Assessment of Acoustic Travel-Time Tomography of the Atmospheric Surface Layer

Acoustic tomography uses the travel time of acoustic pulses (which depends on wind velocity and temperature) to reconstruct slices of the medium through which they propagate. The feasibility method of acoustic tomography of the near-ground atmosphere has been previously demonstrated in several field studies. However, a quantitative evaluation of the accuracy of the methodology is difficult, because the tomographic observations are averages along lines and therefore cannot be directly compared to point observations in the atmosphere. Here we assess the validity of acoustic tomography by numerically simulating both the direct and inverse problems. Realistic, high-resolution turbulence fields are synthesized in the tomographic plane using the quasi-wavelet method. Sound is propagated through these fields to obtain travel-time data. Various inverse methods are then applied and the accuracy in imaging the original turbulence field is studied. Several inverse methods, based on grid-cell partitioning of the tomographic plane as well as on continuous field reconstructions, are studied. Comparisons are made between the abilities of tomography and point measurements to estimate planar averages of fields, as is desired for validation of large-eddy simulations. We also explore optimal placement of the tomographic sources and receivers.