Discussion of picture of the seismic activity given by the automatic grouping

The sizes of the fractures have typically dimensions of several kilometers both laterally and vertically. The number of events on the same fracture is mostly only 6-9 and the estimated fault radii are mostly less than 200 meters. As an sample in Figure 5 we see a view of the earthquakes within the dominating W-E group. The events are fairly well spread over the fracture. The sizes are the real sizes as given by the corner frequency estimates [42]. Even if the total number of events during this day is 5 times more the surfaces of the events will not cover the whole fracture area. The time order of the events over the area is rather random, not like a domino game. Figure 6 shows the same events as Figure 5 but now scaled according to slip size (the area of the event circles are proportional to slip size). The conclusion is that the size of the seismic slip during this day is very unevenly distributed over this fracture and that the activity starts at several places not neighbouring to each other.

**Figure:** This figure shows the 28 events of the largest group striking N79 $^\circ$ E. The view to the right is normal to this strike (N21 $^\circ$ W). The sizes are now not scaled but the true ones as estimated from the corner frequencies. The slip directions (marks at the peripheries) are not perfectly consistent but the slips are mainly in the same direction. Note that the fracture surface is not covered by these events which may indicate stable slip between the events.

**Figure 6:** As Figure 5 but now the events have been scaled according to slip size. The largest slip is 7 mm, the next largest are around 3 mm. The diameters of the event "coins" are proportional to the slip size. Note that the three largest events (slips) have very consistent slip directions. If the slip on this fault has been constant over the whole area most of the slip must have been stable (non-seismic).

If the pore pressure is increased in a rock mass the pore pressure will first increase along certain fractures with highest permeability. If there remains shear stress over this fracture the pore pressure increase will trigger seismic events at locked asperities. The pore pressure will also increase in fractures crossing the main fracture and trigger slip on those fractures. This activity is initially expected to occur close to the first fracture and later spread out.

The pattern we see here with a seismic activity spread along a main fracture and with activity on crossing fractures is very similar to what can be observed during hydraulic fracturing of rock masses through water injection. Thus it is possible that increase in pore pressure is affecting the period of seismic activity.

The seismicity is the unstable slip on the fracture. Stable brittle slip is not seismicly observed but is a possible cause to seismic activity at different places over a fault area within a short-time window. The reason is simple that if a fault starts to slip stably as a whole it is likely that some asperities will remain locked, accumulate stress concentration due to surrounding fault slip, and then break seismicly when the stress is too large. The size of the slip depends then on the strength of the locked part. In Figure 6 showing the peak slips of the events on the possibly primary fault there are three events having significantly larger slip. The peak slips of those events are 3-7 mm. The first events on this fracture have all small peak slips, less than 1 mm. Then the deep 3 mm event appears while smaller slip events continue to occur at "normal" (4-7 km) depth. After 7 hours the upper 3 mm event occurs and after 10 hours the 7 mm event comes. This picture is in reasonable agreement with what one might expect if the fracture started to slip stably at the time of the main event. Possibly the deeper part has started to slip earlier or at a higher rate.

It is of course not the purpose of this simple example to reach any conclusion, the reason of presenting this discussion is to indicate what eventually will come out if more extensive and intensive analysis of this type continues in the area.

The assumption of a general stable slip process triggering seismic events at asperities indicates that the stable slip over this single fracture during the day is at least 7 mm which is the largest slip of a single event on the fracture. One must, however, remember that it may be possible that the 7 mm event was already prestressed and was triggered by a smaller stable slip. This is, however, something that can be analyzed, a 7 mm movement on a large fault area will affect the stability in its close surroundings and can be expected to be consistent with occurrences and mechanisms of the surrounding microearthquake.

Only six of the 15 fracture groups contain events having peak slips exceeding 1.5 mm. Of these 4 shows a rather strict decay of the slip sizes with the largest slips within 2 hours of the main shock. One group has a rather irregular time development with the largest slip event about 5 hours after the main shock. This is one of the N-S striking fractures. Only the largest group, 28 events on the W-E fracture, shows a rather clear increase in the slip sizes with time and with the largest slip event at the end of the day and at the east end of the fault (east of the main event which is on a N-S fracture on the north side of the W-E fracture).

The N-S group having an unclear slip development with time is far from vertical and is at the east end of the activity. It may be that this activity is affected by the possibly increasing slip on the W-E fracture where the slip increases with time.

The activity prior to this day migrated from west along a western extension of the activity of Figure 4. Together with the indications that the W-E fault started to slip at the time of the main event it seems possible that the main event was associated with some locking of the W-E fracture of the 28 events.