Marine Electromagnetic Survey of Scarborough Gas Field

Transcription

1 Marine Electromagnetic Survey of Scarborough Gas Field R.V. Roger Revelle, May 22nd June Preliminary Cruise Report Steven Constable Kerry Key and David Myer October 13, 2009 A Research Project Funded by BHP Billiton Petroleum 1

2 SUMMARY In May/June 2009 we carried out a 32 day cruise aboard the Scripps research vessel Roger Revelle in order to collect marine magnetotelluric (MT) and controlled source electromagnetic (CSEM) data over the Scarborough gas field on the northwest shelf of Australia. The objectives of the experiment were to test new CSEM equipment systems and to provide a comprehensive data set for the study of various modeling and interpretation methods. A total of 54 marine EM recorders were used, which included two new long-antenna gradiometers and two recently developed towed 3-axis recorders. Two marine EM transmitters were provided, and which were operated at 300 A emission current on a 250 m antenna, for a source dipole moment of 75 kam. The source waveform was a newly designed binary sequence that provides a fairly dense array of harmonics from the 0.25 Hz fundamental to at least 10 Hz. An innovative long baseline (LBL) acoustic system was used to track the transmitter during operation, and standard LBL navigation was used to position all the receivers, providing a position accuracy of about 3 m for both receivers and transmitter. Coupled with compass/tiltmeters mounted on all seafloor instruments, these systems should provide better constraints on the CSEM geometry than normally obtained in commercial surveys. The project was divided into four phases. The first phase collected data to provide 1D and 2D responses over the structurally simple southern part of the gas accumulation, the second collected 3D data over the more structurallydifferentiated northern part, the third was targeted at a deeper structure, and the fourth collected data in ways specifically designed to look at shallow resistivity in an area thought to have shallow gas and hydrate. A total of 144 receiver deployments were made, with a total of 12 days CSEM transmitter tow. All receivers were recovered, with data. Initial inspection of the 700 channels of data collected shows only 4 deemed too noisy to be useful. Cursory inspection of the data shows that there is a clear CSEM signal associated with the reservoir on both electric and magnetic field channels at source receiver ranges of 1 to 6 km. Noise floors on the electric field (10 15 V/Am 2 ) and magnetic field (10 18 T/Am) are good, although apparently limited by fields associated with tidal currents, rather than instrument noise. This may limit the quality of the MT data we can process from these records, although the fairly long deployment times should compensate. No health, safety, or environmental incidents occurred during the experiment. 2

3 INTRODUCTION The objective of this research is to carry out marine controlled source electromagnetic (CSEM) soundings and marine magnetotelluric (MT) soundings over the Scarborough gas field in order to demonstrate the effectiveness of marine EM methods for hydrocarbon exploration and to drive the future development of the method. Specifically, we aimed to: To collect a calibration data set over a known structure. Scarborough has excellent 3D seismic coverage, in addition to five exploration and appraisal wells. Such a data set will provide a test-bed for CSEM inversion methods. To collect a data set suitable for joint MT/CSEM/Seismic interpretation. To indentify how shallow confounding resistors can be incorporated into the interpretation. Scarborough has regions of shallow gas and inferred gas hydrate above the reservoir, as well as a resistive layer in the section above the reservoir. To demonstrate how the background resistivity can be understood from MT data. To investigate how well CSEM data can differentiate the various qualities of the reservoir. Other priorities of the research include: Understanding how to optimize CSEM acquisition geometries and source signatures. Examining sources of noise and the repeatability of CSEM data collection. Extending acquisition technologies by including MOBILIZATION * Long baseline receivers that measure electric field gradients very precisely. * Inclusions of vertical electric and horizontal magnetic data into CSEM. * Towing fixed-offset 3-axis electric field receivers for all transmitter tows. * Using a dedicated long baseline acoustic navigation system and installing compass/tiltmeters on all receiver instruments. The research was carried out using the Research Vessel Roger Revelle from 20 May to 23 June The Revelle, operated by Scripps under charter agreement with the U.S. Navy, is only 18 years old, is 273 long and 52.5 wide, with a 3,512 long ton displacement. It berths 37 scientists, has an endurance of 52 days, and can cruise at a speed of 12 knots. There is 4,000 square feet of laboratory space and a working deck space of 4,070 square feet. It is equipped with P-code GPS navigation, dynamic positioning, and the standard suite of oceanographic instrumentation. It has an A-frame and winches suitable or use with the deep-tow transmitter. Scripps supplied the following equipment for the project: 40 SIO Mark III seafloor electric field recorders, with two axes of horizontal B and three axes of E recording. 10 SIO Mark II seafloor electric field recorders, with two axes of horizontal B and two axes of horizontal E recording. 2 SUESI EM transmitters, with a peak current capability of about 500 amps and GPS controlled waveform, with 2 topside power units. The transmission antennas on board were: * A primary 500 A antenna of 250 m length. * A backup 500 A antenna of 100 m length. * Two 200 A hydrate/shallow gas antennas of 50 m length. 2 Vulcan towed 3-axis electric field recorders with pitch/roll/heading/depth sensors. 3

4 2 Long-wire EM (LEM) receivers configured for gradient field measurement, and LEM launching system. Anchors for 160 receiver deployments and 7 LEM deployments. All necessary topside support equipment. Mobilization commenced 17th May at Victoria Quay, Fremantle, Western Australia. We had two 20 shipping containers of concrete anchors, and three 40 containers of equipment, supplemented by two 1000 lb crates of air freight. The vessel also shipped a 20 container of ship s equipment from Scripps, into which we loaded about 30 spare anchors. Quasar Federal Systems also sent two prototype seafloor instruments and support equipment by air freight, and included a stack of 8 anchors in one of our shipping containers. The original schedule for the Revelle had us departing Fremantle on the 26th May and returning Fremantle on the 22nd June. However, in late April UNOLS made the decision to cancel the two cruises following our project, to be carried out off eastern Africa, because of the increasing offshore extent of pirate attacks in that area. We were re-scheduled to come into Darwin to unload, and then for the ship to sail to Brisbane for maintenance and further work in the Pacific. During our cruise the Scripps vessel R.V. Melville lost a propeller off Taiwan, and so on June 20th the Revelle was re-sheduled a second time in order to pick up Melville s ONR funded projects off SE Asia, but leaving our Darwin port call standing. EXPERIMENTAL DESIGN The plan for this experiment was to carry out instruments deployments in three phases, with an optional fourth phase as time permitted. This plan was executed almost exactly as proposed. PHASE 1. In the first phase of the experiment we deployed 50 seafloor receivers along with 2 LEM receivers (at sites 7 and 30) in two lines across the large, quasi-1d part of the reservoir (sites 1 52 in Figure 1). Site spacing on line 1 was 2000, 1000, and 500 m, with the densest site spacing across the southeast edge of the reservoir. A significant amount of data was collected off-target. Transmitter run-in and run-outs were about 5 km. Line 1 was towed a second time (as Line 2) to test repeatability, although in practice the first tow of line 1 served as a shakedown for the equipment and operation. PHASE 2. In the second phase we recovered all Phase 1 instruments and redeployed 51 of them in a grid pattern over the more 3D part of the reservoir structure. Sites 51 and 52 were redeployments from Phase 1 (using the same instruments), and sites 95 and 98 were LEMs. Receiver spacing along lines was 2 km and line spacing was km. We transmitted along all receiver lines, as well as along the mid-points between the lines and on 3 crossline tows. This phase was designed to collect a 3D data set, as well as examine the utility of off-line tows. Line 8, with the two LEMs, was designed to provide off-target control for Phase 3. PHASE 3. For Phase 3 we recovered 23 seafloor receivers and the two LEMs and redeployed these to the north-west over a deeper structure with a site spacing of 1 km over the structure and 2 km for the outer three instruments. Sites 108 and 113 were LEMs, which are specifically designed for this type of target. We redeployed an instrument at site 39 (Phase 1) and also redeployed instruments on the two LEM sites (95 and 98). We towed Line 12 twice (once in each direction, the second time as line 13) and carried out two off-line tows. PHASE 4. The plan was to finish the survey if time permitted by deploying 25 seafloor receivers at a site spacing of 500 m over an area of shallow gas and inferred gas hydrate. It became clear that we would not have the two days available to carry 4

5 19 35 S S S S S 20 S WA-365-P All sites and lines WA-1-R line 14 Line 3 line 12b line 12, 13 line 12a 101 line 8 line 7a line 7 line 6a line 6 line 4a line 4 Line 11 8 line 5a line 5 Line Line 9 Temporary Deployment of Seafloor EM recorder EM Transmitter Tow Line Well WA-346-P WA-383-P Line 1, E E E E E E E Figure 1. Survey layout as proposed. All sites were collected as proposed, except the Phase 4 sites were condensed to a 16-site line slightly to the east. All transmitter tow lines were carried out as proposed. this out as proposed, and so carried out a smaller survey using 16 instruments. For this survey we swapped the 250 m primary SUESI antenna for a 50 m hydrate antenna, and towed two 3-axis receivers at fixed offsets of 250 and 500 m behind SUESI. SURVEY PARAMETERS All transmissions were carried out at a flying height of 50 m using a broadband binary waveform (below) with a 0.25 Hz fundamental frequency and usable bandwidth up to 20 Hz, except the hydrate line which used a 0.5 Hz fundamental. Transmitter waveform. It is becoming clear from experience and modeling that CSEM surveys should have a broad band of frequency content. The original proposal was to tow transmitter lines several times with varying waveforms, but since the proposal was written we had developed and tested a fairly broadband binary waveform which is also compact (that is, the repetition period is relatively short, allowing high temporal resolution during processing). 5

6 1 Timeseries of Waveform Freq. B amp S amp B amp = SIO waveform S amp = square wave Amplitude Phase (degrees) Figure 2. The transmitter waveform used for this project Time (s) Analytical amplitude of waveform 10 0 Frequency (Hz) 10 1 Analytical phase of waveform 10 0 Frequency (Hz) 10 1 One interesting feature of this waveform is that the second and fourth harmonics are higher amplitude than the fundamental, while at the same time being more than half the amplitude of the fundamental of an equivalent square wave. This allows one to set the optimum transmission frequency at one of these harmonics and have the lower frequency fundamental and higher harmonics bracket the optimal frequency. A total of thirteen harmonics exceed 10% of the peak current amplitude, spanning nearly two decades of frequency, compared with six harmonics spanning a little over a decade for a square wave. Receiver frequency. All receivers recorded at 62.5 Hz, except the hydrate experiment which was set to 250 Hz. Navigation. All receivers (except LEMs) had recording compass/tiltmeters. Dedicated long baseline (LBL) navigation was carried out on all receivers, reduced using sound velocity data collected from twice-daily XBT casts. Transmitter navigation was carried out using dedicated LBL navigation, called Barracuda (below), augmented by recording depth gauges installed on the far electrode and both Vulcan instruments. INSTRUMENT PERFORMANCE Overall, the equipment for this project worked very well and as intended. There were a few problems getting some of the newer instrument systems working early in the cruise (Barracudas; serial loggers) which resulted in some loss of navigation data, but these were fixed by the beginning of Phase 2, and other new equipment (LEMs and LEM launching system; transmitter antenna termination system; antenna winch) worked well first time, every time. OBEM Receivers We used two types of ocean-bottom EM (OBEM) receivers with 10 m electrode arms for this project, each capable of recording both MT and CSEM data. Ten instruments were older style Mark II instruments restricted to 4 channels, and 6

7 so these collected only horizontal E and horizontal B data. Forty instruments were newer Mark III instruments with 8-channel A/D converters, and so these instruments were also fitted with 1.5 m vertical electric field antenna. In terms of data quality and data format, the two instrument types are equivalent, except for minor differences in the least count and amplifier responses. Standard data loss rates for marine free-vehicle experiments are around 5% of deployments. Long ago we reduced this to a long term average of around 1% for our operations, and on this experiment we recovered instruments from all 144 deployments. Our recovery rate may be improving; on our last three projects we have carried out 276 consecutive deployments with 277 recoveries (on our October 2008 experiment in the Gulf of Mexico we returned with one more instrument than we sailed with, by recovering an instrument previously lost by another group). The other area of active improvement is data recovery. While recovering instruments is a necessary first step, and important from an operational point of view, data recovery is the important parameter from the scientific perspective. We have been working hard to improve the quantity and quality of our data recovery by carrying out extensive precruise testing, careful checkout procedures for actual operations, and screening of all data by running spectrograms on recovered instruments before re-deployment. All instruments recovered data for the full duration of the deployments. Appendix 3 presents a summary of data quality, based on inspection of the spectrograms. Only four channels of data are deemed unusable. Two electric field channels are unusable from the Phase 1 deployments (on Glider and Dingo, sites 25 and 29), probably because of bad electrodes. Electrodes were swapped out before subsequent deployments and both these instruments collected good data on Phase 2. On Phase 4 one instrument had extremely noisy magnetic channels (Goanna, site 125). This instrument collected good data on all three previous deployments, and CSEM signal from the tows suggests that the sensor cables were correctly connected. The noise is consistent with instrument motion, which is a persistent issue from this experiment (see below), but adjacent instruments performed well. The instrument may have landed on unstable seafloor (either a rock or unconsolidated ooze). Sediment samples returned on two instruments shows the seafloor to be fairly competent sandy clay, at least in the northern half of the survey area. One instrument (Croc, site 10), with a data logger board recently returned as repaired from the manufacturer, had bursts of bad data lasting 36 minutes occurring 5 times per day. This is clearly a fault with the digital electronics, and may be recoverable. Meanwhile, 87% of the data from this instrument is usable. It was not re-deployed, and since we budgeted for one lost instrument during Phase 1, this had no impact on our experiment. B E Figure 3. Two examples of spectrograms from Phase 2, Quokka (site 54) on the left, with fairly good data, and Roo (site 89) on the right, with noise from tidal currents. Much of the rest of the data is affected to some extent or other with noise from tidal currents, either as a result of 7

8 instrument motion or induced electric and magnetic fields (or both). The noise varies from very minor to severe see Figure 3 for examples. The magnetometers are most sensitive to instrument motion, but some electric field channels also exhibit noise from tidal currents, suggesting that to some extend the noise is electric and magnetic fields caused by water motion from turbulence and/or internal waves, an interpretation supported by the LEM data (below). The noise is variable both spatially and temporally, and was most severe during Phase 1 on sites These sites are all in a slight gully, which is clearly either perturbing or concentrating bottom currents. CSEM noise floors are likely to be proportional to this noise, and the MT data from Phase 1 may be marginal for some sites. Only a few channels were severely affected in Phases 2 4, with the exception of Goanna, mentioned above. CSEM Transmitters Because the transmitter is such a critical part of CSEM operations, we took two complete and nearly identical systems on this project. The Scripps transmitter (SUESI), antenna, and topside power supplies are nominally rated to 500 amps, and have indeed been operated at this current in the past, but for this project we aimed not for the largest possible emission current, but the largest dipole moment (which is really what counts). To this end we made a new 250 m transmitter antenna with a new termination scheme and modular design. The longer antenna meant that the maximum emission current was reduced to 300 A, since the higher antenna resistance both reduced the current produced from the available voltage and meant that our transformer ratios were no longer optimal. This still provides a fairly large dipole moment (75 kam), and also allows the topside power supply and SUESI to operate well within their power ratings, making them a lot cooler and more reliable. Our last antenna flooded as a result of a failed termination, and since the conductor is aluminium, this ruins the antenna quickly. Our new termination scheme uses a water-blocking design and a separate, quick-connect near-end strain relief and connector system, allowing rapid replacement if we had any more problems, and also simplifying electrode replacement. The new system worked well during 12 days of transmitter towing, and no replacement or repair was needed. Electrodes were 100 lengths of 3/4" soft copper pipe. Figure 4 shows an actual measurement of the transmitted waveform Current, amps Time, seconds Figure 4. SUESI current output, as measured by the onboard current meter. The rounded corners after switching are a result of the capacitance built into the system to absorb the back EMF of the antenna inductance. We deployed our most recently built underwater unit (SUESI #2) at the beginning of Phase 1 in order to get some run time on the new instrument. However, the downgoing serial link over the towing cable was somewhat noisy, and we seemed to have some problems with the communication link to the Benthos acoustic navigation unit mounted on SUESI, so we switched to SUESI #1, our established transmitter with over 200 hours of operation logged. The acoustic navigation needed more work (see below), but otherwise SUESI #1 performed flawlessly for all of Phase 1 and Phase 2. However, at the very start of Phase 3 this unit failed to transmit any current at all, even though we could see voltage on the output being switched correctly. This was, and remains, extremely puzzling, but meanwhile we swapped SUESI #2 back in, but used the Benthos unit from #1. SUESI #2 performed perfectly apart from some glitches in the downgoing telemetry, which were, at most, a minor nuisance. 8

9 The Phase 2 tow was 6 days 10 hours of continuous operation, and we kept the transmitter running a 300 A during the turns. The total tow time for the four phases was about 12 days. The maximum temperature we saw on either the topside power supply or the underwater units was 37 C. We could have transmitted about 350 A if we ran the topside power unit at maximum, but reliability was preferred over 17% more current, and apart from the mysterious failure of SUESI #1, reliability was what we obtained. SUESI continuously telemeters altitude, depth, pitch, roll, heading, RMS transmitter current, sound velocity, water conductivity, water temperature, and 10 internal temperatures to the ship. The 400 Hz, 2,000 VAC transmitted down the tow cable and used to power SUESI is phase locked to GPS time, allowing SUESI and the transmitted waveform to be phase locked and accurate. The transmitter waveform is started on the minute mark, and has a fixed 1.5 ms offset associated with timing delays in the tow cable. SUESI sends minute markers up the cable, allowing this offset to be monitored and tagged by our GPS clock throughout transmission, to ensure that SUESI time remains locked to GPS. Scarborough CSEM Survey m Scale 1: Figure 5. Navigated positions for all instruments plotted on shaded relief bathymetry. Navigation Systems Navigation is probably the limiting factor in marine CSEM data quality. Short baseline (SBL) acoustic navigation can work well when installed permanently on survey vessels and kept well calibrated, but even then errors of up to 9

10 CRIPPS INSTITUTION OF OCE ANOGRAPHY UCSD 50 m for transmitter positions in commercial data are estimated by some. Our experience with temporary installations of SBL systems on research vessels has been disappointing, with the systems sometimes failing completely. As part of the research carried out in this project, we implemented long baseline (LBL) acoustic navigation surveys for both receivers and transmitter. All receiver instruments contain a LBL navigation and release unit manufactured by Scripps but built along industry lines. On board ship we have a Benthos DS7000 acoustic ranging system operated under computer control, which allows us to merge ship s position with digital ranges on the instruments. Time was allocated during the project to carry out dedicated navigation of deployed receivers. Marquardt nonlinear parameter estimation was used to solve for receiver positions, using sound velocity profiles obtained from twice daily expendable bathy-thermography casts and the Valeport sound velocity sensor on SUESI. Instrument positions obtained thus are accurate to about 3 m (Figure 5). Appendix 2 provides navigated receiver positions for all deployments. Drift on deployment was less than 160 m in all cases and averaged less than 100 m (Figure 12). S board Barracuda; Listen 12.5, Reply 9.0 Ship s Benthos; Ping all, Listen all Port Barracuda; Listen 12.5, Reply 8.5 SUESI relay; Listen 12.0, reply 14.5 SUESI Benthos; Ping all, Listen all Tail buoy; depth, Listen 9.0, Reply 14.5 Vulcan; depth, pitch, roll, heading SUESI; depth, current, temperature, pitch, roll, heading altitude, sound velocity, conductivity OBEM; Listen , Reply 12.0 Figure 6. Diagram showing Barracuda navigation system in relation to SUESI and Vulcan. All OBEM instruments carry an external recording compass/tiltmeter instrument which is mounted as far as possible from the perturbing effects of the magnetometer coils. Two compasses failed to record correctly one from a bug in the compiler which cannot be fixed but we plan to implement a work-around in the future, and one from operator error. One of the most important, and challenging, navigation system was our Barracuda towed LBL system used to obtain SUESI locations (Figure 6). The scheme is to tow two acoustic transponders on the surface, behind the ship and kept apart by paravanes. The Barracuda transponders have GPS receivers and 900 MHz radio modems, which telemeter positions back to the ship in real time. The Barracudas fly m apart. Mounted on SUESI, and integrated into its serial I/O stream, is another Benthos DS7000 system, packaged by Scripps in a pressure case for sub-sea use. SUESI interrogates the Barracudas at approximately 6 second intervals, and telemeters the ranges to the ship along with all the other information. After much struggling we managed to make this system work towards the end of our Gulf of Mexico cruise in October Our problems were associated with electrical and mechanical failures of the paravanes and GPS/modems, all of which had been addressed between cruises and tested, and so had some confidence that the system would work here. However, we had very little success ranging on the Barracuda floats during Phase 1 line 1. We realized that we were operating at an acoustic frequency that was exactly a harmonic of the 400 Hz power that SUESI operates from, and which generates electromagnetic and acoustic noise. We changed the frequency that the Barracudas transpond on by 0.5 khz, and started getting good ranges. The system then worked for the rest of the experiment, including the re-run 10

11 Figure 7. Positions of the two Barracuda transponders during the Phase 2 transmitter tow. of line 1 (line 2). Figure 7 shows the positions of the Barracudas during Phase 2. Instantaneous transmitter positions estimated from Barracuda ranges and wire length, again using Marquardt inversion with water sound velocity structure, have a standard deviation of about 10 m. After averaging we expect positions to be accurate to about 3 m, similar to the receivers. Figure 8 shows the navigated transmitter positions for Phase 2. Another part of the navigation system is a recording depth gauge/tail-end relay transponder mounted on the far electrode of SUESI s antenna. This system had worked flawlessly in the GoM last year as well as a cruise off California. Unfortunately, when we built a backup of this system we introduced a wiring error, which we discovered during a test of the system before starting Phase 1. By the time we solved the problem, the scientist and engineer working on this (Constable and Souders) were overly tired, and we infer that they failed to start the correctly re-wired recording systems when they were deployed for Phase 1, resulting in a loss of data. The units were thoroughly tested between Phase 1 and Phase 2, and deployed successfully for both Phase 2 and Phase 3. SUESI s antenna and the towed Vulcan system (see below) were configured exactly as for Phase 1, and so we are confident that flying attitude will be the same for all surveys. Figure 9 shows an example of the depth data from SUESI, the tail transponder, and Vulcan for one Phase 2 line. All other Phase two lines were similar to this, indicating that the flying attitude is indeed stable. The antenna is about 30 m below SUESI, a 6.5 droop, although the near electrode is probably below SUESI by 5 10 m, which would lower the angle. Rather than try to trim this out during this experiment, we kept everything consistent. One other unfortunate operator error occurred during Phase 4, in which the secondary tail-end transponder was 11

12 Figure 8. Screenshot of Fledermaus navigation software, showing all LBL navigated instrument positions and Phase 2 SUESI Barracuda navigated positions. Vertical exaggeration is 50 times. Figure 9. Depths of SUESI (green), the tail-end transponder (blue), Vulcan (red), and the seafloor (black) for one line of Phase 2. mounted on the antenna, rather than the one that had been prepped and started for deployment, again resulting in loss of navigation data for this tow. Vulcan Another new piece of equipment is the Vulcan system developed for fixed-offset recording behind the transmitter antenna. This instrument records three orthogonal axes of electric field data as well as pitch, roll, heading, and depth. Although developed mainly for hydrate, and really too close to the 250 m antenna to be measuring dipolar fields, we decided to tow a Vulcan 250 m behind the antenna for all surveys. If nothing else, it will provide a calibration of the transmitter output. However, for the Phase 4 hydrate tow, two Vulcans were flown at 250 m and 500 m behind the 12

13 50 m transmitter antenna, and should provide useful information on near-surface conductivity structure. All Vulcan data are good. With the exception of Phase 1, all pitch, roll, heading, and depth sensors worked correctly. LEM/OBEM Amplitudes, Frequency = 0.25 LEM/OBEM Stacking Error, Frequency = 0.25 Mk III OBEM Mk III LEM Mk III OBEM Mk III LEM Amplitude, V/m/Am Amplitude, V/m/Am Source Receiver Range, km Source Receiver Range, km LEM/OBEM Amplitudes, Frequency = 3.25 LEM/OBEM Stacking Error, Frequency = 3.25 Mk III OBEM Mk III LEM Mk III OBEM Mk III LEM Amplitude, V/m/Am Amplitude, V/m/Am Source Receiver Range, km Source Receiver Range, km Figure 10. CSEM data for two transmission frequencies (left panels) and noise estimates (right panels) for a Phase 1 OBEM with 10 m electrode arms (red) and an adjacent LEM instrument with a 100 m antenna (black). Noise estimates were obtained from the variance in the 15 four-second windows which were averaged to generate the 1-minute data shown here. Broken lines are expected noise floors computed from the amplifier and electrode noise. LEMS Although Scripps has been using long-antenna EM receiver (LEM) instruments for over 25 years, this is the first time such an instrument has been deployed over a hydrocarbon prospect. The LEM should have much lower noise floor, as a result of the increased antenna length, and allows accurate measurements of amplitude and phase gradients. However, comparison of the LEM data and adjacent OBEM data (Figure 10) shows that at 0.25 Hz we are not limited by voltage noise from the amplifiers and electrodes, but environmental electric field noise, probably from fields induced by water currents through the E=V B Lorenz force. At higher frequencies (here 3.25 Hz), electric field noise is lower and the LEM has some of its expected advantage, although it is still limited by the environmental noise rather than sensor noise. Gradients obtained from differencing the two 100 m antenna attached to each LEM should reject some of the environmental noise, and so we are optimistic that these instruments will prove useful. All six LEM deployments collected good data. 13

14 19 35 S S S S S 20 S WA-365-P Line 3 WA-1-R Temporary Deployment of Seafloor EM recorder EM Transmitter Tow Line Well Line 9 WA-346-P WA-383-P Line 1,2 SUMMARY AND CONCLUSIONS With 144 receiver deployments and 12 days of transmitter tow, we have collected a huge data set that will provide many research opportunities. Figure 11 shows just two sites from one tow, on and off reservoir. The reservoir signal is clearly visible in the data. The noise floor, which, as we have seen, is limited by noise from water currents, is still respectable at around V/Am 2 and T/Am, and will allow all the objectives of the project to be met. Importantly, the data are not limited by instrument artifacts, transmitter instability, or navigation errors, which, unfortunately, are sometimes given second priority to maximizing transmitter current. The additional data sets represented by the LEMs and Vulcan will add richness to the project having LEM data has already proved useful in estimating the source of noise on the OBEM instruments, which otherwise might easily be mistaken to be instrumental. Frequency: 0.75 Hz line 12b line 12, 13 line 12a line 8 line 7a line 7 line 6 line 6a line 4a line 4 line 5a line Line 10 Line Electric Field s36 line 14 Amplitude, log10 V/m/Am, T/Am s36 s E E E E E E s E 16 Magnetic Field West Horizontal Range (km), in tow(+), out tow(-) East Figure 11. Data from two instruments deployed on the Phase 1 main line (60 second stacks), showing clear signals associated with the target structure. ACKNOWLEDGEMENTS. First and foremost we would like to thank BHP Billiton Petroleum for funding this project. This is the first fully academically collected CSEM and MT data set over a known hydrocarbon reservoir, and it is likely to drive considerable new research and development in the area of marine EM for many years. We would also like to thank Michael Glinsky and Guimin Liu for their help in the experimental design and logistics. Much kudos goes to the scientists, students, technicians, and engineers of the Scripps Marine EM Laboratory who helped prepare the instrumentation and carry out the field work for what became a hugely successful data collection enterprise. The work of our guests Ian Fraser, Jeff Markel, and Rachel Maier in lending a hand with the deployments and recoveries is appreciated. Reece Foster and Cathy Higgs did a wonderful job of organizing the ASEG tour of our vessel and equipment, and we can t imagine how we could have coped with 80 people showing up just before sailing without their able assistance. And thanks for lunch! Finally, of course, the Captain and crew of the R.V. Roger Revelle did their usual tremendous job in supporting the science for this project. Without their professionalism, skill, and enthusiasm this experiment would not have been possible. And cheers to Bruce, Joan, Liz, and Rose, who got our ship in the right place at the right time with the right permits and insurance. 14

15 Appendix 1: Personnel Steven Constable Kerry Key Arnold Orange Yuguo Li Karen Weitemeyer John Souders Chris Armerding Cambria Colt Jake Perez Jennifer Shelstead Rachel Maier David Myer Brent Wheelock Ian Fraser (BOS) Jeff Markel (QFS) Ben Cohen Matt Durham Professor, Chief Scientist Research Geophysicist, Co-Chief Scientist Associate Researcher Project Scientist Postdoc Engineer Technician Technician Technician Technician Student Student Student Observer Techician Computer Technician Resident Technician Tom Desjardins Paul Mauricio Rob Widdrington Joe Ferris Melissa Turner Steve Lewis Kevin Moran Edmond Warren Pete Piscitello Philip Crisfield Matt Peer Jack Healy Frank Oathout David Cobb William Brown Joe Hawkins Antje Galbraith Eddie Angeles Jake Halvorson Jay Erickson Richard Buck Captain Chief Enginneer Chief Mate Second Mate Third Mate Boatswain OS AB AB AB Assistant Engineer Assistant Engineer Assistant Engineer Oiler Oiler Oiler Electrician Oiler Wiper Senior Cook Cook 15

16 Appendix 2: Navigated Positions for all Instruments Site Longitude, W Latitude, S UTM x UTM y Depth S S S S S S S Q S S S S S S S S S S S S S S S S S S S S S S S S Q S S S S S S S S39p S S S S S S S S

17 S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S

18 S S S Q S39p S95b S98b S S S S S S S S S S S S S S S S S S S S S S S124a S125a S126a S127a S128a S129a S130a S131a S132a S133a S134a S135a S136a S137a S138a S139a

19 0 Scarborough Phase 1 Site drift on deployment Mean drift: 107 +/ 28 m Scarborough Phase 2 & 3 Site drift on deployment Mean drift: 62 +/ 32 m 120 Scarborough Phase 4 Site drift on deployment Mean drift: 148 +/ 13 m North (m) 50 North (m) North (m) East (m) East (m) East (m) Figure 12. Drift vector rose diagrams showing difference between drop locations (0, 0) and navigated positions. 19

20 Appendix 3. Data Quality Summary Site Instrument Ch 1 Ch 2 Ch 3 C 4 Ch 5 Remarks PHASE 1 1 shark All sites exhbit periodic (tidal) noise, emphasized on Ch 1&2 Good 2 taipan Occasional 1 Hz noise spots on 1&2&5 Acceptable 3 quindal Periodic spikes on 1&2, ~9-10/day Marginal 4 koala Occasional 1 Hz noise spots Bad channel(s) 5 currawong Ch 1&2&5 noisier "tidal" bursts, Ch 3 LF noise? Not recorded 6 land1 ~2 Hz noise periods all channels 7 lem1 8 wallaby Periodic noise spikes on 1&2, ~9-10/day, possible on Ch 3 9 camel Strong "tidal" noise 1&2&5, also 3&4 but still good data. Some spikes on 4 near en 10 croc 11 goanna "tidal" noise almost renders this site only fair. One noise area coincides with a tow 12 kooka Strong ~2 hz noise periods on 1&2&5. strumming? Clean between noise periods 13 rosella Strong ~2 hz noise periods on 1&2&5. Clean between noise periods Clean between noise periods. Noise periods may contaminate 14 devil CS tows p y ( ) 15 cass channels. Broad spectrum HF noise on 1, LF noise on 3& Hz noise on 3 (why 3 and not 4?) 16 occie Tidal noise may interfere with CS tows 17 redback Tidal noise and "strumming" may interfere with CS tows on mags and Ez 18 mantis Some tidal noise and "strumming" 19 bullant Some noise spikes on mags 20 yabby 21 land2 Narrow band ~2 Hz noise on E's probbly won't interfere with anlysis, some noise on mags 22 brolga Narrow band ~2 Hz noise on ch 3 probbly won't interfere with anlysis, some noise on mags 23 bunyip 24 tazz Some noise spikes on mags 25 glider Bad ch 4, some 2 Hz noise 26 rabbit Some 2 Hz noise 27 brumby Some spikes Ch 2. Why not ch 1 also? 28 cocky Some noise on 3&4 during third tow Site Instrument Ch 1 Ch 2 Ch 3 C 4 Ch 5 Remarks 29 dingo Bad ch 3, spikes on 1&2 30 corella Good 31 LEM2 Acceptable 32 stingray Ch 1&2 just a bit noisy Marginal 33 echidna Spikes 1&2 (disk writes) Bad channel(s) 34 magpie Ch 5 may be marginal, some sharp noise spikes Not recorded 35 bilby 36 roo 37 potoroo Some noise during the tows 38 lerp Disk writes on 1&2, ch 3 noisy towards end of recording and noise bursts around tows 1&2 39 quokka 40 budgie 41 cuscus Tidal noise 42 mozzie Ch 2 bad 43 emu Ch 4 very noisy 44 ibis Tidal noise on mags 45 numbat Tidal noise on mags, some spikes before tows begin 46 fruitbat Ch 1&2 just a bit noisy at tows 47 quoll 48 marron Noise burst during 5/30 may affect mags 49 skink Noise burst during 5/30 may affect mags 50 possum Noise during tow? Spikes on mags. 51 penguin 52 lorrie Ch 1&2 a bit noisy vulcan tests??????????? 20

21 Site Instrument Ch 1 Ch 2 Ch 3 C 4 Ch 5 Remarks PHASE 2 51 penguin Good 52 lorrie Just a bit of LF noise on 1&2 Acceptable 53 camel Just a bit of LF noise relative to the top sites Marginal 54 quokka 1&2 just a bit l\lf noise Bad channel(s) 55 goanna Not recorded 56 kooka Just a bit of broad band noise on 1&2 57 land 2 Some tidal noise 58 koala 59 yabby Just a bit of broad band noise on 1&2&5 60 magpie Strange noise on 5. Just a bit of broad band noise on 1&2 61 redback 62 corella 63 shark Ch 3 broad band noise may interfere with tow 64 land 1 65 budgie 66 currawong Ch 1&2 almost good 67 taipan Ch 1&2 barely good 68 glider Ch 1&5 barely good 69 occie 70 bunyip 71 devil Ch 1 almost good 72 bullant Some spikes ch 1&2 73 brolga LF noise ch 3. Ch 1&2 barely good 74 possum Spikes ch 1&2 75 rosella 76 mantis a bit more noise than usual on the mags 77 quidal 78 cass disk writes on mags, and mags just barely good 79 ibis Tidal noise on mags probably won't interefere 80 dingo Periods of obvious on ch 3noise don't interfere with the tows. Ch 1 noisy 81 rabbit 82 cocky Ch 4 noisy first third or so of the record. Ch 2 barely good 83 lerp Some noise on 1&2 probaably won't interfere Site Instrument Ch 1 Ch 2 Ch 3 C 4 Ch 5 Remarks 84 fruitbat Some noise on 1&2 probaably won't interfere 85 wallaby Ch 3 noisy first third or so of the record. 86 brumby 87 echidna Disk writes on mags. Some motion noise on mags. 88 emu Ch 4 noisy first third or so of the record. Disk writes and motion noise on mags 89 roo Lots of motion noise. Possibly marginal, but not instrument related 90 potoroo Motion noise on mags and ch 5 Good 91 tazz Acceptable 92 bilby LF noise ch 4 Marginal 93 mozzie Ch 1&5 motion noise Bad channel(s) 94 stingray Not recorded 95 lem 1 96 cuscus 97 numbat Motion noise mags & ch 5 98 lem 2 99 skink Motion noise mags & ch marron Motion noise mags 101 quoll Motion noise mags Vulcan?????????? 21

22 Site Instrument Ch 1 Ch 2 Ch 3 C 4 Ch 5 Remarks Phase 3 39 shark Ch 3 noise greatest first third or so of record, problem at start 95 brolga Ch 6, no signal? Ch 1&2 look noisy, possible problem? Scaling? 98 ibis Ch 1&2 barely good, some motion noise 102 corella Good 103 land 2 Acceptable 104 camel Marginal 105 koala Bad channel(s) 106 kooka Not recorded 107 lorrie some noise at start ch 1-4. An exceptionally clean site! 108 lem land quokka noise bursts at start of record 111 glder a few noise bursts at start of record 112 occie a few noise bursts at start of record. Ch 5 bad last half or so 113 lem budgie Probably some good data ch 1, but there's a problem 115 currawong Some motion noise Ch 5. Ch 1&2 some noise at start of record 116 taipan Some noise at start of record. Some motion noise. 117 redback huge noise burst near start to record 118 penguin huge noise burst near start of record, very similar to redback above 119 magpie Ch 5 noise spikes not correlated with disk writes, tow data may be OK 120 devil Ch 4 LF noise, esp at start 121 yabby Ch 1&2 barely good 122 bunyip 123 goanna Ch 6? Vulcan?????????? Site Instrument Ch 1 Ch 2 Ch 3 C 4 Ch 5 Remarks Phase quokka Mags a bit noisier than electrics Good 125 goanna Mags NG Acceptable 126 bunyip Marginal 127 occie Bad channel(s) 128 kooka Not recorded 129 mantis 130 roo 131 skink 132 numbat 133 koala 134 taipan 135 penguin Narrow band noise ~2 Hz, mags motion noise 136 corella Ch 4 a problem? 137 marron 138 stingray Ch 1&2 noisy 139 rabbit Ch 1&2 almost poor vulcan 1???????? vulcan 2???????? 22