Differential evolution adaptive metropolis sampling method to provide model uncertainty and model selection criteria to determine optimal model for Rayleigh wave dispersion

摘要: The near-surface S-wave velocity is important tool for environmental studies. This parameter can be derived by inverting of Rayleigh wave dispersion. Inversion dispersion has a nonunique solution. Thus, solving inverse problems not only done to find the fittest model but also characterize uncertainty result. In this paper, we applied and tested Bayesian inversion method using developed differential evolution adaptive metropolis (DREAM(ZS)) approach provide posterior distribution parameters (PDMPs). consists Markov chain Monte Carlo (MCMC) simulation which rapidly estimates PDMP. After obtaining resulted posterior, could estimate representative (such as mean, mode, median, covariance, percentile model, maximum model), probability distributions individual parameters, curve these models. For real data, number or layer (degrees freedom (DoF)) propagated unknown. Therefore, needed accurately subsurface parameters. problem, membership function fuzzy (MFF) criteria proposed selection various such information (BIC), Akaike’s criterion (AIC), generalized cross-validation (GCV), Kullback (KIC), finite prediction error (FPE), complexity (ICOMP) are compared optimal model. DREAM(ZS) well seven methods select used investigate influence noise on values three synthetics, with linear increase, low-velocity (LVL), high-velocity (HVL). Our results demonstrated that effective thickness each quantify estimates, while all approaches able determine from different layers noise-free slightly noisy MFF obtain data. typically close true both produce best We data collected Ljubljana site in Slovenia. compare our those estimated blow count standard penetrating test (N-SPT) data; it shows good correlation toward other.