Using waveform similarity to constrain earthquake locations

We demonstrate that the numerical value of the cross-correlation between P or S waveforms, commonly computed to determine differential travel times fo r earthquake relocation, itself contains useful information about the relat ive location of the events. This information can be used to supplement trad itional travel-time inversions (especially when few stations are available) and to investigate the heterogeneity that underlies the multiple scatterin g process.