You are on page 1of 11



This document may contain copyrighted materials. These materials have

been made available for use in research, teaching, and private study, but
may not be used for any commercial purpose. Users may not otherwise
copy, reproduce, retransmit, distribute, publish, commercially exploit or
otherwise transfer any material.

The copyright law of the United States (Title 17, United States Code)
governs the making of photocopies or other reproductions of copyrighted

Under certain conditions specified in the law, libraries and archives are
authorized to furnish a photocopy or other reproduction. One of these
specific conditions is that the photocopy or reproduction is not to be "used
for any purpose other than private study, scholarship, or research." If a
user makes a request for, or later uses, a photocopy or reproduction for
purposes in excess of "fair use," that user may be liable for copyright

This institution reserves the right to refuse to accept a copying order if, in
its judgment, fulfillment of the order would involve violation of copyright




Simple Models of Frictional Heating

by an Earthquake

2 . QJJg:ZE f.
'9. --•:...,"#•:,/<'•.".,).'.I-

0." i •:'*t·...f · '··:)%

6 1 .27 4
C .54 -

*Ir., l, 23•+•


Menlo Pork, California

1 ' 0./.
1 6 (1

1 97;..7 .....I .

14 18 MAR 1985
1I 1..'t 3_.__...._ .__...



Simple models of frictional heating by an earthquake


Arthur H. Lachenbruch

Open-File Report 85-131

This report is preliminary and has not been reviewed for conformity
with U.S. Geological Survey editorial standards and stratigraphic nomenclature .


Introduction. When fault surfaces move past one another an amount 2u

against a frictional ( or other dissipative ) resistance T, work is converted to
1/ heat in the amount 2Iu per unit area of fault. This results in a local
coseismic temperature rise (e) that will subsequently dissipate, presumably by
conduction. This effect has been discussed in terms of analytical results for
an infinitesimally thin fault by Sibson ( 1975 ) and McKenzie and Brune ( 1972).
Cardwell et al. ( 1978 ) considered the problem in terms of a numerical result
for the fault of finite width illustrated in figure 1. A simple form of the
solution for the fault of figure 1 has been applied by Lachenbruch ( 1980 ) to a
related problem and here we adapt it to a convenient general scheme for
estimating the magnitude of the coseismic temperature rise and its post-seismic
dissipation. For intuitive convenience, graphical results are presented in a
general dimensional form.

Coseismic frictional heating. Let t* be the duration of the slip event

and 2v = 2u/t* be the slip velocity ( see fig. 1 ). We assume heat is
transferred exclusively by conduction, deformation is homogeneous, and slip
velocity and fault width are constant during the event. The simplest case
occurs when the fault-width 2a is so large that little heat escapes by
conduction from the fault zone during the short time t* of the event. ( This
condition is a > 44at* where the right side is the effective distance heat can
diffuse in time tm in a medium of thermal diffusivity a which we assume to be
0.01 cm2/sec ). Denoting density by p and heat capacity by c and assigning
them reasonable values ( p = 2.65 g/cm3, c = 0.95 J/g ° C whence
pc =2.5 J/cm3 =0.60 cal/cm ) ° C 425 bars/°C ),we can write the temperature
rise 8 in terms of stress (T) and strain ( u/a).

I vt_I u
e= (la)
pc a pc a

# 4 ° C x B if T = 100 bars (lb)


% 40 ° C x 2 if I = 1000 bars (lc)


Thus for fault strains greater than unity ( u/a t 1 ) coseismic heating can be
I large.

If the fault is narrower than the distance heat can diffuse during the

I event ( i.e., if a < 44-at*) or equivalently if the event lasts longer than the
time constant ( A = a2/401 ) for the fault, then the temperature rise on the axial
plane of the shear zone depends explicitly on the duration of slip t*, and
( la ) is replaced by the more general expression ( Lachenbruch, 1980, equation
45a )

8(0,t ) = L vt [1 - 4i2 erfc --a-] , all a, 0<t<t* (2a)

I pc a 4=

-Tvt , a/44at > 1 (2b)

I pc a

I V./t
/ + 2- -1-_
pc 1-
4 irck
, a/44at; o (2c)


..... j It .....
....:Il- .11 It

1 ..-- la-r#.Mi

I 1 1•131 11< i 11
tio« t=o

1 •11 1 1' 1•,1,j
. .".....=tnt
. .Ill.-
I she dr , 1)•1
Wall..: 1 1 1...
i 2one ' 1 2 wall

Figure 1. The simple fault model. Dashed line (t=0) is deformed into
dash-dot line ( t=t*) during slip of amount 2u across shear zone of width 22
in an event of duration t*.


This is iUustrated in figure 2 which gives the rise in axial plane temperature
as a function of time during faulting for a range of event durations ( up to
100 sec ) and fault widths (0 to 40 m ) that probably brackets most
earthquakes. The temperature scale, based on a slip velocity 2v of 25 cm/sec
and friction T of 100b, can be adjusted proportiona•y for any other choice,
e.g., for 2v = 50 cm/sec and T = 1000 b multiply the temperature scale by

Figure 2 inustrates how wide-fault behavior ( e a t, equation 2b ) can

1: 15
pass to narrow-fault behavior ( 0 a t , equation 2c ) as the event proceeds .
With uncertainties of an order of magnitude in friction r, and at least that
much in fault-width 2a, an enormous range of temperatures is possible even if
slip velocity 2v and duration t* are known. If friction, fault width, and the
slip 2u were known, but duration and velocity were not, a range of
temperature is stiU possible on narrow faults. For example, slip of 50 cm
may have resulted from a 1-second event slipping at 50 cm/sec or a 10-second
event at 5 cm/sec. For I = 100b, a 4-mm-wide fault would warm by 500 ° C in
the first case and by 260° C in the second according to figure 2. ( In the
first case, the ordinate scale is multiplied by 2, and in the second, it is
divided by 5 to accommodate the two different velocities.) It is seen that the
curve labeled "4 cm" remains linear to t 4100 sec, probably longer than the
duration of local slip in most earthquakes . Hence equation (1) should be
useful for predicting coseismic temperature rise on the axis of faults
whenever shear zones are at least a few cm wide. In the Parkfield
earthquake, if the slip (2u) were 50 cm and fault width (2a) were 5 cm, then
according to (1) or (2) e would be 40 ° C for 100 bars friction, 400 ° C for
100Ob. Narrower faults attain much higher temperatures, but as we shall
see, they decay very rapidly after the event and are soon indistinguishable
from effects of wider faults with the same slip.

Postseismic decay of temperature disturbance. The postseismic conduc-

tive decay of temperatures in the fault zone depends upon two time constants:
the duration of local seismic slip t*, and the time A required for heat to
diffuse across the half width of the heated shear zone ( A = a2/4ot ). The
complete expression for temnerature decay ( t>t*) within a uniform fault zone
C |x• <a ) and beyond it ( • x1>a ) is ( after minor modifications of Carslaw and
Jaeger, 1959, equations 9 and 10, p. 80 )

0(x,t ) = I- v {t[l - 212 erfc _a-x . 2i2 erfc -aixl

pc a 4#3t

a-x a+x
-(t-t*)[1 - 2i2 erfc - 2i2 erfc ]} x<a
44£Y ( t-t*) 44a ( t- t*)

I v x-a x+a
= - - {t[i2 erfc - - iz erfc -1
pc a 44£YE 443-t

X- a x+a
-(t-t*) [i2 erfc - i2 erfc ]} x>a
440(t-t*) 440(t-t*)

1•i-- - - CH00
C 0].0 · (D N
9-15g.Q >3
' • 4-3 0 01
0421-01 ] 0
1a) Ei1 al 4410•.
.•8 1160 •0 .N@' Ee.
I ejo M-
0• 03.h>,-1
b 8a LA.•
0 -1 01
CHO Ou] 0,R1 41r-|0 4-3aj042H
11 \di LOES
4 :1310 • &1 -

431(29 5>«031
M 0 CO >hO (1)
-0C 9..0
0 . 6000 00 4
WU %09'-IS
0]r-1> 1 042r-1
. 4- Ce .8 9 9 3
li i
'41 Z .i 111 r-IrCIr-1
Arnt9 •11'
Ul r-1
.li K .11 33. CO83
4' 1.- 0-1-)0
C.C 0LnA·A
r ZI *55511 0 X \\ 3 0#\4 r (0 5 S 1 0U
F. 1 0... . . 11 0 .H.=44>fl
g-1 40 0 0.1 4-3
. 00
0 -4 0 0000
4 HA -1 - Q)'NPLE
0UC Eidy =28.
L I-E4'. cd4 >(D ·A4 0-042
00, 060 0042r-160042
l • 0, • olt 0 A 4
42 • F-, U•••
/1 //-. m
.• 5-0 SESSSes 0
12.1 1 4(11
0 E
5I. 03*0
60*** S ,», T .C A 0
8 E °18 k
" .i • 254 2
1 T Zol 40
L 6--' ' ' Ar-· · ·
1, 042 1.r,
t0 i #-i.: •
M •4 r,4 r..1 9#11 •2•
04! N d 51 @"III .p
2,«W' SZ=AZ '9001-1-'°3 (J,) e 2-In+Di,dulg-L Duold ID!xv 4-1d000
E-·i u) EH

( The general expression for the coseismic ( t<t*) temperature rise is obtained
by deleting the second lines in equations (3a) and ( 3b), e.g., by formally
setting t E t*.)

The slip duration t* win generally be only a few seconds and can
therefore be neglected for computing the temperature hours, days or weeks
after faulting. Neglecting slip duration t* relative to post-seismic observation
time ( t-t*), we can treat the faulting as an -instantaneous source of strength
ru/pca distributed throughout the shear zone (| x| < a) and (3) is replaced
by ( Carslaw and Jaeger, equation 3, p. 54 )

8 (x,t) = 1 L u [erf -a-x + erf _aixl all x, t >> t* (4)

44aE •SE

Finally, if our observation time is large relative to the time constant X of

the shear zone ( for a 4-cm wide shear zone A is %100 seconds, but for a 4-m
shear zone, it is a10 days, see inset, fig. 1 ) the fault can be treated as an
instantaneous plane (zero-width) source of strength 2Tu/pc. This leads to
the further simplification ( Carslaw and Jaeger, i959, equation 4, p. 259 ):
u -4£YE
8(x,t ) = I- e all x, t>> t*, t>> A (5)
pc i-

Temperature rise computed from (5) is shown in figure 3 for a fault slip 2u
of 50 cm opposed by frictional resistance T of 100b. The figure shows that
the high axial plane temperatures in the previous numerical examples are
short-lived. If the slip were 100 cm, the temperatures of figure 3 would be
multiplied by 2, if the friction were 1kb, they would be multiplied by 10.
The one-day curve is valid only for faults for which the time constant
( A E a2/40 ) is much less than a day, e.g., for 28 <40 cm in which case A J3
hrs ( see inset, fig. 1 ). Temperatures earlier in the postseismic period can
be obtained for a fault of any width from the more complete expressions (4)
or ( 3). As the area under curves of the type shown in figure 3 is the
source strength 2'cu/pc, it is conserved. Hence a post-seismic temperature
profile across the fault combined with an estimate of slip u provides an
opportunity to recover the friction I.


e (t) for 7=100 b, lu= 50 cm

I day
1/00- '
0.50- I week
3 months
1 1 l l J
5 4 32101234 5
Distance from fault axis, ixl Cme+ers)

Figure 3. Post-seismic temperature distribution. Temperatures are

shown for friction T of 100 bars and slip 2u of 50 cm; they can be scaled
proportionally for any other values. Curves are good approximations as
long as the post-seismic time they represent is large relative to the
fault's time constant X(a): see inset, figure 2.



/ Cardwell, R. K., Chinn, D. S., Moore, G. F., and Turcotte, D. L.,

1978, Frictional heating on a fault zone with finite thickness: Geophysical
Journal of the Royal Astronomical Society, v. 52, p. 525-530.

I •
Carslaw, H. S., and Jaeger,
Oxford University Press, 510 p.
J. C., 1959, Conduction of Heat in Solids:

I Lachenbruch, A. H.,
resistance to fault motion:
1980, Frictional heating, fluid pressure, and the
Journal of Geophysical Research, v. 85, p. 6097-

I McKenzie, D., and Brune , J. N., 1972 , Melting on fault planes during
large earthquakes: Geophysical Journal of the Royal Astronomical Society,
v. 29, p. 65-78.

I Sibson , R. H., 1975 , Generation of pseudotachylyte by ancient seismic

faulting: Geophysical Journal of the Royal Astronomical Society , v. 43,
p. 775-794.





t. •

L 1

( 1


l.- 1