oa The Effect of Smoothing First Arrival Times and the Initial Velocity Model on Refraction Tomography Results

Typically refraction data are collected to characterize a 2D heterogeneous subsurface. to accommodate a 2D subsurface, refraction Tomography has become a useful alternative to the more commonly used layer approach. Common to both methods is the problem of dealing with noisy data which makes picking first arrival times a subjective process. for a layered solution, break points related to refracting layers are chosen which is a subjective process as well. for the Tomographic approach, an Initial velocity model based on less than perfect Information must be chosen as the starting point in the iterative process. <br> using the data set consisting of P-wave travel times for a synthetic seismic refraction experiment created for the ‘refraction shootout’, we investigate the effect of 1) smoothing noisy first arrival time picks and 2) smoothing the Initial velocity model used in the refraction Tomography approach. Smoothing is applied to the first arrival times from the synthetic data set which have uncorrelated Gaussian noise added. We also investigate the effect of smoothing the Initial 2D velocity model produced by the delta T-V method on the final model results. We use a standard refraction Tomography package using Fresnel volume Tomography to produce the P-wave velocity images. Trials are run for varying degrees and types of smoothing of the first arrival times to investigate the time smoothing effect on the final models. Similarly, trials are run for smoothing of varying degree applied to the 2D delta T-V Initial model(s) to determine the Initial model smoothing effect on the final velocity models.