There are several indications that the automatic grouping of the events applied after multievent location (based on relative timing) and after single-event fault plane solutions (based on spectral amplitudes and first motion directions) results in groups of events associated with fractures that may be real. It is not possible to come to definite conclusions from only one investigation of this type. The PRENLAB projects produce however much more data, this example was based on only one day of activity. As not only the grouping algorithm is automatic, but also the fault plane solution algorithm and the multievent algorithm there are no real problems to submit all data to such a detailed analysis as given here. This little study is part of the preliminary work of how to implement this multievent analysis to all data. The point is not whether the tentative possible results of this example are true or not. The point is that studies like this do not demand a lot of work. When all data are interpreted in this way it is likely that altogether the significance will increase so there will be obvious results and not only indications.
Although there seems to be no chaos among the small earthquakes (for instance the very high horizontal stress consistency) the picture seems to be complicated as so many small fractures are involved. However, with use of computers the complexity is no real problem as it easily can be handled.
The coming work will partly go to develop the type of studies here to routine algorithms for analyzing all data, and partly to analyze the physical interaction of the different fracture slips. The hope is that the physical interaction will be of great help in reducing the uncertainty of the tentative interpretations.