CHAPTER 4

Hypocenters and focal mechanisms of the Arctic earthquakes

Two possible solutions to the problem of describing the pattern of earthquakes in the Arctic Region are discussed. In previous summaries on seismicity of the Arctic, a number of seismically active zones of different intensity, shape and dimensions were used. The main zone was the Mid-Arctic Earthquake Belt extending from Iceland through the Norwegian-Greenland Basin to the Lappet Sea shelf, and farther penetrating into the Asian continent in the form of a wide band. Lower order zones have been established in some marginal seas, Northeastern Russia, Canadian Arctic Archipelago, etc. (Fig.6). In our view, outlining the data, such as hypocenters and focal mechanism solutions, is convenient and instructive if it is described by subregions. These subregions are based on a number of tectonic and geophysical indicators. In addition, each study region elucidates the tectonic setting and deep structure to the extent necessary for further understanding of its geodynamics and estimation of the tectonic nature of modern seismicity.

4.1. Basic principles of data selection

The Arctic Seismological Data Bank was used as a source of information on earthquakes. A map of Arctic epicenters (Fig.6) was computer generated on the basis of the General Catalogue. The data selection were based on the principle "better less but better". This principle was implemented so that earthquakes occurring since 1970 have been used primarily because the network of Arctic stations was more reliable. Only the strongest were taken from the earliest events. Usually, earthquakes epicenters recorded by at least 10 stations were plotted on the map. However, it should be noted that in two cases in order to reflect the true level of seismic activity in different Arctic regions this condition was modified.

Using the uniform approach the map summarizing all earthquakes including northern Alaska would suggest a predominance of seismic activity, which according to data on strong earthquakes, is not true. The better observational network in Alaska prejudices the data distribution. Therefore, for northern Alaska, the threshold value of the number of recording stations was raised to 15. According to ISC, location error for earthquakes epicenters meeting these conditions (min.10 and 15 stations) does not exceed 15-20 km.

For areas with poor coverage such as Baffin Island, Baffin Bay, the Queen Elizabeth islands, and western Greenland, the fulfillment of the condition of at least 10 recording stations would lead to distorted vision of fairly low seismic activity in these regions or even complete absence thereof. Thus the limit was reduced to 4-5 stations. Location error here reaches 50 km or greater.

It should also be noted that the above map does not contain weak earthquakes from northern Fennoscandia, Yakutia, and southern Laptev Sea. This information is included in the regional not in the General Catalogue. Earthquakes from the Chukchi Peninsula and adjacent offshore areas recorded only by the Iultin station are also not drawn on our map. All these earthquakes are shown on maps compiled for epicenters of the appropriate regions.

Some 300 focal mechanism solutions are known for Arctic earthquakes. However, the number of events is less than 200 because for some of them up to 5-6 determinations have been made by different scientists. There are several solutions based on application of different wave types.. In our national practice (and abroad until recently) most values were determined by A.V.Vvedenskaya's method (1972) where P-wave first motions are used. It should be noted that this method requires a vast reliable coverage (at least, 30-40 stations) at a wide range of epicentral distances and sufficiently even distribution of stations through all quadrants around the earthquakes. For the Arctic region where the observation network is limited and irregularly distributed, this method has produced many unreliable and frequently opposite solutions for the same earthquake. Therefore, to reduce the ambiguity we have reviewed the information on focal mechanisms. Based on the review we selected the most completely recorded earthquakes. Signs of first motions were checked using national and foreign bulletins with computerized redetermination or determination of the mechanisms through a consistent method designed at the Institute for Physics of the Solid Earth of the Russian Academy of Sciences, and employed for bulk determinations (Aptekman and others,1979). It is noteworthy that for strong earthquakes, dozens of alternative solutions are often possible of which each has different probability. For some events, when the above procedure failed, the best results were taken from various sources based on the following criteria: additional information recorded by the author, particularly on the data obtained from the local observation network and information on signs of first motions directly from seismograms rather than from bulletins. For instance, for earthquakes of northern Yakutia and the Laptev Sea, on which determinations by foreign and Yakutian seismologists are available, the latter were preferable because they had data collected from local stations. All available solutions using the first motions method are shown on the maps.

Fig.6 The map of earthquake epicenters of the Arctic and adjacent regions

Fig. 7 Earthquake epicenters and major structural bottom features of the Norwegian-Greenland Basin and adjacent regions.

Since 1982, the ISC Catalogue has been publishing data on focal mechanisms obtained using centroid-moment tensor method (CMT). Similar determinations have been also published for earlier periods beginning 1977 (Dziewonski and others,1981). CMT method is based on conversion of the complete package of waves recorded by a digital network, ranging from compressional to surface waves, so the process in the source is averaged and the major stage in the fracture is characterized whereas the mechanism determined from first motions of compressional waves corresponds to its initial stage. Therefore, the data obtained from both determinations will only coincide if the process is uniformly developed in time, and their comparison without taking into account the above would be incorrect. Thus, all solutions using CMT method have been plotted on separate maps. A similar approach has been employed for outlining the material regionally.

4.2. Norwegian-Greenland Basin and surrounding areas

A specific feature of earthquake distribution in this area is the presence of a narrow epicentral band which is coincident with the mid-oceanic ridge and related faults. The contemporary axis of the ridge is clearly defined by a positive gravity anomaly of more than 50 mGal within which, however, an axial minimum is observed on the sites having distinct rift valley.

Major intersectional faults split the ridge into three basic segments separated from each other either by a distinct lateral offset or by a drastic azimuthal alteration and geomorphologic rearrangement (Fig.7). This causes the separation of each ridge segment and fracture zones dividing them into independent structural geomorphologic units. Each of them has characteristic seismological features and may be regarded as a specific seismically active zones.

Fig 8a. Focal mechanism of earthquakes in the Norwegian-Greenland Basin (first motions method).

Apart from the central axial zone, sites with higher seismicity are reported from the periphery of the basin as well as from its coastal surroundings.

The Iceland-Jan Mayen (Kolbeinsey) Ridge is the southernmost fragment of the mid-oceanic ridge restricted by the Tjornes and Jan Mayen Fracture Zones. The lateral offset with respect to its southern and northern extensions is 150-200 km westward. Bathymetry of the ridge is complicated and in general, is very poorly known. It is made of a series of highs and lows widening from 40 km in the southern part between Iceland and 69°N (Spar Fracture Zone) to 100-110 km in its northern part. Within the ridge, the most notable axial offset is observed along the Spar Fracture Zone; indications of other less notable offsets are established from detailed aeromagnetic data (Geophysical characteristics.., 1985). The axial rift valley is missing from the southern part of the rift whereas in the north it is represented by several en-echelon valleys with depth reaching 500-700 m with respect to surrounding mountain ridges (Udintsev,1987).

The pattern of earthquake epicentral distribution near the ridge changes laterally. North of the Spar Fracture Zone where the ridge is widest, the axial band of epicenters is evidently linear, while in the narrower south section, a considerable dissemination of epicenters is obvious with offsets from the axial line up to 100 km. In the Spar Zone the line of epicenters is clearly broken. Depths of hypocenters are typical of mid-oceanic ridges and do not exceed 25 to 30 km.

In this area, focal mechanisms solutions have been obtained through the first motions method for 10 events (Table 12, Fig.8a) of which 3 (2, 4, and 9) took place in the Tjornes Fracture Zone.

According to the data compilation on the ridge, the pattern of tectonic movement corresponds mainly to the normal fault with the fault fissure plane dipping around 60° to the horizon, which is typical of rift mid-oceanic ridges. Noteworthy are considerable variations in subhorizontal extension axis strike azimuths, which permitted A.V.Drumya (VNIIOkeangeologia, unpublished reports, 1988) to conclude the non-orthogonality of this type of stress in the ridge axis and suggest the presence of geodymamic factors complicating the geodynamic process. However the seismological data indicate azimuth relationships of stress axes with the epicentral line which characterizes only averaged, generalized strike of the structural axis which is not necessary coincident with the strike of its separate

Table 12

Focal mechanisms of earthquakes of the Iceland-Jan Mayen Ridge (First motions method)

* - in, or close to one of the fault zones + - CMT method solution available

Axes of principal stresses: Nodal plane:

T - tension; P - compression STK - strike azimuth; DP - dip; SLIP - slip

PL - plunge; AZM - azimuth Dislocations:

Quality N - normal fault; S - strike slip; T - thrust

G - good; F - fair; P - poor; U - mechanism type only

N (n) -general quantity of values available(including inconsistent)

Table 13

Focal mechanisms of earthquakes of the Iceland - Jan-Mayen Ridge (CMT method)

*- in, or close to one of the fault zones + - CMT method solution available

Axes of principal stresses: Nodal plane:

T - tension; P - compression STK - strike azimuth; DP - dip; SLIP - slip

VAL - value; PL - plunge; AZM - azimuth Dislocations:

N - normal fault; S - strike slip; T - thrust

Fig. 8b Focal mechanisms of earthquakes in the Norwegian-Greenland Basin (CMT method)

fragments. Therefore, it is possible that an essential contribution to the axial strike azimuth oscillations was made by the difference between the strike of those sites of the ridge where the earthquake occurred.

The only focal mechanism solution that dramatically differed from the others was that for the earthquake 7 which provided a thrust mechanism. Estimating the accuracy of this solution it should be noted that the above event occurred within an anomalous site of the ridge, namely in the area of its intersection with the Spar Fracture Zone. The nearest earthquake 1 provided a strike-slip mechanism.

The strongest events in this region (2 and 9) occurred in the Tjornes Fracture Zone. Various scientists are rather confident about the dextral strike-slip along subvertical planes. NP 1 strike is practically coincident with that of the fracture zone. A similar mechanism has been established for earthquake 4.

The solutions through CMT method (see Table 13, Fig. 8b) have provided a normal fault and strike-slip mechanisms. For event 2 strike-slip could have been expected.

The Jan Mayen Fracture Zone consists of two separate segments: northwestern where the shear of mid-oceanic ridge and epicentral belt occurred to 200-210 km and southeastern intersecting deep sea part of the Norwegian Basin and penetrating into the Norwegian shelf. (Fig.7). Of these two segments only the northwestern one is seismically active only in the part located between the offsets of the ridge. In its active area the Jan Mayen Fracture Zone is represented by a downwarp of 10-15 km wide and up to 2.2 km deep.

Focal mechanism solutions obtained through the first motions method (Fig.8a, Table 14) show most often sinistral strike-slip along the plane coincident with the proper zone (NP2). The fact of incidence of the plane which can be regarded as a fault plane to the south-west under Jan Mayen Island and the presence of a certain thrust component at the hanging wall have permitted L.A.Savostin and A.M.Karasik (1981) to conclude that the uplift of the Jan Mayen Island block was caused by the lighter weight of its continental crust in comparison with surrounding oceanic crust.

Considering that seven determinations using the CMT method (Fig.8b, Table 15) have also been shown the sinistral strike-slip, the aforesaid tectonic model can be adopted for the Jan Mayen Fracture Zone.

The Mohns Ridge is a fragment of a mid-oceanic ridge of the Norwegian-Greenland Basin north-east of the Jan Mayen Fracture Zone to the Knipovich Ridge (Fig.7). This is the only segment of the ridge symmetrical with respect to the surrounding continents. It has the most typical characteristics of mid-oceanic ridges throughout its entire length, such as sharp topography, distinct rift valley with a depth of 3 km, i.e. 1-2 km lower than the surrounding axial ridge and the general absence of sediments near the axis that thicken toward the periphery.

The typical pattern of the Mohns Ridge is verified by the seismological data analysis showing the linearity and continuity of the epicentral belts and their clear association with the axial zone of the ridge.

Fairly unexpected are results of focal mechanism solutions through the use of first motion method. Presently on the Mohns Ridge, nine events had been processed. As Table 16 shows the data obtained are rather ambiguous and provide the full range

Table 14

Focal mechanisms of earthquakes of the Jan-Mayen Fracture Zone (First motions method)

See Table12 for the legend

Table 15

Focal mechanisms of earthquakes of the Jan-Mayen Fracture Zone (CMT method)

See Table 13 for the legend

Table 16

Focal mechanisms of earthquakes of the Mohn Ridge (First motions method)

See Table 12 for the legend

Table 17

Focal mechanisms of earthquakes of the Mohn Ridge (CMT method)

See table 13 for the legend

of mechanisms from normal fault to thrust. The thrust mechanisms are inexplicable in the extensional structure. The analyses we have carried out indicate this is due to insufficient reference data that did not allow reliable solutions to be obtained. For 6 of the above earthquakes the overall number of signs of first motions is less than 15-20, some of them being unconfident and non-compliant; two solutions have 35 and 40 signs, respectively with more non-compliant ones. Besides, in all cases irregular distribution of signs by quadrant is observed. Even the most confident determination (4) has yielded two solutions by two authors where the position of one of the nodal planes differ by 80 degrees. These determinations were admissible 20-25 years ago due to the shortage of seismological information on the Arctic, and even necessary for initial understanding of the Mid-Arctic Earthquake Belt geodynamics. However, in the second half of 1980s, the solutions like those for 3, 6-9 are inadequate. It should be noted that similar solutions are available for other Arctic regions.

To enhance the data base on focal mechanism solutions on the earthquakes on the Mohns Ridge we have carried out more reliable determinations (10-13) which provided fairly uniform normal fault and normal fault - strike-slip mechanism with the extension axis suborthogonal both to epicentral line and the axis of the ridge.

All eleven determinations made by the CMT method have provided solely normal fault mechanism (Fig.8b, Table 17). Data compared from the three latest earthquakes processed using both methods show that they are nearly identical in the orientation of the extension axis, but showing enhanced dipping in the compression axis according to CMT. Further acquisition of similar information is the only way to judge whether this is caused by determination errors or it makes any geological sense related to change in fault plane orientation throughout the time.

The Knipovich Ridge is the northernmost part of the mid-oceanic ridge in the Norwegian-Greenland Basin changing its strike by nearly 90 degrees and extending almost meridionally to the Spitsbergen Fracture Zone (Fig.7). In contrast to the Mohns Ridge it occupies an asymmetrical position in the Greenland Sea clearly shifting eastward. Change in the strike of the ridge is accompanied by a similar radical rearrangement of its geomorphology. It is represented by a series of mountainous topography divided by depressions of which the most prominent is regarded as a rift valley (Sykes,1965). The cross section of the ridge shows definite asymmetry: its western flank with respect to the rift valley is made of 6 to 7 topographic highs and is much wider than the eastern one consisting of 3 peaks and gradually merging with the continental margin of Western Spitsbergen Island and the Barents Sea.(Desimon and Karasik,1979). The rift valley is usually narrow and 1 to 2 km across with its bottom contour at 3 km, and steep walls up to 1 km high. Seismic investigations (Vogt and others,1979) have reported thick sedimentary units both in the rift valley (up to 1000 m) and along both sides, particularly eastward (up to 3000 m).

Linearity of the seismic belt above the Knipovich Ridge is sharply disturbed. A more adequate suggestion might be made about the three higher seismicity zones between which weaker seismic activity is observed (Fig.7). One of these zones is located at 74-75°N at the junction with the Mohns Ridge. In the second one, located at 76-77°N fewer epicenters are observed and these show an isometric swarm that is denser in its center. A large number of earthquakes occur east and west of the nominal rift valley. North of 77° the epicentral belt after a break becomes more linear up to the Spitsbergen Fracture Zone.

Focal mechanism solutions have been found for 16 earthquakes at the Knipovich Ridge: 10 solutions were made through first motions method ( Fig.8a, Table 18) and 8 by CMT (Fig.8b, Table 19). Solutions 3, 4 and 6 made by first motions method are irratic .

Data on focal mechanisms indicate peculiarities of the tectonic process in this segment of the Mid-Arctic Ridge. In contrast to the normal fault mode, a strike-slip component is essential and sometimes predominates despite a distinct normal fault component. Almost complete coincidence of the solutions made by both methods with respect to the earthquakes date April 25, 1988 is noteworthy giving evidence on the uniform nature of the fault development in the focus at different stages of the seismic process.

Profound en-echelon pattern of the ridge frequently divided by faults, lateral dissemination of the epicenters provides no evidence to associate each particular earthquake to any element of the ridge, thus making difficult the selection of a nodal plane which is the fault plane. Complication of the tectonic setting at the Knipovich Ridge is frequently manifested in notable oscillation of subhorizontal extension axis strike in strike-slip solutions, and available reliable solutions showing no remarkable strike-slip component. This normal fault mechanism obtained through CMT method for an earthquake in close vicinity to the Mohns Ridge ( 2), thrust mechanism (first motions method) on the earthquake dated 1981 (7) which occurred near the earthquake of 1983 with strike-slip mechanism, and also thrust mechanism (CMT) on the earthquake dated 1995 (7).

The Spitsbergen Fracture Zone is a link between mid-oceanic ridges of the Norwegian-Greenland basin and Eurasian Subbasin (Gakkel Ridge) in the Arctic Ocean. It covers a fairly large area between Svalbard and Eastern Greenland, and is represented by a number of en-echelon segments resulting in stepwise displacement of the zone 70-80 km northward to the Gakkel Ridge (Fig.7). Near its junction with the Knipovich Ridge (78-79°N) the seismic belt displaced 60-70 km northward is traced up to 80° N where it sharply displaced again 70-80 km westward with farther north-northwesterly strike up to 83° N to the junction with the Gakkel Ridge. It should be noted that scattering of epicenters within the Spitsbergen Fracture Zone is much less than at the Knipovich Ridge.

At present, about 20 focal mechanism solutions with allowance for those made in this paper, are known for this fragment of the Mid-Arctic Earthquake Belt as obtained through first motions method (Fig.8a, Table 20). However, some of them, for instance 12-14 are barely reliable.

In general, great uniformity of the vast majority of the solutions showing the presence of subvertical nodal planes should be emphasized. One of them (NP2) that has northwesterly strike coincident with the general strike of the fracture zone and adopted for the fault plane allows dextral strike-slip mechanism to be confidently

Table 18

Focal mechanisms of earthquakes of the Knipovich Ridge (First motions method)

See Table 12 for the legend

Table 19

Focal mechanisms of earthquakes of the Knipovich Ridge (CMT method)

See Table 13 for the legend

Table 20

Focal mechanisms of earthquakes of the Spitsbergen Fracture Zone (First motions method)