PROCEEDINGS OF SPIE

SPIEDigitalLibrary.org/conference-proceedings-of-spie

Fracture imaging and saline tracer

detection by crosshole borehole
radar data migration

Hui Zhou, Motoyuki Sato

Hui Zhou, Motoyuki Sato, "Fracture imaging and saline tracer detection by
crosshole borehole radar data migration," Proc. SPIE 4084, Eighth
International Conference on Ground Penetrating Radar, (27 April 2000); doi:
10.1117/12.383581

Event: 8th International Conference on Ground Penetrating Radar, 2000, Gold

Coast, Australia

Downloaded From: https://www.spiedigitallibrary.org/conference-proceedings-of-spie on 2/15/2019 Terms of Use: https://www.spiedigitallibrary.org/terms-of-use

 FRACTURE IMAGING AND SALINE TRACER DETECTION BY CROSSHOLE BOREHOLE
RADAR DATA MIGRATION

Hui Zhou, Motoyuki Sato

Center for Northeast Asian Studies, Tohoku University

Sendai 980-8576, Japan
satocneas.tohoku.ac.jp

ABSTRACT limitation of tomography: the resolution and sensitivity of

tomography are not high enough.
Because the resolution of tomography is not high, crosshole
tomography is suitable for imaging large weakly inhomoge- Key words: Crosshole radar measurement, Fracture imaging,
neous bodies, but not suitable for bodies of sinai! size, for Tracer monitorin& Migration
example, thin aperture fractures.

We try to use crosshole migration method to obtain underground INTRODUCTION

small aperture fractures between two boreholes. Since fractures
are filled with water, although the apertures offractures are small, As we know that crosshole radar measurement data are mainly
the electromagnetic property of fractures is very different from utilized to do tomography. In most cases, the result of
bedrocks. Field experiments show that reflection and diffraction tomography enables us to learn the characteristics of
from water filled fractures can be received by radar sondes. underground media between two boreholes very well.
Therefore, it is possible to get images of fractures by crosshole
migration. However, tomography is only suitable for relatively large,
weakly inhomogeneous massive media. When we apply the
We apply the crosshole migration technique to crosshole radar crosshole radar to the situations such as fractured rocks, due to
data sets in frequency and time domain collected both by our the small apertures of fractures, tomography will not be so
research group and by USGS in Mirror Lake area, USA. successful. Tomography, without doubt, can image fracture
zones (Lane et al., 1998), but the detailed structures of fracture
By comparing the migration result of crosshole data recorded in zones are difficult to be detected (Zhou et al., 1999).
frequency domain with USGS's results ofcrosshole tomography
and directional single-hole measurements, we find that the If fractures are filled with water, although the apertures of
migration result is consistent with USGS's results very well. fractures are not large, the differences of electric properties of
Two known fracture zones are correctly imaged. fractures and surrounding rock are great, the fractures Set off
reflected or refracted waves while they are excited by incident
The migrated section ofbackground data illustrates two fracture radar waves. Therefore, it is possible to image subsurface
zones. The difference images between migrated profile from fractures and to reveal the transportation of saline tracer by
USGS's background data and these from USGS's crosshole migrating crosshole radar data after certain essennal data
time-lapse data indicate that the transportation of saline tracer processings are camed out.
inside a trans-missive zone can be known during the water and
tracer injection and pumping. This result is also consistent with In this paper we discuss data processing methods for imaging
USGS's time-lapse difference-tomography results. Moreover; fractures, and the comparison of migrated sections from
the detail structures inside the fracture zones could be seen from crosshole data sets collected by a network analyzer based
the migration profiles and the difference sections. However; stepped-frequency and pulse radar system in Mirror Lake, USA
differences between two results exist. This may indicate the with the results of time-lapse attenuation-difference radar

In Eighth Intl. Conference on Ground Penetrating Radar, David A. Noon, Glen F. Stickley,
Dennis Longstaff, Editors, SPIE Vol. 4084 (2000) • 0277-786X/O0I$15.OO 303

Downloaded From: https://www.spiedigitallibrary.org/conference-proceedings-of-spie on 2/15/2019

Terms of Use: https://www.spiedigitallibrary.org/terms-of-use
 tomography performed in the same place by USGS. all four boreholes; the time-lapse data processed here were
collected using FSEI and FSE2. We operated a network
FIELD EXPERIMENT analyzer based stepped-frequency borehole radar system to
record frequency-domain data. USGS used the RAMAC system
The crosshole experiments were carried out at the U.S. to record time-domain data
Geological Survey (USGS) Fractured Rock Research site in the
U.S. Forest Service Hubbard Brook Experimental Forest in the For our experiment, a transmitter was set in FSE3 from 25 to 45
Mirror Lake area near West Thornton, Grafton County, New m in depth and stopped every meter. When the transmitter was
Hampshire. fixed at a certain position. the receiver was moved from 25 to 45
m continuously in FSE1 radar data were recorded every 1 in.
Figure 1 shows four boreholes and the injection and pumping There were 101 frequency sample points of each trace; the
boreholes at the experiment site. The arrangement of four sample interval offrequency was 2 MHz.
boreholes FSE1-4 is approximately a 9 m square. All the four
boreholes are more than 70 m deep. According to Lane et a!. (1998), a iransmissive zone at a depth
ofabout 40 m was isolated in three ofboreholes (FSE1-3) using
FSE2
9m FSE1 special PVC packers. The tracer tests were conducted by
Measuring Injection pumping the isolated interval in FSE4 at about 4 L/rnin while
Measuring injecting freshwater into the isolated interval in FSE1 at about 2
L/min. After establishing a steady-state flow regime, the
injection of freshwater was halted to allow injection of 20 L
NaCI tracer a concentration of 50 g/L into FSEI over a
10-minute interval. After termination of the tracer injection, the
FSE4 FSE3 freshwater injection continued. Time-lapse radar data were
Pumping Measuring collected using 100 MHz sondes every 10 minutes for 5 hours
(a)
after the start ofinjection using FSE1 and FSE2 from 20 to 70 m.
FSE1 FSE4 The packers could allow radar sondes to pass through. The
intervals of receiver and transmitter were 0.25 and 10 iii. Data
Injection Pumping were received both in FSE1 and FSE2.

42.5 m DATA PROCESSING

43.5 iii
45.0 in Direct wave suppression
47.0 ni
Figure 2 is geometric spreading and attenuation compensated
crosshole data received by the network analyzer based
(b)
stepped-frequency borehole and pulse borehole radar system.
Figure 1: (a) Experiment boreholes form apexes of a square.
FSEI and FSE2 had been used to record time domain time-lapse The direct wave, which cannot be used in migration, is strong. It
data by a pulse system. FSEI and FSE3 were used to collect should be greatly decreased before migration is performed.
frequency-domain data by a stepped-frequency system. (b)
Injection and pumping boreholes in the time-lapse experiment. For vertical radar profile with sources or receivers on the surface
and for layered structures, in time-depth domain the directions of
Two kinds of experiments were done by us and by USGS apparent velocities of direct and reflected waves are different.
respectively. We did crosshole measurements using FSE 1 and Correspondingly, in frequency-wavenumber domain the direct
FSE3. USGS had done time-lapse crosshole experiment using

304

Downloaded From: https://www.spiedigitallibrary.org/conference-proceedings-of-spie on 2/15/2019

Terms of Use: https://www.spiedigitallibrary.org/terms-of-use
 arid reflected waves are separate. the downgoing and upgonig \Vhen fLk nietliod is used to separate upgonig arid downgoing
waves are located in different quadrants. ayes, we should pa attention to the highest frequency and
spatial sai1pl ing rate to avoid spatial al iasi i ig fl le s; )at ial
But for crosshole data it is different from conventional \RP. the saniphiig late WaS I in along the receiving borehole for the
directions of direct wave are both upward and donward. fiequeiicv—doinain data. It is shghtlv coarse If we use f—k
Another important fact is that for our experiment situation there method, we should lower the highest fieqnencv of the data. I hits
is no layered structures. difli'acted waves due to fractures are the resolution of iiiigrted profiles is loereil severei.
doiiiinant. The travel-time trajectories of diffracted waves are
hyperbolas containing both upward and downward waves. Afier Here we do not use f—k method to suppress direct wave fur the
the wave fields are separated, the full-diffiacted waves are frequency—domain data, but attenuate direct wave within a small
divided into two parts. If only the upgoing waves are used in time window along the travel-time trajectory of the (lirect wave.
miiu'ation. some useful information in the downgoing wavefield For the time—domain data, the direct wave is suppressed iii up
about underground media is lost. We should use the full wave and downgonig wavefields.
fields but direct wave. To weaken the influence of direct wave
suppression on useful wavefields. the f-k domain separation To suppress the direct wave and to migrate crossliole data, we
method can also be used. must know the velocity of tile niediuiii. At this experiment site.
the rock is granite and quite homogeneous except for fractures.
We fit the travel tinies oldirect waves using (lie constant velocity
0.13 m/ns. As shown in Figure 2 the thin black inies along the
direct wave are calculated travel-time curves

Migration of crosshole data

Migration is perfoniied by two steps wave field extrapolation

from received non—direct waves and au unagmu.r procedure In
100 i 200 300
Time (ns)
kirchhofinueration the 2.5 dimensional wavelield extrapolation
for single source and mntilti—receivuig positions is achle\ ed using
Kircluhofi integral formula ( Rowbothani et al.. I

(i(x)= [dl. )co) 11f51.1

' 2rI V j
vhere the subscript i' and rekr to the recei\ em and sonice
respectively. i is the velocity of the medium. R /m is the
travel—time froni the source at s to an uiuaoe point p( v.:
and U, /m is the travel—tune from p(x.: )to a receiver at r (1, is
time angle between the nonuial of element Ime ill and U
A its. r, (ii + ii, )/' is obtaimied by cons oltititig the trace

r
recorded at recei 'er r fmoni source s

12 L
iii i the liii iction

v - (I?,
' ''
) (2)

Fiizure 2: (a) Crosshole data received by stepped-frequency fur 1 > 1?, m , and taking the values of the output trace at
system. Tx: 3-I in in FSE3. Rx: from 25 to 45111 iri FSE I. (h) I U, r + U1 i'.
Crosshole data received by RAMAC before injection and
pumping. Tx: 50 in in FSEI. Rx: from 201070111 in FSE2, If crosshole data are only received in one borehole, and there are
niultiple transmitter positions. comirioii—receiver data gathers can

305

Downloaded From: https://www.spiedigitallibrary.org/conference-proceedings-of-spie on 2/15/2019

Terms of Use: https://www.spiedigitallibrary.org/terms-of-use
 be selected from common-transmitter data gathers. Siniilarlv. the results from directional borehole radar measurements
wavefield extrapolation for a common-receiver data gather is
conducted by 1. SOS (Lane et al.. 19%).
expressed as

/ ( + 1<) co4r.s
'
+ 1<1.
1<
U(x.:) = (3) FS F 2
JdI 211< L ] 1 Fracture

and i
where 9 is the anele between the nomial of element lme dl,
t.I[r,s. (!? + R )/v] is obtained by convoluting the 30
trace recorded at source r horn receiver s with tlìe function
similar as (2)

_(R/v))' (4)
L35 '35
40 Fracture
d, R Iv zoiie
45
for i > v arid taking the values of the output trace at
t= v+ v 0 4 8
Position (m)
For crosshole inigranon, the final result is the combination of all (a)
uligratlon images of common-transmitter and common-receiver u
gathers

(T(x. z) = JJiai
Ii? (I?.+R) cosOM [s, R+R,
r,
j
t 2R,. L

tI 2 4 i 8
rr 11? (R + I? ) 1? 1? 1 -
(:)
Position (iii)
+ J J d/,dl cos0A1 r.s. .
(b)
t 2R
Figure 3: (a) Migrated image between FSF3 and FSFI obtained
If crosshole data are received in two horeholes. the final floin the data collected by stepped—f requencv svsteni. (b)

migration profile is the stack of migration results of all Migrated section between FSE I and FSI'I2 obtained from the
background data collected by a pulse svsteiim
common-transmitter gathers.

Figure 4 depicts the differences of migration results of'

RESt. LTS AND COMPARISON background data and these of the data sets i'eceived at 10, 60.
100. 160. 200 and 260 minutes after the beginning of water
Fieure 3 is the niierated sections. Figure 3(a) is the migration injection and pumping. From this figure we can clearly see the
image between FSE3 and FSEI obtained from the data collected transportation of saline tracer. At 10 mimiutes, the ditkrences are

by Network Analyzer system. Figure 3(b) is the iriigration weak. It means that the tracer did not accumulate in the
section between FSEI and FSE2 denved from the background isolated fracture zone. With tune increasing, the differences
data collected by RAMAC system before water injection and become obvious and obvious. At 160 nuilutes, the differences
are the most evident At this time the amount of' tracer iii the
punipino,
fractures was time largest. Afier that time. the tracer tended to he

Accordnig to Lane et al. (1996) dealing with the same site, there less and less and was pumped out

are t o hydraulically condueti e zones connect the borelioles (in

the paper. the boreholes are FSE I and FSE4). The upper zone is These immigratiomi difference results aie, iii soiiie aspects. simimilam

near the bottom of casing (20—25 in) and the lower zone is with the time—lapse ditference—attennation tomography results by

approximately 40 in below the top of casing. In Figure 3 we can I,JSGS (Lane et al.. 1998). Moreover, e can see sonic detailed
see these two zones. Zhou et al. (1999) showed that the migrated changes within the transiiussive fracture zone. F3ut theie are
results from the frequency domain data were also consistent with great differences between Figure 4 and the attenuation-

306

Downloaded From: https://www.spiedigitallibrary.org/conference-proceedings-of-spie on 2/15/2019

Terms of Use: https://www.spiedigitallibrary.org/terms-of-use
 difference tomography results of the same section (Lane et al., CONCLUSIONS
2000). This may indicate the limitation of tomography: the
resolution and sensitivity oftornography are not high enough. Crosshole borehole radar migration is an effective technique not
only to image underground small aperture fracture zones, but
also to monitor the transportation of saline tracer in transmissive
25 25 25
fracture zones during water and tracer injection and pumping.
30 30 30
The migration results are coincident with the results of single
!,35 35 . 35
hole directional radar measurements and crosshole tomography.

40 40 40 Moreover, migration results can give detailed information of

subsurface fractures.
•45
0
45 . -.. 45 .
.
•0
50
bD
50 50
ACKNOWLEDGMENTS
0
'55 55 55
This research work is supported by Grant-in-Aid (10450389 and
60 60 - 60 10044122), Ministiy of Education, Science, Sports and Culture
of Japan. Authors thank J.W. Lane of USGS for proving
65 65 65
time-lapse data
70 - 70 - - 70
J2468 02468 02468
10 mm 60 mm 100 mm
REFERENCES

Lane, J.W., Haeni, F.P., et a!., 1996, Use of borehole-radar

25 25 25 methods to detect a saline tracer in fractured crystalline
bedrock at Minor Lake, Grafton County, New
30 30 30
Hampshire, USA, Proc. 8th liv'. Conf on GPR, 185-190.
,35 35 35 Lane, J.W., Haeni, F.P., et al., 1998, Use of time-lapse
attenuation-difference radar tomography methods to

: I:
40 40 40 monitor saline tracer transport in fractured crystalline
bedrock, Proc. 7th in!. Conf on GPR, 533-538.
Lane, J.W. Jr., Day-Lewis F.D., Hanis J.M., Haenil F.P., and
:: SM. Gorelick, 2000, Attenuation-difference radar
tomography: Results of a multiple-plane experiment at
•55 55 55
the Mirror Lake, New Hampshire fractured rock research
60 60 - site, Proc. 8th ml. Coiif. on GPR, this proceeding.
Rowbotham, P.S., Goulty, N.R., 1993, Imaging capability of
65 cross-hole seismic reflection surveys,Geophysical
Prospecting,41, 927-941.
71] 70 70
02468 02468 02468
160 mm 200 mm 260 mm
Zhou, H., Sato, M., 1999, Estimation of subsurface fracture
extension by using crosshole radar measurement,
Technical Report ofIE.IC'E, SANE99-73, 8 1-86.
Figure 4: Differences of migrated results between FSE1 and Thou, H., Sato, M., 1999, Fracture detection using crosshole
FSE2 of background data and these of the data sets received at borehole radar in Kamaishi. &panded Abstracts, SEG
10,60, 100, 160,200, and 260 minutes. Intemational Exposition and Sixty-Ninth Annual
Meeting, Oct.31-No5, Houston, USA, 480-483.

307

Downloaded From: https://www.spiedigitallibrary.org/conference-proceedings-of-spie on 2/15/2019

Terms of Use: https://www.spiedigitallibrary.org/terms-of-use