Special Section: Remote sensing36 THE LEADING EDGE January 2017

Short- and long-term ground deformation due to cyclic steam stimulation in Alberta, Canada, measured with interferometric radar

AbstractShort- and long-term ground deformation at the cyclic steam

stimulation (CSS) enhanced oil recovery sites in the Cold Lake region of Alberta, Canada, was measured with RADARSAT-2 differential interferometric synthetic aperture radar (DInSAR) during 2011–2016. The interferometric processing of 84 high-resolution spotlight images revealed extensive short-term ground uplift and subsidence greater than 15 cm over the 24-day satellite revisit cycle. Time-series analysis based on the multidimensional small baseline subset (MSBAS) methodology demonstrated that while the predominant deformation signal is cyclical in nature, a long-term subsidence and uplift with maximum deformation rates greater than 5 cm/year is also present. These measurements suggest that CSS injection and production operations result in the long-term semipermanent deformation in the area of reservoir operations, and that the response to injection and extraction can differ from the previous cycles.

IntroductionEnhanced oil recovery (EOR) techniques are employed to

increase the amount of extremely heavy crude oil or bitumen extracted from subsurface reservoirs. The steam-assisted gravity drainage (SAGD) EOR technique consists of a pair of horizontal wells, one drilled a few meters above another (Butler, 1994; Pearse et al., 2014). High-pressure steam is continuously injected into the upper well to heat the oil and reduce its viscosity. The heated oil drains to the bottom well from which it is pumped out. The cyclic steam stimulation (CSS) or “huff-and-puff” EOR technique requires only one well, operated in three different phases (Alvarez and Han, 2013). The injection phase consists of injection of high-temperature (i.e., >200 degrees) and high-pressure (i.e., higher-than-formation fracture pressure) steam. High-temperature steam reduces the viscosity of the bitumen. High-pressure steam dilates the reservoir, which induces fracturing, resulting in in-creased permeability. The soaking phase allows the heat, which is mainly conveyed by the steam, to seep into the formation due to the combined effects of conduction and convection. The in-creased mobility of bitumen accelerates this process because of enhanced convection. During the production phase, fluid flow in the well is reversed, and the heated bitumen, along with the condensed steam, is extracted from the reservoir. The production phase can last from a few months to a few years (Vittoratos et al., 1990). Injection, soaking, and production phases comprise one cycle. A new cycle is initiated after the production rate from the well becomes too low as the reservoir is cooling down. Cycles are repeated until the reservoir operation is no longer economical.

SAGD is used in thin reservoirs with high vertical permeabil-ity, and CSS is used in moderately thick reservoirs with poorer vertical permeability. CSS can also be employed in less “clean”

Sergey V. Samsonov1

reservoirs, where bitumen-bearing sandstones are interbedded with thin layers of low-permeability shale, but this requires the reservoir to be “dryer” (i.e., less water content). Thin shale baffles, which would prevent efficient steam circulation in SAGD opera-tion, are broken during CSS operation because of higher injection pressure (Richardson et al., 1978; Jiang et al., 2010). Consequently, higher confining pressure typically is required to contain the reservoir fluid in the operation zone.

The Alberta, Canada, oil sands contain the third largest oil reserves in the world, after Venezuela and Saudi Arabia. As of 2015, Alberta’s proven reserves were 166 billion barrels, produced at a rate of 3 million barrels per day (Alberta Energy Regulator, 2016). Crude bitumen in Alberta is contained within three large oil sands areas: the Athabasca, centered on Fort McMurray; Peace River to the west; and Cold Lake to the southeast.

The Cold Lake Oil Sands Area (CLOSA) is known for its rich bitumen deposits (about 34 billion barrels) hosted within the Lower Cretaceous Mannville Group (AER, 2016). The CSS operations have been ongoing in this region since the mid-1970s. The Clear-water Formation, which is buried at roughly 400 m depth in the Cold Lake region, is the most frequently targeted stratigraphic unit for CSS (Walters et al., 2000). The inexpensive wet steam (higher water saturation), instead of a dry steam (less water satura-tion) that is often used in SAGD, is the most common injection fluid in Cold Lake CSS operations. The liquid solvent addition to steam for EOR has been tested in order to enhance in-situ mobility of the bitumen and ultimately the recovery rate (Jiang et al., 2010). In recent years, there have been attempts to utilize SAGD for the recovery of bitumen from the Grand Rapids Formation in the CLOSA (Willmer and Quinn, 2015). The Grand Rapids Forma-tion, which consists of mostly shoreline sandstones, has a high water content that limits the effectiveness of CSS (Barson et al., 2001). The overlying low permeability shales of the Upper Creta-ceous Colorado Group act as a caprock that can hold the steam beneath, making SAGD even more viable (CNRL, 2010).

Other production or pilot projects using techniques similar to CSS include operations targeting the Bluesky Formation in the Peace River region (Mohiddin and Koci, 2007) and Grosmont Formation in the Athabasca region (Yang et al. 2014). Pilot projects in the Athabasca region also test the combination of both cold and warm solvent injection at higher pressures to achieve reduced bitumen viscosity and enhanced formation permeability (Jiang et al., 2010). Outside Canada, CSS is employed in EOR operations in Santa Barbara in Venezuela and Liaohe in China (Alvarez and Han, 2013).

The studied area (Figure 1) is one of the oldest EOR sites in Alberta (Buckles, 1979). Although recovery to date has barely exceeded half of the projected ultimate recovery, the production rate is roughly at its peak, as reported by Imperial Oil (2010).

1Natural Resources Canada. http://dx.doi.org/10.1190/tle36010036.1.

Special Section: Remote sensing January 2017 THE LEADING EDGE 37

Wells in the early stages of CSS operation may have cycles lasting less than three months, but, at this site, wells have longer than usual operation cycles due to the lengthy production history. At some well pads, diluents are injected along with steam during the first couple of cycles to further reduce in-situ viscosity and increase ultimate recovery.

Differential interferometric synthetic aperture radar (DInSAR) is used for measuring ground deformation with high spatial resolu-tion over large areas by utilizing repeatedly acquired airborne or spaceborne synthetic aperture radar (SAR) data (Massonnet and Feigl, 1998). DInSAR is routinely used for mapping ground de-formation due to earthquakes and volcanic eruptions (e.g., Hayes et al., 2014; Sigmundsson et al., 2015) and anthropogenic deforma-tion due to mining and fluids extraction and injection (e.g., Sam-sonov et al., 2014; Samsonov et al., 2015a,b). Stancliffe and van der Kooij (2001), Pearse et al. (2014), Samsonov, (2014), Samsonov et al. (2015a), and Shen et al. (2015) have studied ground deforma-tion due to EOR operations in Alberta’s oil sands. Pearse et al. (2014) demonstrated that at SAGD sites in the Athabasca Oil Sands Area ground uplifts with steady rates of 1–4 cm/year due to pressure buildup and thermal expansion, and that uplift rates at these sites correlate with the steam injection rates. Very large ground deformation, sometimes exceeding 30 cm over the 24-day

satellite revisit cycle, was observed previously by Samsonov (2014) and Samsonov et al. (2015a) at CSS sites in the CLOSA. Stancliffe and van der Kooij (2001) suggested that, after the CSS production phase, the ground returns to its initial condition.

While individual interferograms that were used in the previous studies captured separate cycles of injection and production phases very well, the complete multiyear time series of ground deformation at a CSS sites has never been shown, likely due to the lack of high-quality SAR data and the complexity of processing. Here, we address this gap by presenting 2011–2016 deformation time series derived from RADARSAT-2 high-resolution spotlight SAR data. We discuss data, processing methodology, and spatial and temporal distribution of the deformation signals. In this study, we aim to record both long- and short-term subsurface deforma-tion. Given the cyclic nature of CSS, it is expected that periodic ups and downs would be observed. Through this study, we observe that, on top of the periodic deformations, long-term deformation with periods much longer than that of injection/production cycles are present. Such hysteresis may result in unexpected subsurface deformation when initiating new cycles. An assessment of sub-surface integrity is complicated further by the fact that the produc-tion histories of the oldest CSS sites date back multiple decades with few deformation records available.

Data and methodologyFor measuring ground deformation at the cyclic steam stimula-

tion enhanced oil recovery sites in the CLOSA, Canada, we processed 84 synthetic aperture radar images from RADARSAT-2 satellite. The data were acquired from 20100422 to 20160321 (in YYYYMMDD format) by the high-resolution spotlight 75 beam with a footprint of 8 km in azimuth and 18 km in range (Figure 1), and with an incidence angle of 26 degrees and an azimuth angle of 347 degrees (i.e., ascending track).

The initial interferometric processing was performed with GAMMA software (Wegmüller and Werner, 1997) in the fol-lowing way. A master image, acquired on 20141010, was selected, and remaining images were coregistered and resampled to the master geometry. All possible interferograms with spatial baselines less than 250 m and temporal baselines less than 600 days were computed by applying 4 × 15 multilooking (i.e., spatial averaging). The topographic phase was computed from the 30 m resolution shuttle radar topography mission (SRTM) digital elevation model (Rabus et al., 2003) and subtracted from interferograms. The differential interferograms were filtered (Goldstein and Werner, 1998), unwrapped (Costantini, 1998), and geocoded. A bilinear trend was computed by robust fitting using GMT (Wessel and Smith, 1998) grdtrend command and removed from each inter-ferogram. This detrending procedure produced better results than the standard orbital correction based on a correlation of observed phase and topographic height, due to the presence of large defor-mation signals observed in largely flat areas.

On average, the quality of interferograms was moderate due to significant decorrelation caused by seasonal land cover changes and vegetation. The areas covered by water bodies were masked out, and the average coherence for each interferogram was computed. For time-series analysis, we selected 287 interfero-grams with the average coherence after filtering above 0.45 (Figure 2). Spatial gaps in selected interferograms were partially filled by interpolation over short distances (8 × 8 pixel window),

Figure 1. A 30 m resolution shuttle radar topography mission (SRTM) digital-elevation model of cyclic steam stimulation enhanced oil recovery sites in Alberta, Canada. The 8 × 18 km footprint of RADARSAT-2 spotlight 75 images is shown in black. Major water bodies are marked.

Figure 2. Spatial and temporal baselines of interferograms used in multidimensional small baseline subset (MSBAS) analysis.

Special Section: Remote sensing38 THE LEADING EDGE January 2017

Figure 4 are shown in Figure 5. The short-term cyclic deformation observed at seven out of nine regions is clearly due to CSS opera-tion. The long-term deformation trend is also observed at these regions. In particular, long-term uplift is observed at P1, P3, P5, P7, and P8, and long-term subsidence is observed at P2, P4, P6, and P9. At P4 and P6, subsidence started in 2012, and at P2 subsidence started in 2013. The amplitude of cyclic deformation seems to increase over time.

DiscussionCyclic steam stimulation involves injection of heated and

pressurized steam into the reservoir that heats oil and fractures reservoir (Alvarez and Han, 2013). The injection of steam causes the dilation of soil matrix creating regions of enhanced permeabil-ity. Consequently, reservoir deformation extends to regions not hydraulically connected to the EOR reservoir, resulting in the observable pressure changes due to poroelastic effects in an oth-erwise static system (Walters et al., 2000).

Injection and extraction changes the pore pressure and stress field in the overlying caprock, therefore producing observable deformation at the surface. Shape and magnitude of surface deformation can be used for indirect measurement of pressure change in the reservoir, for optimizing steaming operations by minimizing steam production, and maximizing the amount of the recovered oil (Shen et al., 2015). The shape and magnitude of ground deformation, if large enough, may help to detect steam migration through fractures, naturally occurring or induced by reservoir stimulation (Del Conte et al., 2015). Surface deforma-tion is a prominent indicator of the location and extent of pres-surized chambers in a reservoir. With the precise deformation measurements on the surface, subsurface strain and stress states can be estimated, to a certain extent, from the pattern of surface deformation. Such estimation provides an indirect means of assessing the integrity of the subsurface structure where direct measurements are difficult/expensive to obtain.

Ground deformation measurement with DInSAR has a number of advantages and disadvantages compared with con-ventional techniques, such as reflectors, in-well or surface tilt-meters, inclinometers, and GPS (Del Conte et al., 2015). While conventional techniques can achieve very high precision at loca-tions where these instruments are installed, DInSAR can provide measurements over a larger area with very high spatial resolution (e.g., Samsonov et al., 2015a,b). For example, RADARSAT-2 spotlight data used in this study covers an 18 × 8 km area with submeter resolution. The precision of each interferogram computed from two SAR images is not particularly high — about 1–2 cm — but when many SAR images are processed together, the preci-sion can be improved significantly. Since the location of fractures or faults is often unknown prior to their activation, the spatially dense DInSAR observations may be the only source of reliable information. Temporal resolution of conventional DInSAR is limited by the satellite revisit period, equal to 24 days in the case of RADARSAT-2. However, multiple data sets can be integrated together using the MSBAS technique to produce temporally dense two-dimensional time series. For example, by integrating ascending and descending data from the same satellite, it is possible to achieve temporal resolution equal to half the satellite

which significantly improved spatial coverage of final results. Selected interferograms were analyzed further with the multi-dimensional small baseline subset (MSBAS) software (Samsonov and d’Oreye, 2012). While the main purpose of this software is computation of two-dimensional, horizontal east-west and vertical, deformation time series from multiple ascending and descending DInSAR sets, the standard line-of-sight deformation analysis can also be performed — if only one DInSAR set is provided. This software can perform interferogram calibration against up to 10 reference regions and apply Tikhonov regulariza-tion of zero, first, or second orders, which is preferable to standard smoothing in case of data that is irregularly spaced in time. Here, we did not perform interferogram calibration since it has already been done by subtracting the bilinear trend, but we applied the first-order Tikhonov regularization for smoothing results and also for compensating the discontinuity observed in 2013 (Figure 2). The line-of-sight linear deformation rate for each pixel was computed by fitting a straight line. This quantity is used for an approximate assessment of the long-term deforma-tion rate.

ResultsTwo sets of highly coherent interferograms are shown in

Figure 3. The first set consists of five consecutive interferograms over the 6 July 2014 to 21 December 2014 period and the second set consists of five consecutive interferograms over the 22 No-vember 2015 to 21 March 2016 period. These interferograms were selected because of particularly high coherence, but the spatial distribution and magnitude of the observed deformation signals is representative of the entire observation period.

The typical deformation sequence starts with greater than 15 cm line-of-sight shortening (i.e., apparent uplift) over an area of 1–3 km in diameter followed by the line-of-sight lengthening (i.e., apparent subsidence) of approximately similar magnitude and shape. Since DInSAR measures a projection of the displace-ment vector on the satellite line-of-sight, it is impossible to distinguish vertical and horizontal components of motion if data acquired in only one orbital configuration is used. However, assuming that the horizontal component is negligibly small, the line-of-site measurements (Figure 3) can be converted to the vertical deformation by dividing by cos(26), where 26 is the SAR incidence angle in degrees. To validate that the horizontal component of motion is negligibly small, one can observe ground deformation from two different viewing angles — for example, ascending and descending. If horizontal motion is negligibly small, than signals observed in interferograms will have similar shape. Multiple uplifting and subsiding regions are observed during the same period, and the largest deformation signals are observed around locations of well pads.

The line-of-sight linear deformation rate map computed with MSBAS is shown in Figure 4. The color scale is saturated at [-5;5] cm/year for improving visualization. Both uplift and subsidence are equally present in this map. While uplifting areas are concentrated on the well pads, subsidence has a broader spatial extent.

The time series of line-of-sight ground deformation computed with MSBAS for nine representative regions marked P1–P9 in

Special Section: Remote sensing January 2017 THE LEADING EDGE 39

Figure 3. Two sets of consecutive unwrapped differential interferograms during (a) 20140706-20141221 and (b) 20151122-20160321 over cyclic steam stimulation enhanced oil recovery sites. Ground uplift is produced by steam injection and subsidence is produced by fluid extraction. Wells are shown as black lines. To improve visualization, the color scale is clipped to [-15;15] cm. Maximum observed deformation slightly exceeds 20 cm.

Figure 4. Line-of-sight linear deformation rate during 20100422–20160321 computed by fitting straight line to deformation time series, produced by MSBAS software. The 5 × 5 pixel regions marked with P1–P9 deformation time series are shown in Figure 5. Wells are shown as black lines.

Special Section: Remote sensing40 THE LEADING EDGE January 2017

revisit period (i.e., 12 days in the case of RADARSAT-2). Previous studies demonstrated that vertical motion due to a deep deforma-tion source is accompanied by horizontal motion of smaller magnitude (Samsonov et al., 2013; Samsonov et al., 2014). Vertical deformation maps computed with MSBAS are complemented by east-west horizontal deformation maps (Samsonov and d’Oreye, 2012). The high deformation gradient observed at CSS sites can be mapped only with high-resolution SAR data, particularly in areas affected by significant temporal decorrelation (Samsonov et al., 2015a).

The interwell interaction, if left uncontrolled, may reduce the effectiveness of a CSS operation (Vittoratos et al., 1990). DInSAR deformation maps may provide invaluable information about interwell interaction. Such data can be used for validation of numerical models that may lack precision due to the variety of inhomogeneous processes in the reservoir.

Ground deformation observed in this study suggests that cycles of CSS operations last approximately one year and that the deformation magnitude of consecutive cycles increases with time. For example, at P1, P5, and P8 (Figure 4), the magnitude of ground deformation is larger in 2015 than in 2014; and in 2014, it is larger than in 2013. These observations do not neces-sarily imply an increased volume of injected steam at particular wells, because a portion of the deformation can also be produced

by injections at nearby wells (i.e., well interaction). Nevertheless, the observed increase in magnitude of deformation suggests larger pressure changes in the steam chamber and/or a weakening of the rock by fracturing and, consequently, higher risk. The Alberta Energy Regulator collects the monthly volume of injected and extracted steam and reservoir fluids from operators for each well. This information seems sufficient for older CSS sites, in-cluding the one being studied in this paper, where periods of CSS cycles are close to one year. However, for CSS sites at an early stage of development where periods are much smaller, weekly or daily measurements would be more valuable. Depend-ing on duration and frequency of injection and production phases, CSS cycles with the same total monthly volumes can produce different ground deformation. The maximum deformation can be produced by a single injection followed by a single extraction, while a smaller deformation can be produced by more frequent injection-extraction cycles.

Long-term deformation was also observed at the studied area. Long-term uplift can be caused, for example, by the shear dilation of reservoir sands, thermal expansion, and net-positive injection (i.e., more is injected than extracted). It is possible that other EOR techniques, such as SAGD or steam flooding, con-tribute to the long-term deformation. Long-term subsidence is somewhat expected in oil extraction areas.

Figure 5. Time series of line-of-sight ground displacements observed at regions P1–P9 in Figure 4. First-order Tikhonov regularization was applied for smoothing.

Special Section: Remote sensing January 2017 THE LEADING EDGE 41

ConclusionsThis study demonstrates that ground deformation measure-

ments of the cyclic steam stimulation operations with DInSAR provide unique information about the duration, magnitude, and extent of steaming operation not available from any other source. The CSS operations in this region cause cyclic ups and downs with magnitudes often larger than 15 cm over the 24-day satellite revisit cycle. Long-term subsidence and uplift with rates greater than 5 cm/year are also observed.

The amplitude of cyclical deformation due to CSS increases over time, which can be caused by the increased pressure in the rock or a weakening of the rock by fracturing, effectively lowering its Poisson’s ratio. Ground deformation monitoring with DInSAR is beneficial for both operators and regulating agencies.

AcknowledgmentsThis work significantly benefited from communications with

Luyi Shen (Alberta Energy Regulator). The RADARSAT-2 data was provided by the Canadian Space Agency. Images were plotted with GMT and GNUPLOT software. This work was supported by the Canadian Space Agency and Natural Resources Canada joint project “The development and delivery of on-demand RA-DARSAT Constellation Mission ground deformation products based on advanced InSAR technology.”

Corresponding author: sergey.samsonov@canada.ca

ReferencesAlberta Energy Regulator, 2016, ST98-2016 Alberta’s energy re-

serves 2015 & supply/demand outlook 2016–2025, http://www1.aer.ca/st98/data/executive_summary/ST98-2016_Executive_Summary.pdf, accessed 5 December 2016.

Alvarez, J., and S. Han, 2013, Current overview of cyclic steam injection process: Journal of Petroleum Science Research, 2, no. 3, 116–127, http://www.jpsr.org/paperInfo.aspx?ID=5755.

Barson, D., S. Bachu, and P. Esslinger, 2001, Flow systems in the Mannville Group in the east-central Athabasca area and implications for steam-assisted gravity drainage (SAGD) op-erations for in situ bitumen production: Bulletin of Canadian Petroleum Geology, 49, no. 3, 376–392, http://dx.doi.org/10.2113/49.3.376.

Buckles, R. S., 1979, Steam stimulation heavy oil recovery at Cold Lake, AB: SPE California Regional Meeting, SPE, paper SPE-7994-MS, http://dx.doi.org/10.2118/7994-MS.

Butler, R. M., 1994, Steam-assisted gravity drainage: Concept, devel-opment, performance and future: Journal of Canadian Petroleum Technology, 33, no. 02, 44–50, http://dx.doi.org/10.2118/94-02-05.

Canadian Natural Resources Limited (CNRL), 2010, Primrose, Wolf Lake, and Burnt Lake Annual Presentation to the ERCB, Janu-ary 26, 2010, https://www.aer.ca/documents/oilsands/insitu-presentations/2010CNRLPAWCSS9140.pdf, accessed 5 December 2016.

Costantini, M., 1998, A novel phase unwrapping method based on network programing: IEEE Transactions on Geoscience and Re-mote Sensing, 36, no. 3, 813–821, http://dx.doi.org/10.1109/ 36.673674.

Del Conte, S., S. Cespa, A. Rucci, and A. Ferretti, 2015, InSAR monitoring in heavy oil operations: SPE Kuwait Oil and Gas

Show and Conference, SPE, paper SPE-175366-MS, http://dx.doi.org/10.2118/175366-MS.

Goldstein, R., and C. Werner, 1998, Radar interferogram filter-ing for geophysical applications: Geophysical Research Let-ters, 25, no. 21, 4035–4038, http://dx.doi.org/10.1029/ 1998GL900033.

Hayes, G. P., M. W. Herman, W. D. Barnhart, K. P. Furlong, S. Riquelme, H. M. Benz, E. Bergman, S. Barrientos, P. S. Earle, and S. Samsonov, 2014, Continuing megathrust earthquake po-tential in Chile after the 2014 Iquique earthquake: Nature, 512, no. 7514, 295–298, http://dx.doi.org/10.1038/nature13677.

Imperial Oil, 2010, Cold Lake Approvals 8558 and 4510 Annual Performance Review 2010.

Jiang, Q., J. Yuan, J. Russel-Houston, B. Thornton, and A. Squires, 2010, Evaluation of recovery technologies for the grosmont car-bonate reservoirs: Journal of Canadian Petroleum Technology, 49, no. 05, 56–64, http://dx.doi.org/10.2118/137779-PA.

Massonnet, D., and K. L. Feigl, 1998, Radar interferometry and its application to changes in the Earth’s surface: Reviews of Geo-physics, 36, 441–500, http://dx.doi.org/10.1029/97RG03139.

Mohiddin, J. G., and P. F. Koci, 2007, Peace River Carmon Creek Project C—Optimization of cyclic steam stimulation through experimental design: SPE Annual Technical Conference and Ex-hibition, SPE, http://dx.doi.org/10.2118/109826-MS.

Pearse, J., V. Singhroy, S. Samsonov, and J. Li, 2014, Anomalous surface heave induced by enhanced oil recovery in northern Al-berta: InSAR observations and numerical modeling: Journal of Geophysical Research: Solid Earth, 119, no. 8, 6630–6649, http://dx.doi.org/10.1002/2013JB010885.

Rabus, B., M. Eineder, A. Roth, and R. Bamler, 2003, The shuttle radar topography mission-a new class of digital elevation models acquired by spaceborne radar: ISPRS Journal of Photogramme-try and Remote Sensing, 57, no. 4, 241–262, http://dx.doi.org/10.1016/S0924-2716(02)00124-7.

Richardson, J. G., D. G. Harris, R. H. Rossen, and G. Van Hee, 1978, The effect of small, discontinuous shales on oil recovery: Journal of Petroleum Technology, 30, no. 11, 1531–1537, http://dx.doi.org/10.2118/6700-PA.

Samsonov, S., and N. d’Oreye, 2012, Multidimensional time-series analysis of ground deformation from multiple InSAR data sets applied to Virunga Volcanic Province: Geophysical Journal In-ternational, 191, no. 3, 1095–1108.

Samsonov, S. V., 2014, Ground deformation observed near Cold Lake, Alberta, by RADARSAT-2 DInSAR during 2008–2013, Geological Survey of Canada, Open File 7527, http://geoscan.nrcan.gc.ca/starweb/geoscan/servlet.starweb?path=geoscan/fulle.web&search1=R=293701, accessed 5 December 2016.

Samsonov, S., P. J. Gonzalez, K. Tiampo, and N. D’Oreye, 2013, Spatio-temporal analysis of ground deformation occurring near Rice Lake, Saskatchewan, and observed by Radarsat-2 DInSAR during 2008-2011: Canadian Journal of Remote Sensing, 39, no. 1, 27–33, http://dx.doi.org/10.5589/m13-005.

Samsonov, S. V., P. J. González, K. F. Tiampo, and N. d’Oreye, 2014, Modeling of fast ground subsidence observed in southern Saskatchewan (Canada) during 2008–2011: Natural Hazards and Earth System Sciences, 14, no. 2, 247–257, http://dx.doi.org/10.5194/nhess-14-247-2014.

Samsonov, S. V., M. Czarnogorska, and F. Charbonneau, 2015a, Selecting optimal RADARSAT constellation mission beams for monitoring ground deformation in Alberta’s oil sands:

Special Section: Remote sensing42 THE LEADING EDGE January 2017

Canadian Journal of Remote Sensing, 41, no. 5, 390–400, http://dx.doi.org/10.1080/07038992.2015.1104632.

Samsonov, S., M. Czarnogorska, and D. White, 2015b, Satellite inter-ferometry for high-precision detection of ground deformation at a carbon dioxide storage site: International Journal of Greenhouse Gas Control, 42, 188–199, http://dx.doi.org/10.1016/j.ijggc.2015.07.034.

Shen, L., D. R. Schmidt, R. H. Dokht, and S. V Samsonov, 2015, Geomechanical modeling of cyclic steaming induced surface de-formation: SPE Canada Heavy Oil Conference, http://dx.doi.org/10.2118/174415-MS.

Sigmundsson, F., A. Hooper, S. Hreinsdóttir, K. S. Vogfjörd, B. G. Ófeigsson, E. R. Heimisson, S. Dumont, et al., 2015, Segmented lateral dyke growth in a rifting event at Bárðarbunga volcanic system, Iceland: Nature, 517, no. 7533, 191–195, http://dx.doi.org/10.1038/nature14111.

Stancliffe, R. P. W., and M. W. A. van der Kooij, 2001, The use of satellite-based radar interferometry to monitor production activ-ity at the Cold Lake Heavy Oil Field, Alberta, Canada: AAPG Bulletin, 85, no. 5, 781–793, http://dx.doi.org/10.1306/8626C9FB-173B-11D7-8645000102C1865D.

Vittoratos, E. S., C. I. Beattie, and G. R. Scott, 1990, Cold lake cy-clic steam stimulation: A multiwell process: SPE, 5, no. 1, 239–245, http://dx.doi.org/10.2118/17422-PA.

Walters, D. A., A. Settari, and P. R. Kry, 2000, Poroelastic effects of cyclic steam stimulation in the Cold Lake Reservoir: SPE, 1–10, http://dx.doi.org/10.2118/62590-MS.

Wegmüller, U., and C. Werner, 1997, GAMMA SAR processor and interferometry software: 3rd ERS symposium on space at the ser-vice of our environment, European Space Agency.

Wessel, P., and W. Smith, 1998, New, improved version of the ge-neric mapping tools released: AGU Eos, 79, no. 47, 579, http://dx.doi.org/10.1029/98EO00426.

Willmer, T. and G. Quinn, 2015. Characterization of the Albian Lower Grand Rapids Formation, Cold Lake, Alberta: CSEG Recorder, 40, no. 1, 26–36.

Yang, D., M. Hosseininejad Mohebati, S. Brand, and C. Bennett, 2014, Thermal recovery of bitumen from the Grosmont Car-bonate Formation — Part 2: Pilot interpretation and develop-ment strategy: Journal of Canadian Petroleum Technology, 53, no. 04, 212–223, http://dx.doi.org/10.2118/171561-PA.