Cao §×nh TriÒu

At 12.18 (Hµ Néi time) of 24 June 1983 in the area of Pu Nhung mountain, about 11 km NW of TuÇn Gi¸o townlet a destructive earthquake took place. Its magnitude reaches 6.6 - 6.7 in Richter scale with a maximum intensity on the land surface reaching 8-9 in MSK- 64 scale. The earthquake caused great losses to TuÇn Gi¸o, Lai Ch©u, Tña Chïa, ThuËn Ch©u, Quúnh Nhai and §iÖn Biªn areas. Besides the losses in terms of human lives, houses, public facilities and crops, the earthquake also caused many geologic hazards, such as land cracking, landslide, slumps, subsidence, changes of springs, etc.

The TuÇn Gi¸o earthquake is one of the two strongest earthquakes occurring in the territory of ViÖt Nam in the last century (the previous one is the §iÖn Biªn earthquake in 1935 with M_s = 6.8). Many domestic and foreign scientific papers touching upon the TuÇn Gi¸o earthquake was published [2,5,6,8-11,14,15] but no complete research has been made on the seismotectonic characteristics and earthquake maximum evaluation of the epicentre area of this earthquake. In the mean time, study on the generation condition as well as the law of seismic activities in the TuÇn Gi¸o area is necessary and of scientific significance. The results of studies on this direction allow us not only to understand better the law of seismic activities in the areas where strong earthquakes have been occurring, but also to use it in studying on the structural characteristics of the seismic sources in the territory of ViÖt Nam.

In previous studies [9-11,14] some authors have touched upon the characteristics of the TuÇn Gi¸o earthquake source. The main parameters referred to in these studies are: the focal depth, the shape of the epicentral area and the kinetic characteristics of the main fault related to the focus. In this paper the author presents some new results of determining the source and focal depth of TuÇn Gi¸o earthquake in 24 June 1983 based on the analysis of the aftershock distribution law.

Usually major earthquakes entail a series of minor earthquakes which occur in sequence immediately after the main shock and have the same focal characteristics, called as aftershocks. An earthquake of Ms = 7.0 in Richter scale may entail thousands smaller earthquakes (aftershocks). The aftershocks reflect directly the relationship between the fault slipping surface and the earthquake. The main shock creates abrupt increase and decrease of stress and impacts on the surrounding environment, making the stress in this area change in a complicated manner. Thus, within the fault zone and the areas adjacent to the main epicenter occurs the process of rebalancing the stress state of the earthquake source, thus generating the aftershocks. Typical aftershocks appear immediately after the main shock and are distributed mainly in the earthquake source area. In general, the magnitude of aftershocks are always smaller than that of the main shock (but they still cause additional destruction). The total amount of energy released by the aftershocks usually does not exceed 10% of that of the main shock. The frequency of earthquake aftershocks decreases quickly with time.

In the 1930s, on the basis of studying on the earthquake aftershock distribution law in Japan, Omori established an empirical formula reflecting the acting process of aftershocks as follows (Omori's law):

where: n - frequency of aftershocks at the point of time t after the main shock; k, c, p - constants depending on the magnitude of the main shock.

 Thus, based on the aftershock distribution the devastation areas of most earthquakes can be determined. The size of the devastation area is directly proportional to the magnitude of the main shock.

In 1954, Utsu and Seki established an empirical formula for determining the size of the earthquake source [16,18]:

lgA = 1.02 Ms + 6.0 (2)

where A is measured in cm².

Actually, the earthquake sources are determined by aftershocks occurring 1 - 2 days after the main shock. This has been proved by studies on focal mechanism of earthquakes worldwide. The aftershocks occurring after 2 days or later are considered to be generated by sources beyond that of the main shock.

The area of TuÇn Gi¸o earthquake source was determined in this study on the following bases:

1) The distribution of aftershocks occurring in the two days 24 and 25 June 1983 (Table 1). The result of this analysis shows (Fig. 1a) that the source of the TuÇn Gi¸o earthquake has a length of L = 33.1 km and a width of W = 21.2 km. Thus, its area is:

2) Based on the empirical formula (2), for TuÇn Gi¸o earthquake with Ms = 6.7 we have:

The difference of the source areas determined by the two above methods is as follows:

According to Utsu and Seki, the free coefficient of the empirical function (2) is a constant characterizing the tectonic conditions of each study area.

 Applying formula (2) to the case of TuÇn Gi¸o earthquake we get the following expression for calculating the source area:

lg(686.2).10¹⁰(cm²) =1.02 Ms + b. with Ms = 6.7.

lgS = 1.02 Ms + 6.01. (3)

In a published work [4], Cao §×nh TriÒu established an empirical formula characterizing the relationship of the length (L_km) and the width (W_km) of the focus with the magnitude of the earthquake, corrected by the new result of this work as follows:

lg L(km) = 0.6 Ms - 2.50

lg W(km) = 0.25 Ms - 0.35 (4)

Within 6 months after the TuÇn Gi¸o earthquake of 24 June 1983, in the area adjacent to the focus 25 aftershocks with Ms > 3.0 were recorded (Table 1).

The aftershocks of TuÇn Gi¸o earthquake are mainly distributed in the depth limit from 3 to 20 km. If the focal depths of aftershocks are considered as reflecting the structural characteristics of the earthquake source, the height of the source of TuÇn Gi¸o earthquake can be determined as 20 - 3 = 17 km (Fig. 1b).

Table 1. List of aftershocks in the area of TuÇn Gi¸o earthquake in 1983

The combination of geophysical methods used in studying the structural characteristics of the earth crust in the TuÇn Gi¸o area and its surroundings includes [2,3,7,14,16]: 1) Magnetic and gravity field variation method, of which the upward field continuation, downward field continuation, anisotropy analysis and high order derivative methods play the main role; 2) Gravity back analysis method; 3) Georadar method, which is also used by the author in this study.

On the basis of the existing data (Bouguer gravity map of 1/50,000 scale; aeromagnetic map of 1/200,000 scale; 10 profiles of detailed gravity, ground magnetic and georadar survey of 1/50,000 scale) and the methods of study presented above, the author has compiled an earth crust structure map of the study area. The result is as follows [3]:

The faulting scheme of the study area is described by the author based on the subdivision of the faults into various orders. It should be noted that this subdivision is only qualitative and based mainly on the structural characteristics: the extents, the expression on the Bouguer gravity anomaly map, the role of the faults in respect to the geologic structure and the basic boundary surfaces of the earth crust such as the crystalline basement surface, the Conrad surface, the Moho surface as well as the lateral heterogeneity of the crust density. Some main faults are described in Table 2.

Table 2. Some main faults in the TuÇn Gi¸o area and its surroundings

The structures are subdivided after their age as follows: a) Early Paleozoic: S«ng M·, ThuËn Ch©u; b) Late Paleozoic - Early Mesozoic: M­êng TÌ, S«ng §µ. The S«ng M· zone was initiated by rifting during the destruction of the Late Riphean continent crust, leading to the formation of basins with oceanic crust.

1) Structures with low seismicity: S«ng C¶ zone;

2) Structures with moderate seismicity: Fan Si Pan; Tó LÖ; S«ng §µ and NËm C«;

3) Structures with medium seismicity: Pu Si Lung and M­êng TÌ zones;

4) Structures with strong seismicity: S«ng M·, ThuËn Ch©u and SÇm N­a – Hoµnh S¬n zones

a. Area distribution of earthquake epicentres

A distribution map of earthquake epicentres of TuÇn Gi¸o area and its surroundings is shown in Fig. 2. The most prominent feature in this map is the concentration of epicentres in zones coinciding with the major tectonic fault zones of the earth crust, namely:

3) The S«ng M· - §iÖn Biªn - TÜnh Gia earthquake zone extending in NW-SE direction. Within this zone the §iÖn Biªn earthquake of 1935 occurred with Ms = 6.8.

In general, the focal depth of earthquakes in TuÇn Gi¸o area and its surroundings has not exceeded 25 km. The maximum depth of over 15 km pertains to the earthquakes with large magnitudes such as: §iÖn Biªn earthquake with Ms = 6.8 and H = 23 km; TuÇn Gi¸o earthquake with Ms = 6.7 and H = 18 km.

With the aim to find out the distribution law of earthquake foci, within this study the authors determined the average density of foci with the use of a grid with interval of 1km. From the result of this calculation (Fig. 3) we can easily recognize that the earthquake foci in the TuÇn Gi¸o area and its surroundings are concentrated mainly in the depth from 4km to 21km. The thickness of the seismic active layer is about 17 km.

a. Time distribution characteristics

The most prominent feature of this distribution is the manifestation of seismic activities with low frequency in the period before 1914 and from 1945 to 1976 (Fig. 4). Before the occurrence of TuÇn Gi¸o earthquake in 1983, there was in 1977 a manifestation of intensive seismic activities, and then it decreased. If this is a premonition of a strong earthquake, we can use it in the future for long-term earthquake prediction in ViÖt Nam.

The seismic quiescence in an area is determined clearly in the spatial and temporal distribution curve of the earthquake. On the basis of the earthquake spatial and temporal distribution of the TuÇn Gi¸o area and its surroundings one can make some remarks as follows (Fig. 4): 1) The average seismic quiescence time before the maximum earthquake is about 50 years; 2) Intensive seismic activities before the §iÖn Biªn earthquake took place in the period from 1914 to 1945, i.e. about 30 years. During this period of intensive seismic activities the most prominent thing is the occurrence of the §iÖn Biªn earthquake in 1935 (Ms = 6.8); 3) If the previous period of intensive seismic activities is deemed to characterize the present, then at present the TuÇn Gi¸o area and its surrounding is being subjected to a period of intensive seismic activities which may continue for 10 years.

An extremely important task in seismotectonic study, that is the premise for medium- and long-term earthquake forecast, is the study and prediction of seismic sources. This is a long-term and urgent direction of research of a nation, the result of which will serve as the primary basis and foundation for the earthquake forecast in the future. The main methods used for evaluating the maximum earthquake occurrence in this study include the forecast of maximum earthquake based on the identification problem, and the time - magnitude problem in earthquake forecast [1-12,14,15].

- The data on vertical component vector of gravity field (along the z-axis) calculated in the area of earthquake with corresponding magnitude.

- The data on lineament density and lineament intersection density established from the results of space image and geological data interpretation.

The standard sample selected for the identification problem is an area around the epicentre of an earthquake with a radius of about 8 km.

Usually the selected standard samples are based on the task set forth for the identification problem. For the purpose of forecast the seismic maximum of the TuÇn Gi¸o area and its surroundings the author uses two types of standard samples:

The result of earthquake forecast by identification problem is shown in Fig. 5. According to this result some remarks can be made as follows:

The time - magnitude model has been adopted by the author for long - term prediction of earthquakes in TuÇn Gi¸o area and its surroundings. The solution of this new prediction problem is carried out in three steps as follows:

1) Subdividing the study area into seismic areas with different seismic characteristics. The subdivision is based on the geologic structure, tectonic conditions and characteristics of seismic activities along earthquake-generating faults most reliably identified. Based on these criteria and on the previous studies we have subdivided the study area into two seismic areas, namely: §iÖn Biªn and TuÇn Gi¸o.

2) Identifying the representative earthquakes of each seismic area. It is required that these earthquakes are representatively distributed in time and magnitude in the respective seismic area.

3) Establishing a calculation program on PC, carrying out the calculation and presenting the results.

The time intervals between earthquakes with M_min thresholds in each seismic area have been taken into account respectively. Totally in seismic area 1 there are 10 time intervals and in seismic area 2 there are 23 time intervals. Coefficient b = 0.86 and c = 0.09 of the study area have been established for the equation after iterated calculations, whereas the coefficients = -3.51, = -3.38 are the average value of coefficient a of the seismic areas 1 and 2.

Thus, for §iÖn Biªn seismic area the empirical formula representing the relationship between time and magnitude of the main earthquakes is:

log T = 0.86 M_min + 0.09 M_p - 3.51 (5)

The similar formula for TuÇn Gi¸o seismic area is:

log T = 0.86 M_min + 0.09 M_p - 3.83 (6)

The standard logarithmic distribution is correct for the seismic areas. The occurrence time and magnitude of the last major earthquake in the seismic area are taken into consideration. The occurrence probabilities of earthquakes with Ms = 5.0; 5.5; 6.0 and 6.5 in Richter scale respectively after 10, 30, 50 years from 2002 have been calculated by formulas (5) and (6).

The calculation results show that in 50 coming years from 2002 the occurrence probability of earthquakes with different magnitudes in the 2 seismic areas will not be high. The occurrence probability of earthquakes with Ms = 5.0; 5.5; 6.0 and 6.5 after various periods of time from 2002 have been calculated by formulas (5) and (6) for both seismic areas. It can be seen that at all 4 magnitude levels of earthquake used for prediction, the occurrence probability of earthquakes within the prediction period of the seismic area 1 is always higher than that of the seismic area 2.

The occurrence probability of earthquakes with Ms = 5.0 after 80 - 100 years is very high (0.957 and 0.976 for the seismic area 1; 0.864 and 0.908 for the seismic area 2 respectively). Similarly, the interval for earthquakes with Ms = 5.5 is 200 - 300 years, for those with Ms = 6.0 is 400 - 500 years and Ms = 6.5 is 1000 - 1250 years.

On the basis of the results obtained in this study, some conclusions can be made as follows:

1) Before the occurrence of the TuÇn Gi¸o earthquake there was in the epicentral area a long lasting seismic quiescence.

2) There is a manifestation of quiescence - activation law in the TuÇn Gi¸o earthquake. Before its occurrence in the neighbouring area an earthquake with lower magnitudes appeared.

3) The area of the TuÇn Gi¸o earthquake source is about 676.1 km². The parameters of the source have been determined as follows: length, L = 33.1km; width, W = 21.2 km; and focal height = 17 km.

4) The crust - upper mantle boundary (Moho surface) in the area varies in a rather complicated manner and has a depth varying between 32 km and 36 km, increasing from east to west.

5) The faults in the study area have a relatively great penetration depth (crust penetrating faults) and divide the earth crust into block structures with a clear variation of average density in lateral direction. Most of these faults have NW - SE direction (S¬n La, S«ng M·, S«ng §µ, ThuËn Ch©u, Fan Si Pan, Tó LÖ) and sub-meridional direction (Lai Ch©u - §iÖn Biªn, TuÇn Gi¸o).

6) In the TuÇn Gi¸o area and its surroundings exist three earthquake generating zones: Lai Ch©u - §iÖn Biªn, S¬n La and S«ng M·- §iÖn Biªn - TÜnh Gia. The thickness of the seismic activity layer is 17 km; the earthquake epicentres are concentrated mainly in the depth from 4 km to 21 km.

7) There is a clear manifestation of two seismic areas of strong seismicity: §iÖn Biªn and TuÇn Gi¸o. Earthquakes with magnitude 6.5 may occur in these two areas after 1000 - 1250 years with probabilities 0.879 and 0.797 respectively. The §iÖn Biªn seismic area has a recurrence interval of earthquakes with magnitude 5.0; 5.5 and 6.0 as 80-100 years (with probability 0.976); 200-300 years (0.949) and 400 - 500 years (0.896) respectively. The TuÇn Gi¸o seismic area has an occurrence probability of earthquakes with magnitude 5.0 as 80-100 years (0.908), magnitude 5.5 as 200-300 years (0.930) and 6.0 as 400-500 years (0.823).

6. Cao §×nh TriÒu, NguyÔn Thanh Xu©n, 1997. Main seismic faults in T©y B¾c region. TC Khoa häc Tr¸i ®Êt, 19 / 3 : 214 - 219. Hµ Néi.

14. NguyÔn §×nh Xuyªn, Cao §×nh TriÒu et al, 1990. The TuÇn Gi¸o earthquake, 24th June, 1983. Nxb KHKT, Hµ Néi, 108 trg.

15. NguyÔn Thanh Xu©n, Cao §×nh TriÒu, 1997. The application of Geographic Information System (GIS) to seismic hazard analysis in §iÖn Biªn - S¬n La area. TC Khoa häc Tr¸i ®Êt, 19 / 2 : 119-123. Hµ Néi.