Chapter 4

Seismicity patterns from focal mechanisms, earthquake clusters,

multiplet analysis, and the search for surface fault ruptures

Index

4.1 Introduction

In the CSZ, most earthquakes are assumed to be spatially related to preexisting faults of the St. Lawrence Paleozoic rift system (Adams and Basham, 1989). This relationship is based on a projection of the hypocentres to the surface up the dip of the regional faults. When projected, the earthquakes are located between a paleo-rift fault and a possible fault-controlled bathymetric feature. In support of the paleo-rift model, some focal mechanisms have one nodal plane oriented similarly to the paleo-rift faults (Adams and Basham, 1989). It was recognized, however, that the highly variable focal mechanisms are not easy to interpret, possibly due to the complications introduced by the impact structure.

To evaluate the CSZ earthquake-fault connection, four sources of information are examined: possible surface ruptures, focal mechanisms, hypocentre alignments and earthquake multiplets. The best evidence of the earthquake-fault connection would be to find a surface rupture caused by some of the numerous M > 6 CSZ earthquakes. Section 4.2 describes the search for such a rupture under the St. Lawrence River. Focal mechanisms are a representation of the earthquake rupture process and, as such, constitute an essential source of seismotectonic information (Section 4.3). They provide the strikes and dips of two nodal planes (one of which being the rupture plane), the trends and plunges of the P, T and B axes, and the nature of faulting that produced the earthquake. Fault zones can also be located from hypocentre groups and alignments (Section 4.4). Another source of information on fault orientations is provided by earthquake multiplets, which represent earthquakes with similar locations and focal mechanisms (Section 4.5).

In the CSZ, hypocentres are not distributed randomly. There are areas where earthquakes are shallow (< 10 km depth), intermediate (10-15 km), deep (> 15 km), or dispersed. To better define the local factors that determine the hypocentre distribution, earthquakes are divided into a series of sub-zones in Section 4.6. The seismotectonics of these sub-zones are examined using the available focal mechanisms, earthquake groups and multiplets. Based on the findings, regional seismicity aspects of the CSZ are discussed in Section 4.7. In the Appendix, recommendations for future work are presented.

4.2 The search for surface fault ruptures

The best possible correlation between the CSZ earthquakes and the regional faults would be a surface rupture caused by one or by a series of magnitude > 6 earthquakes. In the CSZ, at least five earthquakes of this size are known to have occurred since the XVIIth century, and numerous prehistorical earthquakes are suggested by local landslides and lake disturbances (Filion et al., 1991; Doig, 1986). Despite these large earthquakes, no surface rupture has ever been reported in the geological maps (Rondot, 1979; 1989), in the historical accounts, or in the scientific reports on the 1925 earthquake (Hodgson, 1925; 1950). Based on the source parameters for a moment magnitude (M) 6.0 (Table 3.3) and the focal depth (10 km; Bent, 1992), the 1925 earthquake is unlikely to have ruptured the surface. However, the recurrence of similar or larger events over thousands of years may have led to a surface rupture.

It can be argued that the likelihood of a surface rupture is greater under the St. Lawrence River where most magnitude > 4 earthquakes of the XXth Century occurred (Stevens, 1980; Chapter 1). Under the river, however, thick sequences of unconsolidated Quaternary deposits cover the basement. These sequences represent multiple glacial and inter-glacial periods covering more than 250,000 years (Occhietti et al., 1997). The seismic lines by the Société Québécoise d'exploration pétrolière (SOQUIP) suggest that some river-parallel valleys include up to 600 m of sediments. At present, the complete sequence has not been drilled and, consequently, the Quaternary age of the basal units can only be assumed. A model was defined for a surface rupture caused by a thrust earthquake occurring in the Precambrian basement (Figure 4.1). A rupture from the bedrock causes offsets or truncations of the basement-sediment interface, propagates into the sedimentary sequences, giving displaced, or truncated sub-horizontal horizons on seismic sections. A bathymetric feature can be evident in the depth soundings, such as those done during the 1994 gravity survey. To consider this model, a series of factors were defined (Table 4.1).

The SOQUIP seismic profiles, acquired in the early 70's (described in Chapter 3) can be used to infer the position of sub-river ruptures. Most profiles show continuous horizontal reflectors in the Quaternary deposits, in the underlying Precambrian basement and in the Appalachians (Figure 4.2). It is assumed that a vertical throw of 20 m (0.02 s offset) through the sediments would be evident on most SOQUIP lines. This corresponds to the accumulated slip of seven shallow M 6.0 earthquakes. In addition, a series of shallow high resolution seismic profiles were acquired jointly by the Université du Québec à Rimouski (UQAR) and the Université du Québec à Montréal (UQAM), with partial financial support by the GSC. More than 200 km of shallow seismic reflection profiles were acquired on the St. Lawrence and Saguenay rivers during the summers of 1995 to 1998. Up to 800 ms of analogue data were acquired (corresponding to about 800 m of penetration with a P-velocity of 2 km/s). The resolution power of the method is very high, and provides images of sedimentary facies within the Quaternary sequences.

Ruptures with meter-size slips should be seen. As of May 1998, the careful examination of the shallow seismic lines is ongoing at the UQAR (B. Long, pers. comm.). One of these lines (line 13) is shown in Figure 4.4B and is discussed below.

On most SOQUIP seismic sections, one or more continuous sub-horizontal reflectors is evident in the Quaternary sequences, in the basement or in the interpreted interface between the two (Figure 4.2). On some profiles of lesser quality, a few small, although ambiguous, offsets were seen southwest of the Ile-aux-Coudres and within the impact structure. Elsewhere, no unambiguous ruptures were found on any profile, and most areas have clear uninterrupted markers (Figure 4.3). The NE end of the CSZ, where most magnitude > 4 earthquakes concentrate, shows no evidence of a rupture on Lines 13, 35 and 37, including in the thick Quaternary sequence near the north shore (Figure 4.4A).


Table 4.1 Factors defining the nature of the potential ruptures

Nature of discontinuity Quaternary Precambrian

Basement

Appalachian Basement Interface Quaternary-Basement

Bathymetric Feature
Neotectonic Fault Rupture Yes Yes Possible if SE of Logan's Line. Yes Possible
Syn-sedimentary Movement Yes No No No Possible
Ancient and Inactive Fault No Possible Possible No No

Yes: Should rupture the given layer.
No: Should not affect the given layer.


Similarly, at the SW end, where some magnitude > 4 earthquakes cluster, shallow seismic profiles do not show any abrupt disruptions that could be interpreted as a rupture. The seismogenic fault suggested by correlating surface-projected hypocentres and the six fathom isobath (Anglin, 1984), was not evident on the section that crosses the area (Figure 4.4A). The high resolution seismic line did not reveal any rupture. A series of reflectors can be seen near the North Shore, but none appear to be disrupted by a fault. Except for what appears to be a slump, most deformation appears to be related to compaction.

The analysis of some 200 km of SOQUIP seismic reflection lines did not reveal any unambiguous >20 m (20 ms) discontinuity (possibly a surface rupture), under the St. Lawrence River. High resolution seismic profiles support this finding. This implies that the offshore CSZ was probably not the site of a series of M 6 shallow earthquakes in the Quaternary. These large earthquakes would have almost certainly produced a surface rupture detectable on the seismic profiles. Consequently, the CSZ earthquake activity may be relatively young, perhaps a few thousands of years old instead of hundreds of thousands of years. It is possible that the larger earthquakes occur at > 10 km depth, with the near-surface displacement being dissipated on a series of sub-parallel faults. The six-fathom bathymetry line, supposedly making a SE boundary to the seismicity (Anglin, 1984), is not linked with a deep seismogenic fault. Hence, the reference to "a possibly fault controlled fault evidenced in the bathymetry" (Anglin, 1984), should be dropped.

4.3 Focal mechanisms of CSZ earthquakes

4.3.1 Introduction

Traditionally, focal mechanisms of micro-earthquakes have been determined with P-wave first motions, due to the availability of the data (most stations are vertical component only) and to the simplicity of the method. Unfortunately for CSZ studies, events below magnitude mN 4.0 rarely provide enough P-wave first motions for a well-constrained mechanism. With the present rate of mN 4.0 events (seven during the 1977-1997 period), decades would be needed to obtain a comprehensive focal mechanism database. To compute focal mechanisms of smaller earthquakes, two approaches are possible: adding S wave information and/or increasing the number of seismograph stations. Both solutions were attempted in this study. First, the information contained in the S wave was added. SH first motions were preferred to Sv/P ratios, which are usually difficult to assess due to the imprecise onset of the S phases. SH first motions are generally better defined and only require the conversion of the horizontal traces into the transverse component of the S wave. This rotation can be easily applied to the three-component Charlevoix data. Using P and SH first motions, CSZ focal mechanisms can be computed for events larger than mN 3.0. The second approach is to increase the number of seismograph stations, which was done during the 1996 summer field survey (Lamontagne et al., 1997). The following sections review the methodology and the results of these studies. Some re-computed solutions of previously published focal mechanisms are also given. More details on these aspects, including the full descriptions of the mechanisms, can be found in Lamontagne (1998).

4.3.2 Input data and methodology

In the CSZ, a focal mechanism can be determined if the event is sufficiently large (i.e. yields enough first-motion observations) and if three component data are available (i.e. to determine SH first motions). Consequently, only events of magnitude mN 3.0 recorded on the three-component CSZ network were selected (between November 1988 and April 1997). Earthquake locations and focal mechanisms were calculated primarily with the data of the CSZ permanent network (LMQ; A11; A16; A21; A54; A61; A64; Figure 1.8). When available, first motions were added from five additional stations located within 150 km of the active zone (DAQ; SHQ; SLQ; CIQ; QCQ; Figures 1.11). During the period from mid-June to mid-November 1996, smaller events (mN 2.0) were also considered, with data from up to eight additional analogue and digital field seismographs (Lamontagne et al., 1997; Figure 1.12).

All hypocentre locations of this study were routinely computed with CLTN phases only, except between June to November 1996, when additional phases from field stations were added. Earthquakes were located using the "standard" GSC velocity model, which assumes a 36 km thick crust with 6.2 km/s Pg velocity and 3.57 km/s Sg velocity. The lateral velocity contrast introduced by the Appalachian nappes does not change significantly the takeoff angles and azimuths for south shore stations (Lamontagne, 1987).

P-first motions on the digital and analogue records were mostly read directly by the author. SH first motions were read on rotated traces from the three-component stations. In general, SH first motions were less consistent than P-first motions. For example, two focal mechanisms had sufficient P first motions to give a reliable solution (events 960714 18:46 and 960924 23:41, both recorded during the 1996 summer field survey). In the first case, three SH out of seven did not fit the solution, while two out of nine did not fit the second solution. From these results, it was decided that the P first motions were the primary constraints to the focal mechanisms, while SH would further refine the solutions.

The mechanisms were determined with the program FOCMEC of Snoke et al. (1984). The program uses a grid search algorithm that finds all mechanisms that match a given number of P and SH first motion misfits. The program was run interactively, using a one degree B-axis increment. In cases where too many solutions existed, two degree, then, five degree increments were used. The variations in the P, T and B axes were used to quantify the constraints on the nodal planes. The method was inspired by the work of Moustafa (1992). The P, T and B axes were treated as vectors for which average direction, plunge and length were computed. The average length is a measure of the dispersion of the axes, i.e. the best solutions have lengths close to 1. A quality factor was assigned to the solutions using the scheme shown in Table 4.2. The rating scheme agreed with a visual inspection of the focal mechanisms. In a few cases, the P, T and B axes were well constrained but the solutions were rejected (quality "X") based on the poor distribution of first motions. For every event, the description of the mechanism and three figures were provided: the focal mechanisms, the first motion picks and the epicentral map (Lamontagne, 1998; Appendix 1).


Table 4.2 Rating scheme of the focal mechanism

Quality

Considered as:

Average

length

P, T, B

Corresponding variation

in degrees

around average position

Maximum

P

errors

(%)

Maximum SH

errors

(%)

A Reliable

> 0.994

+- 10 <10

<50

B

Reliable

> 0.988

+-15 <25 <75

C

Weakly constrained

> 0.979

+-20 <50 100

X

Rejected

<= 0.979

> +- 20. >50 100

O

Cannot be computed

---------

------- ------- -----

4.3.3 Results

Some 20 new focal mechanisms were computed for the period November 1988 to December 1997. Out of the 27 magnitude mN 3.0 events of that time period, 12 resulted in quality A and B mechanisms (Figure 4.5; Table 4.3). The addition of up to eight seismograph stations during the 1996 Summer field survey, provided data for 15 additional mechanisms, for earthquakes as small as mN 2.0. Of these, eight were of quality A and B (Figure 4.6; Table 4.4). While most mechanisms were calculated for the first time, some had been calculated by other authors (referenced in Table 4.5) and are discussed below. Since Lamontagne (1998), a focal mechanism was computed for the 19971028 mN 4.7 event (Appendix 1). This event was the largest CSZ earthquake since the 1979 mN 5.0 earthquake. Although the oblique-slip mechanism is fairly well defined, it was quality C due to the poor constraints on the P, T, and B axes.

Some previously published focal mechanisms were re-examined. The six mechanisms of Leblanc and Buchbinder (1977) were re-computed using the original data. Two of the six mechanisms are well constrained (74/06/09 and 74/06/23; Figure 4.7), and similar to the solutions of Leblanc and Buchbinder (1977) and Adams et al. (1989). Two mechanisms have two main families of solutions (74/06/20 and 74/06/30) and the last two are very poorly constrained (74/07/02 and 74/07/13).

Due to its seismotectonic importance, the focal mechanism of the 19 August 1979, mbLg 5.0 earthquake was recomputed using the program FOCMEC. Since the original data set was lost, first motion data were assembled from GSC analogue and digital playouts,


Table 4.3: Focal mechanisms of earthquakes of magnitude 3.0 and larger

Q      Date   Time        Latitude  Longitude Depth  Mag  Sta/Pha     Reference

       yymmdd hhmm ss        (N)       (W)    (km)     


A  890131 1439 48.20  47.4426    -70.6710   19.69  3.1mN     7/012  
X  890309 0941 32.26  47.7171    -69.8569   10.52  4.3mN     6/011    1 2 3
X  890311 0831 52.16  47.7182    -69.8699   10.41  4.4mN     7/012   1 2 3
B  891013 1404 42.80  47.3926    -70.1330   22.74  3.2mN     6/012  
A  891122 2302 51.72  47.4559    -70.3420    7.39  3.4mN     7/014  

A  900303 0206 03.38  47.8559    -69.9765   20.85  3.6mN     8/015  	4
A  900313 1910 39.34  47.5338    -70.1366   15.38  3.2mN     7/014  	4
C  900421 0123 04.12  47.5532    -70.0698    9.56  3.1mN     7/014  	5
X  900423 0028 04.78  47.4143    -70.1787    8.04  3.0mN     7/014  		 
A  901021 1338 43.20  47.3975    -70.3644   15.85  3.3mN     7/013  	5

X  901026 0913 51.51  47.5692    -69.9848   10.96  3.1mN     6/012  
X  901106 1130 10.76  47.3943    -70.1506   14.19  3.4mN     6/011  	6
X  901218 0710 46.23  47.2627    -70.3359    9.38  3.3mN     7/013  	6	
0  910703 0926 42.32  47.5290    -70.1464   18.44  3.0mN    CLTN data not saved.
B  911208 0300 30.12  47.7792    -69.8643   23.05  4.3mN     7/013  

X  920310 0545 32.64  47.7167    -69.8574    9.96  3.3mN     7/013  	3
X  920501 0037 51.49  47.4463    -70.4069    2.67  3.2mN     6/011  
A  930304 2202 21.84  47.5145    -70.3621    4.39  3.1mN     6/011  
X  930807 2125 31.92  47.6681    -69.8893    7.75  3.1mN     7/014  
0   931201 1247 15.89  47.4671    -70.1584   18.0g 3.5mN   No CLTN data.

X  931230 2301 47.69  47.4532    -70.3609    5.94  3.8mN     6/012  
A  940925 0053 29.46  47.7518    -69.9612   12.18  4.3mN     6/012  
B  941201 1302 47.14  47.4374    -70.3138   10.77  3.0mN     5/009  
C  960512 1153 21.91  47.5161    -70.0281   14.82  3.1mN     7/014  
X  960607 0941 42.83  47.5299    -69.9417   13.32  3.1mN     7/014  

A  970110 1927 27.56  47.5094    -70.1965   17.06  3.2mN     7/014  
X  970114 0447 32.39  47.6574    -69.8765   14.96  3.1mN     7/014  

1- Drysdale et al. (1989); 
2- Wetmiller and Adams (1990).
3- Li et al. (1995).
4- Drysdale et al. (1990); 
5- Drysdale et al. (1991a).
6- Drysdale et al. (1991b).

Q: Quality.

	A: Very good    
	B: Good
	C: Fair
	X: Rejected 
	0: Cannot be computed.

Total: 8 A; 2 B; 3 C; 12 X; 2 O


Table 4.4: Focal mechanisms of earthquakes recorded during the 1996 summer field survey

Q      Date   Time        Latitude  Longitude Depth  Mag  Sta/Pha

       yymmdd hhmm ss        (N)       (W)    (km)     

A  960617 1118 30.66  47.5328    -70.1463   14.11  1.9mN     8/016  
X  960624 1311 16.55  47.4862    -70.1239   14.36  1.6mN     7/014  
X  960704 1227 07.48  47.6197    -70.1352    5.28  2.0mN     7/014  
A  960714 0715 02.89  47.4829    -70.0503   13.71  2.2mN     7/014  
A  960714 1846 49.22  47.6938    -69.9927    7.26  3.3mN     7/013  

C  960726 1438 45.14  47.6097    -69.9468   10.36  2.3mN     7/014  
X  960819 1706 09.66  47.3002    -70.2448    6.44  2.1mN    11/022  
C  960913 2355 35.92  47.5008    -70.2134   12.09  2.2mN    11/022  
B  960923 0526 54.22  47.6591    -69.8932   14.04  2.2mN    12/024  
A  960924 0644 45.55  47.5875    -70.1451   21.40  2.0mN    12/024  

A  960924 2341 02.88  47.5475    -70.2417   12.83  3.1mN    11/021  
C  960925 0834 24.87  47.8528    -69.7504   22.35  2.2mN     9/018  
C  961011 0228 50.34  47.4761    -70.0551   15.56  1.9mN    11/022  
A  961025 0947 24.43  47.4281    -70.3887    4.06  2.2mN    13/025  
A  961028 0245 39.23  47.5536    -70.0414   11.56  2.3mN    12/024  

Q: Quality.

	A: Very good
	B: Good
	C: Fair
	X: Rejected /  0: Cannot be computed.

Total: 7 A; 1B; 4 C; 3 X.



Table 4.5 Other mechanisms. (1974; 1979 + events near La Malbaie)

Q      Date   Time        Latitude  Longitude Depth  Mag  Sta/Pha

       yymmdd hhmm ss        (N)       (W)    (km)     

B    19740609 2324 30.82   47.3431  -70.2393 10.12  0.6ML 16 / 27 
X    19740620 1836 57.48   47.4041  -70.1802 17.30  1.5ML 18 / 22 
A    19740623 1406 57.40   47.5127  -70.2144 14.95  0.4ML 16 / 28 
X    19740630 1655 11.21   47.7155  -69.8409 15.54  2.0ML 20 / 27 
X    19740702 0730 18.62   47.5641  -70.2271  4.43  0.2ML 13 / 22 

X    19740714 0029 56.78   47.4919  -69.9720 12.71  0.5ML 13 / 24 
B    19790819 2249 30.60   47.6720  -69.9010 10.00  5.0mN  8 / 13 
X    19891208 1720 34.44   47.7009  -70.0644 10.37  2.6mN  6 / 12 
C    19910723 0103 14.31   47.6869  -70.1031 11.31  1.9mN  7 / 12 
B    19930330 0215 18.48   47.6844  -70.1040  4.36  1.8mN  7 / 14 

Q: Quality.

	A: Very good
	B: Good
	C: Fair
	X: Rejected 
	0: Cannot be computed.


from paper playouts of Lamont-Doherty and Weston Observatory and from the ISC Bulletin for regional and teleseismic phases. A total of 45 P-first motions was gathered, slightly more than the 39 used in Hasegawa and Wetmiller (1980). Although both the original and the new solution show predominantly reverse faulting (Figure 4.8), the positions of the nodal planes are somewhat different. The new solution shows one plane steeply dipping to the west and one plane more shallowly dipping towards the southeast.

Some focal mechanisms of mN 3.0 events had already been published in the Canadian Earthquake Summaries (Drysdale et al., 1989; 1990; 1991a; 1991b). Four of the new mechanisms are similar to those previously calculated (890309; 890311; 900311; 900313) while four others are significantly different. The original solutions relied on P first motions and on Sv/P ratios. The differences arose due to the use of emergent P first motions at regional distances in the original solutions.

The four mechanisms of Li et al. (1995) were re-computed using P and SH first motions. In Li et al. (1995), the plotting of the first motions for stations within 150 km was wrong. For direct P and S phases, the azimuths were not reversed, which imply that the mechanisms are questionable especially for the 901021 and 920310 solutions that depend heavily on the CLTN data. Three of the four mechanisms calculated by Li et al. (1995) did not meet our quality criteria and were rejected. Differences were mainly due to our conservative picking of the SH first motions. Our quality "A" mechanism is sensibly different due to the differences in plotting the CLTN data.

For a sub-zone near station A61, Lamontagne and Ranalli (1997) published three focal mechanisms for three small events (magnitude mN 2.6: 891208; mN 1.9: 910703 and mN 2.6: 930330). While the general tectonic style is well defined for the three mechanisms, only the 930330 event is rated B.

4.3.4 Discussion of the focal mechanisms

With the current CSZ seismograph network, earthquake focal mechanisms can be computed for events as small as magnitude mN 3.0, when additional SH constraints are added. Approximately 50% of the mechanisms in that magnitude range can be considered reliable. The eight additional seismograph stations of the 1996 summer field survey added much needed P and SH first motions. They allowed the calculation of focal mechanisms for events as small as magnitude mN 2.0. Some published mechanisms were found to be poorly defined and should be discarded (Figure 4.9). More important, a total of 25 well-defined mechanisms is now available for the CSZ (Figure 4.10).

The faulting style that emerges from the focal mechanism study is mainly thrust to oblique-thrust faulting on fault planes with highly variable orientations. Out of 22 mechanisms of quality "A" and "B", none was of a pure normal-faulting type. The nodal planes do not have any specific trend, only a weak concentration in the NE quadrant (Figure 4.11A). The stress system appears to consist of a shallow-dipping P axis with heterogeneous orientation (Figure 4.11B).

4.4 Hypocentre groups and alignments

4.4.1 Hypocentre groups

A series of earthquakes can reactivate various portions of a given fault. If sufficiently numerous, well-located hypocentres of these earthquakes can define the strike and the dip of the reactivated fault. Various methods have been proposed to compare hypocentre alignments with faults (e.g., Rieken and Thiessen, 1992; Jones and Stewart, 1997; Chapman et al., 1998). For this study, a simple method to define hypocentre alignments was developed. Given a group of hypocentres, each event is associated with its neighbour, located within a given distance. These associated hypocentres are grouped into strings of events (Figure 4.12). A string can be defined as a volume within which each event has at least one neighbour within the given radius. The analysis can define earthquake clusters, alignments, and concentration of activity.

As a first step, areas that include most CSZ events were defined. A data set that insures location completeness was used, i.e. events with ML 0.2 and mN 1.5 (as discussed in Section 1.4.3). The data set included some 1205 hypocentres of the period October 1977 to December 1997. To examine associations between events, a maximum inter-event distance of 2 km was selected, slightly larger than the uncertainty in relative hypocentre positions. Some 918 events (76%) have at least one neighbour within 2 km hypocentre distance, while 287 (24%) occurred as "isolated" events. Eleven groups include ten events or more, representing 54% of all CSZ events (Figure 4.13). Out of the eight magnitude 4.0 events of the 1977-1997 period, two occurred in the most active string of events, one was within a 15-event string; three had between two and six events; and one was isolated. Except for the region surrounding the mN 5.0 hypocentre, most magnitude 4.0 events occur in weakly active areas. About 30% of all CSZ events regroup into one group of hypocentres (Figure 4.14). This group of events defines a faulted volume under the St. Lawrence River with a length of about 40 km, a width that varies between 2 and 11 km, at depth between 6 and 18 km. Variations in strike and dip occur along the length of the fault zone: whereas it is well defined near the magnitude 5.0 earthquake (steeply dipping towards the SE; Figure 4.14D), it becomes progressively shallowly dipping (Figure 4.14C), and diffuse in the middle of the impact crater (Figure 4.1A-B). Inside the group, various smaller alignments appear to exist. Other groups of events define a series of alignments either at an angle with the strike of the St. Lawrence River or as isolated clusters.

In order to delineate more detailed hypocentre concentration, a similar test was conducted for all CSZ hypocentres of the period 1977-1997, recorded on 5 stations with 8 phases, with a shorter inter-event distance (1.5 km). Out of some 1643 events, 517 (31%) had no associated neighbours. About 38% occurred in groups of 10 events or more. Compared with the 2 km groups, smaller groups become evident and display a variety of orientations.

4.4.2 Discussion

Over a 20 year period, approximately one-fourth of the CSZ events with mN 1.5 occurred as isolated events, i.e., without neighbouring events within a 2 km radius. We can infer that stresses take a long time to rebuild along some faults and fractures. On the other hand, under the St. Lawrence River, one-third of all CSZ events, including some magnitude 4.0 events, occurred within a faulted volume of 50 km length, 10 km width, located between 5 and 20 km depth. Although the faulted zone is parallel to the St. Lawrence River, earthquakes define multiple orientations, especially towards the centre of the CSZ. Earthquakes surrounding the magnitude 5.0 event are discussed in detail in Section 4.6.3.

Grouped hypocentres can define a volume, which may be a fault or a weak zone where failure conditions exist. The orientation of these volumes can be compared with the focal mechanisms. Focal mechanisms of events within groups of 2 or more events are shown in Figure 4.15 (groups defined with a 1.5 km minimum inter-event distance). For most focal mechanisms, it is difficult to pick the best nodal plane by looking at the groups of events, either due to their clustered nature, to the small number of events or to the similar orientation of the two nodal planes of any given mechanism. In a few cases, the orientation of the group reflects one of the two nodal planes, providing a way to infer the actual fault trend. In other cases, however, the orientation of the whole group does not correspond to one fault plane, implying that the group defines a weak volume.

4.5 Earthquake multiplets

4.5.1 Introduction

In many seismic areas of the world, earthquakes occur in spatial clusters. Events within these clusters (also called families or multiplets) display highly-correlated seismic traces, caused by repeated slip on the same fault planes. The identification of such clusters represents a real advance in seismology: what appeared as clouds of events can become hypocentres a few hundreds of meters apart (see, among many others, Deichmann and Garcia-Fernandez, 1992; Nadeau et al., 1994 and, for some Charlevoix multiplets, Li et al., 1995). Due to their spatial closeness, the hypocentres of a multiplet can be used to derive the source parametres of the main shock (the effects of the path, site and instrument response being removed; see e.g. Li et al., 1995) and to infer the orientation of the reactivated fault via precise hypocentre locations and possible comparisons with the nodal planes.

Earthquake multiplets were sought among the 915 hypocentres recorded with at least 8 phases from 5 seismograph stations between November 1988 and August 1997 (the data format prior to November 1988 is not currently readable). All existing time series were extracted from the archives of the GSC (a few files were missing before the beginning of continuous archiving, in August 1994). All time series files, originally in GSC MKII or SEED formats, were reformatted in SAC format. The original time series files and the SAC files now exist on one 1.2 Gbyte optical disk.

All available seismic traces were systematically cross-correlated to find earthquake multiplets. The procedure consisted of cross-correlating each of the three-component seismic traces for two events and calculating the normalized correlation coefficient. Two highly similar traces would yield a coefficient approaching one and inversely, two dissimilar traces would give a coefficient approaching 0 (generally below 0.7). A time window was defined to include 0.5 s before the P onset and 0.5 s after the S onset. In most cases, the total trace duration varied between 2.5 and 6.5 s, depending on the hypocentral distance. Stations A61 and A54 were generally chosen due to their lower background noise levels. Figure 4.16 displays a series of traces for events that define a 9-event multiplet.

4.5.2 Results

Out of the 915 CSZ events tested, 170 (19%) turned out to have at least one associated event. These events constitute 67 multiplets, of which 51 are doublets, and five are five-tuplets or more (see Appendix 2 for complete list). During the time period considered, the majority of CSZ earthquakes (81%) occurred as isolated (i.e. non-multiplet) events. For events of the multiplets, the inter-event time varied between seconds and years, with 25 to 30% within the first 0.1 year (Figure 4.17). Spatially, about 90% of the events are within 2 km distance of the first hypocentre (Figure 4.17). Since two events of a multiplet must lie within a few hundred metres of each other (to have similar waveforms), the inter-event distance found is related to the precision of the routine hypocentre locations. When relocated with carefully re-picked phases, most events cluster within 1 km hypocentral distance. It was found that multiplets are more common in some areas than in others. The area near the 1979 magnitude 5.0 earthquake has the highest percentage with 38% of events that had at least one associated event (15 events out of 39). The most active areas of the CSZ, such as the Central and the Ile-aux-Coudres sub-zones, do not have higher proportions of multiplets. This seems to imply that most events, even if spatially associated, are not related to the same fault plane, and consequently, reinforces the conclusion that seismicity occurs in highly fractured volumes.

Unlike grouped events which are only spatially associated, multiplets are events that occurred on the same fault plane or on nearby sub-parallel fault planes. If a focal mechanism exists, this fault plane is one of the two nodal planes. With high precision hypocentres, one can define which nodal plane is the actual fault plane. Ideally, one would have liked to find numerous multiplets with focal mechanisms, but only five multiplets included a quality A or B focal mechanism (Figure 4.18). Of these, only one had more than two associated events (911208 03:00 U.T mN 4.3). This event is discussed in more detail in Section 4.6.3.2.

4.6 Integration of focal mechanisms, earthquake groups and multiplets

The previous four sections have provided the basis of a search for a correlation between earthquakes and faults in the CSZ. Without a surface rupture to directly correlate earthquakes and faults, 25 focal mechanisms can be compared with earthquake groups and multiplets. A similar approach was used in the eastern Tennessee seismic zone by Chapman et al. (1997).

To examine the local variations in CSZ seismotectonics, 19 sub-zones were defined on the basis of epicentral clustering and on common depth distribution (Figure 4.19). These sub-zones are re-grouped in the next three sections according to the value of the median of their respective focal depth (Table 4.6): shallow (z < 10 km; Section 4.6.1; Figure 4.20A); intermediate (10 z < 15 km; Section 4.6.2; Figure 4.20B); and deep (z 15 km; Section 4.6.3; Figure 4.20C). The weakly seismic sub-zones are examined separately (Section 4.6.4).

4.6.1 Sub-zones with shallow activity (A61, COU, EBO)

4.6.1.1 Earthquakes of the A61 sub-zone

A sub-zone near seismograph station A61 (thereafter named sub-zone A61) was defined for a pilot-test of our study of other sub-zones. The A61 sub-zone is a cluster of events within a few kilometres of station A61. The boundaries correspond to a change in focal depths to the southeast and to a decrease of earthquake activity elsewhere. The following section updates the description of the sub-zone given in detail in Lamontagne and Ranalli (1997).

The sub-zone presents many favourable characteristics for a detailed seismotectonic study: (1) good constraints on relative focal depths (probably better than ±1 km due to the proximity of station A61); (2) a large proportion of shallow hypocentres (90% at less than 12 km depth) located onshore providing the potential for stronger correlations with remote sensing lineaments than in most of the CSZ; (3) location within the sub-zone of an onshore earthquake of magnitude mN 4.0 (most others mN 4.0 have occurred under the St. Lawrence River in areas known for their concentration of larger events); (4) two focal mechanisms available; (5) occurrence of faults related to both major structural systems, i.e. paleo-rift faults which are assumed to bound the earthquake activity (Anglin, 1984) and faults of the Devonian meteor impact structure.


Table 4.6 Statistics of the various sub-zones

Zone Number % Depth Magnitude No. No. of of of (km) Min Max Median M4+ Multiplets (1) events(2) CSZ Min Max Median (2) = doublets

(3) = triplets, etc.

CSZ 1638 3.4 0.0 30.0 11.5 -1.5 5.0 0.8 8 A61 88 5.4 0.0 18.2 9.4 -0.8 4.0 1.1 1 1(2); 1(3), 1(4) COU 182 11.1 1.4 22.3 8.3 -0.8 3.8 0.9 0 10(2); 2(3) EBO 53 3.2 2.1 21.6 6.3 -1.3 3.1 0.9 0 5(2) MA5 83 5.0 2.0 16.2 10.6 -1.0 5.0 1.6 4 6(2); 1(3)
CEN 433 26.5 1.1 27.1 11.0 -1.0 3.6 0.6 0 11(2);1(3); 3(4) KAM 153 9.3 6.5 24.9 12.5 -1.0 3.1 0.7 0 6(2); 1(5) POC 116 7.1 0.0 30.0 13.2 -1.1 3.4 0.8 0 1(2); 1(6); 1(9) IRE 202 12.3 0.0 22.4 11.3 -1.5 3.7 0.6 0 7(2); 1(3); 3(4) OIE 65 4.0 8.0 23.0 15.7 -1.0 3.9 0.4 0 0
DNE 91 5.6 6.4 27.5 21.6 -1.0 4.3 1.2 2 4(2); 1(5); 1(6) SHI 56 3.4 1.4 28.5 10.7 -1.0 3.1 1.0 0 0 AIG 9 0.5 16.8 28.7 21.2 -0.3 2.7 1.2 0 0 SGA 13 0.8 6.1 17.7 11.4 -1.2 2.6 0.0 0 0 CGA 19 1.2 1.5 26.5 19.6 -0.6 2.5 0.8 0 0
NGA 13 0.8 7.2 23.0 14.2 -1.3 3.4 1.1 0 0 UPS 18 1.1 6.4 29.0 13.5 -0.2 2.2 1.4 0 0 DOW 4 0.2 18.6 24.0 22.9 0.1 2.4 1.6 0 SOS 17 1.0 9.9 25.4 14.7 -0.5 2.4 0.5 0 LOU 30 1.8 3.1 26.7 9.1 -0.6 4.2 0.8 1

A total of 88 seismic events were located in the A61 sub-zone, or about 5% of some 1644 CSZ hypocentres. In the sub-zone, earthquake epicentres can be divided into an eastern concentration (all located within 5 km of station A61), and a diffuse western group (Figure 4.21). In contrast with the CSZ as a whole, which shows a continuous increase of activity between 0 and 9 km depth, focal depths in the A61 sub-zone show a bimodal distribution which peaks at about 5 and 9 km (Figure 4.22). Compared with the activity of the CSZ, the 5 km group can be considered shallow, while the 9 km group is average. It is not known if this bimodal distribution is due to the short period of observation (i.e. the 8 km low might disappear over a longer time period) or to physical processes that cluster events at 5 and 9 km.

As a first step to evaluate clustering, routinely-determined hypocentre locations were compared to evaluate their spatial proximity. It was found that about 45% of events have at least one other hypocentre within 1.5 km. The rest, about 28%, occurred as isolated events. Ten clusters of two events or more are found, with one group of 31 events and one of nine. Surprisingly, the largest earthquake (mN 4.0) has no event within 1.5 km of its focus, which implies that it was not preceded by foreshocks or followed by aftershocks.

Three multiplets were found in the sub-zone. In general, the normalized coefficients were lower for the vertical component than for the horizontal components, suggesting scattering of the incident Pg wave. The small magnitude of these events makes P arrivals emergent on most other stations. One quadruplet (891208; 891221; 940614; 960215 with a focal depth of 10 km) includes an event with a computed focal mechanism (event 891208). Routinely-computed hypocentres are very similar for these events. While most traces are consistent for the three events, some are different implying variations in source mechanisms. Although it can be assumed that most of these variations are due to variable focal mechanisms, the similarity of traces on some stations and differences on others suggests that velocity complexities near the hypocentres also play a role. Hence, for some azimuths, small changes in hypocentral locations can produce travel path changes. This quadruplet may have occurred near a lithological complexity, such as a geological contact or a wide fault zone. Traces are extremely similar for events of the triplet (dates of events: 930706; 950303 and 950820 with a focal depth of 9 km). Routine hypocentre locations also show that they are within 1 km of each other. One doublet was also found.

The multiplets represent a total of 9 events out of 47 events tested (or about 19% of the total). Thus, over the seven year period, only a minority of events occurred as consistent reactivations of the same fracture planes. In the A61 sub-zone, very few events could have occurred on the same fractures with similar focal mechanisms. A number of isolated events was found. The time intervals of up to 4.5 years between events of multiplets may indicate a slow strain build-up on a given fracture.

An interesting offshoot of the method described above is the discrimination between seismic events and quarry blasts, as previously applied in Mount St. Helens (Frémont and Malone, 1987). Three quarry blasts were detected by their highly similar waveforms (average normalized correlation coefficients greater than 0.9; Figure 4.23).

In the area surrounding the A61 sub-zone, a very dense system of faults of various orientations and nature exists (Figure 4.24), most of them corresponding to faults either mapped in the field or photo-interpreted by Rondot (1989). In the A61 sub-zone, the focal mechanisms of the two largest earthquakes (11 January 1986, mN 4.0 and 18 March 1987, mN 3.3) show mostly reverse faulting in agreement with most other solutions of the CSZ. In both cases, a nodal plane is sub-parallel to the conspicuous WNW-ESE set of lineaments (Figure 4.24). Thus, it is plausible that these earthquakes represent reactivations of this series of faults especially when one considers the shallow focal depths of these two events (in the 4-5 km range) and the well-constrained nodal planes based on first motions.

Three additional focal mechanisms have been computed for small magnitude events based on first motions and SH polarities. These focal mechanisms are relatively well constrained and show strike-slip-reverse, reverse, and normal faulting respectively. While the reverse faulting type agrees with most mechanisms of the CSZ, dominant strike-slip components and normal faulting are less common. First motions on A61 indicate that the strike-slip component and/or normal faulting types are not unusual: many events had dilatational first motions on station A61. Dilatational first motions in the centre of the focal sphere are inconsistent with pure reverse faulting mechanisms. In addition, a composite mechanism of all earthquakes of the sub-zone shows a large scatter in first motions incompatible with a unique focal mechanism for all events (78 out of 264 or 30% of all readings are misfits to the best solution). Faulting complexity appears to be present everywhere in the sub-zone; events located within a few kilometres of each other are found to have very different first motions and transverse SH on A61.

For the period 1978-1982, hypocentres along the north shore of the St. Lawrence define steeply-dipping volumes bound by paleo-rift faults (Anglin, 1984). The present study suggests that within the A61 sub-zone, numerous preexisting fractures are being reactivated by local stress and/or strength variations. This WNW-ESE fault system is older than the paleo-rift faults (Rondot, 1979), but was reactivated by the post-impact readjustment as were many faults of the CSZ (Roy, 1978). These peripheral impact structure faults are part of an inward-dipping system of normal faults. In the A61 sub-zone, the south-dipping nodal planes should be the fault planes. Hence, it appears that some impact structure faults may be reactivated in the A61 sub-zone.

When earthquake concentrations are seen in a data set, the hypocentre precision is always questioned. The earthquake concentrations of this study are not an artifact of location uncertainty. In the A61 sub-zone, highly-correlated traces were found only for events that had routinely-determined hypocentres within 1 km. Thus, these hypocentres appear reliable, probably enough to examine possible correlations between earthquake alignments and lineaments.

Unlike in most of eastern Canada, focal mechanisms of the A61 sub-zone are not pure reverse faulting. While focal mechanisms of two earthquakes with magnitude larger than 3 show reverse faulting, smaller magnitude earthquakes indicate variations in the faulting style, evidenced by the mechanisms of some events (some strike-slip component and normal faulting), and by the rather inconsistent first motion distribution on A61 for the composite mechanism.

4.6.1.2 Ile-aux-Coudres sub-zone (COU)

The hypocentres of the Ile-aux-Coudres sub-zone occur from 2 km down to 22 km depth, with peaks at 5-7 km and at 10 km depth (Figure 4.20A). The four focal mechanisms display a mixture of pure reverse faulting on NE planes to strike slip mechanism (Figure 4.10). The area appears to be heterogeneous in terms of stress and reactivated fault planes. No nodal planes are oriented parallel to the NE trending paleo-rift faults with their steep dip to the SE. The deepening of the SE activity in respect to the NW may correspond to the Charlevoix fault, described in Chapter 3.

4.6.1.3 Les Eboulements sub-zone (EBO)

The Les Eboulements sub-zone is only midly active (3% of CSZ) and is the shallowest sub-zone. The sub-zone includes one of the best epicentre alignment of the CSZ. Most events occur within a 4 km wide N-S corridor which does not correlate with any lineament in the Radar imagery. It is possible, however, that the earthquakes occur within a highly fractured layer, sub-paralell to the N-S magnetic lineaments to the North of the sub-zone. The only focal mechanism has one nodal plane oriented N-S and steeply dipping to the west, consistent with a group of 10 earthquakes located within 1.5 km of each other. Hence, in this sub-zone, most earthquakes occur within a N-S volume, possibly lithologically controlled, which includes a fault with similar orientation. Earthquake groups suggest that this zone may continue under Ile-aux-Coudres.

4.6.2 Intermediate depth sub-zones (MA5, POC, IRE, CEN, KAM)

4.6.2.1 The Magnitude 5 sub-zone (MA5)

The Magnitude 5 sub-zone was named after the largest earthquake of the 1977-1997 period, the August 19, 1979 mN 5.0 earthquake. The sub-zone is bound to the South by a cluster of earthquakes and by deeper activity elsewhere. With only 5% of the CSZ activity, the sub-zone is not the most active, but four of the eight mN 4.0 earthquakes of the 1977-1997 period occurred there. In addition, an anomalously high proportion of multiplets were found: 15 events out of 37 (40%). Four mechanisms are available, all for magnitude > 4.0; the mN 5.0 event with its pure reverse faulting mechanism; a strike-slip solution for the 19971028 mN 4.7 event; and two reverse faulting solutions for the mN 4.3 and 4.4 doublet of March 1989. On a cross-section perpendicular to the River, the hypocentres define a steeply-dipping plane (Figure 4.14A).

Based on available information, it is likely that the 1925 M 6.2 earthquake occurred in this sub-zone. It is known that the earthquake occurred within the NE portion of the CSZ at a 10 km focal depth, that its magnitude defines fault dimensions of about 8 length by 4 km width, and that it was a pure-reverse faulting event on a plane oriented NNE to NE. The M5 sub-zone matches all these parameters: it lies within the epicentral zone of the 1925 earthquake, earthquakes concentrate at 10 km depth, and hypocentres define a plane with a trend and dimensions that match what is expected of the 1925 event. In addition, the sub-zone is subject to some of the largest earthquakes of the CSZ, and three out of four mechanisms have more-or-less the same reverse faulting mechanism as the 1925 event.

On October 28, 1997, a mN 4.7 event occurred, the largest CSZ earthquake since the 19 October 1979 mN 5.0 earthquake. The main shock was followed by a series of ten aftershocks (as of May 1st 1998; Table 4.7), all within 3 km epicentral distance of the main shock. Within the aftershock sequence, a doublet and a triplet were found. The largest aftershock of the sequence (magnitude mN 3.2) occurred 90 seconds after the main shock. Its exact location in respect to the main shock is uncertain due to overlapping traces. Most of the other events were well recorded by the four nearest stations (A16, A21, A61 and A64), and were carefully picked to relocate the main shock and aftershocks (Figure 4.25). The aftershocks are all 0.9 to 3.4 km shallower than the main shock, and are well outside the dimensions of the rupture for an event of this size (in the 600 m range; Table 3.3). Interestingly, aftershocks of the nearby 1979 mN 5.0 event were also shallower than the main shock, and also outside the immediate rupture area (Hasegawa and Wetmiller, 1980). A possible upward migration of fluids to trigger events outside the rupture zone, and possibly on other fault planes can be envisaged, especially if the surrounding volume is highly fractured.


Table 4.7 Aftershocks of the 19971028 event

Date Time Lat Long Depth Mag (UT) (km) 97/10/28 11:44:18 47.67N 69.91W 11.3 4.7mN
97/10/28 11:45:48 47.67N 69.91W ??? 3.2mN
97/10/28 15:14:10 47.67N 69.91W 9.4 1.5mN
97/10/28 16:54:03 47.68N 69.91W 10.5 2.5mN
97/10/29 13:25:15 47.67N 69.91W 8.2 1.3mN
97/10/31 10:19:18 47.67N 69.91W 8.6 -1.0ML
97/11/02 00:00:30 47.66N 69.92W 9.2 1.8mN
97/11/28 10:56:35 47.68N 69.91W 9.1 2.4mN
98/01/12 02:49:25 47.66N 69.90W 7.9 1.3mN
98/03/10 13:28:43 47.68N 69.91W 10.4 1.0mN
98/03/18 18:21:06 47.67N 69.92W 8.3 1.2mN


4.6.2.2 The St-Irénée sub-zone (IRE)

The St-Irénée sub-zone is fairly active with 12% of the CSZ earthquakes. The earthquakes, however, are of low magnitude with only two events exceeding magnitude mN 3.0. The area is bound to the SE by the St-Laurent fault. Only one focal mechanism exists for the sub-zone, displaying pure-reverse faulting on planes oriented ENE. About 20% of events (20 out of 98) were multiplets. The low magnitude of the events suggests a highly fractured area, possibly related to the meteor impact.

4.6.2.3 The Central (CEN) and Kamouraska (KAM) sub-zones

The CEN and KAM sub-zones group approximately 35% of all CSZ activity, with 25% only for CEN. Contrasting with this high level of earthquake occurrence, both areas did not witness any magnitude > 4.0 for the period 1924-1997. The CEN sub-zone is an area of diffuse activity, where earthquakes concentrate but do not define planar features (Figure 4.14A-B. The KAM sub-zone, on the other hand, was defined primarily on the more homogeneous and slightly deeper depth distribution than the one of the CEN. The hypocentres of this sub-zone define a SE dipping plane, very similar to one of the nodal planes of the 1925 earthquake (Figure 4.14 C-D).

4.6.2.4 The La Pocatière (POC) sub-zone

The POC sub-zone, an area of mild activity, is made up of two main clusters of events. As opposed to the CEN sub-zone, it does not show one large hypocentre volume. It appears to mark a transition between the highly active central zone to a less active area to the SW. A total of 17 events out of 66 were part of multiplets (26%), including a nine-event multiplet (the largest earthquake swarm found), and a six event multiplet. All events of these two multiplets were of small magnitude (mN 2.2), hence none has a focal mechanism.

4.6.3 Deep activity sub-zones (OIE, DNE)

4.6.3.1 The Cap-aux-Oies sub-zone (OIE)

The Cap-aux-Oies sub-zone lies along the trend of the midly active corridor that parallels most of the north shore of the St. Lawrence see (Section 4.6.4). The sub-zone has a radial trend from the central peak of the impact crater. The hypocentre groups suggest that the NE planes may be the reactivated structures. Four focal mechanisms exist for this sub-zone with two showing a reverse-oblique-slip mechanism on NE to SW oriented planes. No multiplets were found in this sub-zone.

4.6.3.2 The Deep Northeast sub-zone (DNE)

The Deep Northeast sub-zone includes a series of deep earthquakes located at the northeast extremity of the CSZ (Figure 4.26). The earthquakes are markedly deeper than in any other sub-zone of the CSZ. The depth of these earthquakes is not due to its location in respect to the seismograph network: the emergence angles on station A64 agree with deep focii. On land, the shallow depths to the West correspond to the Ste-Mathilde fault of Rondot (1989). Although most events occur as isolated earthquakes spread out over the whole sub-zone, there are a few clusters of activity, including some multiplets that represent 25% of the events. The largest cluster started with the 19911208 mN 4.1 earthquake, with an unusual strike-slip mechanism. It was followed by a series of aftershocks, including two within the first 6 hours, that constitutes a seven event multiplet. The Pg and Sg arrival times at four stations were carefully re-picked and the events were relocated using time corrections defined by the residuals of the main shock (as in Deichmann and Garcia-Fernandez, 1992). The hypocentres suggest that the NNE-SSW plane is the actual rupture. Two other mechanisms show pure-reverse faulting on NE oriented planes. These fault trends correspond to the numerous lineaments of the north shore.

4.6.4 Weakly seismic sub-zones (SHI, SGA, CGA, AIG, NGA, UPS, DOW, SOS, LOU)

A number of weakly active sub-zones exists in the CSZ, especially in a river-parallel corridor (SGA, CGA, AIG, NGA) or in the periphery of the more active zone (UPS, DOW, SOS, SHI). All together, only 11% of all CSZ earthquakes occur there, with most events representing isolated occurrences (no multiplets were found). Three focal mechanisms exist, showing mostly pure or oblique reverse faulting. At present, there is no evidence for a lithological control of the river-parallel corridor, but the NE boundaries of the SGA, CGA, AIG and NGA sub-zones correspond to changes in hypocentre density and depth, coincident with the position of the St-Laurent fault. The relative quiescence of this strong region may be due to an undetected time dependence of micro-earthquake activity (Wetmiller and Adams, 1990). More simply, the seismic quiescence may be due to the absence of high pore fluid pressure or of gouge material with low coefficient of friction.

4.7 Discussion

In this chapter, four sources of information were examined to bring to light the relationships between CSZ earthquakes and faults. No surface rupture was found at the bottom of the St. Lawrence on the seismic profiles. Future work with high resolution profiles will bring a more definite answer on this. An absence of surface rupture implies a lower rate of activity during the past few thousands of years, or that earthquakes were too small and/or too deep.

For larger events of the CSZ, such as the 1925 M 6.2 and the 1979 mN 5.0 earthquakes, focal mechanisms suggest that N to NE trending paleo-rift faults are reactivated in response to regional stresses. On the basis of earthquake depth, focal mechanisms, and hypocentre alignments, the area surrounding the 1979 mN 5 earthquake may represent the source region of the 1925 event. This area has been subject to repeated M > 4 events and is a likely candidate of future large CSZ earthquakes. Although this area is the most likely to have a surface rupture, none was found on the seismic lines that cross it.

Concerning the possible correlations between seismic events and structural trends, some earthquakes do spatially relate to remotely-sensed and/or potential field lineaments, but not exclusively to paleo-rift faults. Although the A61 sub-zone is intersected by a regional rift fault (the Rang-Ste-Mathilde fault of Rondot, 1989; Figure 4.24), nodal planes of the two largest earthquakes are not parallel to it. For lower magnitude earthquakes, smaller fractures with various orientations might be reactivated by local stresses, as shown by the variety of focal mechanisms and the scatter in P and T axes. In the A61 sub-zone, two of these events have clear orientations parallel to a WNW-ESE series of lineaments possibly reactivated by the Devonian meteor impact. Since not all CSZ impact structure faults are currently being reactivated, it is possibly the combination paleo-rift--impact faults that gives rise to the seismicity. The St-Laurent fault, evident in the remote sensing as well as in the Bouguer gravity anomaly, is a clear boundary to the seismicity of the sub-zones IRE and A61. Southeast of the fault, earthquakes are more dispersed. In general, earthquakes do not concentrate on this paleo-rift feature.

For small magnitude events, the varied faulting styles imply that local stress and/or strength variations control earthquake occurrence. This contradicts the general assumption that background seismicity corresponds to regional stress systems. In our opinion, earthquakes controlled by local conditions are more likely where resistance to sliding is low, for instance as a consequence of high background pore-fluid pressures or low friction coefficient. Interestingly, a high electrical conductivity zone has been detected by a magnetotelluric survey close to the eastern concentration of hypocentres of sub-zone A61 (Chouteau, 1985). This anomaly has been explained by the presence of water or solutions in a zone of high porosity at depth deeper than 1400 m. In the CSZ, two aftershock sequences may be explained by the migration of fluids away from the immediate rupture of the main shocks.

The complexity of CSZ seismicity can tentatively be explained by a combination of factors. In the A61 sub-zone, evidence of lithological complexity near faults is accompanied with variations in fault plane trends and faulting styles. This suggests that CSZ fault zones are irregular surfaces, surrounded by highly fractured rocks. These highly fractured zones respond primarily to regional stresses; however, for some smaller events, they may respond to local changes in stress and/or strength. The whole process can be enhanced, especially for deeper events, by high pore-fluid pressures as discussed in Chapter 2 and in Lamontagne and Ranalli (1996). The concentration of CSZ events near 10 km depth (Figure 4.20B) may indicate a highly stressed level or a frictionally weak level, possibly due to high pore-fluid pressures or to highly fractured rocks.

The small proportion of multiplets found in the CSZ (19%) contrasts, for example, with the Parkfield area of California where most earthquakes occur as clusters (63% for the period 1987-1992; Nadeau et al., 1995). This shows that stress takes a long time to replenish along most faults of the CSZ and that most sub-zones of the CSZ are intensely fractured.


1. For November 1988-August 1997.

2. The seismicity numbers refer to events recorded between October 1977 and December 1997 on five stations with a minimum of 8 phases.


> > > To Chapter 5


Figure Captions

Figure 4.1 Schematic cross-section of the type of rupture expected in the CSZ. Layers A, B, C, D represent sediments accumulated during and between glacial episodes. Multiple reverse faulting events originating in the Precambrian basement rupture propagate to the surface (possibly through the Appalachian sequence). The rupture reaches the thick sequence of sedimentary layers. Since more earthquakes ruptured the basal layers (A for example), the amount of slip is larger than for the top layers. The rupture may reach the bottom of the St. Lawrence River, causing a bathymetric feature.

Figure 4.2 Example of a seismic section (SOQUIP line 22). The profile shows continuous, uninterrupted reflectors (layer in blue) that we interpret as an absence of rupture. The interface between the Precambrian (P) and the Appalachians (A) is shown by an arrow.

Figure 4.3 Location map of the seismic lines acquired by SOQUIP and by their partners. Lines in red show portions of profiles where sub-horizontal markers can be seen. These areas are interpreted to be free of > 20 m ruptures.

Figure 4.4 Seismic profiles that show the Quaternary sequence: (A) Enlarged view of the SOQUIP seismic lines near the North shore: Line 35 (top) and line 37 (bottom). These two profiles cross the region with most magnitude > 4 events. The sediment filled valleys and its sub-horizontal reflectors can be seen. (B) The high resolution profile acquired along SOQUIP Line 13. The reflector in red represents the bedrock. Reflectors in colour represent Quaternary layers that can be distinguished in the numerous multiples. The blue rectangle represents the area where the 6 fathom line, supposedly linked to a deep seismogenic fault (Anglin, 1984) lies. No rupture can be seen within the rectangle.

Figure 4.5 Lower hemisphere focal mechanisms of the 12 quality A and B focal mechanisms for events of magnitude mN 3.0. The maximum and minimum pressure axes are shown as (p) and (t), respectively and the null axis as (b). The dates of the events are shown above the mechanisms; ( C): Compressional and (D) Dilatational first motions; (<) and (>) are the SH first motions.

Figure 4.6. Focal mechanisms of the 8 quality A and B focal mechanisms for 1996 summer events. Legend as in Figure 4.7

Figure 4.7 The two well-constrained focal mechanisms for the 1974 field survey (5 b-axis search). Legend as in Figure 4.7

Figure 4.8 Focal mechanisms of the 1979 mN 5.0 earthquake: (A) as published in Hasegawa and Wetmiller (1980); (B) The revised version of Lamontagne (1998).

Figure 4.9 Revised focal mechanisms of Figure 1.24. Poorly constrained mechanisms of 1974 are crossed while the two acceptable ones are checked. Events 890309 and 890311 were found to have one poorly constrained nodal plane (shown as ?) and were rated "X". For the 1979 mN 5.0 event, the two nodal planes found in this study are shown on the previous solution. The other mechanisms were not reexamined.

Figure 4.10 Focal mechanisms of quality "A" and "B", including the mechanisms of the 1979 mN 5.0 event, the two 1974 events, and the event near La Malbaie. In addition, the 1925 focal mechanism (Bent, 1992) is shown in the upper right corner, in the general area where its epicentre was located by Stevens (1980). An arrow shows the location of a few mechanisms.

Figure 4.11 Results from the quality "A" and "B" of Figure 4.8 (A) Nodal planes; (B) P, T and B axes.

Figure 4.12 Method of defining earthquake groups. Hypocentres (dots) are grouped if located within a given distance (horizontal line) from each other. Although this figure shows epicentres for simplicity, the association is made according to the hypocentral distance. (A) and (B) are groups of events, whereas ( C) and (D) are isolated events. The hypocentres of group (A), define a preferential orientation shown by the dotted line. This trend can be compared with the available focal mechanisms of events of group (A).

Figure 4.13 Hypocentre groups with 10 events or more created with an inter-event distance of 2 km.

Figure 4.14 Cross-sections across the largest group of hypocentres. The four cross-sections A, B, C, D are shown in the epicentral map above.

Figure 4.15 Earthquake groups (1.5 km distance) with a focal mechanism (red dot). When possible, the mechanism is located as to line up one of the nodal planes with the orientation suggested by the group.

Figure 4.16 Traces of a nine-tuplet (earthquake swarm) recorded on station A54 between July 1994 and October 1996: (A) Vertical component; (B) North-South component; (C) East-West component.

Figure 4.17 Characteristics of the events of multiplets. (A) Time between the first event and the other members of the multiplet (in blue) and between events of the multiplets (in red). Events are grouped in 0.1 year time intervals. (B) Hypocentral distance between the first event and the other members of the multiplet. Events are grouped by 0.5 km distance, absolute distance in red; cumulative percentage in blue.

Figure 4.18 CSZ Multiplets of the period November 1988 to August 1997 (yellow); focal mechanism are centered on the event epicentre (red).

Figure 4.19 Earthquake hypocentres and sub-zones. The sub-zone numbers refer to the acronyms listed below. Table 4.7 lists the earthquake statistics of the sub-zones.

Figure 4.20 Focal depth distribution for the various sub-zones. The number of earthquakes per 1 km depth intervals are presented as absolute numbers (thin lines) and cumulative percentage (thick lines. Sub-zones with earthquakes mostly: (A) Shallow; (B) Intermediate; and (C) Deep.

Figure 4.21 Hypocentres of the A61 sub-zone in map view (top), W-E (below left) and N-S (below right) cross-sections. The map view presents focal mechanisms hypocentres with depth colour scheme similar to Figure 4.19. Red dots indicate the positions of events belonging to earthquake multiplets.

Figure 4.22 Compared depth distributions in absolute numbers and cumulative percentage between CSZ (in blue) and sub-zone A61 (in red). The cumulative curves show that the A61 sub-zone events are generally shallower than the CSZ earthquakes (maximum depths of 19 and 30 km respectively).

Figure 4.23 Vertical seismic traces of three quarry blasts recorded on station A61 and recognized by their highly similar traces.

Figure 4.24 Airborne Synthetic Aperture Radar (SAR) image of the A61 Sub-zone (green polygon) with earthquake epicentres of the period 1978-1995 (red dots). Lineaments oriented approximately E-W (shown by the dashed lines to the left) are sub-parallel to the strike of two nodal planes (blue lines) for the two largest events of the sub-zone. The WNW-ESE lineaments and their convergence towards the sub-zone are clearly seen. The Rang Ste-Mathilde fault (Rondot, 1979) is shown and is sub-parallel to one of the nodal planes of a focal mechanism.

Figure 4.25 The 19971028 aftershock sequence. Map view (above): epicentres of the 19971028 event (red dot), of its aftershocks and focal mechanism of the main shock. (Below): West-East cross-section that shows the shallower locations of the aftershocks and its possible link with the east-dipping nodal plane.

Figure 4.26 Hypocentres of the DNE sub-zone. (Above) Epicentral map of the DNE sub-zone with focal mechanisms. (Below left) West-East cross-section; (Below right) North-South cross-section. Earthquake multiplets are shown in red on the cross-sections.