«Interpretation of 3D multicomponent seismic data for investigating natural fractures in the Horn River Basin, northeast British Columbia Abdallah ...»
Natural fracture analysis from 3D-3C seismic data
Interpretation of 3D multicomponent seismic data for
investigating natural fractures in the Horn River Basin, northeast
Abdallah Al-Zahrani* and Don Lawton
A 3D multicomponent seismic dataset from the Horn River Basin was assessed for
mapping fractures. The data had good fold, offset and azimuth distributions and several
approaches were used to interpret the distribution of natural fractures. In addition to amplitude mapping, PP and PS curvature maps enhanced the structural interpretation of the data and enabled the lateral continuity of faults and fractures to be mapped across the area of the seismic survey. Both horizon and volume based most negative curvature were effective in mapping fault and fracture trends within both Exshaw and Muskwa shale gas targets. At the Exshaw level, the curvature shows two main fault trends: northwestsoutheast trending normal faults that dip toward the southwest, as well as northeastsouthwest strike-slip faults. At the Muskwa level, the curvature image shows different major fault trends, namely north, northeast-south, southwest (normal and reverse faults), and northwest-southeast faults. Fractures interpreted using curvature attributes are close to the major faults and their dominant trends are generally parallel to the major faults in the area.
INTRODUCTIONShale gas is one of the three types of unconventional gas plays that are growing rapidly in North America (Kuuskraa and Stevens, 2009). What make this play attractive to the oil companies compared with conventional resources is that the shale gas formations are continuous over large areas, with giant reserves, and the gas is enclosed within microporosity of the shales, and the rocks, in this case act as source, reservoir and trap at the same time (Kuuskraa and Stevens, 2009). The challenge in this play is how to flow the gas from these tiny pores. One of the current solutions to this problem is the use of the new drilling technology such as multi-pad horizontal wells and multi-stage hydraulic fracturing (Figure 1). For example, this technology helped to double the gas productivity of the Barnett shale field in Texas and make it the largest gas field in that state (Beaudoin and Shaw, 2009). However this technology alone is not enough for optimum exploration and development. More advanced multi-disciplinary specialized techniques such as spectrographic, petrophysical, geochemical and geophysical analysis are needed in order to get a detailed picture of the shale gas complexity and heterogeneity. Data integration of all these methods will help to minimize the cost of these horizontal wells by locating them in the optimum location (Beaudoin and Shaw, 2009).
The geophysical data investigated in this study was a 3D multicomponent survey which provided optimal fold, offset, and azimuth distribution with the goal to map natural fracture orientations and density within these shale units. Acquiring converted-wave P-S data can provide lithological information and possibly detect fracture networks, whereas conventional P-wave data can reflect matrix and fluid characteristics (Stewart et al., CREWES Research Report — Volume 22 (2010) 1 Al Zahrani and Lawton 2003). Wide-azimuth multicomponent data can provide several methods to detect fracture network. These are seismic attributes (e.g. coherency and curvature), P-wave azimuthal anisotropy, and shear-wave splitting (Jianming et al., 2009). Each method has its own limitations, however the integration of all these methods with well data will lead to better fracture prediction (Jianming et al., 2009).
This report is focused on the interpretation part of a recently acquired wide-azimuth 3C-3D dataset within the Horn River Basin, to map the shale gas reservoirs and predict its nature fracture orientation, and possibly density using seismic attributes.
Figure 1: Multipad-horizontal drilling through the shale gas. Source:
STUDY AREAIn Canada, the Horn River Basin, which is located in northeast of British Columbia (Figure 2), is the area where this shale gas play is being explored. The areal extent of this basin is about 1.28 million hectares and the estimated total gas-in-place is 500 trillion cubic feet (Beaudoin and Shaw, 2009). Intensive core analysis was carried out by CBM Solutions Ltd., for the British Columbia Ministry of Energy and Mines, Oil and Gas Division, Resource Development and Geoscience Branch to highlight the gas shale potential in the area (British Columbia Ministry of Energy and Mines and CBM Solutions, 2005). The analysis was focused on four properties that are important characteristics in each shale gas play. These are the total organic carbonate content, the thermal maturity, sorption capacity, and the mineralogy. The result of this study is shown on Table 1, and it designates the potential shale gas reservoirs (British Columbia Ministry of Energy and Mines and CBM Solutions, 2005). These targets are the Upper
Geologic setting A stratigraphic column of northeast British Columbia is presented in Figure 3. The Devonian-Mississippian strata of northeast of British Columbia is interpreted to be deposited in ramp and basin settings (Richards, 1989). The shaly units represent the basinal deposits (e.g. Besa River, Exshaw, and Muskwa shales), while the carbonate units represent the ramp deposits (e.g. Slave Point, Trout River, and Kotcho carbonates) (Figure 2). The Exshaw and Muskwa shales are thin and present throughout the basin.
They are characterized by high gamma ray reading on the logs due to their high organic content (British Columbia Ministry of Energy and Mines and CBM Solutions, 2005). The vertical thickness of the Exshaw Formation is from 5 to 15 meters and the Muskwa Formation is around 25 to 50 meters based on the wells that penetrated these formations on the study area. The Exshaw Formation can be found over a depth range between 1105m and the Muskwa can be found at a depth of around 2175-2250 m. Structurally, the area was subject to north-south extension followed by a compressional period at the CREWES Research Report — Volume 22 (2010) 3 Al Zahrani and Lawton pre-Proterozoic sediments time which resulted in erosion of these sediments towards the southeast (Ross and Stephenson, 1989). At late Paleozoic time, two main fault trends were documented in the area: northwest-southeast trending normal faults that dip toward the southwest and linked in depth with a detachment fault (McClay et al., 1989), and northeast-southwest transverse faults that are associated with compression (Churcher and Majid, 1989) (Figure 4).
Figure 3: Stratigraphic section of Devonian-Mississippian strata in northern British Columbia.
Dashed blue lines show Horn River Basin stratigraphy and the red solid boxes highlight the Exshaw and Muskwa shale reservoirs (Ross and Bustin, 2008).
Table 1: Summary data for each formation analyzed. The solid red rectangles indicate the main shale gas reservoir on the HRB (British Columbia Ministry of Energy and Mines, 2006).
During Cretaceous-Eocene time, the area was subject to a compressional regime that reactivated old normal faults above a deep detachment in the Upper DevonianCarboniferous section, forming thrust faults, related fault-bend fold and detachment folds (Eaton et al., 1999).
DATA AND SOFTWAREOlympic Seismic Ltd acquired the 3D3C dataset that is used in this study. It was obtained as a multicomponent survey and was processed by Sensor Geophysical Ltd. The data were imported into the Kingdom Suite software and a synthetic seismogram was computed to tie the PP seismic volume to the formation tops at this stage. Seismic reflections were correlated to well formation tops first by applying the seismic datum, and then matching the P-wave and density logs. After that, a 35 Hz Ricker wavelet was convolved with the reflectivity log calculated from the density and P-wave logs to create a PP synthetic seismogram. Figure 5 shows the seismic-to-well tie and the correlation between synthetic seismogram and the PP seismic stack is quite good for shallow target (Exshaw shale, 0.82 seconds) and good at the deep target (Muskwa shale, 1.52 seconds).
The Exshaw Fm. in this well is only 5 m thick and is characterized seismically by a strong peak that is tuned with the event from the top of Kotcho Formation below it (Figure 5). In comparison, the Muskwa shale is about 51 m thick, and is characterized seismically by a weak trough due to the low impedance contrast between the overlying Fort Simpson shale and the underlying Muskwa shale (Figure 5). Figure 6 shows the seismic-to-well tie for PS data. The correlation between synthetic seismogram and the PS seismic stack is good for Exshaw shale (1.128 seconds) and fair to good at the top of Muskwa shale (2.178 seconds). Seismically, the top of both Muskwa and Exshaw shales are characterized by relatively moderate peak.
INTERPRETATIONThe PP and PS seismic data were interpreted on a workstation using Kingdom Suite software version (8.5). Seismic interpretations started after performing the well-toseismic tie (Figure 5 and Figure 6). Figure 7 shows the ten key seismic horizons picked (from top to bottom). These are the tops of the Debolt Formation, the Banff Formation, the Exshaw Formation, the Tetcho Formation, the Trout River Formation, the Kakisa Formation, the Fort Simpson Formation, the Otter Park Formation (the Base of Muskwa Formation), the Evie, and the Lower Keg River Formation. All these seismic horizons were picked at every 40 inline and cross line spacing. After that, a 3D interpolation were performed for these picks for each seismic horizon using the 3D hunt feature that provided by Kingdom Suite. Figure 8 shows the two-way time structure maps for both top Exshaw Formation and the base of Muskwa Formation with interpreted faults intersecting them. Three normal faults dip southwest and their strike direction is northwest-southeast. Two probable strike-slip faults intersecting at 90° are interpreted at the top of Exshaw Formation in the middle of the survey (Figure 8). The interpretation is quite different in the deeper Muskwa Formation. One normal fault and one reverse fault have been interpreted at this level. The reverse fault dips northwest direction and trends northeast-southwest, forming a small four way anticline closure. The anticline axis is parallel to the strike direction of the fault (Figure 8). The normal fault has small throw and its direction is running north-south (Figure 8).
CREWES Research Report — Volume 22 (2010) 5 Al Zahrani and Lawton Figure 4: Map of the Slave Point Formation which is the base of the Muskwa Formation in the area of Horn River Basin. This map, shows the Bovie Fault and several other dominant faults that trend in the area (SW-NE and NW-SE) indicated by the blue dashed lines.
Source:http://www.wcsbgroup.org/meetings/pdf/2009/Jurisdiction%20Updates/BC_Update%20an d%20Shale%20Gas_Levson.pdf Figure 5: Seismic-well tie for PP data showing from left to right: Interval velocity (black), Sonic log (red), density log (blue), calculated impendence log (black), reflectivity(black), the 35 Hz Ricker wavelet, the synthetic seismogram with five traces (blue) and nine traces that have been extracted from the 3D dataset near the wellbore (black).
6 CREWES Research Report — Volume 22 (2010) Natural fracture analysis from 3D-3C seismic data Figure 6: Seismic-well tie for PS data showing from left to right: Interval velocity (black), Sonic log (pink), density log (blue), calculated impendence log (green), reflectivity (black), the 20 Hz Ricker wavelet, the synthetic seismogram with five traces (blue) and nine traces that have been extracted from the 3D dataset near the wellbore (black).
Figure 7: Inline from PP data showing the seismic-well tie for well A and the ten key interpreted seismic horizons. The two red arrows on the top right corner show the PS section correlated approximately with the PP data.
Figure 8: Exshaw shale two-way time structure map (upper) and base of Muskwa shale two-way time structure map (lower) showing the interpreted faults (black solid lines) that cut these surfaces in the area.
CURVATURE ANALYSISRecently, the curvature attribute have been used to recognize subsurface geological features such as faults/fractures and channels (Chopra et al., 2006). Mathematically, it is the computation of the second order derivative of the curve (Chopra et al., 2006). In 2-D curve, it is the reciprocal of the radius of a circle that is tangent to the curve at any point on it (Figure 9a) (Chopra et al., 2006). While in 3-D, the curvature analysis can be computed by applying the same curvature in 2D analysis on the intersections of two orthogonal planes with the 3D surface (Figure 9b) (Chopra and Marfurt, 2007). This
computation can be made on two scenarios:
2- Volumetric based Since the volume-based curvature is computed from a time window of seismic data, the results are statistically less sensitive to backscattered noise and has a higher signal-tonoise ratio than horizon-based curvature (Chopra and Marfurt, 2007). Horizon based computations require high quality seismic data and a strong impedance contrast on the
horizon of interest, whereas volume-based curvature do not need a picked horizon to do the analysis (Chopra and Marfurt, 2007).