Conclusions

Considering the above cases leads to the following conclusions:

• Before performing an experiment where array data will be used to locate sources in the way described here, some simple computations should be performed. This ``array calibration'' tells us, given the array geometry, how accurately we can expect to locate sources.
• When we use incorrect velocities in determining the moveout trajectory used for computing semblance, we can still locate the energy source. However, in the result, the source location will be shifted from its true position (mostly in the vertical direction) and the signal to noise ratio decreases. Here I considered the case of a 10% velocity error and found that the magnitude of the semblance maximum decreased approximately by a factor of three.
• Computations have been performed for spatially aliased data for two different seismic arrays. The results show that even with aliased data we can still do a good job of locating sources. Pictures of better quality were obtained using broadband signals. Using narrowband high-frequency signals can result in serious artifacts (sidelobes).