79

3
Geometryof seismic
wave
paths

fft"tffiter usesa geometrical-optics

approach
to
derivethe basicrelationshipsbetweentraveltimeand the
locatio$s of reflecting/refracting interfaces; *ori rt*"-
tural ifterpretationrelieson suchan approach.
The accurateinterprctation of reflection data re-
quiresd knowledgeof the velocityat all points alongthe
reflectibnpaths.However,evenifwe had sucha detailed
knowleflgeof the velocity, the calculationswould be
tedious$andoften we assumea simple distribution of
velocitfwhich is closeenoughto give usehbleresults.The
simpleslassumption,which is made in g3.1,is that the
y.l1ilqr-t constantbetweeathe surfaceand the reflecting
bed. Afthough this assumptionis rarely even approxi-
matelyfrue, it leads'tosimpleformulaewhich give answcrs
which$Jewithin thercquiredaccuracyin a"tty inrtunor.
The basicpr6blem in reflection seisniii surveyingis
_
to deteimine the position of a bed which gives rise to a
rcflection on a seismicrecord. In generalthis is a prollem
in thru_e dimensions.Hower,er,the dip is often very gqtle
a1f th€ dhection of profiling is frequently nearty ilong
eitherthedirectionof dip or the direaion of strike.In suc[
cascs:atwo.dimen5ionalsolutiod.is generallyused.The
arrirdl:tirne versusoffset rclation for a plifre.rcflector and
coostantvelocity is hyperbolic. to the rcflec-

dT*{+ a $ophone is mowd away from the shotpoint,

callednornitl
.hlmoveout,
moveout-provides
orovidesthemosf
mostimflorrrntftir€-
importantaite-
 GeometryoJ'seismicwaoepaths 80

rion for ideotifying re{lectionsand a method of determin- geophonedistance(offse\, we can write

ng rrclociE'.The dip is found from difTcrqncesin ariival
:imesof a reflection at dilferent locationsafter correction V 2t2-x2+ 4h2, (3.1)
br normal moveour; dip groreout is rclated to dip and, or
tlso to the angle of ap,prroach of wavefrontsat the surface
md to apparcnt velocity. Reflector dip and strfte cdn be (v2t214h21 - g2l4h\: t. (3.2)
bund from the compoients of dip morrcbutseenat the Thus the traveltime curve is a hyperbola as shown in the.
ntersectionof seismiclioes. upper part of fig. 3.I .
Section3.2 dealswith reflection ralpa.thswhercthe The geophoneat R will also record the direct waoe
alocity,clrangesverticaily; this resultsin chatges iri ray- which trarrclsalong the path SR. Since SR is always less
lath dire$ion. One soltttion is to use cquivalenf a\€rage than (.SC* CR), the direct waryearrives first. The trarrcl-
elocity. Fot parallel velocitylayersthe cff€cton thetra\El- timc is t, = xlV and the traveltime curvesate the straight
ime curve is to girre a root-mean:squarc(rms) velocity. lines OM and ON passingthrough the origin with slopes
lertical vbloc-ityis oftc,gexpressedas a function of arriral of tllv.
lme or dgpth. l$here;iiLlocity is linear with depth, para- When the distancex becomesvery large the differ-
ronts lre spheriealadd raypa.tbs8re arcs.ofcirclesifacts ence betweenSR and (.SC+ CR) becomessurall and the
rtich can Ueusedin gpphical ptoning of depth s"4ions. reflection traveltime approachesthe direct wavetravcltime
Section 3.3 concernsthc geometry of headlfarre asymptoticalll,';.
aths as usedin refraction explo4qtion.In most casgswe The location of the reflecting bed is detcrmined by
ssumea seriesof bedswith the sane strike and eact with measuringlo, the traveltimefor a geophoneat the shot-
constarit velocity, the ri:tocity increadingas we go to
eeperbeds,and,tlren we derirre.formulaer€lathg travel-
me, ofhs-t, dapth, dip and velosities.,'Thec15gs,,,c6n- Fig.3.l. Traveltime curve for horizontal reflector,
dered inehde a singlehorizontal rcfractgr, senerdlifrori-
rffa! reifi{iritor.s,a'singtd ilipinll refractor, and several
ipping nftactors. Velociticscan h founit from slopcsof
retrayrltim€ vers,usoffset curves,dcpth$from the inter:
pts of projectionsto the Sotpoint, and Sp by differ,inces
dlepth'at adjacent shotfuints. Refradibn paths * the
Ee of a linear increasein owrburdm velocity are.also
uisidered. ,
i
I i' Geonetryof relleftion pgtls.for sfirstant veto&y
l.l Elorizontalre;flector;nornul npocilrt .I
T,h€ simplest twodimensional pro,blemis thft of
ro dip illustrated in the lower Frt of tig. 3.l. fhe refoect-
gtr.d, AB, is at a depth } bclon *fre shrrydnt S. Erkgy
rving S along the dircction SCwill be ne,flccted in sucha
rection that the angle of:.re{lectioncqr{alsthe'angh of
sidence.
Although the reflectedray CR canbc determinedby
nng offan angleequal to d at C, it is easierto make.use
tfu imagepoit f $rhi;h is locatedon thesamenormalto
e reflestor as S and as far belowthe bed'as.gis abow. If
: join f to Cand prolong the strlight line to & CR is the
llected ray (since CD is paralhl to Sd making,:allthe
gfesmarkedaequal). ' ' '
D€FotingJe€alieragr+elocity by Vnthetravehirncr
r th"reflected w,aveis(fC:F Hdwlvds, Sg',:.ffi-
-CfttZ.,
lhat IR is egual in lenflifi-io ihe aihr4l path;-S(k."--
crefore, , : IRIV and in terms of x,,lhe shotpoint-t{"
3.1 Geometryof reflectionpathsJbr constantoelocity
,81
ooint.Settingx: 0 in (3.1)we seethat r, !o and Ar"; this forms the basisof the l-Almethod of
h: r vto. velocity (see97.3.30)and also of veloCityanalysis
(3.3) t{rg
(98.2.3).Brown (1969)discusses refinementsto handledip
Equation(3.1)can be fritten and ldng offset.
f :(x2lV2) +(4h2lVr: + tzo.
1x21v21 (3.4)
3.1.2 Dippingrcflector; dip moaeoat
lf we plot !2 againstx2 (insteadof , versusx as in fig. 3.1), Whenthe bed is dipping in the direction of the pro-
ve obtain a straight line of slope(llV2) and interceptrfr. file, we havethe situationshownin fig.3,.2,( beingthe dip
fhis forms the basis of a well-known schemefor deter- and i the distancenormal to the bed.To draw the raypath
nining Z, the'X2-7"2 method'; this will be describedin for the reflectionarriving at the geophoneR, wejoin the
t.J.5 A . imagepoint l to R by a straight line cutting the bed at C.
We can solve(3. I ) for l, the traveltimemeasuredon The path is then SCR and r is equalto (SC * CR)lV;since
he seismicrecord.Generally2h is appreciablelargerthan (.SC+ CR) = /R" application of the cosine law to the
: so that we can usea binomial expansionas followi: triangle S1Rgives
t: (2hlv){l + (xl2h;2}I: to{t,,* (xlVtol?1+-- V2t2 :|I i2
: ro{1+i\lvtd' - : x2 + 4h2- 4hxcosllr + e)
fi(xlvtda+ .. .}. (3.5)
I tr, tr,r, uoi aretwo traveltimesand offsets,we have : x2 + 4h2+ 4hxsin{. (3.9)
",
) the lirst approximation On completingthe squares,we obtain
Lt: tz - t, x (xl - xl11zv2to. (3.6) V2tz _(x + 2hsint)z _ ,
-(z,cosljz-:
r the specialcasewhereonegeophoneis at the shotpoint, Ona;* ''
I is known asthe normalmoueoat(NMO) which we shall
Thus, as beforethe traveltimecurveis a hyperbolabut the
enoteby Al". Then,
axis of symmetryis now the line x: -Zhsint insteadof
Alnx x2f2v2to. (3.7) the l-axis.This meansthat r has different valuesfor geo-
phonessymmetricallyplabedon oppositesidesof the shot-
t timeswe retain anotherterm in the expansion(seealso point, unlike the casefor zero dip.
roblem3.lc):
A,t^= (xzlTV2t) - (x4lgV4t3o) Fig.3.2.Traveltimecurvefor dippingreflector.
: (x2l2Vz ts){l - (xldk)2}. (3.8)
From (3.7) we note that the normal moveout in-
€asesasthe squareofthe offsetx, inverselyasthe square
'the velocity,inverselyasthe first powerof thetraveltime
r depth - see(3.3)).Thus rellectioncurvatureincreases
.pidly as we go to morc distant geophones,at the same
ne the curvature becoming progressivelyless with in-
easingrecord time.
The concept of normal moveout is extremely im-
rrtant. It is the principal criterion by which we decidc rQ
rether an event observedon a seismicrecord is a reflec-
Fa.'i+ar .-l
,n or not. If the normal moveout differs from the value
ren by (3.7) by more than the allowableexperimental R
for, we are not justified in treating the event as a reflec-
rn.Oneof the most importantquantitiesin seismicinter-
etation is the changein arrival time causedby dip; to
rdthisquantitywe musteliminatenormal mov€out.Nor- tZ
tt
rl moveout must also bc eliminated before .stacking' \
ti,
lding together) common-depth-point records (see
3.1). Finally, (3.7) can be usedto find Z by measuring
I

x/
I
 Geometryof seismicwavepaths 82

Settingx equalto 0 in (3.9)givesthe samevaluefor differencebetweenro at differentshotpoints.As showni

& as in (3.3); note, however,that l, is not measurcdvcr- fig. 3.3,A/u : lo, - ?o2and
tically as it was in the earlierresult.We call the points C,
C' , C" in frg. 3.2, wherethe anglesof incidenceand reflec- .ioi:12(f), (3.r2
tion are equal, refiecting points. (Ihese are sometimes
called'depthpoints' but this term is alsousedfor the point where Ax is the distancebetweenshotpoints.When w
on the surfacc midway between source and receiver; we measuredip on a recordsection($5.6.3),Ax is the distanc
call the latter a midpointand to avoid confusionwe shall betweenany two convenientpoints.
avoid the term 'depth point'.) The updip displacementof It shouldbe noted that normal moveoutwaselimi
reflectingpoints comparedto midpoints for dipping re- natedin the derivationof (3.1l). The termsin (Ax)2whicl
flectors is important in migrating data ($5.6.3)and in the disappearedin the subtractionrepresentthe normal move
common-depth-pointmethod($5.3.l). out.
To obtain the dip (, we solvefor t in (3.9)by assum- Fig. 3.4illustratesdiagrammaticallythe relationbe
ing that 2lr is geater than x and expandingasin the deriva- tween normal moveout and dip moveout. The diagram a
tion of (3.5).Then the left representsa reflection from a dipping bed; th
(- alignment is curved and unsyrnmetrical about the shot
r : 72h x2 + 4lxsin At
[ , +----4 F -) point. Diagram (,B)showswhat would havebeenobserver
if the bed'i':ad'bien horizontal; the alignmentis curve<
(. x2 + 4&xsint\
-,ot +--w-) (3.10) symmetricallyabout the shotpoint position owing to thr
normal moveout. The latter ranges from 0 to 13 rn
(1 millisecond: l0-3 s : I ms,the unit of time commonll
using only the first term of the expansion.The simplest
usedin seismicwork) at an offset of 400 m. Diagram (C
method of finding ( is from the differencein traveltimes
was obtainedby subtractingthe normal moveoutsshowr
for two geophonesequallydistant from, and on opposite
in (8) from thearrival timesin (l). The resultingalignmen
sidesof, the shotpoint.Letting x havethe values*Ax for
the downdip geophoneand -Ax for the updip geophone showsthe effectof dip alone; it is straightand hasa timr
differencebetween-theoutsidecurvesof l0 ms, that is
and denoting the equivalenttraveltimesby l, and rr, we
ger Aro : 19ms whenAx: 400m. Thus,we find that the diJ
is 2500(10x 10-3/800): 0.031rad : 1.8o.
,, =," (, -, (ax)'z+jl4axsin1, The rnethodof normal-moveoutremovalillustratec
in fig. 3.4cwasusedto demonstratethe differenceberween
normal moveoutand dip moveout.If we requireonly thr
,,-,"(-eu=gry),
Lto: 1,- t2 .(na*, xzfsint. Fig.3.3.Ceometryinvolvedin Cipmoveoutmeasured
- ) between shotpoints or on record sections.
The dip i is given by

,in(**r(9. (3.r l)

The quantity Alu/Ax is calledthedip moueoul.(Note

that dimensionallydip moveout is time/distancewhereas
normal moveoutis time.) For small angles,f is approxi-
matelyequalto sin { sothat thedip is directlyproportional
to Ato under thesecircumstances.To obtain the dip as
accuratelyas possible,we useas largea valueof Ax as the
data quality permits; for symmetricalspreads(95.3.2),we
measuredip moveoutbetweenthe geophonegroupsat the
oppositeendsofthe spread,Ax then beinghalfthe spread
length(fig. 3.2).
Dip moveout can also be measuredby the time
3.1 Geometryof retlectionpathsfor constanraelocity
,83
dip moveout Ato we meretysubtract the traveltimesfor the The apparentoelocity I/" of 4 wavefront is the ratio
two outsidegeophonesin (l). of the distance (Ax) between two points on a surface
Frequently we do not have a symmetrical spread (usually the surface of the ground) to the difference in
and we find the dip moveout by removing the effect of arrival times (Ar) for the same event at the two points.
normal moveout. As an example,refer to fig. 3,4 curve It is given by
(D), which showsa reflection observedonb spreadextend-
ing from x: - 133m to x: *400 m. kt to = 1.225s, 4: LxlLt- Vlsina, (3.r3)
t r : 1.223s€c,12: 1.242s, V : 28ffi m/s. From (3.7)we wherec is the angleof approach(fig. 3.5); a is sometimes
get for Atn at offsets of 133 and 400 m respectivelythe calledapparentdip.Thts equation is somewhatsimilar to
values I ms and 8 ms (rounded off to the nearestmilli- (3.11)and (3.12)but it-has different significance,sinceit
secondsincethis is usuallythe precisionof measurement givesthe direction oftravel ofa plane wave as it reaches
on seismicrecords).Subtractingthesevalueswe obtain for the spread, Vbeing the averagevelocity between C and
the corrccted arrival times tr 1.222,tz: 1.234i hence the surface.In (3.1l) and (3.lZ) Vistheequivalentvelocity
-
the dip moveout is 12 msl(533m/2). The corresponding down to the reflector and ( is the angle of dip. Because
dip is ( :
- 2800(12x t0-3/533) 0.063rad : 3.6o. $in d can be very small, Z"can be very large and for enerry
An alternative to the above method is to use approachingvertiially, V"- q.
the arrivaltimesat ,r : - 133m and.x : * 133m, thus
obtaining a -S;1:mnistdgal spread and eliminating the 3.1.3 Crossdip
need for calculating normal moveout. However, doing When the profile is at an appreciableangle to the
this would decreasethe effective spread length from direction of dip, the determinationof the latter becomes
533 m to 266 m and therebyreducethe accuracyof the a three-dimensionalproblem and we use the methodsof
ratio (Alo/Ax). solid analyticalgeometry.In fig. 3.6 we take the,ry-plane

Fig.3.4.Relationbetweennomralmoveoutanddip
moveout. For curves{A),(B)and(C),ro: 1.000 s,
V : 2500m/s.For curve(D), to = 1.225s, t, : 1.223s,
tz: l-242s, ,/ = 2800m/s.

--1r-
,?v
l 0 ms

I '\44,
t4{x)n
+383 Jw
+267
+200
+\^
jl, ^
4t-
+ t33 -nJ^
JA,^
J\r-
J\t-
+67

0
tc
\y'l\h" \4r-
to

-67
r\ . r\i-
- r3l
-200
J\'-
-'r\^
-261
J\^

+
-333

-400 m
Geometryof seismicwaoepaths

Fig.3.5.Findingtheanglcof approachof a $,ave. as horizontal with the z-axis extendingrcrtically down-

ward.The line OP of lengthI is perpendicularto a dipping
pftne bed which outcrops (that is, intersectsthe xy-plane)
along the line Mffif eKendedsufliciently.
We write er,02, e3 for the anglesbetweenOp and
x-, ,r- and z-,axes,I, m, n for the direction cosinesof OP.
The angleEbetweenMIf and the x-axis is the direction
of strike of the bed while 83 - (, the angle of dip.
The path ofa reflectedwave arriving at a geophone
R on the x-axis can be found using the imagepoint .L The
line joining .I to R cuts the reflectorat Q, henceOe.R is
the path. SinceOp : O/, the line /X is equal to Vt, tbeing
the trarreltimefor the geophone.atR. The coordinates of
Fi93.6. Three-dirnensional view of a reflection path for .I and R arerespectively(2/r/, 2lvt, 2lm) and(x, 0, 0), hene
a dipping bed. we have
ORMx yzf :eR)2
= (x - 2ht)2 + (0 - 2hm)zi (O- 2hn)2
: x2 + 4h2(12+ m2 + n21- 4hlx
= x2 + 4h2- 4hlx,

since12+ m2 + n2 : l.
When x: 0, we obtain the samerelation between
i and roasin (3.3).Proceedingasin the derivationof(3. l0)
,ta o lr
\r"l we get for the approximatevalueof !,
\l
\ .-.. (, , x2- hl x\
tl tol l + ----;;--f.
I u/t'
\ /
By subtractingthe arrival timesat two geophoneslocated
on the x-axis at x : * Ax, we find ,
' Lt,x tn(lA,xlh)
x2lA,xlV,
Fig.3.7.Determination of strike.
/: cosgr
=Iv(+\. (3.r4)
\a*,/
If we also have a spread along the 1.alr 1"rrrr-
spread),we get

m: cosoz
" +, W), (3.15)

where Al, is the time difference ('cross-dip') between

geophonesa distanceZAy apart and symmetricalabout
the shotpoint.Since
n: cos(: {l _ (/2I mt)} t,
si n(:(l -r\+ :Q7+ mz1t

:+u(t*)'.
(*)'l (3.l6)
 3.2 Vertical uelocitygradientand raypath curu^ature 85

The componentsof dip moveout,A,tJA,x and AtrlLy, are changesin seismicvelocity as we move horizontally are
also called apparmt dips. for the most part small, being the result of slow changes
To find the strike 3, we start from the equationof a in densityand elasticpropertiesofthe beds.Thesehori-
plane (that is, the reflector) which has a perpendicular - zontal variations are generally much less rapid than the
from the origin of lengthlr and directioncosines(1,m, n),. variations in the vertical direction where we are going
namely from bed to bed with consequentlithological changesand
I x* my * n z :h . increasingpressurewith increasingdepth. Becausethe
horizontal changesare gradual they can often be taken
Settingz : 0 givesthe equationof the line of intersection into account by dividing the surveyarea into smaller areas
of the reflector and the surface: this strike line has the within each of which the horizontal variations can be
equation ignoredand the samevertical velocity distribution used.
Such areas are often large enough to include several
lx * mY: 11'
structures of the size of intercst in oil exploration so that
The interceptsof this line on the x- and y-axesarc hll
and hlm. Referringto fig. 3.7, we find that Fig.3.8.Determining dip andstrikefrom non-
hlm l perpendicular observations. (a) Relationbetweenthe
fa nS:
-hll
:
- point of observationO andthe reflectingpoint ,il
m (,tlis alwaysupdipfrom O); (6)exarnple ofgriilhical
solution.
-(AtJA,x) (3.17)
( t/Ly)' t2

Consider the case where the profile lines are not (a)

perpendicular,for example,wherethey are in the r, and

rr directionsof fig. 3.8aand the dip is in the ro direction.
We expressthe dip moveout as the vector (d/d-r)ro:
AO; the componentof dip moveout on the line in the 12
directionis thus (drldx)rs.12: (drldx)cosp: lOBl (see
problem 3.?a).The converseproblem of finding the total
dip moveout from measurementsof the componentsof dip
moveout OB alaidOC can be done graphically as shownin
frg. 3.86 (seealso problerh 3.2b), or mathematicallyas
follows. We tak€ one profile along the x-axis and the other
along the yraxis at an angle a to the x-axis. Taking the
length of a symmetrical spread along the y'-axis as Ay,,
thc coordinates ofthe ends of thc spread (relative to the
x-, y-axes)are fAy'cosc, !Ay'sinc. Then
V2 t2*: Qhl + Ly' cosa)2+ (?imt Ay' sinc)2 + (2t ri2
: (Ay')z + 4hz+ 4hLy'(toosd* zsiuc). (b)
Apparcnt dip 20 msikm

The dip moveout along this line, A{Ay', is

Lt' / AY' : 2(l cosa* m si', a)/ V. (3.l8) \

\a
\eo
Sincec is known, /can be found from At/Ax and m from \76-
.(3.r8). \?4
\a-
-- \?
3.2 Vertical velocity gradicrt and raypatt currafirre \-
-- \
3.2.1 Effect of oelocity sodation --
The assumption of constant rrclocity is not valid in ---
general, the velocity usually changing as we go from ohe
point to another. In petroleum exploration we arc usually
l- - r : - - ---ilt
 ' Geometrl,of seismicwaoepaths 86

changesfrorn on'e velocitl'function to another do not this reflector; usingthe ,raluesof lo, the dip moveout Ard
necessarilyimposea seriousburdenupon the interpreter. and V, we calculatethe depth fr and the dip { using (3.3)
and (3.11).
3.2.2 Eqaioalentaoerageoelocity
Vertical variations in velocity can be taken into 3.2.3 Velocity layering
account in various ways . ODe of the simplest is to usea A method which is commonly used to take into
modification of the constant velocity model. We assume accountvelocity variations is to replacethe actual velocity
that the actual sectionexistingbetweenthe surfaceand a distribution with an approximateone correspondingto
certain reflecting horizon can be replacedwith an equiva- a number of horizontal layersof different velocities,the
lent singlelayerof constantvelocity Zequal to theaverage velocitybeingconstantwithin eachlayer.Simpleequations
velocity betweenthe surfaceand the reflectinghorizon; suchas(3.3)and (3.1l) areno longerappropriatebecause
Zis the equioalmtauerageoelocity.This velocity is usually rays are bent at each interface. A graphical method
given as a function of depth (or of to, which is nearly the can be used to fmd the depth and dip. The method uses
same except when the dip is large). Thus the section is a \ravefront chart' ; the preparation and useofthese charts
assigrreda different constant velocity for each of the will be desgribedin 95.6.3.In effect this method replaces
reflectorsbelow it. Despitethis inconsistenrythe method the actual raypaths with a seriesof line segmcntswhich
is usefuland is extensivelyapplied. The variation ofthe are straight within eachlayer but undergo abrupt changes
averagevelocitywith to is found usingone of the methods in direction at the boundariesbetweenlayers.Dix (1955)
describedin $7.3.For the observedvaluesof the arrival shows that the effect of this on (3.4) is to replacethe
time !s we selectthe averagevelocity Zcorrespondingto averagevelocity 7 by its rms (root-mean-square) value,
I/. His derivationis as follows. Referringto lig. 3.96,the
X2-72 curvefor the reflectionshownin fig.3.9ais curved
Fig.3.9.Derivationof formulafor rmsvelocity7 in and for a given offset x we write the equation of the
twoJayermedium.(a) Reflectionpath;(b)X2-72 tangentas
curve.
t2: x21Vz
+ifi
.
(a) hence
dtld,x: xlV2t. (3.le)
The angleofapproach, ir, is given by

si ni , : v,dt
' :v,* . (3.20)
'dx V,i
using (3.19). Also, writing A/, for the one-way vertical
traveltime through the ith bed, and keeping x small, we
have
jx: Ax1 * Ax2 : hrtani, + hztaniz
x I\L,trsini, * VrLt"sini,
(b) f
x (VlN, t VzrLtr)sinirlV,
= (vlN, + vlN,)glV2l
from (3.20).Sincer I 2(N, + Arr)qwe get
2 12
,' * N,
zv?NtlL
This equation can be generalizedfor r horizontal beds,
giving
t2 z x2l V t + 7?. (? 7r\
3.2 Vertical oelocity gradient and raypath curuature 87
n
72- \' v?d,t,lLtt,. sinr, sinro
(3.22) -T: :P'
i=I vo
(SeeShah and Levin, 1973, for higher-order approxima- 4:4k) ,
tionsnecessary
to givemoreaccuracyfor largevaluesofx.)
Axn : Azntan io,

3.2.4 Yelacityfunctions
Ltn::t' . .
At times the assumptionis made that the velocity /, cos Ir
varies in a systematiccontinuousmanner and therefore
can be representedby a velocity function. The actual The raypathparameterp(see(2.103))is a constantwhich
velocity usually varies extremely rapidly over short dependsupon the direction in which the ray left the shot-
intervals,as shown by sonic logs (see57.3.2):however,if point, that is, upon ro.
we integratethesechangesover distancesof a wavelength In the limit when z becomesinfinite, we get
or so (30-100 m), we obtain a function which is generally sini sinr"
smooth exceptfor discontinuitiesat marked lithological :fr:n, V=V(z), (3.23)
changes.If the velocity discontinuitiesare small, we are
V
often able to representthe velocity distribution with suf- dxdtl
: ran I'
ficient accuracyby a smooth velocity function. The path l! dr: v*"i'
of a wave traveling in sucha medium is then determined
by two integralequations. *: I't*rdr, ,: l" :Z-:
Jo Vcosi' lo
To derivethe equations,we assumethat the medium
is divided into a large number of thin beds in each of
which the velocity is constant; on letting the number of pV dz
beds go to infinity, the thicknessof each bed becomes *: f' I
infinitesimaland the velocitydistribution becomesa con- l"p:1rvvY'1 (3'24)
tinuous function ofdepth. Referringto fig. 3.10,we have a, ''
, - f' l
for the nth bed Jo v{t - (pv)'i}' )
Since Z is a function of z, (3.24) furnishestwo
integral equationsrelating x and t to the depth z. These
Fig.3.l0.Raypathwherevelocityvarieswith depth.
equationscan be solvedby numericalmethodswhen we
havea table ofvalues of V at variousdepths.

3.2.5 Linear increaseof uelocitywith depth

Sometimeswe can expressZ as a coutinuousfunc-
tion of z and integrate(3.24).One caseof considerable
importance is that of a linear increaseof velocity with
depth,namely
V: Vo* az,
where Zo is the velocity at the horizontal datum plane,
/ the velocity at a depth z below the datum plane and
a is a constaatwhosevalue is generallybetween0.3 and
1.3/s.
If we introducea newvariableu : pV: sinr, then
du - pdV : pade. and we can solve for x and r as
follows (p is the raypathparameter):

, : ! f" rruour.r:
I
_ u-), pa tt -
r\tl":!"oril'"
pa J uo\t pa
l, I,
| ,^^.,
-
 Geometryof seisiic row poih, 88
I

Fig.3.l l. Circular ray leavingthe shotpoint at the angle

b.

hence,
i :2tat-r(eo'tanlro;, (3.27)
z:(V - Vs)la= (sini- snis)lpa. (3.28)

The parametric equations (3.25)and (3.28) give the coor-

dinate$x and z, the paramet€r i being related to the one-
way traveltime t by Q.2A or Q.27).
The ralpath grvcn by (325) and (3.28) is a circle;
this can be shown by calculating the radius of curvrature
p which turns out to be a constant:
p: (l + x'2fl{,
Fig.3,l2. Construction of wavefrontsfor linear
where . increaseof velocity.

*' :*: tanr,using(3.25),(3.28),

oz
. . dz x d, ^dt ,.d i
X" : -7- - =(tan t)--: S€C-l-
oz- or oz oz
- pasec3i, using (3.28).

flence
(l + tan2rlt 1 /v"\ I
P:-: - = l - l -:: c o n s l a n t.
ws ec - t p a \,t/s rn ro
Fig. 3.ll showsa ray leavingthe shotpoint at the
angle ro. The center, O, of the circular ray lies above the
surfacea distancep sin i6, that is, Vsla. Sincethis is inde-
pendent of lo, the centers of all rays lie on the same
horizontal line. This line is located where the velocity
would be zero if the velocity function were extrapolated
up into the air (sincez: -Vola at this elevation).
To determine the shapeof the wavefront, we make
use of fig. 3.12. The raypaths Sl and SB are circular
arcs with centers Oy and 02 respectively.If we continue
the arcs upwards to meet the vertical through S at the
point S,, the line O1O2bisects S'S at right angles.Next
we selectany point C on the downward extensionof ,S'S
and draw the tangentsto the two arcs, CA and C8. From
plane geometry we know that the square of the length of
a tangent to a circle from an external point (for example,
CA\ is equal to the product of the two segmentsof any
chord drawn from the samepoint (CS.C.S' in fig. 3.12).
Using both circles we see$at
CS ' C. S ' : CA i:C 8 2 .
 3.3 Geometryof refractionpaths 89

hence Cl : C8. Thus a circle with center C and radius 3.3 Geometryof refraction paths
R: CA cuts the two raypathsat right angles.SinceSl 3.3.1 Single horizontal refractor
and SB can be any raypaths and a wavefront is a surface Refraction seismology ihvolves the study of head
which meetsall rays at right angles,the circle with center waves,which have beenintroduced in $2.4.7.In the case
C must be the wavefront which passesthrough A and B. of a single horizontal refracting horizon, we can readily
Even though the arc Sl is longer than SB, the greater derive a formula expressingthe arrival time in terms of
path length is exactly compensatedfor by the higher the offset, the depth and the velocities.In fig. 3.13 the
velocityat the greaterdepth ofthe raypath Sl. lower part shows a horizontal plane refractor separating
We can draw the wavefront for any value of r if two bedsof velocitiesV, and V2, where V, > V1. For a
we can obtain the valuesof ^EIand R in fig. 3.12.Thus, geophoneat R, the path of the refracted wave is OM?R,
the quantities If and R are equal to the values of z and @-beingthe critical angle.The traveltimer can be written
x for a ray which has i: lr at the time r, that is, SD in
the diagram. Substitutionof f : *z in (3.25),(3.27) and , OM,MP .P R MP ^OM
(3.28)yields ' :T - T - vr : k- ' T
tanllo = e-"', sin16: sechat,cosro: tanha!, _x -2htan@ * 2h
Y2 Zrcos@
1t: (llpa)(l - sinro)
: (Vola){(llsinro)- l} = i 1+
Vz
%
Vrcos@\-
(t
-2'"")
: (Zo/a)(coshat - l), (3.2e) x +-. 2icos@
:- (3.30)
i : (llpa)as io : (Vsla)cotis V 2' Yr '
: (Vola)sir.hat. wherewe have usedthe relation sin@ : Vrly2in the last.
step.This equationcan also be Written
Equation (3.29) showsthat the c€nter of the wavefront
moves downward and the radius becomeslarger as time t: (xlV) * ty (3.31)
increases.
Field measurements yield valuesof the arrival time Fig.3.13,Relation betweenrcflection and refraction
at the shotpoint ro and angle ofapproach Al/A.t. Sincl raypathsand traveltimecurvcs.
the ray which returns to the shotpoint must havj
encountered a reflecting horizon normal to the raypath
and retracedits path back to the point oforigin, the dip
is equal to the angle rr at the time t : |ro. Thus, to locate
the segmentofreflecting horizon correspondingto a set
of values of lo and Lt/Lx, we make the following
calculations:
Llo'
(a)t = !to, (b)io:r"-' (*#), I
(c) rr : 2tan-r(edtanlio),
I
(d) H = (Zela)(cosha- l),
I
(e) R: (Vola)sinhat.
With thesevalueswe hnd C, lay off the radius R at the
I
R"
ingle ir, and draw the reflecting segmentperpendicular
to the radius as shown at the point :{ in fig. 3.12.This
method is easily adapteC to a simple plotting machine
(Daly, 1948)or to wavefrontcharts(Agocs,1950).
Refractionstudiesinvolvinglinear increaseofover-
burdenvelocityare discussedin 93.3.5.
 Geometry oJ'seismic waoepaths 90

where V, and the same measurementfor the refraction evenr

givesV2.We can then calculatethe critical angle@ from
tr: (2hcrrrs@)|V1,
the relation O : sin-r (VrlV), and usethe intercepttime,
or (3.32) tr, to calculatei from (3.32).
In fig. 3.13the time-distansecurvesfor the re.flec-
h:tyrtrlcos@.
tion from the interface AP" and for the direct path are
Obviously the head wave will not be observedat representedby the hyperbola CDE and the straight line
offsets less than the critical distance,OQ in fig. 3.13; O"E,respectively.Since the path OMe can be regardid
writing.r/ for the critical distance, either as a.reflectionor as the beginningofthe refracted
wave, the reflection and refraction time-distauce curves
x' : OQ = 2htan@:2htan{sin-r (VJY)} must coincideat x : x', that is, at the point D. Moreover,
:2h{(vzlv)2 - l}-} (3.33) differentiating(3.1) to obtain the slope of the reflection
time-distancecurveat x : x', we find
The relation betweenx'lh and VrlY, is shownin fig. 3.14.
As the ratio Y2lY, increases,x' decreases.When VrlY, fd,-l : :+l#frl]
equals1.4;x'is equalto 2ft. As a rule of thumb, offsets
should be greater than twice the depth to the refractor to
Ld"l,=,, [on,J,=.,
r /+ oo\
:-l I ^ I
observerefractions without undue interferencefrom shal- l ::si,1n@ :* . y2
V 1\Ou'1
lower headwaves.
Equations(3.30)and (3.31)representa straightline We seethereforethat the - reflectionand refractioncurves
of slope llV2 and intercepttime tr. This is illustrated in havethe sameslopeat D, andconsequentlythe refraction
fig. 3.13 wbere OMQ, OMP'R', OMPR and OMP"R" curve is tangentto the reflectioncurve at x: x,.
are a seriesof refractionpaths and DWS the correspond- Comparing reflected and refracted wavesfrom the
ing time-distance curve. Note that this straight-line samehorizon and arriving at the samegeophone,we note
equationdoesnot have fhysical meaningfor offsetsless that the refraction arrival time is always lessthan the re-
than x'since the refractedwave doesnot exist for such flection arrival time (exceptat D). The intercepttime tr
valuesofx; nevertheless we can project the line back to for the refraction is lessthan the arrival time ,o for the
the time axisto find lr. reflectionat the shotpoint,since
The problem to be solved usually is to frnd the
tr: (2hlVr)coso, to: (2hlV); hence!, < te.
depth lu and the two velocities V, and Vr. The slope of
the direct-wavetime-distance curve is the reciorocalof , Startingat the point
Q, we seethat the direct wave
arrives ahead of the reflected and refracted waves since
its path is the shortestof the three.However,part of the
Fig.3.14.Relationbetweencriticaldistance refraction path is traversedat velocity Zr, so that as x
x,,
crossoverdistancex. andvelocitycontrast. increases, eventuallythe refractionwavewill overtakethe
direct wave.In fig. 3.13thesetwo traveltimesare equalat
the point W. If the offset correspondingto W is x., we
\ have

x- x^ 2h
E I ;:;
/1 12
i ;cos@
/1
\ ko*"
{
X
@
o
:ter drstarrc,
/deprh) h:+(-3)l*'"
-x2 F \ -
- (vr- vr\
_x"
2 \ h ) vt- /7 i
vz
z
3i,;,,
Ye/depth) n:* " (V ' = -V t\' . (3.34)
2 \v2+ vrJ
This relation is sometimesusedto find ft from measure-
mentsof the velocitiesand the crossooer
distancex". How-
I
3.3 Geometryof reJiactionpatta 9I
I
F

and hence(3.32)provides a bettermethodof determiningi. the anglesQ also) by measuringthe slopesof the various
The relation betweenx.lh and vrlV, is shownin fig.3.14. sectionsof the time-distancecurveand then getthe thick-
nessesof the layersfrom theintercepts:
3.3.2 Seoeralhorizontal refractors
Where all layers are horizontal, (3.30) can be gen-
eralized to cover the case of more than one refracting (3i.37)
horizon. Considerthe situation in fig. 3.15wherewe have
threelayersof velocities,Vr, V, and I/r. WheneverVz) Vr, 3.3.3 Elfect of refrauor dip
we have the refraction path OMpR and corresponding The simple situations on which (3.30)_(3.37)are
time-distancecurve lIr.S,just as we had in fig. 3.13. If basedare frequently not valid. One of the most serious
Vt> Vz> Vr,travelby a refractionpath in Z, will even- defectsis the neglectofdip sincedip changesthe refraction
tually overtakethe refractionin Vr. The refractionpaths time-distane curvedrastically.The lower part of fig. 3.16
suchas OM'M"P"PR'are fixedby Snell'slaw: showsa vertical dip-section through a refracting horizon.
Let r be the traveltime for the refraction path OM4O,.
sin0, sin@, I Then, we have
vr -:_Y2 V3
,_OM+ O' P , MP
where@, is the critical anglefor the loyer horizon while '- v r: - q
0, is lessthan the critical angle for th€ upper horizon. The
hd+ h,
expression for the traveltime curve SZ is obtained as _
- r-Oe - (hd+ h")ta\@
before: (coso ------E-

^ OM'+ R'P . M'M" + PP" M"P' : "":t6 + io I i'"or@. (3.3s)

vz Vr
vr vz vr
%r 2hz If we placethe shotpointat O anda detectorat O,,
: * + x -2 h rta n Ar-Zhrtan@ , we are'shooting downdip'. In this caseit is convenientto
Vrcos9, ' Vrcos@2' Vg

:*+ ' ,=*'= (,-!rr^o)

Z, Zrcos@, ,3 \- / Fig.3.l5.Raypathsandtraveltimecurvesfor two-
refractorcase.
+ ZrcosO,
= = h ' ^6 - 5V;'r * r , )
\ /
: t'***rr,
x .2h,
E*
-fi"o"
(3.35)
:fr+u. l
Thus the time-distancecurve for this refraction is also a
straight line whoseslopeis the reciprocalof the velocity
just below the refractinghorizon and whoseinterceptis
the sum of terms of the form Qh,cosl,lv), each layer
above the refractinghorizon contributing one term. We
can generalizefor n layers:

t:fi+\ff"rle,, (3.36)

whereQ = sin',(Vtl\). This equationcan be usedto find

the velocitiesand thicknesses
ofeachofa seriesofhorizon-
tal refractinglayers,eachofconstant velocityhigherthan
any ofthe layersaboveit, providedeachlayercontributes
enough of the time-distance curve to permit it to be
analvzed corre-cJlv \ile can ffnd all nf ttra .,-l^^i+i-- /L--^-
 Geomety of sekmil waoepaths 92

haw t in terms of the distance froil the shotpoint to the denotedby r,. The conceptthat traveltimealong a path is
refractor16; henccwe eliminate/ruusingthe relation the sameregardlessof the direction of trarrelis an example
of the principle of reciprocity.
h,: ha* xsin (.
Theseequationscan be expressedin the sameform
Writing to for the downdip traveltime, we obtain as(3.31):

to = (xl V2)cose + @l Y) cos@sin { * (2h61V) costD tu: (xlVu)+ tr6, (3.41)

: (x/I/r) sin (@ + o + (2hulv) 9s @ tu: (xlVu)* tru, Q.a)
= (x/I/r)sin(@+ {) + tra, 'l where
where I t:.lll Va: hlsin(@ * €), Y": (1lsin(@- e). (3.43)
t,u: (2h6lY)cos@. ) Voand V" are apparent velocities and are giwn by the re-
The result for shooting in the updip direction is ciprocalsof the slopesof the time-distancecurves.
similarly obtaised by eliminatingfto: For rerrersedprofrles such as shown in fig. 3.16,
(3.43)can be solvedfor the dip ( and the critical angle@
t, = (xlVr)sin(O - O * tru, l (and hencefor the refractorvelocity Zr):
where It €.40) 9 : |{sin-t(4 lv) + sin-r(vtlv")\,1
Q'44)
tru: QhulVr)cos@. ) 6 : {{sin-r(K lvo) - sin-t(vrl1),}.J
Note that the downdip traveltimefrom O to O'is equalto The distancesto the refractor,lruand &., canthenbe found
the updip traveltime from O' to O; this shotpoint-to- from the interceptsusing(3.39)and (3.40).
shotpoint traveltime is called thtreciprocal time and is Equation (3.43)can be simplifredwhere { is small

Fig.3.l6.Raypathsandtraveltimecurvesfor a dipping
refractor.

r+ l
3.3 Geomety of refiaction paths

enough that we can approximate by letting cos( = I and 3.3.4 SeoeraldiWioC refractors with the sarrrestrike
sin { r 6. With this simplification,(3.43)becomes Equations similar to {3.36) have beenderived for
V1lV6: sin(@* O arsin@* {cos@; the caseof severalbedswith the samestrike and different
dips. An interesting formula due to Adachi (1954)departs
VrlV": sin(@- () :: sin@ - ( cos@; from the usual parameters and usesvertical thicknesses
and angles of incidenceand refraction measuredwith
hence
respectto the vertical (seefig. 3.17).The derivation of
sin@:(\lv) xlvr{(rlvu)+ Qlv")|, Adachi's formula is straightforward but involves lengthy
trigonometric manipulation (seeJohnson,1976),and we
so that
merelyquote the result:
rlv,xt{glv)+1t1u"17. (3.45)
x sin B, "^-rh,
tn:-# + );(cosa; + cosp,), (3.47)
An even simpler approximateformula for'V, (al- t/r r-=rYi
though slightly lessaccurate)can be obtained by applying
where lo is the traveltime of the refraction at the ath inter-
the binomial theorcm to (3.43) and assumingthat ( is
face (separating layers of velocities V, and V,*r), u,, and.
small enoughthat higherpowersof ( are negligible:
p, are the anglesbetweenthe vertical and the downgoing
Vu: (Vrlsii@)(cosf+ cot@sin{)-1 and upgoing rays in the ith la;er, i,. the vertical thicknesi
x V2(l- (cot@), of the fth layer.We definethe anglesa;, g (seefig. 3.17)as
anglesofincidence,ai,Dianglesofrefraction, all measured
Vux l/r(l + (cot@), relative to the normal, and (i*r : dip of the ith interface.
Then
henoe
ai : sin-t {(4 alvl)sina,},
Yrxt(Vu+ V"). (3.46)

Fig.3.17.Notation usedin Adachi's formula.

 Geometryof seismicwauepaths 94

b! : sin-t{(4*, I V,)sinb,\, surfaceagain,as shown in fig. 3.1g. For this case(3.25)

d, i: 4; * ( 1a 1 , fi : a i - ti * t and (3.26)give(seeproblem3.tg):
di+ r : a! * (i * r, F i * t 4 - 4 ,* t. x: (2Vola)cotio: eVola) sinhlar,
- J
For the refraction along the zth interface,ao: b,: @n, t:(2la)rn(cot!io). €'48)
J
the critical angle.
The maximum depth of penetration,ft-, is reachedwhen
Assumingreversedprofiles,we measureVr,the ap- iequalslz, hence
parentvelocitiesnZru and V2d,and the intercepts,tru and
trd, as usual.For the hrst interface, h^: (Vsla)(coshlar- l) (3.4e)
q : sm-r(l/rl v2d,), h : sit-r (vtl vz") using(3.28)and (3.48).
The case of a high-velocity layer overlain by a
@r : 4r - br : i@ r+ fr),
layer in which the velocity increaseslinearly with depth
€z: *@r - F), (cf. (3.2t4)), (fig. 3.19) is of considerablepractical importance.The
Vz.: Vtlsin@1, hr'- VlrJ(cosdr + cosps). relation betweenr and x for a horizontal refractor can be
found as follows:
To solvefor the secondinterface, we calculatenew values
of ar, Br and then find the other angles(note that {, is . l: tsn+ txp + tpx
now known): :2/rn * @ -2MN)IV^.
dt: sin-1(YtlVta), f t: sin-l(VtlYs),
ar : dt - €2 , h : frt * tz ,
aL : sin't {(vzl vr) sina},
bL: sin-| {(Vzl Vt) sinb rl,
dz : aL* €2 , fz : b 5 - tz , Fig.3.t8. Linear increasein velocity applied to
refraction interpretation.
az: bz- @z: i@z + fr): +(ai + b),
V t : V z ls in@ 2 ,* :I@ z -
F ),
t2": (h1lVr)(cosa,
* cosfr)
* (h2lV2)@osa,* cosf),
h2 being found from the last relation. In principle this
iterative procedurecan be continued indefinitely but in
practic,e,as with all refraction schemes,the errors and
difficultiesmount rapidly asthenumberof layersincrease.
Adachi'sformula is bestsuitedto simplecaseswhere
the refractorsareplane,no velocityor structuralproblems
existand the refractorsareshallow.Whentheseconditions
are not met, the formula, in common with other si.rnilar
ones,may be of limited value.Often one is not surethat Fig.3.l9. Refraction with linear increasein velocitv in
formulaeare appticableto a specificreal situation.Where upper layer.
there are morb than two refracting horizons it is often
diffrcult to identify equivalentupdip and downdip seg-
ments,especiallyif the refractorsar€ not plane or if the
dip and strike change.
Yc =Vo+ahc <Vm
3.3.5 Linear increaseof oaerburdenoelocity
The conceptsof a linear continuous increaseof
velocity have beendiscussedin g3.2.5.It is obviousfrom
frg 3.12 that the circular rays will eventuallyreach the
 Problems 95

Noting that sin iolyo:(sin@)/( =llV^, we find from Depth Velocity

(3.26)
0-1.00km . 2.00km/s
1.00-2.50 3.00
2.50-2.80 6.00
2.80-4.80 4.00

t ( /.,\ /- .\\ (6) A reflectionfrom the 2.5km horizonhasadip moveout

: l { c o s h -' { 5 1 - (Ie l l - of 104 ms/km. Calculate the dip using Z and V and
a| \Vo / " o rr,-' \V./J' comparewith the angle of approach on a split spread.
where use has been made of the identitv cosh-l x: Why are the threevaluesdifferent?
In {x + ("t - t)}}. From (3.25),we get 3.4. Show that (3.9) becomes
= (l Ipa)(cosio - cos@)
11,111 (Yt)z : (2xcosQ2*4h2"
: (rlpa\f{r- (volv)z!,- {r - (V.lV)2}r7. (Gardner, 1947)wherelr is replacedby lr", the slant depth
Substitutingin the first expressionfor r gives at the midpoint betweenthe shotpoint and receiver (see
Frs.3.2l).
t: (xlV^) + to, (3.50) 3.5. (a) Using (3.11)and the r3lulrs"cf problem 3.4,
where/6 : intercepttime:
to: Qla)l{cosh-r (Y^lV) - cosh-l (VJV.)} Fig.3.20.Combiningdip components.
- {1- (volv)2l|+ {l - (vclv)z}rf.Q.sr) YAt,
The slope of the traveltime curve gives Z-. A curve is ^=6 y
plotted of te againsth" (or V.) for given valuesof I/6 and
a, ard h. and V" are read from this curve for particular
measurements of to.

Problems
3.1. (a) CalculateAtn for a geofhone 600 m from the
shotpoint fOr a rehectionat to:2.358 s, given that
V:2,90 kmis. (b) Typical errors in x, V, to might be
0.6 m, 0.2 km/s, 5 ms; calculatethe correspondingerrors
in Atn approximately.What do you concludeabout the
accuracyof Aln calculations?(c) Show that (3.8) can be
written
Arnn-Arn(l -L,t"l2to) Fig.3.2l. Derivation of X2-72 relation for a dipping
bed.
whereAtn is givenbV(3.7)and Arnois the secondapproxi-
mation. Taking into accountthe errors it x, V, fo, when
is this equationuseful?
3.L (a) Show that the quantity dr/dx can be consideted
as a vector or component of a vector according as dr
correspondsto the total dip or componentof dip. (b) Using
fig.3.20, verify that the constructionof fig. 3.8 givesthe
same results as (3.18).fHint: Express/, m and OC in
terms of Ol.]
3.3. (a) Calculate V and,V down to each of the beds
in the following table. Why do they differ'(givc a geo-
metricalexplanation)?
 Geometryof seismicwaoepathi 96

verify the following result (due to Favre,'accordingto 3.g. Using the data in problem 3.3a, plot V and V
Dix, 1955): versusdepthand versustraveltimeand determinethe best-
fit straight lines in the four cases.What are the main
tan( x tl(t)"- tf,)*,
problemsin approximatingdata with functionalfits?
where ( :dip, / - tsz- tsb tts= traveltime between 3.10. (a) Assumingflat bedding,calculatedepthscorre-
shotpoint I and receiverB, to: traveltime at shotpoint spondingto lo : 1.0,2.0, 2.1 and 3.1 s usingthe following
S (seefig. 3.21).(r) Using (3.9),showthat velocity functions: (i) Z from problem 3.3a,(ii) 7 from
sin{ = V' (t3^- t2r"11ttt"x. problem 3.3a, (iii) thb best-fit functions determinedin
problem 3.9. (r) What errors are introducedin the above
(c) Under what condition is the result in (b) the sameas relativeto the depthsgiven in problem 3.3a.
(3.11)and also consistentwith (aX
3.11. (a) Repeatthe calculationsof problem 3.10a foi
3.6. The expressionfor dip in terms of dip moveout, flat velocity layering but dip moveout of 104 ms/km
(3.11), involves the approximation of dropping higber- and find the dip in each case.(D) Using the velocity data
order terms in the quadraticexpansionusedto get (3.10). tabulatedin problem 3.3a trace a ray down through the
What is the elfecton (3.1l) if an additionalterm is carried various layersand find the arrival time, reflectingpoints
in this exBansion?What is the percentagechangein dip? and dips ofreflectorslocatedat the depthsofeach velocity
3.7- (a) Shotpoints.Band C arerespectively600m north boundary.
and 500m eastof shotpointl. Traveltimesat,4, B and C 3.12. ShotpointsI and .B are located at the ends of a
for a certainrefl€ciionzta ts:1.750, 1.825and 1.796s. 225 m spread of 16 geophones.Using the data below,
What are the dip and strike of the horizon, Z being find the velocities,the dip and depth to the refractor.
3.25km/s? (r) What are the changesin dip and strike if
the line AC hasthe bearingN80'E? l^ Xg tA aB

3.8. (a) Two intersectingseismicspreadshave bearings 0m Oms 98 ms 225 m 120m 70 ms 52 ms 105m

Nl0'E and Nl40"E. If the first spreadshowsan eventat 15 l0 92 210 135 73 46 90
to : I.760 s u'ith dip moveoutof 56ms/km while the same 30 2l 87 195 150 78 43 75
eventon the secondspreadhasa dip moveoutof 32ms/km, 45 30 8t 180 165 8l 37 60
find the true dip, depth and the strike, assumingthat 60 4l 75 165 180 85 3l 45
(i) both dips are <iownto the south and west, (ii) dip on 75 50 7t 150 195 89 2l 30
the frrst spread is down to the south while the other is 90 59 65 135 2r0 94 l0 l5
down to the southeast.Takethe averagevelocityas3 km/s. t05 65 60 l?0 225 98 0 0
(b) Calculatethe position of the reflectingpoint (migrated
position) for eachspreadin (i) as if the crossinformation 3.13. (a) Showthat the two geologicalsectionsillustrated
had not beenavailableand eachhad beenassumedto be in ftg. 3.22 produce the same time-distance curves.
indicating total moveout; compare with the results of (6) What would be the appar€ntdepth to the lower inter-
pafi (a). Would the errors be more seriousor lessserious face in fig 3.22a,b if V3: 3.15km/s insteadof 6 km/s?
if the calculationsweremadefor the usualsituationwhere 3.14. Fig. 3.23 showsa refraction profile recordedas a
the velocityincreaseswith depth? ship firing an airgun steamedaway from a sonobuoy.

Fi9.3.22.
Two differentgeologicsections
whichgivethe
samerefractiontime-distancecurve.
Surfae

frr =3 00 m / r = 3k m / s
-
ftr =300m Zr =. l'5 km/s
,l,
/3 = 6 km/s

Vt = 6kmls
 Problerns 97

Identify the direct wave through the water and use its Identify multiplesand explaintheir probabletravel paths.
slope to give the source-to-sonobuoy distances(assume (fhc data in the upper right corner result from paging
1.5 km/s as the velocity in water). (a) Identify distinctive and actually belongbelow thc bottom ofthe record.)
head-wavearrivals, determinetheir velocities,intercept. 3.15. The time-distance observationsin hg. 3.24 con_
times and depthsof the refractorsassumingflat bedding stitute an engineeringrefraction problem. (a) Solve for
and no velocity inversions.(b) What is the water depth? the first layer for both setsofreversedprofiles and show

Fig.3.23.Sonobuoyrefractionprofilein BaflinBay
shotwith 1000in3airgun.(CourtesyFairfield
Industries.)
 Geomettyof seismicwauepaths 98

that the layer has a thicknessof about 2.9 m and 3.8 m 3.18. Prove(3.48)and (3.49)from (3.25)-(3.28).
at shotpointsA and C (dip about 0.3"). (b) Apply (3.35) 3.19. (a) Given the velocityfunction V:1.60 * 0.60z
to get approxirnate thicknessesof the second laycr. (c) km/s (z in km), hnd the dip of the reflector and the depth
What is the dip of the deeperinterface? (d) Why are the and offset of the point of reflection when to :4.420 s
answersin (D) and (c) approximate? and Lt/A,x = 0.155s/km. (D)lVhat interpretationwould
3.16. To find the depth to bedrockin a damsitesurvey, you give ofthe result in part (a)? Ifthis reflectordid not
traveltimes were measuredfrom the shotpoint to 12 exist and the ray continuedwithout reflection,when and
geophoneslaid out at 15m intervalsalong a straightline wherewould it emerge?What moveoutwould be observed
through the shotpoint. The offsetsd range from 15 to at the recordingspread?(c) If you were given the results
180m. Determinethe depth of overburdenfrom the data of part (b) insteadof the valuesof re and AtlA,xin (a),
in the following table.By how much doesthe depth differ calculatethe maximum depth of penetration.
if a two.laycr solution is calculated instead of a three- 3.20. If the velocity function in problem 3.19 applies
layer solution? above a horizontal refractor at a depth of 2.40 km where
the refractor velocity is 4.25 km/s, plot the traveltime-
distancecurve.
3.21. Given that situations (a) through (i) in fig. 3.25
l5m l9m s ?5m 59ms 135m 72ms involve the same two rook types, draw the appropriate
30 29 90 62 150 76 time-distancecurves.Diagram (c) showstwo casesfor
.- 45 39 105 65 165 78 dip in oppositedirections.In frgs Q) and (J), the velocity
' 60 50 l2o 68 180 83
in the lower medium varies laterally according to the
densityof the shading.
3.17. Given the following data for a reversedrefraction 3.22. Barton (1929)discussesshooting into a geophone
profile with shotpoints,4, 8, useAdachi's methodto find
placedin a borehole(frg.3.26)as a meansof determining
velocities,depthsand dips.
where the bottom of the boreholeis located. (a) Given
that A, B, D, E are equidistantfrom W in the cardinal
.r 0 0.5 l.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0km directions and assuming straightJine travel paths at the
velocity V and,thatthetraveltimesfrom shotpointsD and
rr 0.00 0.25 0.50 0.74 0.98 t.24 1.50 1.70 l.8l 1.91 2.02s
E in fig. 3.26a are equal,derive expressions for CC, and
r, 3.00 2.90 2.80 2.68 2.52 2.4t Z.3r 2.20 2.0j l.9l 1.80s
CW infi9.3.26b in termsof the traveltimesfrom shotholes
.r 5.5 6.0 6.5 7.0 7.5 8.0 8.5 9.0 9.5 10.0km A and,B, tag'and tps,. (6) What arethe valuesof 1,4c,, and
tps, for V :2.500km/s, AW : BW: CC: 1000m,
t^ 2.16 2.28 2.38 2.44 2.56 2.& 2.72 2.8A 2.89 3.00s CW:200 m? (c) How sensitiveis the method, that is,
t, 1.65 1.50 1.40 t.25 r.l2 t.00 0.?S 0.49 0.23 0.00s whatareA(CC')|Lt^c,andL,(CW)IL,rr..?For thespecific

Fig.3.24.Enginceringrefraction profile.
Problems 99

Fig.3.25.Time-distancecurvesfor various two-layer situationin part (b), how muchchangeis madein WC and
configurations.This figure is adaptedfrom figuresin CC' per millisecond effor il tac.?(d) Modify the as-
Donald C. Barton's paper'The seismicmethod of sumptionsin part (D) by taking the velocity as 1.5 km/s
mapping geologicstructure' (1929),the first publication
for the first 500m and 3.5 for the lower 500m; what are
in English on the seismicmethod.The upper part
(above O) of each diagram provides spacefor a curve
the actual traveltimesnow and how would thesebe inter-
of arrival time versusdistancefor the model shown in pretedassumingthe straight-pathassumptionin part (a)?
cross-sectionbelow (below O). Part (c) has been
completedto show what is expected.In (c), two
Fig.3.26.Mappinga crookedboreholeby measuring
alternativesare given so two setsof curvesshould b€
traveltimeto a geophone at C'in the borehole.
(From
drawn. In (r) and (7), velocitiesare assumedto vary
Barton,1929.) (a) Planview;(D)verticalsectionAWB.
horizontally proportional to the densityofthe shading.
In eachcasethe velocity in the cross-hatchedportion is
higher than that above.
{}
,E
,
,
I
t
l ^V
A ----!-:------
ci
I
I
I
I
I
Vn

cau

\ It7c'

(6)

(h)