Geo[. Int. Vol. 21-2, 1982, pp.

157 - 175

THE EARTHQUAKE SWARM OF FEBRUARY 1981

IN MEXICO CITY

J.HAVSKOV*
(Received: March 26, 1982)
(Accepted: June 1st., 1982)

RES UM EN

Entre el 4 y el 15 de febrero de 1981 ocurri6 en la Cd. de Mexico una serie de temblores parti-
cularmente bien registrados. Los 5 eventos mayores aqu{ analizados (2.7 ~ML ~3.2) tuvieron
hipocentros dentro de un volumen de 1 km 3, con localiz~ci6n media en 19.3780N - 99.196°W
y una profundidad de 1.2 km. Las localizaciones relativas de estos eventos mostraron una mi-
graci6n de 0.7 a 1.6 km en un piano casi vertical. Se obtuvo una soluci6n de piano de falla nor-
mal con el eje tensional en la direcci6n casi Este-Oeste, usando estaciones locales y regionales.
El analisis espectral dio momentos si'.smicos en el rango 0.5 - 2.0 x 1020 dinas-cm y ca{das de
esfuerzo de 1 a 5 bares.

Se obtuvieron parametros similares para el evento mayor ocurrido en una serie de temblo-
res en la misma area en febrero de 1980. Parece ser probable que estos eventos ocurran como
resultado de la acumulaci6n de tension regional; no obstante, el hundimiento del Valle de Me-
xico podrfa posiblemente originar tensiones orientadas similares, las cuales, si no ocasionan los
sismos, los pudieran disparar.

• lnstituto de Ingenierfa, UNAM, Mexico 04510, D. F.

157
 GEOFISICA INTERNACIONAL

ABSTRACT

Between February 4 and 15, 1981, a particularly well recorded earthquake swarm occurred in
Mexico City. The 5 largest events analyzed here (2.7 ,$'. ML ;:S; 3.2) had hypocenters within a
volume of 1 km3 with mean location at 19.3780N-99.1960W, and a depth of 1.2 km. Rela-
tive locations of these events showed a migration from 0.7 to 1.6 km on a nearly vertical plane.
A normal fault plane solution with the tensional axis almost east-west was obtained using local
and regional stations. Spectral analysis gave seismic moments in the range 0.5 - 2.0 x 1020
dyne-cm and rather low stress drops (1 to 5 bars). Similar parameters were obtained for the
largest event occurring in a swarm in the same area in February 1980. It seems most likely that
these events occur as a result of accumulating regional tension; however, the sinking of the
Valley of Mexico could possibly produce similar oriented tensions, which, if not causing the
earthquakes, might trigger them.

INTRODUCTION

On February 4, 1981, at 1340 GMT occurred the first in a series of local earth-
quakes widely felt in Mexico City. These events were rather large by local standard;
magnitude estimates for the initial and a, couple of later shocks are in the range of
3.0 to 3.4. Consequently the events were recorded as far as Oaxaca and Acapulco,
Mex.

Several portable seismographs from the Institute of Geophysics and Institute of

Engineering were mobilized capturing many of the later events. In general our
knowledge about the local seismicity in the Valley of Mexico is rather poor and im-
precise due to small number of local stations and to a large extent unknown crustal
structure. However no other local event has been recorded with such a large num-
ber of local and distant stations, both digitally and analog, and this data set is de-
finitely the best ever obtained for any event in Mexico City. In the present study
only the 5 largest events, hereafter called 1 to 5, occurring between Feb. 4 and Feb.
9, 1981 will be analyzed, since they have the most complete data set. A study of
all the events between Feb. 4 and Feb. 15 is under preparation at the Institute of
Geophysics (R. Mota, personal communication). A comparison will also be made
with the largest event of apparently a similar earthquake sequence, felt in the same
region of Mexico City in February 1980.

DATA

In Table 1 all the arrival time readings used in this study have been summarized,
and Fig. 1 and Table 2 show the stations used. The two nearest permanent stations
recording all 5 events were the Institute of Engineering (I.I.), station ABC and the

158
 T A B L E

Arrival time (sec) relative to P-arrival times at station IIM. Also shown are average relative times used for locating
the group of events and corresponding residuals.

P - Time at station UM
Station Feb 20, 1980 Feb 4, 1981 Feb 7, 1981 I Feb 7, 1981 Feb 9, 1981 Feb 9, 1981 Average for Residuals
Phase 4:23 10.64 13:40 28.70 22:40 25.05 22:46 56.45 2:37 06.63 5.53 48.40 1981 events

UM p o.oo 0.00 0.00 0.00 0.00 0.00 o.oo -0.08

s 2.05 2.40 2.30 2.45 2.40 2.32 -0.03
ABC p -1.10 -0.98 -0.82 -0.83 -0.85 -0.92 0.06
s 0.23 0.30 0.43 0.55 0.42 0.39 0.28
SAH s 2.46 2.15 2.35 2.20 2.05 2.24 0.14

-
Vl
\0
TAC p

s
-1.20
-0.05
-1.05 -0.85
0.40
-0.93
0.22
-0.95
0.15
-1.00
0.18
-0.17
-0.23
~

J
UC p 6.66 6.02 6.05 6.00 5.97 5.90 5.98 -0.14

UP p 3.92 4.05 3.95 3.90 3.87 3.80 3.91 -0.09

CA2 p 28.60 28.75 28.6~ 28.67 1.96

III p 17.30 17.30 0.42

OXM p 7.90 7.95 8.05 8.37 7.90 8.03 0.55

TPM p 6.60 6.60 -0.09

CRX p 7.50 7.50 0.26

· CGA p -0.83 -0.95 0.89 0.01

})
 GEOFISICA INTERNACIONAL

National Seismological Service Station TAC, both at epicentral distances of about

2.5 - 3.0 km.

From TAC conventional 3 component short-period (SP) records were used; how-
ever S-arrival was difficult to read because of saturation.

The ABC station is a 3 component acceleration station transmitting to I.I. and

recording in analog form on magnetic tape. Thus there is no saturation problem
and playback at high speed permits very accurate readings of P and S arrivals. Fig.2
shows an example and it is seen that S-arrivals are identified unambiguously.

I
/
/
/
/

/
/
,,,,,,, ..... - \
\
\
I
I /
I
I /
/ I
I I I 19.40° N

IV /
/
/

IV-V
- I
I
I
: ·~~~
(\ • Feb•*\
Bl I
I
I
I
\ I
CED '\ I
• '\
,. CGA I
4v. Rio Churubusco
CAP
• ''' ..... I 19.35°N
I
I
I
I
I
I
I
I
I
' ' ' .......
- ....
0 2
_,.,./
/
• Stations
• Epicenters
* Average epicenter
of 1981 swarm
kt lometers
99.2L5__w_ _ _ _ _ _ _ _ _ _9_9-.2.L0__W_ _ _ _ _ _ _ _ _ _ _9_9...c.1L....5°=-w-___, 19 · 300 N
0 0

Fig. 1. Location map showing stations and epicenters. Isoseismic map of the l3!gest event is
also shown.

160
 J. Havskov

FEB 7 1981 22 :46

NS

EW

0 2 3 4 5

Time, sec
Fig. 2. Accelerograms from station ABC. Traces are plotted to show P arrival; they' are clipped
for S-waves. The peak acceleration was 24 gals on the EW component.
161
 GEOFISICA INTERNACIONAL

The I. I. has similar acceleration stations at SAH and IIM (Fig. 1) at epicentral
distances of 5 to 6 km; however due to attenuation and low gain only S-readings
could be used. These S-arrivals were of lesser quality than the ABC-readings. At sta-
tion IIM there is also a SP seismometer. Thus P-readings could be picked very ac-
curately; however S-waves saturated the tape. Fortunately Red Sismol6gica Mexi-
cana de Apertura Continental (RESMAC) also operates a digital seismic station
(SP-Z) close to I. I. thus permitting unsaturated recordings of all 5 events. In addi-
tion a large number of stations at epicentral distances of more than 25 km recorded
the events (Tablel,2); however these were only used in determining the fault plane
solution.
TABLE 2

STATION COORDINATES

Station code Latitude N Longitude E

IIM * 19.326 99.182

ABC * 19.400 99.205
SAH * 19.393 99.147
TAC ** 19.405 99.194
CGA ** 19.358 99.212
ICC * 19.767 99.258
IIP * 19.347 98.918
CA2 * 17.975 100.048
III * 19.376 99.468
OXM ** 19.294 99.688
TPM ** 18.983 99.057
CRX *** 19.406 99.680
CAP ** 19.352 99.231
CED ** 19.359 99.243
CSM ** 19.393 99.147

* Instituto de Ingenieria
** Instituto de Geoffsica
*** RESMAC

162
 J. Havskov

Of the 5 largest events, only the last two were recorded clearly by one portable
field station CGA operated by the Institute of Geophysics. A large number of smal-
ler events were recorded by various field stations in the period Feb. 4 to Feb. 9.
For the Feb. 20, 1980, event readings from stations UM, UP and IIC were used
(Table 1). Both the ABC and TAC recordings are missing.

EPICENTERS

From Table 1 it is seen that the arrival times relative to station IIM have a variation
of less than 0.4 seconds indicating that these 5 events occur within a small region.
The same is indicated from lhe seismograms which are quite similar for the 5 events
(Fig. 3,5). Thus to get an accurate average location average relative readings were
used (Table 1). The HYPO 71 location program (Lee and Lahr, 1978) was used
with the following model:

Velocity, km/sec Depth to interface, km

2.9 0.0
4.7 2.0
6.6 4.0
7.1 30.0
8.1 35.0

The two first layers of this model is taken from a study of the crustal structure
in the Valley of Mexico (Havskov and Singh, 1979) while the rest of the model is
based on minimizing residuals while locating Mexican earthquakes. Since only sta-
tions within an epicentral distance of 10 km are used, the two first layers are the
most important. The velocity ratio Vp/Vs was determined from the S and P read-
ings at stations HM, TAC, and ABC and resulted in the value 2.11. This high ratio
is probably caused by water saturation lowering the S-velocity.

With the above mentioned model the average epicenter was located at 19.378°N
and 99.196°W (5.8 km N and 1.4 km W of station IIM) and the depth was 1.2 km.
Changing the model did not change the epicenter significantly, however the depth
was sensitive to model changes, especially changing the depth of the first interface.
It seems likely however that the events were very shallow considering the large sur-
face waves generated (Fig. 3, Fig. 5).

The epicenter can be compared with the isoseismic map (Fig. 1) (J. Figueroa,
personal communication) and there is good correlation between the epicenter and
the region of maximum intensity.

163
 ·-
'iii-
u
!!..-

~--

~--

o __
C')

o_
N

...
ao ...
...
OJ ao
...
0
.....
.,,
C')
~ II')
N

a,
.a
...
o_
....•

Fig. 3. Seismograms of some of the smaller events registered at station IIC.

164
 J. Havskov .

Changing the Yp/Vs ratio to the nonnal value of 1.78 shifts the epicenter 1.1 km
to the west, the depth to 1.0 km and increases the RMS error from 0.15 to 0.30.
It thus seems that the high Yp/Vs ratio is real and not a result of misidentifying S
arrivals.

Table 1 also gives the travel time residuals from all stations recording any of the
5 large events. It is seen that the readings from TAC and ABC do not agree well.
There may be a clock difference between I.I. and TAC; however eliminating TAC
only changes the epicenter 0.3 km to the south and depth remains unchanged.

The small changes in the relative arrival times (Table 1) indicate small but sys-
tematic changes in hypocenter location. To detennine these relative changes accur-
ately the master event technique was used with stations IIM, IIP, IIC, ABC and
TAC. The residuals from the average event (Table 1) (or master event) were subs-
tracted from the readings of the 5 individual events which then were relocated.

The event of Feb. 20, 1980, was also located using the same residuals but em-
ploying only P readings from stations IIM, IIP and IIC. Thus the depth for this
event was fixed at 1.2 km. Table 3 and Fig. 4 show the results. It is seen that the
hypocenters 1-5 are contained within a volume of less than 1 km3. The epicenters
show relatively large standard errors (Fig. 4) making it difficult to decide on any
alineation. Although the two best located events ( #4 and 5) and the event of Feb.
20, 1980 could indicate a N 45°W alignment, it should be pointed out that the epi-
central error of the 1980 event may be large ( unknown but at least of the order of
1 km).

The two near stations TAC and ABC receive the first arrivals as direct waves
while at the more distant stations, the first arrivals are refracted rays. This gives a
very good constraint on depth detennination since travel time-depth derivative aT /aH
will have different signs for near and far stations. The same is true for the epicenter
detennination in the north-south direction due to the predominantly north-south
alignment of the stations while east-west control is poor being essentially controlled
by only station IIP. Thus the spread in the epicenters in the east-west direction as
seen on Fig. 4 is most likely due to error in relative location and in general it is· to
be expected that depths are better constrained than epicenters, as also indicated by
the estimates of standard errors of location (Table 3). Thus the clear migration to-
wards deeper hypocenters seen for the last events in the series is most likely real.
This could explain higher intensities and greater damage reported for the first event
as compared to the subsequent events although magnitudes were similar.

165
1-

T ABL E 3

The table gives the average location of events 1-5, the location of each of the 5 events and the location of the
1980 event. X and y are linear coordinates relative to station !IM, z is the depth, RMS are the RMS travel time
residual, and ERRH and ERRZ are the horizontal and vertical standard errors in location, respectively, as given
by the HYP071 program. Cl
~
"Tl
F;.;
Event Latitude N Longitude E x(km) y(km) z(km) RMS(sec) ERRH(km) ·ERRZ(km) ()
>
°'°' Average 19.378 99 .196 -1.5 5.8 1.2 0.00 0.0 o.o ~
t'1
. :=
19. 379 99 .195 -1.3 5.9 0.7 0.05 0.5 0.2 z
2 19. 376 99.204 -2.3 5.5 1.0 0.06 0.8 0.4 ii;
0
3 19. 379 99 .199 -1. 7 5.9 1.5 0.04 0.4 0.2 z
>
t""
4 19. 378 99 .195 -1.3 5.8 1.4 0.01 0.1 0.1
5 19. 381 99.195 -1.4 6.1 1.6 0.02 0.3 0.1
FEB 80 19. 366 99 .172 1.0 4.4 1.2* 0.00

* Depth fixed
 J. Havskov

North (km)

2
•

Feb eoe
East ( km)
-2 -1

j, *
•
*
Hypo or epicenter
Average location of
1981 swarm

Depth
(km)

Fig. 4. Relative locations of individual events (top) and corresponding E-W profile (bottom).
The origin of the coordinate system is at station IIM. The circles and error bars indicate HYP0-
78 standard errors in location.

167
 GEOFISICA INTERNACIONAL

SPECTRAL ANALYSIS

P-wave spectra were made for all 6 events using the digitally recorded data from the
RESMAC station MEX, which is located less than 200 m from IIM. The signals
used and the time window selected (1.7 sec) are shown in Fig. 5.

P waves were used instead of S waves since the latter were difficult to identify.

Moment, source radius and stress drop were calculated using the extension of
Brune's model (Brune, 1970) to P waves (e.g. Hanks and Wyss, 1972). AP-wave
velocity of 2.9 km/sec and a density p of 2.5 g/cm3 were used. Spectra were cal-
culated both with and without correction for Q (Fig. 6). Since Q is unknown, a
value of 100 was chosen which is probably reasonable for the upper crust in Mexico
City which consists of water-saturated sedimentary layers.

Table 4 gives the moment, source radius, stress drop, magnitude (ML) and dis-
placement on the fault, with and without correction for Q. It is seen that selection
of Q is not critical in determining the spectral parameters. The magnitude ML was
estimated from the relation (Hanks and Kanamori, 1979, Singh and Havskov, 1980)

(1)

and displacement D was obtained from

Mo
D=-- (2)
µ1rr2
In Eq 2 the rigidityµ was calculated fromµ = (32/p, where {j is the S wave velocity.
Using {j = 1.4 km/sec and p = 2.5 givesµ= 4. 7 x 1ol O dyne/ cm2. Table 4 also gives
the coda length r(sec) measured at station UC and the corresponding coda magni-
tude Mc calculated from (Real and Teng, 1973)

Mc = -1.01 + 1.89 log r

Relative consistency in the magnitudes estimated from M0 and coda duration sug-
gests that M0 estimates are not in gross error. Since the moments and the magni-
tudes of all events are nearly equal, the sequence has been called an earthquake
swarm. Stress drops are· rather low, which could be expected for events occurring
in sedimentary layers and although the stress drop is the least well determined para-
meter, it seems that later (and deeper) events have lower stress drop.

The source parameters for the 1980 and 1981 events are similar, indicating a
similar cause for their origin.

168
 •

0 2 3

Time, sec
Fig. 5. The fust 3 seconds of the seismograms as recorded at RESMAC station MEX. Traces
are scaled such that the plotted peak amplitude of all the events is the same. Also shown is the
time window used for spectral analysis.
169
.r.>

•
T ABLE 4

FOCAL PARAMETIRS

~ is the local magnitude, estimated from moment, MC coda magnitude and u the displacement on the fault

Q= m Q = 100

20 20
Time of event Corner Source Moment xlO u(cm) Moment xlO Stress drop
freque.!!. Radius (km) dyne-cm
Stress drop!
bars
~
dyne-cm bars ~ u(cm) Coda
l,ength
Mc

Feb 4, 13:40
cy (H )

4.0
1

0.27 2.01 4 3.1 1.9 2.42 5 3.3 2.1-

(sec)

220 3.4
i
Fil
()
>
-
-..J
0
1

9
Feb 7, 22:40 4.0 0.27 1,03 2 3.0 1.0 1.50 3 3.1 1.4 220 3.4
~
Feb 7, 22:46 3.8 0.28 1.00 2 3.0 0.9 1.59 3 3.1 1.4 205 3.4 z
8 ~
2.8 0.3 150 3.1
0
~
Feb 9, 2:37 3.4 0.32 0.38 0.5 2.7 0.3 0.43 0.6
1
Feb 9, 5.53 3.5 0.31 0.48 0.7 2.8 0.3 0.61 0.9 2.9 0.4 180 3.3

1
Feb 20, 4.23 4.1 0.26 0.65 2 2.9 o. 7 1.03 3.0 3.0 1.0 150 3.1
9
R
0
 J. Havskov

,~,,...... ...........
1981
\ ' , .................-1
(.)
Cl)
FEB 4
\ ''
\ ' ',-2
UI \
I \
E
(.) \
\
Cl)
\-3
-
"O
:,

a.
E
FEB 7 22:40 \

c
C>
.3

FEB 9

1980

Q =Cl)

-0.60 -0.20 0.20 0.60 1.00 -0.20 0.20 0.60 1.00

Log frequency , Hz

Fig. 6. The displacement spectra with and without correction for Q. The selected low frequen-
cy spectral levels and roll-offs are indicated by dashed lines. Decay rates of in, n = 1,2 and 3
are indicated in the upper left hand corner. The vertical scale of individual spectra are shifted,
however they are all plotted with the same scaling factor.

171
 GEOFISICA INTERNACIONAL

FAULT PLANE SOLUTION

First motions for the first event was initially used in the fault plane solution and
projected on to the lower hemisphere in a stereographic equal area projection using
stations TAC, IIM, IIP, IIC, III, ABC, TPM, OXM, and CA2 (Fig. 7). Since the first
motions of events 2-5 on the same stations remained almost constant, it seems that
the focal orientation for the individual events in the swarm did not change appre-
ciably. Therefore, data was added from some of the later events recorded on field
stations. Events 4 and 5 were recorded at station CGA while the polarities used for
stations CAP, CED and CSM were read from smaller events occurring before event 5.

0
I IC
0
CSM

w 011 P E

s
Fig. 7. Fault plane solution. Solid circles are compression and open circles dilations. All data
is from the largest event except the fust motions for the field stations CSM, CGA, CAP and
CED which are from later events (see text).

172
 J. Havskov

In Fig. 7 fault plane solution is shown. Compression at III is inconsistent with

the focal mechanism. Ignoring III it is possible to draw extreme fault plane solu-
tions (not shown in Fig. 7). In any case they all show a normal fault on a nearly
vertical N-S plane or a nearly horizontal E-W plane.

DISCUSSION

The seismic history of the Valley of Mexico is not known with sufficient precision
to say whether the present series of earthquakes, in terms of magnitude and loca-
tion, is a normal occurring sequence or something unexpected. Two possibilities
for the origin are (a) the events are of tectonic origin and occur as a result of accu-
mulation of regional stress and (b) the events are related to the sinking of the Val-
ley of Mexico.

· The fault plane solution shows a normal fault with the tensional axis going near-
ly east-west. Several events in the Mexican volcanic belt show similar normal fault
solutions (Molnar and Sykes, 1969; Jimenez and Ponce, 1978; Singh and Wyss,
1976; Astiz, 1980) and it is possible that the February swarm was caused by this
regional tectonic tensional regime. In the valley there are both older north-south
and younger east-west going vertical faults (F. Mooser, personal communication)
and it is possible that some of the N-S aligned faults have been activated by the re-
gional tension. This seems to favor the near vertical fault in the fault plane solution.
Although hypocentral locations are not accurate enough to assure that all events
occurred on the same plane this seems likely considering the migration of the hypo-
centers. •
The area of maximum sinking of the Valley of Mexico constitutes a nearly north-
south going trough located a few kilometers to the east of the epicentral area, as
known from levelling data until 1974 (Wade et al., 1975) and the sinking in the per-
iod March 1970 to August 1973 was of the order of a few centimeters a year.

In Long Beach, California, a subsidence rate of a maximum of 71 cm/yr due to

extraction of oil and water caused earthquakes of similar magnitudes as observed in
Mexico City (Kovach, 1974). The earthquakes occurred about 1 km away from the
area of maximum subsidence, and it was concluded that they were caused by hori-
zontal tension originating from the sinking. Depths were about 500 m and bore-
hole samples showed that slippage had occurred on horizontal planes.

173
 GEOFISICA INTERNACIONAL

In Mexico City the subsidence rate does not seem to reach similar high values;
however, the sinking could cause tension in the epicentral area in the direction ob·
served from the fault plane solution. Although the near vertical fault plane is fa.
vored, the possibility of slip along the second possible plane dipping down towards
the east cannot be ruled out. Since the earthquakes are located close to the boun-
dary of the transition zone, slip might have occurred on the contact between dif-
ferent geological layers.

Related to the sinking is the water pressure changes measured in wells. From
1970 to 197 4 changes were largest in the western part of the valley near the epicen-
tral area and reached values of about 1 bar (Wade et al., 1975).

For the period January 1978 to October 1981 the maximum change of water
pressure in wells near the epicentral area was about 0.5 bar (X. Haro, personal
communication). Near dams, it has been observed that pore pressure changes of
the order of 1 bar can trigger small earthquakes (Talwani, 1976) and the pore pres-
sure changes under the valley might have some influence on the time of occurrence
of tectonic earthquakes which would have occurred anyway.

CONCLUSION

The earthquake swarm recorded in Mexico City in February 1981 constitutes the
best data set obtained so far for any event in the valley. The 5 largest events occur-
red within a volume of 1 km3. The fault plane solution and relative hypocenter
determination suggest that the hypocenters were most probably located on a single
nearly vertical N-S plane. The events seem to be caused by a near E-W tensional re-
giq:le which most likely has its origin in the regional tectonics, although the sinking
of the valley could cause similarly oriented tension which also could influence the
seismicity.
•

ACKNOWLEDGMENTS

I thank S. K. Singh, R. Mota and 0. Rasc6n for helpful discussions. R. Mota kindly
supplied us with all available data at the Institute of Geophysics. Most signals were
plotted using the RESMAC computer, which also was used for the spectral analysis
employing a program by A. Nava.

174
 J. Havskov

BIBLIOGRAPHY

ASTIZ, L., 1980. Sismicidad en Acambay, Estado de Mexico. El temblor del 22 de

febrero de 1979. Thesis, Facultad de Ingenieria, UNAM, Mexico.
BRUNE, J. N., 1970. Tectonic stress and the spectra of seismic shear waves from
earthquakes. Journ. Geophys. Res., 78, 4997-5009.
HANKS, T. C. and M. WYSS, 1972. The use of body wave spectra in the deter-
mination of seismic source parameters. Bull. Seism. Soc. Am., 62, 569-590.
HANKS, T. C. and H. KANAMORI, 1979. A moment magnitude scale. Journ.
Geophys. Res., 84, 2348-2350.
HAVSKOV, J. and S. K. SINGH, 1978. Shallow crustal structure below Mexico
City, Geo[. Intern., 17, 223-229.
JIMENEZ, Z. and L. PONCE, 1978. Focal mechanism of six large earthquakes in
Northern Oaxaca, Mexico, for the period 1928-1973. Geo[. Intern., 17, 379-386.
KOVACH, R. C., 1974. Source mechanisms for Wilmington oil field, California,
subsidence earthquakes. Bull. Seism. Soc. Am., 64, 699-715.
LEE, W. A. K. and J. C. LA.RHR, 1978. HYP0-71 (revised 1978): A program for
determining hypocenter, magnitude and first motion pattern oflocal earthquakes.
U. S. Geological Survey, open file report 75-311.
MOLNAR, P. and L. R. SYKES, 1969. Tectonics of the Caribbean and Middle
American regions from local mechanisms and seismicity. Bull. Geol. Soc. Am.,
80, 1639-1684.
REAL, C. R. and T. TENG, 1973. Local Richter magnitude and total signal dura-
. tion in Southern California. Bull. Seism. Soc. Am., 63, 1809-1827.
SINGH, S. K. and M. WYSS, 1976. Source parameters of the Orizaba earthquake
of August 28, 1973. Geo[. Intern., 16, 165-184.
SINGH, S. K. and J. HAVSKOV, 1980. On moment-magnitude scale,Bull. Seism.
Soc. Am., 70, 379-383.
TALWANI, P., 1976. Earthquakes associated with the Clark Hill Reservoir. South
Carolina, a case of induced seismicity. Engineering Geology, 10, 239-253:
WADE, L. R., A. J. B. MOLL and L. R. CABELLO, 1975. Boletin de Mecdnica de
Suelos, No. 2, published by Secretaria de Recursos Hidraulicos, Comisi6n de
Aguas del Valle de Mexico.

175

,~