June 2013

PANACEA

Predicting and monitoring the long-term behavior of CO2 injected in deep geological formations

Project Number : 282900

Work-Package: WP02

WP Title C02 Sites and Analogues

Deliverables D202 & D204

Status (DRAFT, FINAL) Version 1.00

Review level (WP, EU, ALL) Planned delivery date Month 18 Actual delivery date Month 18 Leading participant UEDIN

Contributing participants P. Pezard, N. Denchik, J. Lofi, H. Perroud, G. Henry, J.

Paris, M. Geeraert, A. Deleau, O. Jaafar

Dissemination Level

PU Public X RE Restricted to the consortium members, the SIRAB, the end-users and the EU

officers

CO Confidential (only the consortium and the EU officers)

Deliverable number D202 & D204 Deliverable name Geo-process facies analysis of experimental injection sites

Work-package WP02: Field quantification techniques Lead participant UGOE

Version Submitted by Review level

Submitted Reviewed

1.00 CNRS WP June 21, 2013

Editors Name Participant email

Philippe Pezard CNRS ppezard@gulliver.fr Nataliya Denchik CNRS denchik@gm.univ-montp2.fr Johanna Lofi CNRS lofi@gm.univ-montp2.fr Herv Perroud UPPA perroud@gm.univ-montp2.fr

Executive summary A shallow experimental site has been set-up for shallow CO2 injection at Maguelone near Montpellier (Mediterranean coastline, gulf of Lions, France) and the development of integrated surface and in-situ (downhole) monitoring methods. The site will be further developped in the context of PANACEA later in 2013 with 4 new boreholes, improved injection facilities and the downhole set-up of new monitoring instruments down to a maximum of 20 m. Keywords Shallow experimental site for CO2 injection, hydrogeophysical monitoring

methods. Clastic sediments (sand-shale sequences).

Table of content

INTRODUCTION ......................................................................................................................... 1

1. "SIMEX" AND THE MAGUELONE SHALLOW INJECTION EXPERIMENTAL SITE ......................... 2

1.1. The Maguelone experimental site ..................................................................................... 2 1.2. Field experimental set-up ................................................................................................ 4

2. HYDROGEOPHYSICAL MONITORING FROM SURFACE AND DOWNHOLE .................................. 5

2.1. Pressure monitoring and fluid sampling from the WestBay. .................................................. 5 2.2. Seismic monitoring ......................................................................................................... 6 2.3. Downhole electrical monitoring ........................................................................................ 8 2.6. Downhole sonic monitoring .............................................................................................. 9

CONCLUSION AND PERSPECTIVES ........................................................................................... 10

REFERENCES ............................................................................................................................ 11

Figure captions

Figure 1. Geographical location (left) and aerial photograph (right) of the Maguelone experimental site located to the NW of the photograph (with the MAG1 and MAG4 boreholes) and one km to the east of the Maguelone island (with the MAG3 borehole). .................................................................................. 2 Figure 2. Compilation of data set in borehole MAG. From Lofi et al., (2012). (A) simplified lithological log. RES: regional erosion surface. G: Gravelly horizon; S: sandstone; Stars: radiocarbon datings (years cal. B.P.) after Raynal et al. (2009); (B) Vertical grain size distribution (fraction

Introduction

An experimental setup for shallow subsurface hydrogeophysical monitoring has been installed at the Maguelone site, located along the Mediterranean lido of the Gulf of Lions near Montpellier, France. This experimental site was created first in the context of ALIANCE EC project (FP5), then further developped as part of MUSTANG and PANACEA EC projects (FP7), as well as from CNRS and regional funding (Languedoc-Roussillon ARPE program), whether for drilling, instrumentation or operational matters. "SIMEx" (Shallow Injection Monitoring Experiment) provides a unique opportunity to test in a cost effective manner a full suite of coordinated monitoring techniques, either from surface or downhole. The field spread includes at this point an injection hole, a time-lapse logging hole (TLL), a downhole hydrodynamic observatory based on a pore fluid sampling completion from WestBay (SWS), two permanent downhole electrical resistivity observatories (imaGeau), a downhole seismic observatory, plus surface seismic and electrical resistivity monitoring arrays. This coordinated set of observatories will be further developped as part of PANACEA and should lead to the design of integrated sensors and methods for the monitoring of gas injection in deeper reservoirs (Pezard et al.,2010). Prior to CO2 injection, N2 injection experiments were undertaken at Maguelone as part of SIMEx project in order to measure the site response to gas injection. Nitrogen was chosen because of its nonreactive nature and because of the reducing nature of the in-situ environment present in the shallow subsurface at Maguelone, precluding oxygen injection to avoid massive bacterial developments. Nitrogen injection experiments consisted in an injection in the dedicated hole at a depth of 13-16 m (fluvial conglomerates and sands). For the three nitrogen injection experiments in March, June and November of 2012, the gas plume was successfully detected by all monitoring techniques (pressure, electrical resistivity and seismic monitoring), either from surface or in-situ. The detection of gas propagation was immediate. The pressure values increased during injections and decreased between and after injections. The increase in resistivity was measured during all injection periods. The seismic monitoring appears to be extremely sensitive to the underground gas storage, with very clear modifications of the recorded traces as soon as the injection begins. The nitrogen storage was observed closer to the surface than expected (~7 m) under the Late-Holocene lagoonal sediments (mostly impermeable dark green clays), forming an impermeable seal overlying homogeneous fine-grained Pliocene continental deposits. During nitrogen injection experiments, a clear noise of the gas bubbles was heard directly in the imaGeau and TLL holes just after the begging of injection. Also a clear seismic noise anomaly was detected from surface monitoring in the vicinity of the TLL hole. In order to avoid the gas leakage and in preparation for the CO2 injection experiment in 2013, all holes were cemented as part of PANACEA down to 7 m by drilling 3 small holes around the original ones. The next phase of SIMEx within MUSTANG was that of CO2 injection, with similar surface and downhole hydrogeophysical monitoring than for N2 injection. For the CO2 injection experiment in January 2013 concluding SIMEx operations as part of MUSTANG, the gas plume was successfully detected by all monitoring techniques (pressure, electrical resistivity, seismic monitoring and fluid sampling, including on-site pH measurements). The reaction to gas propagation was immediate. The pressure values increased during injections and decreased between and after injections. Electrical resistivity monitoring and induction logging showed a similar reduction in resistivity due to CO2 dissolving into pore water below 8 m. A derease in pH values was also measured. The analysis of pore fluid samples taken during the experiment will permit to confirm the CO2 dissolving during injection.

MUSTANG - 227286 P a g e | 1

1. "SIMEx" and the Maguelone shallow injection experimental site

1.1. The Maguelone experimental site

The Maguelone experimental site is located along the Mediterranean lido of the Gulf of Lions passive margin, 10 km to the south of Montpellier. Limited to the north by the Prevost coastal lagoon and to the south by the Mediterranean Sea (Figure 1), this site offers a natural laboratory to study porous coastal reservoirs in a clastic and clay-rich context saturated mostly with saline fluids.

Figure 1. Geographical location (left) and aerial photograph (right) of the Maguelone experimental site located to the NW of the photograph (with the MAG1 and MAG4 boreholes) and one km to the east of the Maguelone island (with the MAG3 borehole). Continuous geological samples and geophysical data from shallow boreholes at Maguelone have leaded to identify two depositional sequences (Figure 2): Near the ground surface (0-9 m), a thin Late-Holocene sequence (< 5000 yrs B.P.) is

constituted with lagoon sediments with impermeable dark green clays topped by grey shelly beach sands. This sequence forms an impermeable seal overlying the Pliocene sequence with an unconformity.

Pliocene sequence, from ~ 9 m to the base of MAG1-DSO (60 m). This sequence consists

mainly in relatively homogeneous fine grained continental deposits (clays, silts, and clayey silts). Locally, some marine incursions (grey clays) and lacustrine levels (white carbonates clays) are visible. The clayey fraction is relatively high all along the sequence, making those deposits relatively poorly permeable.

Gulf of Lions

MUSTANG - 227286 P a g e | 2

Figure 2. Compilation of data set in borehole MAG. From Lofi et al., (2012). (A) simplified lithological log. RES: regional erosion surface. G: Gravelly horizon; S: sandstone; Stars: radiocarbon datings (years cal. B.P.) after Raynal et al. (2009); (B) Vertical grain size distribution (fraction

Figure 3. Fluvial deposits at 13-16m depth. At left: Geological sample taken at Maguelone study site (depth 15.40-16.50 m), consisting in a porous and permeable conglomerates with sands (low sea level stand fluvial deposits). At right: H2S encountered during coring operations near 15m depth. In the Pliocene sequence, a single remarkable depositional unit is located from about 13 to 16 m depth and consists in a porous and permeable conglomerates and sands interpreted as fluvial deposits (Figure 3). The conglomerates, clearly identified downhole from low natural gamma radioactivity values, can be correlated laterally with boreholes located a t a km distance, showing the lateral extension of this unit. Sedimentary facies, geophysical measurements and hydrological testing indicate a high permeability (K = 4.10-3 m/s), salinity (34 g/l) and porosity for these conglomerates, also bounded above and below by clay-rich horizons. Hydrogen sulphite (H2S) encountered in a repetitive fashion during coring and drilling operations near 15m depth confirms the presence of a small reservoir at this depth. The possibility of the conglomerate forming a 3 m-thick gas-rich reservoir may consequently be envisaged, this anomaly resulting from lagoonal organic matter decomposition. From a hydrological point of view, the induction electrical resistivity (opposite to electrical conductivity, Figure 2) data in MAG1 show that the sedimentary column is saturated with seawater to brackish water from surface down to 32 m. Below, a gradual increases in electrical resistivity can only be explained by a gradual freshening in pore fluid with increasing depth. Below 40 m, values as high as 8.0 .m and above are reached. During drilling, these horizons were found to be artesian for about 30 minutes, with an initial fountain a few meters high. The hole is now equipped with a PVC casing perforated at the base only (from 59 to 62 m).

1.2. Field experimental set-up

The field spread (Figure 4) for gas injection included an injection hole GIH, a time-lapse logging hole TLL, two downhole electrical observatories imaGeau (DEO), a downhole hydrodynamic observatory based on a multipacker completion from WestBay (SWS) DHO, a downhole seismic observatory DSO, plus surface seismic (SSO), downhole electrical resistivity observatories (imaGeau), and a surface electrical resistivity monitoring set-up. Prior to gas injection, a whole set of pre-injection experiments were conducted in order to prepare, test, and calibrate the Maguelone site for later experiments. The initial petrophysical and geophysical properties of the study site were also characterized both in the laboratory from core and from downhole geophysical measurements (gamma ray, electrical and acoustical logging), followed by surface (electrical and seismic tomography) and surface-toborehole (seismic) surveys.

MUSTANG - 227286 P a g e | 4

Nitrogen injection experiments were undertaken in order to measure the site response to gas injection. Nitrogen was chosen because of the reducing nature of the in-situ environment present in the shallow subsurface at Maguelone, precluding oxygen injection to avoid massive bacterial developments. Three nitrogen injection experiments were realized at the Maguelone study site in March, June and November 2012 and consisted in an injection of nitrogen in the injection hole GIH (Figure 4) at a depth of 13-16 m (fluvial conglomerates and sands). It revealed a series of downhole leaks corrected by shallow cementation (down to 7 m) of several holes around injection and monitoring holes.

Figure 4. Field spread at the Maguelone experimental site for the SIMEx integrated monitoring experiment.

2. Hydrogeophysical monitoring from surface and downhole

2.1. Pressure monitoring and fluid sampling The downhole hydrodynamic observatory (DHO, Figure 4) is a multilevel groundwater device based on a multi-packer completion from WestBay (SWS), including packers in order to isolate the different hydrological horizons and to provide fluid samples, temperature and pressure data before, during and after injection, thereby time/space calibration points from tracers (i.e. precise boundary conditions) to numerical modelers (Figure 5). At the Maguelone site, eight Westbay zones were equipped for fluid sampling and monitoring down to a depth of 49 m. For experimental purpose as part of SIMEx, two of these zones were located within the reservoir where gas is injected (at 13.9 and 15.5 m), one was located above (at 7.9 m) and a fourth one below (24.9 m). For long-term survey of underlying groundwater, another set of four sampling zones was installed in front of sand-rich layers to allow probing at 32.1, 36.7, 39.8 and 49.0 m.

MUSTANG - 227286 P a g e | 5

Figure 5. Multilevel groundwater device based on a multi-packer completion from the Schlumberger "WestBay" downhole hydrochemical monitoring system.

2.2. Seismic monitoring

The seismic time-lapse monitoring is performed before, during and after gas injection, both downhole and at surface, as follows : both downhole and surface seismic records were obtained at 10 to 15' time intervals. downhole 3 components 28 Hz geophones were installed at 7, 12 and 17 m respectively.

The surface seismic antenna is composed of 28 vertical geophones (50 Hz), spaced every 1 meter, along a line spaning from DSO to TLL and beyond, starting at 8 m from the DSO (Figure 4 & 25). A surface geophone, at the distance of 1m from DSO, was used as a reference for synchronization (cross-correlation).

The source (hammer blow) was kept at a fixed position 2m from the well head. Sequences of 5 successive shots were acquired for staking, and recorded individually. The ambient noise before the first arrival was analyzed first on the seismic traces. Figure 7 presents RMS amplitude of the seismic noise level along a surface antenna (channel numbers on vertical axis) according to the shot numbers (vertical axis). Just after the beginning of the injection a strong noise anomaly was detected in the vicinity of the geophone channels n19-20 near the TLL hole (MAG6). This noise was related to the presence of gas bubbles heard in this hole and corresponding the gas leakage along the hole from the injection reservoir and all the way to surface. Figure 27 shows an example of results for vertical and surface geophones: a borehole geophone at 7m depth, the most affected by the gas injection with very clear modifications of the recorded traces as soon as the injection begins; and a surface geophone with 22m offset, one of the surface geophones affected by gas injection.

MUSTANG - 227286 P a g e | 6

Figure 6. Seismic time-lapse monitoring set-up (2012). Above: acquisition setup. Red: seismic source (hammer blow). Blue: a channel used as a reference for synchronization, surface and borehole geophones. Below: monitoring timing during the injection experiment. In blue: injection periods.

Figure 7. Seismic time-lapse monitoring (March 2012): seismic rms noise level along surface antenna close to the leaking TLL hole later repaired with cement injection aroud the hole down to 7 m depth.

MUSTANG - 227286 P a g e | 7

2.3. Downhole electrical monitoring The subsurface observatories (DEO7&DEO9, Figure 4) were built from a PVC pipe along which downhole electrodes were located on the external surface of the pipe. The electrodes are made of gold plated copper to reduce the corrosion process. The technical specificities for the DEO9 (MAG9) apparatus are: 17.80 m depth capacity, with 0,35 m electrode spacing. The technical specificities for the DEO7 (MAG7) apparatus are: 18.15 m depth capacity, with 0.70 m electrode spacing from 0.0m to 13.0 m depth, 0.10 m electrode spacing from 13.0 m to 16.4 m depth, and 0.70 m again depth from 16.4 m to 18.15 m depth. The observatories are set-up in automatic way for data acquisition, with electrical resistivity profiles recorded once a day outside of experimental periods. During injection experiments, profiles are acquired every hour. Just before and after injection (pre- and post-injection monitoring), profiles are acquired every two/three hours. The high frequency acquisition during injections leads to the following in 2D (z,t) of downhole resistivity changes over time related to nearby gas injection. The results are presented here in terms of resistivity difference relative to the baseline measurement acquired before injection.

Figure 8. Time-lapse single hole electrical monitoring in June of 2012 in two nearby holes with differences overtime in electrical resistivity relative to the baseline measurements. Above: imaGeau observatory (DEO9). Below: time-lapse induction logging (TLL). The obtained resistivity are free of any clay contribution to the electrical signal and relate to the variable gas content in the pore space as nitrogen is being injected and slowly dissipates within the formation. An increase in resistivity is obtained at ~ 8m depth just after the start of injection in the two holes (imaGeau observatory and TLL). This increase due to N2 storage in two thin sand layers identified from core is observed to extend at this depth during the first and second injection periods, as well as between and after ~25 hours after the end of 2nd injection for imaGeau observatory and ~70 hours for the TLL hole. The increase in resistivity

MUSTANG - 227286 P a g e | 8

due to the gas injection is more significant for the imaGeau observatory then for the TLL hole. A second very slight increase in resistivity is obtained at ~15 m depth in the TLL hole, starting later than the first one, after about 4 hours of injection (Pezard et al., 2012).

2.4. Downhole sonic monitoring Repeated downhole sonic measurements were performed occasionally (1-3 times per day) in the Time-Lapse Logging borehole (TLL) and the DSO hole before, during and after nitrogen and CO2 injection experiments (June, November 2012 & January 2013). This work has been realized with Full Waveform Sonic logging probe built by Mount Sopris. The sonic measurements were recorded every 5 cm while logging up the hole. Figure 9 shows changes over injection time in sonic signal amplitude while Vp determined from the first arrival during the CO2 injection in the TLL hole (January 2013) does not appear to be modified. We can notice the changes in amplitude over the injection period, especially at the depth of 15 m.

Figure 9. Full waveform dual sensor sonic well logging (20 kHz) during CO2 injection in January 2013 (MAG6 - TLL hole). Left: changes over injection time in sonic signal amplitude. Right: Vp determined from first arrival.

MUSTANG - 227286 P a g e | 9

Figure 10. Downhole velocity model at the site in the DSO (MAG1) and TLL (MAG6) holes from full waveform sonic (20 kHz), seismic and core measurements (GeoTek bench). Please note the change in sonic velocity in MAG6 before and after cementation to ensure the sealing of the experimental box from potential leaks along the injection and monitoring holes.

The downhole sonic data also helped to verify the quality of the cementation performed after gas injection in March 2012 to minimize leaking of the experimental set-up along several monitoring holes such as MAG6 (TLL), MAG9 (DEO) and the injection hole itself. Faster P-wave velocities were obtained in MAG6 after cementation (Figure 10 - black curve, middle track), in November 2012, with respect to the initial profile (Figure 10 - red curve, middle track) recorded in 2010.

Conclusion and perspectives The validity of the experimental set-up developped at Maguelone for gas injection monitoring as part of MUSTANG has been verified in 2012 and leaks along a series of holes have been reduced or eliminated from cementing along them down to 7 m depth. The site will be further developped as part of PANACEA in 2013 and 2014, with the drilling of 4 new monitoring or injection holes, the set-up of new monitoring instruments, followed by phases of CO2 injection in order to : - get a more complete monitoring spead, - test new and particularly integrated and multimethods downhole monitoring instruments, - provide improved constraints to the modelers.

MUSTANG - 227286 P a g e | 10

References

Lofi, J.; Pezard, P.; Bouchette F.; Raynal O.; Sabatier P.; Denchik, N.; Levannier A.; Dezileau L. & R. Certain. (2012). Integrated Onshore-Offshore Investigation of a Mediterranean Layered Coastal Aquifer. Groundwater. DOI: 10.1111/j.1745-6584.2012.01011.x Pezard, P.; Denchik, N.; Lofi, J.; Perroud, H. & Neyens, D. (2010), Maguelone (Languedoc coastline, France): a shallow experimental site for CO2 storage. Hydrogeophysical monitoring methods.MAGUELONE. EGU General Assembly 2010. Pezard, P.; Denchik, N.; Lofi, J.; Perroud, H. & Neyens, D. (2012), SIMEx : a shallow integrated, multi-method hydrogeophysical monitoring experiment for CO2 storage conducted at Maguelone (Languedoc coastline, France). EGU General Assembly 2012.

MUSTANG - 227286 P a g e | 11

D202 & D204Introduction1. "SIMEx" and the Maguelone shallow injection experimental site1.1. The Maguelone experimental site1.2. Field experimental set-up

2. Hydrogeophysical monitoring from surface and downhole2.1. Pressure monitoring and fluid sampling2.2. Seismic monitoring2.3. Downhole electrical monitoring2.4. Downhole sonic monitoring

Conclusion and perspectivesReferences