Weight Coefficients for Representing Data Uncertainities in Refraction Microtremor (ReMi) Inversion

In the ReMi (refraction microtremor) method, the phase velocities are determined at the frequencies where a slope or a peak occurs in the spectral ratio map. Once, a dispersion curve is obtained as a velocity values versus periods, then S-wave velocities of the subsurface can be derived by using some inversion methods. We use the damped least-squares method with the singular value decomposition for the matrix manipulations. An appropriate weighting scheme is suggested for the ReMi data for describing data uncertainties. First, the measured data are smoothed by a combination of simple fitting functions that simulates the behaviour of a dispersion curve. Then the weight coefficients are derived from the differences between the measured and the smoothed data sets in order to make the interpretation algorithm stable.