Noise source imaging with correlation asymmetry

We have developed a new method for mapping ambient noise sources using the asymmetry of noise cross-correlations. The approach is to measure the energy ratio of the causal to anticausal signal windows (instead of signal-to-noise ratios or absolute amplitude). This approach has the advantage that the absolute noise correlation amplitudes need not be known or modeled. The sensitivity of the measurement to the space-dependent power spectral density of sources is quantified by considering the correlation as a signal from spatially uncorrelated noise sources. Applications of the measurement to global and regional vertical-component datasets, combined with simplified sensitivities along ray paths, show that it yields results which are consistent with earlier studies. Moreover, first synthetic inversion tests demonstrate that, at least for simple cases, an inhomogeneous source distribution can be well retrieved using this measurement. Our goal is to finally include the measurement in a combined noise source-structure inversion.