Geophysical Monograph Series Properties and Processes of Earth's Lower Crust Vol.

51

SYNTHETIC SEISMOGRAM MODELING OF CRUSTAL SEISMIC RECORD SECTIONS

FROM THE KOYNA DSS PROFILES IN THE WESTERN INDIA
V. G. Krishna, K. L. Kaila and P. R. Reddy

National Geophysical Research Institute, Hyderabad 500007, India

Abstract. Two deep se ismic sounding profiles, wave s from the Moho are also not observable at
each about 200 km long, were recorded in the the appropriate time s in the se record sections
east-we st direction in the Koyna region of the thus substantiating the transitional nature
We stern India by the National Geophysical of the Moho, which is at least 2 km thick, in
Re search Institute acquiring wide angle crustal the Koyna region. The Pn phase, well recorded
se ismic data in the analog form. The analog as the first arrival beyond 165 km record ing
DSS records, acquired by continuous profiling, distance, reveals an uppermost mantle P velocity
have been digitized and assembled into trace of 8.25 km/ sec in this region.
normalized record sections which are displayed
in the form of reduced travel time with reduction Introduction
velocity of 6 km/ sec . Travel time s and relative
amplitude s modeling, with the aid of synthetic The continental crust forms the most important
se ismograms, of the record sections obtained in zone of study for understanding the relationship
the eastern and we stern directions on the be tween the processes in the mantle and geo-
nor thern profile as well as a record section physical phenomena observed at the earth's
in the eastern direction on the southern profile surface. In-situ information on the properties
reveal consistent models of the crustal velocity and processes of the lower crust and the nature
structure with minor lateral variations in the of the crust-mantle boundary can essentially
Koyna region. The prominent feature s of the be obtained by indirect methods, although ultra-
crustal velocity models inferred in this region deep drilling may eventually provide answers
include low velocity layers (LVL) in the upper to some key questions such as continental volca-
crust (6.0 to 11.5 km depth ) as well as in the nism and intraplate seismicity. Seismic investi-
lower crust (26.0 to 28.0 km depth) and at least gations by refraction and reflection profiling
2 km thick transitional Moho at 35.5 to 37.5 km using controlled source s provide the most valua-
depth. However, the Moho is 1-2 km deeper along ble information on the properties of the conti-
the southern profile. The upper crustal LVL, with nental crust and the subcrustal lithosphere.
its top at 6-7 km depth in the se models, is Since 1972, the technique of deep se ismic sound-
consistent with the observed se ismic ac tivity ing (DSS) has successfully been employed in India
concentration at 4-5 km depth and an appreciable by the National Geophysical Re search Institute
reduction of se ismic activity at greater depths ( NGRI ) acquiring wide angle crustal se ismic
in the Koyna region. The inferred crustal velo- data in the analog form over more than 4000 line
c ity mode ls are also consistent with the essen- km by continuous profiling along 17 profile s in
tially ase ismic nature of the lower crust in various geological se ttings in the peninsular
this region. The two low velocity layers, in shield and in the Himalaya. Recently NGRI has
the upper and the lower crust in this region, may also acquired the DFS V digitial recording
suggest rheological stratification of this part systems and currently DSS profiling in India
of the lithosphere with levels of increased is accomplished by digital data acquisition
ductility at those depths. The reflected phase and processing.
PMP , from the Moho boundary, is relatively strong Two DSS profile s, each about 200 km long,
only bey ond 90-100 km recording distance while were recorded during 1975-77 in the east-we st
it is almost suppressed in the subcritical direction in the Koyna region (Figure 1) of the
distance s on the record sections considered in We stern India by the NGRI acquiring wide angle
the pre sent study. Further the PS converted crustal se ismic data in the analog form. This
region is also well known for the se ismic acti-
Copyright 1989 by vity in this par t of the subcontinent. Kaila
International Union of Geodesy and Geophysics et al [1981a,b] constructed the crustal depth
and American Geophysical Union. sections from the DSS data along the two profile s

143
Geophysical Monograph Series Properties and Processes of Earth's Lower Crust Vol. 51

144 MODELING OF KOYNA DSS RECORD SECTIONS

7ed ped 75°0' is ° field tape s have subsequently been used for
30 30'
reproduction of se ismic records with the desired
gain and filter se ttings. Each spread 11.6 km
long with geophone spacing of 200 m utilizes
a shot with varying charge size ( 50 to 800 kg )
'so 18 of chemical explosives from the shot point,
O'
O
depending upon the shot point to recording
spread distance and the quality of the shot
point be ing used. It may be mentioned here
that the shot points considered here on both
the profile s have been exceptionally good in
terms of energy conversion/ transmission and
a charge size of only about 800 kg was adequate
for recording spread distance s of 180-200 km.
The analog DSS data available in the form of
17 * playback record s (with a high cut frequency
e 73°d 74re of 15 Hz) on photographic paper, each covering
Fig. 1. Location map showing the KOYNA I and 5.8 km distance, have been used for digitization.
KOYNA II DSS profiles with shot points. The analog magnetic tapes, however could not
be used for this purpose due to compatibility
problems. Although the original data was recorded
which reveal essentially horizontal layering with 200 m geophone spacing, we have used every
in the crust. Krishna and Kaila [1986] , referred alternate trace about 1 400 m apart for digitiza-
here as paper I, initiated digitization of analog tion with a sampling interval of LI msec . As
DSS records and re interpretation with the aid the entire record section for a particular
of synthetic se ismograms. In paper I the analog shot point has to be assembled from se ismograms
DSS records for SP 0 on the KOYNA II profile, of a large number of DSS record s obtained on
after digitization and amplitude normalization, different spread s from shots of vary ing charge
were assembled into record sections and displayed size, recording gain and other instrumental
in the form of reduced travel time. These record se ttings, we assembled and plotted the digitized
sections were used to derive a kinematically data by trace normalization with a reduction
and dynamically well constrained velocity model velocity of 6 km/sec . In th is normalization
for the region by modeling of travel times and process, the maximum amplitude in each trace
relative amplitudes data and successively refin- is scaled to a fixed value. In this type of
ing the model. In this paper we present the presentation, the re lative amplitude s of differ-
crustal se ismic record sections assembled after ent phase s in the record section can be used for
digitization and amplitude normalization, for comparison with synthe tics for mode ling. In our
SP s 0, 20, 40, 180, 200 on the KOYNA II profile opinion, the models are be tter constrained when
and for SP 10A on the KOYNA I profile, and the observed amplitude ratios of different
the synthetic se ismogram modeling of these phase s in the record sections are matched with
sections producing internally consistent crustal those obtained in the synthetics. Therefore we do
velocity models in this region. The final velo- not consider it as a serious disadvantage that
city models thus derived reveal the fine struc- true amplitude variation with distance of indi-
ture and nature of the se ismic boundaries, vidual phase s cannot be extracted from this
lead ing to considerable improvements over the type of DSS record s. The data from SP 20 and
earlier velocity models based on inversion SP 1 40 on the KOYNA II profile towards east,
of travel times data alone [Kaila et al, 1981a,b]. which are available in the distance ranges from
0 to 70 km and from 80 to 150 km respectively,
are combined together to form one continuous
Digitization and Presentation of record section (SP 20-40) from 0 to 150 km
the Record Sec tions distance. Similar ly the data from SP 200 and
SP 180 on this profile towards we st, which are
Typically, the DSS data acquisition has available in the distance range s from 0 to
been made by continuous profiling using two 70 km and from 75 to 170 km re spec tive ly , are
se ismic units of Russian make, type POISK-I-48- also combined together to form one continuous
KMPV-OV , with a frequency range of 5 to 30 Hz. record section (SP 200-180) from 0 to 170 km
The dynamic record ing range and reproducing distance. Although we have plotted the record
with high cut at 30 Hz filter is about 140 db . sections for various shot points with trace
A high cut filter with cut off frequency of interval of 400 m, 800 m, 1.6 km and 3.2 km,
22 Hz has been used for recording in the Koyna we present here only those with about 1.6 km
region [Kaila et al, 1979] . The low cut filter trace interval to illustrate the possible
was kept out always. Since there is provision correlations of different phases identified.
for field record ing on magnetic tapes, the Nevertheless we have used the record sections
Geophysical Monograph Series Properties and Processes of Earth's Lower Crust Vol. 51

KRISHNA ET AL. 145

2
0 20 40 160 180 200
VERTICAL COMPONENT
GRIN 0.2

KOYNA II PROFILE -- SP 0 DATA SET -- OT = 4 MSEC

Fig. 2. Observed record section for SP 0 on the KOYNA II profile assembled from the
digitized data of DSS records (trace interval - 1.6 km). Theoretical travel time
curves, corresponding to the final velocity-depth model, are also shown for various
phases correlated in the record section.

plotted with 400 m trace interval for phase Analysis and Interpretation of
correlations and the identification of various the Record Sections
prograde and retrograde branches was rather The record sections for SP 0 (Figure 2),
unambiguous due to larger data density. SP 20W (Figure 8), SP 20-40 (Figure 9), SP

PIP P9
10

t 20
Pe

30
PIP
w 40

50

2
0 20 140 160 184 201.
WI AL COMPONENT
GAIN N 0.75

KOYNS- - SP 7 -- MODEL. WITH FIRST ORDER MO1i0 36.5 KM

Fig. 3. Normalized record section of ray synthetic seismograms computed by using the
velocity-depth model, with a first order Moho, shown for SP 0 on the KOYNA II profile.
Various phases corresponding to the correlations in the observed record section are
also labelled for comparison.
Geophysical Monograph Series Properties and Processes of Earth's Lower Crust Vol. 51

146 MODELING OF KOYNA DSS RECORD SECTIONS

.4

=
oy
0

10

20 oI
12 °
w40
O
-I
50

50
5 6 7 0 1 20 140 160 ISO 200
VELOCITY IN KM/S E 71041. CONPONEN
CAIN N 0.76

KOYNA II -- SP 0 -- MODEL WITH 2.0 KM TRANSITION MONO 35.5-37.5 KM

Fig. 4. Same as Figure 3 with 2.0 km transition Moho.

200-180 (Figure 11) on the KOYNA II profile P1P and P2P - reflected waves respectively
and SP 10A (Figure 13) on the KOYNA I profile from the top and bottom of the upper crustal LVL,
are shown in various Figures with the correla- PcP - reflected wave from the midcrustal boundary,
tions made by us for all the prominent phases Pc - refracted wave through the midcrustal
observed. It can be seen from these Figures boundary,
that the DSS data presented in the form of record P3P and P4P - reflected waves respectively
sections with relatively small trace interval from the top and bottom of the lower crustal LVL,
offer the great advantage of reliable phase PMP - reflected wave from the Moho boundary,
correlation especially for later arrivals. Pn - refracted wave from the uppermost mantle.
The correlations shown in Figure 2 for SP 0 The travel time curves for various phases
record section are essentially the same as correlated, as shown in Figures 2, 8, 9, 11 and
those interpreted by us in paper I. We use 13 correspond to the final velocity models
here the following nomenclature for various inferred from these record sections. We have
crustal phases interpreted: computed for each record section a large number
Pg - refracted wave from the uppermost part of travel time models and every successful
of the crystalline basement, velocity model giving an acceptable travel

10
171
20 0
us
4.1. 4..4

—* 30
WO
t.s
O

50

BO
20 4 so so leo Ito 140 160 180 200
TERTI OIL COMPO0ENT DISTANCE IN KM
ORIN M 0.11

KOVNO II -- SP 0 -- MODEL 111TH 4.0 KM TRANSITION MONO 34.5-30.5 KM

Fig. 5. Same as Figure 3 with 4.0 km transition Moho.

Geophysical Monograph Series Properties and Processes of Earth's Lower Crust Vol. 51

KRISHNA ET AL. 147

8 8

6 6
a.
a_
4 4
a_ a

2 2

0 0
0 20 40 60 80 100 120 0 20 40 60 80 100 120
X(Km) X(Km)
8
1
1
6
a_ 1
1
1
4 1
a_U 1
111
2

0
0 20 40 60 80 100 120
X(Km)

Fig. 6. Variation of the amplitude ratios with distance (X) for various combinations
to the phases from the observed ( ) and the synthetic ( _ ) record
sections for SP 0.

time s fit was tested for reproducing the re la- Upper Crustal Model
tive amplitude s match by computing ray syntheticc
se ismograms [ Cerveny et al, 1977 ] . Thus the In the later arrivals a prominent wave group
final velocity models are der ived by successively de signated as P1P , shortly after the first
refining both the travel time s and relative arrivals, can be seen (with 180° phase shift)
amplitude s match with the observed record in the distance range from about 20 to 65 km
sec tions. in various record sec tions (F igure s 2, 8, 9,
As can be seen from various record sec tions 11 and 13 ). Another wave group, de signated
(Figures 2, 8, 9, 11 and 13), the first arrivals as P2P , can also be seen in the later arrivals
out to 20-25 km distance reveal continuous in the distance range from about 60 to 110 km
variation of apparent velocity of P wave s from in the se record sec tions. The prominent P 1P
4.4 to 6.0 km/ sec due to the Deccan Trap s cover. phase indicates reflection from a velocity
The first arr ivals in the distance range from discontinuity in the upper crust. However,
20-25 km to 160-170 km reveal an apparent velo- the extensive travel time s and amplitude s mode l-
city of 6.0-6.2 km/ sec and appear to be rela- ing re sults indicate that it is almost impossible
tively very weak bey ond 80 km distance. We to fit any viable model with a positive velocity
have however used the or iginal monitor records jump at the discontinuity corresponding to
to fit the first arrivals data. We identify the P 1P phase. Because , a very small increase
the se first arrivals, out to 160-170 km distance, of velocity at a supposed discontinuity although
as the Pg phase (basement refraction) re sulting produce s a travel time s fit with the P 1P phase,
from wave s penetrating in to the crystalline the corresponding synthetic amplitude s are
basement. Bey ond 160-170 km distance the fir st comparative ly much we aker th an the observed
arrivals are more prominent with a sharp increase P1P amplitudes. Further the first arrivals,
of the apparent velocity to 8.25 km/ sec as computed for such a model with even a very
can be seen from the record sec tions for SP 0 small velocity jump, are also found to be much
and SP 10A on the two profiles. We identify earlier, revealing higher apparent velocity,
the se fir st arrivals, bey ond 160-170 km distance, than the observations. Therefore, we considered
as the Pn phase re sulting from wave s penetrating the pre sence of a low velocity layer (LVL )
in to the uppermost mantle. in the upper crust which may explain both P 1P
Geophysical Monograph Series Properties and Processes of Earth's Lower Crust Vol. 51

148 MODELING OF KOYNA DSS RECORD SECTIONS

12 12

10 I0

•■ 6

0
40 80 100 120 140 160 180 200 40 60 80 100 120 140 160 180 200
X(Km) X (K m)

12 12

10 10

8 8

a
a 6 6
m
a
4 4

2 2

0 I 0
40 60 80 100 120 140 160 180 200 40 60 80 100 120 140 160 180 200
X (K m) X(Km )

Fig. Same as Figure 6. In the case of PMP/P4P and PMP/P3P amplitude ratios,
7.
corresponds to the model with 2.0 km transition Moho, corresponds to
the model with 4.0 km transition Moho and. . corresponds to the model with a
first order Moho.

and P2P phases. We could successfully model depth, with 0.2 km/sec velocity reduction,
both P1P and P2P as the reflected phases respect- and its bottom as a 3 km thick transitional
ively from the top and bottom of an upper crustal boundary from 8.5 to 11.5 km, velocity increasing
LVL. The velocity reduction of 0.2 km/sec occurs to 6.3 km/sec at 11.5 km depth. However this
at 6.0 km depth where the top of this LVL upper crustal LVL is found to be 1 km deeper
produces a reflection corresponding to the on the KOYNA I profile, with its top at 7.0 km
P1P phase. The top of this upper crustal LVL and the 3 km thick transitional bottom from
which is inferred to be a sharp boundary is 9.5 to 12.5 km depth. The velocity models derived
consistent with the observations of relatively from these record sections also contlistently
strong amplitude P1P phase with 180° phase reveal a strong velocity gradient from ground
shift [Smith et al, 1975]. The relative ampli- surface to about 3 km depth with minor lateral
tudes pattern of the P2P phase however requires variations due to the varying thickness of
the bottom of this upper crustal LVL to be the Deccan Traps along the two profiles [Kaila
transitional. A first order boundary at the et al, 1979]. P velocity at SP 0 (towards east)
bottom of the LVL would produce relatively increases from 4.4 km/sec from ground surface
strong subcritical reflections which are not to 5.8 km/sec at 1.8 km depth and 6.15 km/sec
evident on the observed record sections. The at 3.2 km depth, this model being also consistent
final models thus derived from various record at SP 20W (towards west) as shown in Figure 8.
sections with the aid of synthetic seismograms, The velocity increases from 4.4 km/sec to
indicate the presence of an upper crustal LVL 5.8 km/sec at 1.0 km depth at SP 20-40 (towards
with its top as a sharp boundary at 6.0 km east) and to 5.95 km/sec at 1.0 km depth at
Geophysical Monograph Series Properties and Processes of Earth's Lower Crust Vol. 51

KRISHNA ET AL. 149

w5
(r)
=
c 4
cn

4-- 3

O1

0 1111111111111111111•11•111111111H
0 5 10 15 20 25 30
VERTICAL COMPONENT DISTANCE IN KM
GRIN 0.2

KOYNA II PROFILE -- SF 20W DATA SET -- DT 4 MSEC

Fig. 8. Same as Figure 2 for SP 20W with trace interval " 400 m.

SP 200-180 ( towards we st ). However at SP 10A from the top at 26 km and bottom at 28 km ( at

( towards east ) the velocity increase s from 29 km on the KOYNA I profile) of a lower crustal
4. 14 km/sec to 6.2 km/sec at 2.8 km depth. LVL by travel time s and synthetic seismogram
computations. The velocity in this lower crustal
Lower Crustal Model LVL is inferred as 6.3 km/ sec . Be low this LVL,
the velocity increase s to 6.9 km/ sec . Mode ls
We have correlated another se t of wave groups considered without this LVL did not give accepta-
in the later arrivals, de signated as PcP and ble fits to the P3P and P4P phase s. Attempts
Pc, in the distance range from about 70 to to fit the se two phase s by positive velocity
150 km as shown in Figure s 2, 9, 11 and 13. jumps, however small they may be, increased
The phase Pc appears to be tangential to the the average crustal velocity beyond the accepta-
PcP phase at 100-110 km distance and indicates ble range while match ing the PMP phase , besides
an apparent velocity of 6.7 km/sec . Both travel shifting the PMP critical point to larger
time s and synthetic se ismogram computations distances than that observed at —100 km distance.
have been made to fit the PcP and Pc phase s. On the other hand, the wide angle reflection
The observed amplitude s pattern of the reflected computed from the bottom of th is lower crustal
phase PeP require s a 1 km thick transitional LVL matche s well with the relatively strong
boundary from 19.5 to 20.5 km depth separating P4P phase beyond 140 km distance. Thus we prefer
the upper crust from the lower crust. The to interpret P3P and P4P phase s as reflections
computed wide angle reflections from this transi- from the top and bottom of the lower crustal LVL.
tional boundary, centered at 20 km depth ( at
21 km depth on the KOYNA I profile ), match
well with the correlations for the PeP phase. The Moho Boundary
The velocity increase s to 6.7 km/sec be low
this boundary. There are also another se t of The most prominent wave group correlated
wave groups, correlated and de signated as P3P in the later arrivals on all the record sec tions
and PIP at distance s beyond 70-80 km with a (Figure s 2, 9, 11 and 13) is the phase de signated
relatively low measure of lateral continuity as PMP, which is identified as the wide angle
on these record sections. The se two phase s reflection from the Moho boundary by mode ling.
are however found to correspond to reflections This phase is almost continuously observed
Geophysical Monograph Series Properties and Processes of Earth's Lower Crust Vol. 51

150 MODELING OF KOYNA DSS RECORD SECTIONS

with relatively large amplitude s at distance s prominent only from about 100 km onwards. However
greater than about 100 km on all the record fur ther increase of the transition thickness
sections on the KOYNA II prof ile whereas it is to 2.0 , 3.0 and 4.0 km shifts the prominent
relatively strong beyond about 80 km distance relative amplitude s of the PMP phase respectively
on the KOYNA I profile as can be seen from to 96, 100 and 104 km. Therefore we infer that
the record sections for SP 0, SP 20-40, SP the Moho in this region is a transition zone
200-180 and SP 10A. However at smaller distance s of at least 2.0 km thickness ( from 35.5 to
in the subcritical range, the amplitude s of 37.5 km depth ). Similarly we have also computed
the PMP phase appear to be relatively weak. the synthetic se ismogram sec tions with 2.0 km
By travel time s mode ling and considering a thick transitional Moho corresponding to the
first order Moho at 36.5 km depth, we obtained velocity models for SP 20-40 and SP 200-180
a very satisfactory travel time s fit to the record sections and shown in Figure s 10 and 12
observed PMP time s. However the synthetic se ismo- respectively. It can be seen from the se Figure s
gram section computed by using this model with that the se models for SP 20-40 and SP 200-180
a first order Moho at 36.5 km depth reveals also match quite well with the observed record
relative ly very strong PMP arrivals in the sec tions. Figure 14 shows the synthetic seismo-
subcritical distance s also as can be seen from gram sec tion computed for 1.0 km thick transi-
the synthetic record section in Figure 3. This tional Moho corresponding to the velocity model
is certainly not consistent with the amplitude for SP 10A record section/ on the KOYNA I profile.
observations of the PMP phase. Therefore we Comparison of the synthetic sec tions with the
tried to suppress the subcritical PMP amplitude s observed record sec tion for SP 10A revealed
in the synthetics by considering the Moho as that Moho transition thickne ss of 1.0 km,
a transitional boundary rather than a sharp producing prominent PMP amplitude s from 80 km
discontinuity. There is also another good reason onwards, may be more acceptable than the model
for us to consider the Moho as a transition with Moho transition thickness of 2.0 km on
zone in this region, due to the absence of this profile. We thus infer a 1.0 km thick
otherwise observable and possibly strong PS transitional Moho (from 37.0 to 38.0 km depth )
converted wave s in the record sections at about in the region of the KOYNA I profile and it
5 sec after the PMP arrival time s where one is also quite evident that the Moho is slightly
would expect such wave s. While a first order deeper in this region than towards north in
discontinuity could also explain the observed the region of the KOYNA II profile.
travel time s of the PMP, but at the same time The Pn velocity of 8.25 km/sec observed
would produce such conversions with considerable on the SP 0 and SP 10A record sections has
amplitude s whereas a transition zone suppresses been used in all the models discussed above.
them. Therefore a transition zone at the Moho However we further tried some models by vary ing
is the most likely explanation of the observed the velocity gradient in the upper mantle to
absence of reflected PS wave s from the Moho, produce a comparable PMP/Pn amplitude s ratio
which also appears to be a general feature in the synthetics as that in these observed
of crustal record sec tions [Fuchs, 1975] . Thus record sec tions. By this modeling proceduri
the range of possible models that satisfy the we inferred a P velocity gradient of 0.004 sec
data certainly excludes the possibility of in the uppermost mantle to a depth of about
a sharp discontinuity at the crust-mantle bound- 60 km, producing a satisfactory fit to the
ary in the Koyna region and indicate s the Moho observed PMP/Pn amplitude s ratio. It can be
as a transition zone. We have therefore consider- seen from all the record sec tions pre sented
ed a transitional boundary at the Moho and here that there is reasonably good agreement
progressively increased its thickness in order to be tween the synthetics and the observations
obtain an acceptable fit to the observations from various shot points considered in the
of the PMP phase while modeling. We considered pre sent study. To demonstrate this point more
various models with 1.0, 2.0, 3.0 and 4.0 km convincingly we have pre sented in Figure s 6
thick transitional Moho boundary and in each and 7, plots showing variation of the amplitude
case after a satisfactory travel time s fit ratios with distance for various combinations
has been obtained, we computed the synthetic of the phase s both from the observed as well as
se ismogram section and compared with the observed the synthetic record sections for the SP 0
record sec tion for SP 0 on the KOYNA II profile. data se t. It can be found from these plots
Figure s 4 and 5 show the se synthetic se ismogram that in spite of some random uncertainties,
sec tions for SP 0 velocity models respectively they reveal acceptable match of model outputs
with 2.0 and 4.0 km thick transitional Moho with the observations. Thus the models derived
boundary. It was found in paper I that Moho in the pre sent study are internally consistent
transition thickness of 1.0 km though suppresses with various observations. However the frequency
the subcritical PMP amplitude s considerably, dependent effects may have some influence on
ye t its re lative amplitude s are very prominent the der ived models (e .g. Cerveny ' s effect )
from 84 km onward s, whereas in the observed which could be be tter resolved by application
record sec tion for SP 0 the PMP amplitude s are of the exact methods (e.g. reflectivity method )
Geophysical Monograph Series Properties and Processes of Earth's Lower Crust Vol. 51

KRISHNA ET AL. 151

160

KOYNA II PROFILE -- SP 20-40 DATA SET -- DT = 4 MEC

Fig. 9. Same as Figure 2 for SP 20-40 with trace interval - 1.6 km.

10

z 20

30

W 40

2
10 0 20 120 140 160
VERTICAL COMPONENT
GRIN N 0-75

KOYNA II -- SP 20-40 -- MODEL WITH 2.0 KM TRANSITION MOHO 35.5-37.5 KM

Fig. 10. Same as Figure 3 for SP 20-40 with 2.0 km transition Moho.
Geophysical Monograph Series Properties and Processes of Earth's Lower Crust Vol. 51

152 MODELING OF KOYNA DSS RECORD SECTIONS

1AOp

li i 1111111111111 1111111110111111111i 120 140 160 180

KOYNR It PROFILE -- SP 200-180 DATA SET -- OT = 4 MSEC

Fig. 11. Same as Figure 2 for SP 200-180 with trace interval - 1.6 km.

10

t 20

30
r
w40

50

60
5 6 7 9 9 10 120 140 160 1130
VELOCITY IN KM/S

KOYNA II -- SP 200-180 MODEL WITH 2.0 KM TRANSITION MONO 35.5-37.5 KM

Fig. 12. Same as Figure 3 for SP 200-180 with 2.0 km transition Moho.
Geophysical Monograph Series Properties and Processes of Earth's Lower Crust Vol. 51

KRISHNA ET AL. 153

,;,:t.N7-7-
2,
=44
-....--,•. ----

. - -_ tt ESIli t f
IVNi /11111,4114
1 id _
_ _-.4.117,-/i,„,c, It„
fil:i_FIZ 11, 4 atibij 4
IttiWt1101
1 - Wit ,7lz 1 -

inz1-14iiiiiiilit
„,, ........ ... ___ ..........ifil‘1.41111ill
___f...__________
131'''-'-'
11111111111Liniftimrinamt: "--1.-------z-
itiiiitill-i't41:-2417Itt;II
L--

0 20
VERTICAL COMPONENT
GRIN 0.2
III 1 1 III Ill "111 111111111111111111111111111111111111
KOYNA I PROFILE -- SP 10A DATA SET - DT 4 msEc

Fig. 13. Same as Figure 2 for SP 10A on the KOYNA I profile with trace interval - 1.6
km.

rather than the ray theoretical methods as modeling of the DSS record sections for SP 10A,
applied in the present study. SP 0, SP 20-40 and SP 200-180 are shown in
Figure 15 for comparison. These velocity models
Discussion reveal almost similar structure with minor
The final models of the crustal P velocity lateral variations in this region. The prominent
functions in the Koyna region as derived by features of these velocity models are quite

a■•

•••

4.
1 20 a
z • .4.4. 3.
3. 4. a
30

t40

SO
•
00 2
0 0 20 40 00 00 100 020 140 144 100 200
VELOCITY IN 101/5 VINT [COL CO/V0112111 OISTRNCE IN AM
ORM 11 0.70

KOYNA 1 -- SP 10A -- MODEL N1TN 1.0 KM TRANSITION MONO 37.0-30.0 KM

Fig. 14. Same as Figure 3 for SP 10A on the KOYNA I profile with 1.0 km transition
Moho.
Geophysical Monograph Series Properties and Processes of Earth's Lower Crust Vol. 51

154 MODELING OF KOYNA DSS RECORD SECTIONS

VELOCITY (Km/Sec)
894567894567894
0

10-

E 20-
Y
1 30-
i—
o-
w 40 -
0
50- SP 10A SP 20 - 40
60 1

Fig. 15. Crustal P velocity-depth models for the Koyna region, derived in the present
study.

consistent with a majority of the latest velocity prominent features are quite consistent in
models of the continental crust based on quanti- the two velocity models. A large number of
tative evaluation of seismic refraction/wide studies dealing with detailed interpretation
angle reflection data along a number of profiles of refracted and reflected phases in seismic
over different regions of the earth. Mueller refraction surveys elsewhere, have also reported
[1977] advanced a new model of the continental evidence for an LVL in the upper crust in the
crust and the schematic model presented by depth range of 5 to 15 km, which had originally
him is shown in Figure 16(a) along with the model been proposed by Mueller and Landisman [1966].
derived by us for the Koyna region from SP 0 Mueller [1970] and Landisman et al [1971] pointed
record section shown in Figure 16(b). The simi- out that a semicontinuous laccolithic zone of
larity of these models is very striking and the granitic intrusions could well be the most

V (km/sec) V (km/sec)
4 5 6 7 8 9 5 6 7 8 9
I I I
0
X X X XX
x X x X x X x X 1 11
xx
I0 it..V11X7102121 %. -10

..■■••

^2 - 20 ID
m
\ /
o -

I 30 - 30 I
F-
a
3
cl 40 - 40

- 50
I Fa 5!6 F71 7 :•;•;! 8
60
Fig. 16(a). Schematic model of the continental crust (after Mueller, 1977): 1.Cenozoic
sediments, 2.Mesozoic and Paleozoic sediments, 3.Upper Crystalline basement consisting
of metamorphic rocks (gneisses and schists), LI.Laccolithic zone of granitic intrusions
(sialic low velocity layer), 5.Migmatites (middle crustal layer), 6.Amphibolites (high
velocity tooth), 7.Granulites (lower crustal layer), 8.Ultramafics (uppermost mantle).
(b): Crustal P velocity-depth model for the Koyna region, derived in the present
study, revealing similar features. (c): Crustal features in the Vogtland area of the
southern German Democratic Republic (after Bankwitz et al, 1985). V - P velocity model
(after Knothe, 1972), dots - depth distribution of earthquakes in the area, (al - 73) -
depth distribution of strength, T - depth distribution of temperature.
Geophysical Monograph Series Properties and Processes of Earth's Lower Crust Vol. 51

KRISHNA ET AL. 155

KOYNA 1983 -84 Depth Sections in the depth range be tween 3 to 16 km with a
Computer Located Epicenters SW
I offer Rostogi of ol.19651
NE maximum frequency around 7 to 9 km corresponding
0
73 • 30' 73• 43' 4roal:rilr• "t • : to the uppermost part of the LVL in this region.
.y. • V
4 Similarly, recent analysis of the Koyna earth-
11 •
•• • quake s of 1983-84 by Rastogi et al [1985]
17*.
PL
revealed that the depths are shallow, mostly
A
ALO . E 12
within 4 km, and that there are no hypocenters
• •
s
NEL
. x deeper than 25 km in this region as shown in
I-
o. Figure 17. In the Koyna region, Langston [1976,
KN • W 20
0
• %.
1981] as well as Langston and Franco-Spera [1985]
24
have also found very shallow source depth s,
• ...514, re: . ok • # .
• 8 •• 8• 25 well constrained by wave form inversion studies,
•
.• / • •••• • . 32 for the main shock of December 10, 1967
/• .
•• .. r'
up—
a
ss
•
..., •
•
•
• ,P;
• -. ..
o f
NW (4.5+1.5 km ), for its fore shock of September
13, 1967 ( 5 km ) and also for the aftershock
'41. • . '. 4 of December 12, 1967 ( 3.5-4.0 km ). All the se
studies in the Koyna region thus suggest that
81.,%/ • •
the major se ismic activity in this region is
°• • • • 12
confined to only within 4-5 km depth. The crustal
away*Ittluu • M velocity models derived in the pre sent study
• 3.0
W 20 reveal an upper crustal LVL from 6.0 to 11.5 km
ire
on • 3.0 - 3.9 o 5 10 17• 0
■ 4.0 - 41 1 KILOMETRES
0' depth and also a lower crustal LVL from 26 to
24
A iloWnie Motion 28 km depth consistent with the above findings
73 • 30' 75 45'
on thepossible relation be tween the se ismic
32 activity and the crustal velocity mode ls.
According to Chen and Molnar [ 1983] the
Fig. 17. Spatial and depth distribution of ,

distribution of focal depths discussed above

earthquakes in the Koyna region (after Rastogi et
mainly reflects the dependencies of strength
al, 1985).
of geological materials on temperature and
mineralogy. Bankwitz et al [1985] also concluded
probable explanation for the sialic LVL in the that the crustal structure has an important
upper crust. Smith e t al [1975 ] proposed that a influence on the earthquakes occurrence. In
plausible explanation of the sialic LVL is Figure 16(c) are also shown schematically the
the effect of a high temperature gradient on the depth distribution of earthquake s, ( 0 1 - 03 )
velocities of upper crustal rocks. Mueller [1977] and T depicting depth distribution of strength
however maintained that this upper crustal LVL and temperature respectively. The earthquakes
could be associated with a zone of lower densi- seem to be occurring in the upper rigid layer
ties and a higher attenuation within the granitic of the crust with relatively more strength
intrusions into the surrounding basement rocks. and temperature s within 200°-300°C. At larger
He also pointed out that it is the water content temperature s, the strength rap idly decrease s
and the pore pre ssure in the se granite s which and the ear thquakes seem to be very scanty
play the major role in causing the upper crustal or absent. The se observations are also consistent
LVL rather than the temperature effect. In with the proposed upper crustal LVL in the
this context it may be mentioned here that hot Koyna region where the major se ismic activity
springs occur in the Konkan foothills area is confined to only with in 4-5 km depth and
of the Koyna region, as can be seen from Figure 1, there is absence of appreciable activity at
which are indicative of anomalously high tempera- greater depths. The velocity models in the
tures in the upper crust in this region. lower crust with an LVL and constant velocity
Smith et al [1975] pointed out that there layers, der ived in the present study also seem
seems to exist a correlation of the sialic to be consistent with the essentially aseismic
LVL with a decrease in the frequency of occur- nature of the lower crust in the Koyna region.
rence of earthquakes in the corresponding depth According to Chen and Molnar [1983] , the aseismic
range which is also a tectonically ' sof t' layer lower crust corresponds to a zone of low strength
with decreased rigidity and high Poisson's where ductile deformation predominate s. The
ratio [Braile et al, 1974]. Mueller et al [1973] temperature at the source region is likely
as well as Mueller and Rybach [ 1974] have also to be an important factor de termining whether
shown that in the Rh ine graben area and its deformation occurs seismically or not. Chen
neighbourhood the majority of hypocenters are and Molnar [1983] concluded that the limiting
concentrated close to the top of the sialic temperature s are about 250°-450°C for the crustal
LVL. More recently, Bormann et al [1985] and mater ials. At higher temperature s, the strength
Bankwitz et al [1985] reported that in the decrease s rapidly with increasing temperature
V og t 1 and ear thquake region of the southern and become s so low that no stress accumulation
German Democratic Republic, earthquake s occur is possible for earthquakes occurrence. There-
Geophysical Monograph Series Properties and Processes of Earth's Lower Crust Vol. 51

156 MODELING OF KOYNA DSS RECORD SECTIONS

fore, it can be inferred that temperatures Acknowledgements. We are grateful to Prof.

higher than these limiting temperatures are V. K . Gaur , Director, National Geophysical Research
probably prevailing in the lower crust which Institute, Hyderabad, for his continuing support
is essentially aseismic in the Koyna region and encouragement for this research activity
and those higher temperatures are also in turn and kind permission to publish this paper.
responsible for the proposed LVL and the constant
velocity layers in the lower crust in this References
region. Thus we conclude that the crustal P
velocity models derived in the pre sent study Bankwitz, P. , E. Bankwitz, and Grunthal, Parame-
are viable and are consistent with the observed ters of the Earth's crust and causes for the
seismic activity and the material properties occurrence of ear thquakes in the southern
and the ir dependence on the prevailing tempera- part of the G. D . R. , Gerlands Be itr . Geophysik,
tures. The two low velocity layers within the 94, 294-298, 1985.
crust in the Koyna region, as inferred from Bormann, P., A. Schulze, E. Apitz , P. Bankwitz,
the pre sent study, may suggest rheological and K-D. Klinge , Me thods and re sults of
stratification of this part of the lithosphere interpretation of deep seismic investigation
and they can be identified with levels of in the G. D. R . , Gerlands Be itr Geophysik,
increased ductility at those depths. 94, 259-268, 1985.
Braile, L.W. , R.B. Smith, G.R. Keller, R.M.Welch,
and R. P . Meyer, Crustal structure across the
Conclusions Wasatch front from de tailed seismic refraction
studies, J. Geophy s. Re s. , 79, 2669-2677, 1974
Travel times and relative amplitude s modeling Cerveny, , V., I. A . Molotkov, and I. Psenc ik,
of a series of crustal seismic record sections, Ray me thod in seismology, Univerz ita Karlova,
with the aid of synthetic seismograms, yielded Praha, 214 pp., 1977.
internally consistent velocity models with Chen, W.-P., and P. Molnar, Focal depths of
minor lateral variations in the Koyna region in trac on tinen tal and in trap late ear thquakes
to about 60 km depth. The prominent features and the ir implications for the thermal and
of these velocity models include, ( ) a strong mechanical proper ties of the lithosphere,
velocity gradient at shallow depths in the J. Geophy s. Re s. , 88, 4183-4214, 1983.
Deccan Traps and the upper cry stalline basement, Fuchs, K., Synthetic seismograms of PS-reflect-
(ii) an upper crustal LVL from 6.0 to 11.5 km ions from transition zone s computed with
depth with 0.2 km/sec velocity reduction within the reflectivity method, J. Geophys., 43,
the LVL and a broad transition zone at its 445-462, 1975.
bottom, (iii) a 1 km thick transitional boundary Kaila, K. L . , P.R. Reddy, P.R.K. Murty, and
centered at 20-21 km depth separating the upper K . M. Tr ipath i , Deep se ismic sounding studies
crust from the lower crust, ( iv ) a lower crustal along KOYNA I and KOYNA II profiles, Deccan
LVL from 26 to 28 km depth with 0. 14 km/ sec Trap covered area, Maharashtra state, NGRI
velocity reduction within the LVL, (v ) constant Tech. Report, 64 pp., 1979.
velocity layers with velocities of 6.3 km/sec Kaila, K.L., P.R.K. Murty, V. K. Rao, and G. E.
from 11.5 to 19.5 km, 6.7 km/ sec from 20.5 Kh are tchko , Crustal structure from deep
to 26 km and 6.9 km . / sec from 28 to 35.5 km se ismic soundings along the KOYNA II (Kelsi-
depth, and (vi) ) at least 2 km thick transitional Loni ) profile in the Deccan Trap area, India,
Moho from 35.5 to 37.5 km depth on the KOYNA II Tectonophy sics, 73, 365-384, 1981a.
and 1 km thick transitional Moho from 37. 0 Kaila, K.L., P.R. Reddy, M. M. Dixit, and M.A.
to 38.0 km depth on the KOYNA I DSS profiles, Lazarenko, Deep crustal structure at Koyna,
and a Pn velocity of 8.25 kny se c increasing Maharashtra, indicated by deep seismic sound-
with a gradient of 0.004 sec in the upper ings, J. Geol. Soc. India, 22, 1-16, 1981b.
mantle to about 60 km depth. Knothe , D. H . , Deutsche Demokratische Republik,
The upper crustal LVL inferred from the NKGG Geodat. Geophy s. Veroff. . Re ihe III,
present mode ling study, from 6.0 to 11.5 km 27, 59-68, 1972.
depth, is consistent with the observed seismic Krishna, V . G. , and K L .Kaila, Digitization
ac tivity concentration at 4-5 km depth and of analog DSS records and the ir reinterpreta-
an appreciable reduction of seismic activity tion with the aid of synthetic se ismograms:
at greater depths in the Koyna region. The Application to the Koyna data, in Deep Seismic
velocity structure in the lower crust, with Sound ings and Crustal Tectonics, Edited
an LVL from 26.0 to 28.0 km depth and constant by K.L. Kaila and H.C. Tewari, AEG Publica-
velocity layers, is also consistent with the tion, 99-119, 1986.
essentially aseismic nature of the lower crust Land isman , M., S. Mueller, and B. J. Mitchell,
in this region. The se two low velocity layers Review of evidence for velocity inversions
within the crust may suggest rheological strat- in the continental crust, in The Structure
ification of this part of the lithosphere with and Physical Properties of the Earth' s Crust,
levels of increased ductility at those depths Geophys. Monogr. . Ser. , Edited by J.G. He ac ock ,
in the Koyna region. AGU Monograph, 14, 11-34, 1971.
Geophysical Monograph Series Properties and Processes of Earth's Lower Crust Vol. 51

KRISHNA ET AL. 157

Langston, C.A., A body wave inversion of the of the Earth's crust in continents, I,
Koyna, India, Ear thquake of December 10, 1967, Evidence for a low-velocity zone in the
and some implications for body wave focal upper part of the lithosphere, Geophys. J. R.
mechanisms, J. Geophy s. Res., 81, 2517-2529, astr. . Soc., 10, 525-538, 1966.
1976. Mueller, S., E. Pe ter schmitt, K. Fuchs, D. Emter, ,
Langston, C.A., Source inversion of seismic and J. An sorge , Crustal structure of the
waveforms: The Koyna, India, Earthquakes Rh inegraben area, Tee tonophy sic s, 20, 381-391,
of 13 September 1967, Bull. Se ism. Soc. Am., 1973.
71, 1-24, 1981. Mueller, S., and L. Rybach , Crustal dynamics
Langston, C.A., and M. Franc o-Spera, Modeling in the central part of the Rhinegraben, in
of the Koyna, India, Aftershock of 12 December Approaches to Taphrogenesis, Edited by
1967, Bull. Seism. Soc. Am., 75, 651-660, H. Illies and K. Fuchs, Schweizerbart,
1985. Stuttgart, 379-388, 1974.
Mueller, S., Geophysical aspects of graben Rastogi, B. K. , C. S . P. Sarma, R. K. Chadha, and
formation in continental rift systems, in N. Kumar, Koyna ear thquakes of 1983 and 1984,
Graben Problems, Edited by H. Illies and NGRI Technical Report, 1985.
S. Mueller, Schweizerbart, Stuttgart, 27-37, Smith, R.B. , L.W. Braile, and G. R. Keller,
1970. Upper crustal low velocity layers: Possible
Mueller, S., A new model of the continental effect of high temperature over a mantle
crust, in The Earth's Crust, AGU Monograph, upwarp at the Basin Range-Colorado Plateau
20, 289-317, 1977. transition, Ear th & Planet. Sol. Lett., 28,
Mueller, S., and M. Land isman , Seismic studies 197-204, 1975.