be assumed to be less. The Levitus ocean atl as was employed to explore the effects of nonlinearities when inverting mul timegameter-range acoustic data and to find accurate inversion methods. The se methods were applied to acoustic data obtained in the North Pacific duri ng the acoustic thermometry of ocean climate (ATOC) project using an acoust ic source on Pioneer Seamount off the coast of California. In order to line arize the inversions, the annual cycle was removed by referencing the measu red travel times to travel times computed using the Levitus monthly ocean a tlas, This linearization results in a more accurate time series of range- a nd depth-averaged temperatures, but the solution for range- and depth-avera ged temperature is only slightly different from that using a time-independe nt set of rays. Standard uncertainties for the 0-1000-m depth-averaged temp erature are typically +/-0.012 degrees C, while the annual peak-to-peak tem perature variation is about 0.4 degrees C. Because the travel time data are inherently averaging, the time series of range- and depth-averaged tempera ture is insensitive to different assumptions made in the forward model, suc h as the model parameterization, variances, wavenumber spectra, and the dat a uncertainties.