Paul

To answer exhaustively more information about data acquisition layout, data and the adopted inversion technique should be given.

Nonetheless, I attempt a general answer. The smearing effect in the travel-time inversion (tomography) is always a pitfall. In my opinion, some good practices to avoid it involve:

1) in the acquisition phase: It should pay attention to sample the volume (or plane) so as to illuminate it in as many directions as possible. The source-receiver directions should span several different angles. This implies that the ratio between the distances receiver-source boreholes and max-min depth of receiver or source should lesser than 1. In general, should be avoided the horizontalization of the rays. This should guarantee the illumination of the pixels (or nodes) of the model with a wide range of angles.

2) in the inversion: I advise parametrizing the model with a minimum dimension of the mesh of the model greater than source and receiver spacings.

The pixel or nodes should be characterized by similar (better almost homogeneous) counts of the ray fractions in different directions. This implies that the pixel could be set with different dimension.

Pixels or nodes characterize by a great number of ray fractions with similar directions become significant attractors of velocity perturbations with relevant smearing effects. I advise you to start with a great dimension of mesh and try other inversions decreasing the mesh dimension up to the lowest paying attention to preserve the homogeneity of the mesh hit-counts (as I advised before).

oops..

Tomographic inversion SHOULD fit if you do it right.

Inversion is a model process. Tomography is a model..

SO if you doit right IT WILL WORK. There should be NO BIAS..

If there is THEN you have issues with the assumption, equations, calibration. Develop simply check points along the entire data path and include anisotropy (?)

I would assume your inversion is properly benchmarked?