Exploration Geophysics (2006) 37, 19-29

Butsuri-Tansa (Vol. 59, No. 1)

Mulli-Tamsa (Vol. 9, No. 1)

Estimation of CO2 saturation from time-lapse CO2

well logging in an onshore aquifer, Nagaoka, Japan
Ziqiu Xue1 Daiji Tanase2 Jiro Watanabe3
Key Words: well log, time-lapse, CO2 monitoring, P-wave velocity, Gassmann’s equation, neutron porosity, cased-hole

ABSTRACT target reservoir at 20 to 40 tonnes per day. The total amount of CO2
injected during the experiment period was about 10 400 tonnes.
The first Japanese pilot-scale CO2 sequestration project has During CO2 injection, a series of field surveys and measurements,
been undertaken in an onshore saline aquifer, near Nagaoka in including crosswell seismic tomography, well logging, reservoir
Niigata prefecture, and time-lapse well logs were carried out formation pressure and temperature measurements, and micro-
in observation wells to detect the arrival of injected CO2 and seismicity monitoring were conducted, to improve understanding
to evaluate CO2 saturation in the reservoir. CO2 was injected of CO2 behaviour in a sandstone reservoir (Kikuta et al., 2004;
into a thin permeable zone at the depth of 1110 m at a rate of Xue et al., 2005). One of the main objectives was to determine the
20–40 tonnes per day. The total amount of injected CO2 was arrival times of injected CO2 in observation wells from time-lapse
10 400 tonnes, during the injection period from July 2003 to well logging. The CO2 breakthrough times at observation wells are
January 2005. The pilot-scale demonstration allowed an improved necessary for modifying reservoir model and for describing fluid
understanding of the CO2 movement in a porous sandstone flow in the reservoir.
reservoir, by conducting time-lapse geophysical well logs at three
observation wells. Comparison between neutron well logging Gould et al. (1991) and Butler et al. (1993) successfully
before and after the insertion of fibreglass casing in observation demonstrated the capability of time-lapse well logging for
well OB-2 showed good agreement within the target formation, monitoring a miscible CO2 flooding in Enhanced Oil Recovery
and the higher concentration of shale volume in the reservoir fields in West Texas, USA. Induction resistivity and neutron logs
results in a bigger difference between the two well logging results. were run through fibreglass casing, and saturation of three phases
CO2 breakthrough was identified by induction, sonic, and neutron of oil, water, and CO2 was monitored in carbonate reservoirs.
logs. By sonic logging, we confirmed P-wave velocity reduction Recently Müller et al. (2005) and Sakurai et al. (2005) developed
that agreed fairly well with a laboratory measurement on drilled an effective technique for monitoring injected CO2 in fibreglass-
core samples from the Nagaoka site. We successfully matched cased wells with pulsed neutron logs. Results of induction and
the history changes of sonic P-wave velocity and estimated CO2 neutron logging enable us to quantify fluid saturation within
saturation after breakthrough in two observation wells out of reservoirs. However, in most applications of geological CO2
three. The sonic-velocity history matching result suggested that the monitoring, we usually obtain time-lapse seismic data. There is
sweep efficiency was about 40%. Small effects of CO2 saturation a practical need to correlate seismic velocity change with fluid
on resistivity resulted in small changes in induction logs when saturation in geological CO2 sequestration (Chadwick et al.,
the reservoir was partially saturated. We also found that CO2 2004). In this paper, to correlate sonic velocity change with CO2
saturation in the CO2-bearing zone responded to suspension of saturation (volume fraction) in a partially water-saturated reservoir,
CO2 injection. we performed laboratory measurements of P-wave velocity change
during CO2 injection into a water-saturated sample drilled from the
INTRODUCTION reservoir. We confirmed that the laboratory results explain well
the sonic P-wave velocity reduction due to CO2 breakthrough in
A pilot-scale geological carbon dioxide (CO2) sequestration the observation wells. We also estimated the CO2 saturation in the
project is being undertaken by the Research Institute of Innovative target formation by matching the sonic P-wave velocity change
Technology for the Earth (RITE), in cooperation with the with an application of the rock-properties model proposed by
Engineering Advancement Association of Japan (ENAA). The CO2 Hoversten et al. (2003).
injection site is located at the Minami-Nagaoka gas and oil field,
near Nagaoka, Niigata Prefecture. Injection of 99.9% pure CO2 , Geology and CO2 injection
purchased for the project, started in July 2003 and ended in January
2005. The CO2 was injected into a thin permeable zone of the The pilot-scale CO2 injection site is located at the Minami-
Nagaoka gas and oil fields, where Teikoku Oil Co. produces
natural gas from the deep reservoir (4700 m). Figure 1 shows the
location of the pilot site and the simplified geological setting,
which was well studied during the oil and gas exploration. The
1 Research Institute of Innovative Technology for the Earth (RITE) upper 1000 m of the geological section at the site consists of the
9-2, Kizugawadai, Kizu-cho, Soraku-gun, Pliocene-Pleistocene Uonuma Group and Haizume Formation. The
Kyoto, 619-0292 JAPAN
Uonuma Group is represented by two types of depositional cycles
Phone: +81-774-75-2312
Facsimile: +81-774-75-2313 (Hoyanagi et al., 2000). One cycle is composed of shallowing-
Email: [email protected] upward successions deposited on shelf-to-shoreface environments
of a wave-dominated delta system and is accompanied by
2 Engineering Advancement Association of Japan (ENAA) transgressive lag at the base. The other cycle is repetitious
3
deposition of fluvial-estuary sediments. Depositional systems of
Geophysical Surveying Co., Ltd.
the Uonuma Group changed from deltaic to estuarine in the early
Manuscript received 3 October, 2005. Pleistocene. The Haizume Formation consists of shallowing-
Revised manuscript received 15 December, 2005. upward successions, which resulted from deposition on the margin
Part of this paper was presented at the 75th Annual International of a wave-dominated delta system.
Meeting of the Society of Exploration Geophysicists (2005).

© 2006 ASEG/SEGJ/KSEG 19
Xue, Tanase, and Watanabe Estimating CO2 saturation from time-lapse well logging

Fig. 1. Location and simplified geological structure of the CO2 injection site at Nagaoka city, Niigata Prefecture.

condition. The reservoir interval in each observation well was

cased with fibreglass casing. Time-lapse well logs were conducted
to detect the arrival of injected CO2 in each of the three observation
wells (Figure 3). Open-hole logs were run to investigate lithology
and porosity before casing each well. To evaluate the petrophysical
properties of the geological unit for sealing and retaining CO2,
drilled cores were acquired from the wells IW-1 and OB-2. Along
with conventional core analysis, thin sections were made to
examine pore geometries, mineralogy, and cementing materials. A
photomicrograph is presented in Figure 4, indicating the moderate
porosity (filled with blue resin). Mineralogy from X-ray diffraction
analysis showed the presence of plagioclase and feldspar, and the
results were used to calibrate a lithological model for porosity.
Fig. 2. Configuration of the injection well (IW-1) and the three Pressure and temperature transducers were placed at the reservoir
observation wells (OB-2, -3, -4) at the depth of the reservoir formation. depth in IW-1 to monitor CO2 injection conditions, and in OB-
The arrow indicates the formation dip. 4 to record changes in pressure and temperature, especially
the formation pressure build-up throughout the CO2 injection
experiments. The initial temperature and pressure of formation
At the pilot site one injection well (IW-1) and three injection water were 48°C and 10.8 MPa, respectively. The temperature
wells (OB-2, -3, and -4) were drilled, and the well layout at and pressure were above the critical values (31°C, 7.38 MPa) of
reservoir depth is shown in Figure 2. Observation wells OB-2 and CO2, and therefore the injected CO2 was in supercritical state. The
OB-3 were drilled toward the possible migration and accumulation formation pressure and temperature data are necessary for to apply
of the injected CO2. OB-2 and OB-3 were located 40 m down environmental corrections to cased-hole logs.
dip and 120 m up dip from the injection well, respectively. The
reservoir is a porous sandstone bed of the Haizume Formation, Well logging methods for CO2 monitoring
at a depth of 1100 m. At the injection well the target formation
thickness is 60 m, and the formation was divided into five zones Laboratory studies of porous sandstones showed that wave
(zone 1 to zone 5) based on results of well tests and well logs. velocity decreased significantly due to CO2 injection in water-
A thin permeable zone having a thickness of 12 m (zone 2) was saturated conditions (Wang and Nur, 1989; Xue et al., 2002). Such
selected for CO2 injection. The averaged horizontal permeability laboratory measurements strongly support the prediction from
for zone 2 was estimated to be around 7 millidarcy from the Gassmann’s equation, that a drastic change of P-wave velocity
pumping test. The overlying mudstone above zone 1 is about results from very little CO2 presence. The arrival of injected CO2
160 m thick and forms a regional seal. When monitoring CO2 at observation wells would be expected to cause a great reduction
distribution, both casing and cement are required to be in good in sonic log velocity.

20
Xue, Tanase, and Watanabe Estimating CO2 saturation from time-lapse well logging

Fig. 3. The time-lapse well logging program during the CO2 injection period at the Nagaoka site. The orange and blue lines indicate the CO2 injection
rate and the accumulated volume of injected CO2, respectively. Small squares indicate the well-logging runs, and the green and red squares indicate
the runs carried out before and after the CO2 breakthrough.

At the Nagaoka pilot site, sonic wave velocities of P- and which is indicated by a black broken line in each track in Figure 5.
S-waves have been acquired frequently in all observation wells The relative changes (difference between baseline and monitoring
during CO2 injection (Figure 3). Dual induction logs would also surveys) with time, for induction, neutron, and sonic logs were
be expected to detect CO2 breakthrough, because less-conductive plotted in the right tracks in Figures 6a–6d.
CO2 is replacing the highly-conductive formation water in porous
rocks. Experimental studies showed that resistivity increased Figures 6a and 6b show the changes with time in the medium
up to several times due to CO2 injection into brine-saturated and deep induction logs, respectively. The black broken line
porous sandstones (Myer, 2001; Kubota et al., 2005). At the indicates the baseline, and the initial resistivity was about 5 Ω.m.
Nagaoka site the reservoir interval in each observation well was When the injected CO2 arrived at the depth of 1116 m (zone 2),
cased by fibreglass rather than steel, so that dual induction log the resistivity increased to about 5.5 Ω.m. The continued change of
measurements could be made throughout CO2 injection. The resistivity with time showed that the CO2-bearing zone expanded,
neutron log responds primarily to the amount of hydrogen, which because the peak value and area increased. By the 25th run the
is contained mostly in formation water. It can be used to identify
formation porosity, which is usually called neutron porosity.
Therefore, the neutron log would provide important information
on porosity data, to help in quantifying CO2 saturation from sonic
or induction log data. In this study the neutron log was carried out
both in open and cased holes. The two logs were used to examine
the correlation between open-hole and cased-hole porosities. A
gamma-ray log is used to obtain shale volume of reservoirs. From
the well logs we confirmed the occurrence of CO2 breakthrough at
wells OB-2 and OB-4, but there has been no sign so far showing
that CO2 has arrived at OB-3.

Results of well logging at the observation well OB-2

Figure 5 summarises the results of dual induction, gamma ray,

neutron, and sonic logs obtained the 14th run in the observation
well OB-2. The CO2 breakthrough was discerned after the 13th run,
when a total amount of about 4000 tonnes CO2 has been injected,
as evidenced by (1) a moderate increase in resistivity, (2) a drastic
decrease in P-wave velocity, and (3) a decrease in neutron porosity. Fig. 4. A thin section of the reservoir sandstone showing pore space,
The results of the 1st to 13th runs were averaged to form a baseline, indicated by blue resin.

21
Xue, Tanase, and Watanabe Estimating CO2 saturation from time-lapse well logging

Fig. 5. Results of dual induction, gamma ray, neutron, and sonic logging at the observation well OB-2. The three zones masked in the induction log
are no-data zones, because of steel bridges between fibreglass casings. Averages of runs 1–13 are shown by the black broken lines. The shadow areas
indicate the ranges between maximum and minimum values for all surveys.

increase in resistivity was about 1.6 Ω.m, and the width of CO2- wave velocity, because of indistinct first arrivals. Therefore, results
bearing zone had extended to about 2 m. Both changes in zone of S-wave velocity logging have not been used in the present
width and resistivity values were much larger in the earlier stages analysis, although the S-wave also contains useful information on
after CO2 breakthrough. There was no clear difference between the the CO2 behaviour in the target formation.
medium and deep induction results.
Results of well logging at the observation well OB-4
Figure 6c shows the changes in neutron logs over time. At the
depth of 1116 m, neutron logging suggests a porosity reduction Figure 7 shows the summary of well logging results at
after CO2 breakthrough. In later records, after the 16th run, the observation well OB-4. In OB-4, the temperature and pressure
zone width extended first downward and then upward around the transducers were placed outside the fibreglass casing at the
1116 m depth. A change of neutron porosity also appeared at the reservoir depth. Lead wires from the transducers behind the casing
depth of 1114 m from the 18th run. The width of the CO2-bearing heavily disturbed the measurements of induction log, so induction
zone detected by neutron log was much larger than that shown results above zone 2 are not shown in Figure 7. Breakthrough
by the induction log. By comparison with the induction data, of CO2 at the observation well OB-4 was confirmed at the 17th
the neutron log showed greater fluctuations. There were some run, when a total of about 5500 tonnes CO2 had been injected.
small fluctuations throughout zones 1 and 2, possibly indicating The breakthrough was indicated by the same criteria, except the
the noise caused by fibreglass casing and casing cement in the induction log, as in OB-2. Figures 8a and 8b show the changes
cased holes. However, the changes that appeared in zone 2 are with time in neutron porosity and P-wave velocity, respectively.
large enough to suggest that displacement processes between the From the neutron porosity results, the CO2-bearing zone could be
injected CO2 and the formation water are responsible. confirmed to be around the depth of 1090 m, in zone 2, although the
data showed strong fluctuations similar to OB-2. Neutron porosity
Figure 6d shows the sonic P-wave velocity changes after the decreased by much more than 2% at the 23rd run due to the CO2
CO2 breakthrough. P-wave velocity decreased noticeably, up breakthrough having commenced at the 17th run. In Figure 8b, two
to 23%, due to the presence of injected CO2. The CO2-bearing CO2-bearing zones could be found in the sonic P-wave velocity
zone also shows similar changes to those shown by neutron changes. From the 18th run these zones began to merge into one
logging, corresponding to the CO2 migration pattern in zone 2. zone, with similar width to the zone in the neutron log results.
Unfortunately, sonic S-wave velocity is less well measured than P-

22
Xue, Tanase, and Watanabe Estimating CO2 saturation from time-lapse well logging

(a) (b)

(c) (d)
Fig. 6. Changes in well log responses with time, at OB-2. (a) Medium induction log; (b) deep induction log; (c) neutron porosity log; and (d) sonic
P-wave velocity log. The black broken line indicates the averages of runs 1–13.

Fig. 7. Results of dual induction, gamma ray, neutron, and sonic logging at the observation well OB-4. The three zones masked in the induction log
are no-data zones, because of steel bridges between fibreglass casings and lead wires from pressure and temperature transducers. Averages of runs
1–16 are shown by the black broken lines.

23
Xue, Tanase, and Watanabe Estimating CO2 saturation from time-lapse well logging

(a)

Fig. 9. P-wave velocity changes with respect to time during injection

of supercritical CO2 into a drilled core sample (fine-grain sandstone).
Velocity reduction is about 20%, which agrees with the sonic P-wave
velocity change due the CO2 breakthrough at OB-2.

(22% reduction), which again showed a good agreement with the

sonic log result. Another experiment with a CO2 injection pressure
of 10.5 MPa resulted in a velocity reduction of 20%, which was a
little lower than the sonic log result. The amplitude of first arrival
P waves showed a drastic decrease prior to the velocity decrease:
the amplitude ratio decreased about 10% (Figure 10). This result
suggested that amplitude as well as travel time of P waves also
could be a good signal for the breakthrough.
(b)
DISCUSSION
Fig. 8. Changes with time in neutron porosity and sonic P-wave Calculation of P-wave velocity in a partially saturated
velocity, showing the CO2-bearing zone expanded at OB-4. (a) Neutron
porosity, and (b) sonic P-wave velocity.
reservoir with Gassmann’s equation

Laboratory velocity measurement on drilled cores In this study the pore fluids are water and supercritical CO2.
Gassmann’s equation (Gassman, 1951) relates the bulk modulus
Xue and Ohsumi (2004) carried out experiments on porous of a fluid-saturated rock (K), to the dry rock bulk modulus (Kd),
sandstone to investigate P-wave velocity change during CO2 the solid matrix (Ks), the fluid bulk modulus (Kf), and the porosity
injection. They concluded that P-wave velocity decreased of the (φ):
order of 10% when injecting supercritical CO2 into the water-
saturated Tako Sandstone (23% porosity). In this study drilled
(1)
cores, collected from the depth of 1145 m in OB-2, were used
to confirm velocity reduction due to CO2 injection. Cylindrical
samples, 5 cm in diameter and 7 cm in length, were cut from the
drilled cores. Stainless steel end pieces were attached to both ends The shear modulus G of a rock is expected to be unaffected by
of the sample. Piezoelectric transducers with a centre frequency fluid saturation at relative low frequencies:
of 200 kHz were placed inside the end pieces. P-wave velocity (2)
changes were measured along the axial direction of the sample
during CO2 injection. To simulate in-situ temperature and pressure where Gd is the shear modulus of the dry rock. The density of a
conditions, the sample was placed in a high-pressure vessel. partially saturated rock can be calculated from the densities of pore
Details of experimental procedures are shown in Xue and Ohsumi fluids, solid matrix and the degree of saturation:
(2004). The hydrostatic pressure and pore pressure applied to the
sample were changed alternately. Because of the friable nature (3)
of the sample, the pressures were increased gradually. It took
about one week to reach the hydrostatic pressure of 16 MPa where ρ is the density of partially saturated rock, ρs is the density
and the pore pressure of 10 MPa, respectively. The temperature of solid matrix, and ρw and ρc are the densities of pore water and
was controlled at 39°C by a heating system. The pore water was CO2 in pore spaces. Sw is the water saturation (volume fraction) and
artificial saline water having the same chemical composition and the remainder (1-Sw) is the CO2 saturation (volume fraction) in the
pH as the formation water. Under these pressure and temperature partially saturated rock.
conditions the P-wave velocity was 2.3 km/s, which agreed well
with the sonic P-wave velocity (2.4 km/s). Figure 9 shows the P- Now P-wave velocity in the partially saturated rock can be
wave velocity changes during CO2 injection into the core sample calculated from the bulk modulus (K), the shear modulus (G) and
under an injection pressure of 11 MPa. The P-wave velocity the density (ρ):
decreased quickly after CO2 injection, from 2.3 km/s to 1.8 km/s

24
Xue, Tanase, and Watanabe Estimating CO2 saturation from time-lapse well logging

(8)

(Dvorkin and Nur, 1996). Then the bulk modulus of the fluid-
saturated rock (K) is modelled by Gassmann’s equation:

(9)

Finally, P-wave velocity in the partially saturated rock can be

calculated from equation (4). Following the above processes we
Fig. 10. Changes in P-wave velocity and amplitude with respect to time matched the sonic P-wave velocity change due to the breakthrough
during injection of supercritical CO2 into sandstone (different sample
than Figure 9). Amplitude of the first arrival decreased to about 10%
of injected CO2 at the observation wells OB-2 and OB-4. We used
of the initial value. the corrected neutron porosity from the well logging results, and
estimated the CO2 fraction after CO2 breakthrough by matching the
P-wave velocity change.

(4)

Because of the friable nature of the core samples (weak-

cemented porous sandstone), laboratory measurements of the dry
rock bulk modulus and the solid matrix density were unavailable in
this study. We used a rock-properties model proposed by Hoversten
et al. (2003). They estimated the effective bulk (Kd) and shear (Gd)
moduli of a dry, dense, random pack of spherical grains from the
following expressions:

(5)

(6)

where φ0 is the critical porosity (the porosity above which the

grains become a liquid suspension), Peff is the effective pressure, ν
and Gs are Poisson’s ratio and shear modulus of the grain material,
and l is the average coordination number of contacting grains.

Equations (5) and (6) describe the effective dry-frame moduli

at the critical porosity φ0. The effective moduli (Keff and Geff) at a
different porosity φ are given by:

(7) Fig. 11. Comparison between porosities measured under open-hole and
cased-hole conditions. The two porosities agreed fairly well around the
depth of 1115 m, except that open-hole porosity shows short-distance
fluctuations.

25
Xue, Tanase, and Watanabe Estimating CO2 saturation from time-lapse well logging

Open-hole and cased-hole porosities

In addition, we assumed that the distribution of formation
The reservoir porosity is an important parameter for interpreting water and supercritical CO2 fluids is described by the fine-scale
results of well logs, especially the quantitative evaluation of CO2 (or uniform) saturation model described by Mavko and Mukerji
saturation in reservoirs. Figure 11 shows the comparison between (1998). There is another coarse-scale or patchy saturation model.
open-hole porosity (red) and cased-hole porosity (blue) from the However, we failed to obtain any reasonable results with this
neutron logs in OB-2. The two porosities are in good agreement model. The cause of the failure has been left for future studies.
at the target formation (zone 2), especially the CO2-bearing zone Therefore, we showed only results for CO2 saturation estimated
(1113 to 1118 m), as confirmed with sonic and neutron logs. In under uniform saturation conditions. Figures 13 and 14 show
the cased holes the neutron logging reflects the reservoir porosity changes in CO2 saturation with time in the observation wells OB-2
behind the fibreglass casing and the casing cement. Results from and OB-4, respectively. In OB-2, CO2 saturation ranged from 10 to
the Cement Bond Log showed that the casing and cement were 40%, and there were no signs of CO2 leakage from the reservoir. In
firmly consolidated in each well. This enables us to use the cased- OB-4 the CO2 saturation increased to 37% from 17th through 21st
hole porosity in quantifying CO2 saturation within the reservoir. runs, and then decreased to 20% at the 23rd run. CO2 saturation
We found that a higher proportion of shale (estimated from the within the reservoir strongly responded to the CO2 injection
gamma ray log) in the reservoir results in a bigger difference process. Because of a CO2 supply shortage we stopped injection
between open-hole and cased-hole porosities. several times (Figure 3), and obtained low CO2 saturations at the
subsequent runs. This suggested that CO2 was penetrating new
History matching with time-lapse sonic logging, and areas from the CO2-bearing zone. If CO2 migration is driven by
quantifying CO2 saturation buoyancy, resulting from the density difference between formation
water and CO2, time-lapse well logging may provide essential
Based on the rock-properties model, we successfully matched information on long-term CO2 behaviour in saline aquifers.
the sonic P-wave velocity changes at the two observation wells OB- Injection of CO2 injection ended in January 2005 at the Nagaoka
2 and OB-4. Figure 12 shows an example of the matched velocity site, and the pressure build-up due to CO2 injection has gradually
before and after CO2 breakthrough in OB-2. In this figure Vp_Calc decreased to the time of writing. We have launched a follow-up
indicates P-wave velocity predicted with Gassmann’s equation. program to monitor the CO2 behaviour over the next three years.
The sonic velocity (log data: blue line) and the Gassmann-
calculated velocity agreed fairly well throughout the well- logging Capability of CO2 monitoring by time-lapse well logging
interval. Velocity reduction due to CO2 breakthrough at the 14th
run was also matched below the zone 1 (a little below 1100 m). Time-lapse well logging has been applied successfully to
The results of history matching strongly supported our approach monitor changes in resistivity, neutron porosity, and wave velocity,
of quantifying CO2 saturation from the sonic log data. By mainly resulting from CO2 displacement of formation water.
comparing the matching results between the 13th run (before CO2 Results from neutron and sonic logs are quite similar, when
breakthrough) and the 14th run (after CO2 breakthrough), the CO2 detecting CO2 distribution within the reservoir. The result of
saturation was estimated to be about 30% at the depth of 1115 m. history matching suggested that the maximum sweep efficiency

Fig. 12. History matching on sonic P-wave velocity changes before (13th run) and after (14th run) the CO2 breakthrough. Vp_Calc indicates velocity
values predicted with Gassmann’s equation.

26
Xue, Tanase, and Watanabe Estimating CO2 saturation from time-lapse well logging

Fig. 13. Changes with time in CO2 saturation at the observation well OB-2.

Fig. 14. Changes with time in CO2 saturation at the observation well OB-4.

was about 40%. We found that induction logging detected only CONCLUSIONS
half of the CO2-bearing zone width, compared to the neutron and
sonic logs. The induction log reflects a special smoothed average We have demonstrated that the time-lapse well logging can
of the response from the CO2-bearing zone. Within the partially detect CO2 breakthrough at the Nagaoka pilot site of CO2
saturated reservoir, induced currents still propagate mainly through geological sequestration. The CO2 breakthrough was identified by
the formation-water saturated zone because of the extremely low induction, sonic and neutron loggings. We found that the neutron
resistivity of the formation water. Formation water and CO2 are and sonic logs showed similar CO2 distribution in the observation
two components of a parallel circuit. Lesser contributions by CO2 well OB-2. The induction log detected only half of the width of
saturation to the resistivity resulted in small changes in induction the CO2-bearing zone compared to the neutron and sonic logs.
logs. However, the induction log is still an important method in Our results suggest that small contributions by CO2 saturation to
CO2 monitoring, together with neutron and sonic logs. In Figure resistivity resulted in small changes in induction log responses in
6a we found the resistivity increased gradually in the time-lapse the partially saturated reservoir.
surveys. This result provides us with essential information about
the process of CO2 displacing formation water in the reservoir. The Sonic P-wave velocity decreased significantly with the
detailed results of induction logging will be discussed in another breakthrough of injected CO2 in the observation wells. We
paper. confirmed that P-wave velocity reduction in the sonic logs
agreed fairly well with laboratory measurements on drilled core
samples from the pilot site. Based on a rock-properties model in
conjunction with Gassmann’s equation, we successfully matched

27
Xue, Tanase, and Watanabe Estimating CO2 saturation from time-lapse well logging

Mavko, G., and Mukerji, T., 1998, Bounds on low-frequency seismic velocities in
the history changes in sonic P-wave velocity with estimated CO2 partially saturated rocks: Geophysics, 63, 918–924.
saturation after breakthrough in two observation wells out of three.
The result of history matching in the sonic log data suggested that Müller, N., Ramakrishnan, T.S., Boyd, A., and Sakurai, S., 2005, Time-lapse CO2
monitoring with pulsed neutron logging, paper 22: Proceedings of the 4th Annual
the maximum sweep efficiency was about 40%. From the time- Conference on Carbon Capture and Sequestration DOE/NETL, May 2–5.
lapse results we also found CO2 saturation strongly responded
Myer, L.R., 2001, Laboratory measurement of geophysical properties for monitoring
to intermittent injection into the CO2-bearing zone. During the
of CO2 sequestration [Web document]: Accessed 21 December, 2005. Available at
injection phase, the driving force for CO2 movement could be <http://www.netl.doe.gov/publications/proceedings/01/carbon_seq/6a4.pdf>.
the pressure gradient, combined with the density difference
Sakurai, S., Ramakrishnan, T. S., Boyd, A., Müller, N. and Hovorka, S., 2005,
(buoyancy) between formation water and CO2. We will continue
Monitoring saturation changes for CO2 sequestration; Petrophysical support of
our well logging program from 2005 through 2007, to monitor the Frio brine pilot Experiment: Proceedings of the SPWLA 46th Annual Logging
the CO2 movements mainly driven by buoyancy forces. Our field Symposium, New Orleans, June 26-29.
measurements indicate that the pressure build-up associated with Wang, Z., and Nur, A.M., 1989, Effects of CO2 flooding on wave velocities in rocks
CO2 injection is gradually decreasing. with hydrocarbons: SPE Reservoir Engineering, 4, 429–436.

Xue, Z., Ohsumi, T., and Koide, H., 2002, Laboratory measurements of seismic wave
ACKNOWLEDGMENT velocity by CO2 injection in two porous sandstones: in Gale and Kaya (eds),
Proceedings of the 6th International Conference on Greenhouse Gas Control
We thank Osamu Nishizawa of AIST for his thoughtful review Technologies, 359–364.
comments, which have helped substantially to improve this Xue, Z., and Ohsumi, T., 2004, Seismic wave monitoring of CO2 migration in water-
paper. Thanks also to an anonymous reviewer for his constructive saturated porous sandstone: Butsuri-Tansa, 57, 25–32.
comments. We are grateful to all staff involved in this project
Xue, Z., Tanase, D., Saito, H., Nobuoka, D., and Watanabe, J., 2005, Time-lapse
in ENAA, RITE, and Teikoku Oil Corporation. This work was crosswell seismic tomography and well logging to monitor the injected CO2 in an
supported by Ministry of Economy, Trade, and Industry of Japan onshore aquifer, Nagaoka, Japan: 75th Annual International Meeting of the Society
under the contract for “Research and Development of Underground of Exploration Geophysicists, Expanded Abstracts, 1433–1436.
Storage for Carbon Dioxide”.

REFERENCES
Butler, R., Gerard, M., and Stevens, J., 1993, The use of monitoring logging to
evaluate CO2 foam flooding in the East Vacuum Grayburg San Andres unit:
Transactions of the SPWLA Annual Logging Symposium, 34, O1–O24.

Chadwick, R.A., Eiken, O., and Lindeberg, E., 2004, 4D geophysical monitoring of
the CO2 plume at Sleipner, North Sea: Aspects of uncertainty: Proceedings of the
7th SEGJ International Symposium, Sendai, November, 24–26.

Dvorkin, J., and Nur, A., 1996, Elasticity of high-porosity sandstone: Theory for two
North Sea data sets: Geophysics, 61, 1363–1370.

Gassmann F., 1951, Über die Elastizität poröser Medien: Vierteljahrsschrift der
Naturforschenden Gesellschaft in Zürich, 96, 1–23.

Gould, J., Wackler, J., Quirein, J., and Watson, J., 1991, CO2 monitor logging: East
Mallet Unit, Slaughter Field, Hockley County, Texas: Transactions of the SPWLA
Annual Logging Symposium, 32, PP1–PP23.

Hoyanagi, K., Ikedzu, D., Shimizu, Y., and Omura, A., 2000, Reconstruction of the
delta and estuary systems and sequence stratigraphy of the Plio-Peistcene strata
in the Higashikubiki Hills, Niigata Prefecture, central Japan: Earth Science, 54,
393–404.

Hoversten, G.M., Gritto, R., Washbourne, J., and Daley, T., 2003, Pressure and fluid
saturation prediction in a multicomponent reservoir using combined seismic and
electromagnetic imaging: Geophysics, 68, 1580–1591.

Kikuta, K., Hongo, S., Tanase, D., and Ohsumi, T., 2004, Field test of CO2 injection
in Nagaoka, Japan: 7th International Conference on Greenhouse Gas Control
Technologies, September 5-9.

Kubota, K., Suzuki, K., and Xue, Z., 2005, Experimental study on resistivity and
SP monitoring during CO2 injection into water-saturated porous sandstone:
(Preprint).

28
Xue, Tanase, and Watanabe Estimating CO2 saturation from time-lapse well logging

29