MCD inversion for sound speed using time-distance data

We present an inversion technique based on a MultiChannel Deconvolution (MCD) algorithm that speeds up the calculations by doing the computational work in the Fourier domain. This means that inversions can be done in a matter of seconds, even for high-volume data sets like the ones from the SOHO satellite. This technique has been developed recently within geophysics for inversion of electromagnetic data. It is based on the Born approximation where one inverts for perturbations from a reference model. Because the reference model has horizontal translational invariance, the forward calculation becomes a convolution of the model with the Frechet kernels, which according to the convolution theorem can be performed as a simple multiplication in the Fourier domain. Also inverse calculations speed up significantly in the Fourier domain.