Comparison of relative locations methods and their accuracy for determining fault structures

Precise earthquake locations are a prerequisite for determining real fault structures. To improve the precision of the event location, a few relative locations methods are commonly used to refine event locations. Relative relocation methods reduce effects of an imperfect velocity model and errors due to arrival time measurement. We performed comparative tests of tree different relocation methods: HypoDD (HD), GrowClust (GC) and Master event (ME). We tested the efficiency and differences in the event locations using these three methods on dataset from REYKJANET seismic network operating in Iceland on Reykjanes Peninsula. All these methods provide substantially focused shapes of clusters compared to the absolute event locations but the locations of individual events differ evidently depending on the method used.

We also aimed at an effect of the control parameters of HD, GC and ME on final location results and their optimization as well as computational and memory demands.