By WILLIAM C. PEARSON
The Geophysical Corner is a regular column in the EXPLORER, edited by R. Randy Ray. This month's column is titled "Aeromagnetic Imaging of Faults in a Basin-Centered Gas Play."
Finding Faults In a Gas Play
Editor's note: Bill Pearson is a geophysical consultant and potential field expert in Denver. For the past seven years he has co-chaired the annual Denver 3-D Seismic Symposium and is currently chairman of the Society of Exploration Geophysicists Scholarship Committee.
Figures courtesy of William C. Pearson
Click on image to enlarge.
Recent technologic advances in aeromagnetic acquisition, processing, imaging and interpretation provide effective methods for locating faults in sedimentary basins.
Mapping regional fault patterns within a basin are critical to finding new fields and extending existing fields in basin-centered gas plays. Deeper and older structures are often reactivated with later stage compression and wrench faulting to produce enhanced porosity and permeability in tight sands.
In some instances a pattern of open fractures or intersecting fault trends causes good gas reservoirs. In other instances wrench faults are mineralized and act as seals to adjacent gas pressure compartments.
A consortium of the public-supported Gas Research Institute (GRI) and private exploration companies teamed up to sponsor a study on tight gas sand production enhancement in the Piceance Basin of northwest Colorado. The study combined Landsat imagery, aeromagnetic surveying, 3-D seismic surveying and drilling to determine a method for locating areas of high gas deliverability.
The results of a study in the Rulison gas field indicate that the best estimated ultimate recovery wells are located in a wedge of tight sandstones sourced from underlying coal beds, and sealed laterally by faults seen on both 3-D seismic and aeromagnetic data.
Rulison Field Producing Formation
Gas production in the Rulison Field, near the town of Rifle (Figure 1), has been studied extensively by GRI (1997). In the study Landsat imagery, drilling information, a 3-D seismic survey and a high-resolution aeromagnetic (HRAM) survey complete a picture of a basin-centered, tight gas sand reservoir.
Gas production in the field is from a tight sand in the Williams Fork Formation of the Upper Cretaceous Mesa Verde Group. Individual wells vary from non-commercial to over 3 Bcf estimated ultimate recovery within a distance of a half mile or less.
The more economic wells are on a structural flexure defined by well tops and 3-D seismic data. However, bounding faults to the west, east and south as defined by seismic and magnetic data produce an important seal to the reservoir.
Three-D seismic is relatively effective at locating velocity minima and anisotropic velocity anomalies that correspond to the best wells, but its use is restricted in the Piceance due to limited surface access and cost considerations. Aeromagnetic surveying is unencumbered by surface restrictions, is very fast to acquire and is inexpensive compared to seismic, leasing and drilling.
Improved Magnetic Techniques
Current aeromagnetic technology is a much-improved tool as compared to the ground and airborne magnetic surveys of 10 or more years ago.
The GPS satellite navigation allows flight lines to be positioned and located within a few meters of desired location, providing for improved flight line leveling and interpolating onto map grids. HRAM surveys are optimally flown with closer spaced lines -- in this case, a 250m by 1,000m array of lines.
Recently developed processing techniques remove noise due to surface sources such as wells, pipelines and towns, thus providing a much cleaner signal.
Computer filtering, imaging and interpreting tools are important upgrades to the old familiar contouring and hand interpreting techniques.
Magnetic Block Model
Intrasedimentary magnetic fault anomalies in the Piceance are pre-amplified on the profiles and interpolated from the 250m spaced decultured profile data onto a 100m grid. The total magnetic intensity grid is then filtered by a sequence of steps including a reduction-to-pole filter, Wiener pseudo depth matched filter (focusing on the 5,000 to 6,000 foot depth range), and horizontal derivative filter.
Figure 2 illustrates the nature of magnetic lineaments mapped by the first horizontal derivative conversion. The figure displays a synthetic block model, its magnetic field model and a shaded relief image of the first horizontal derivative amplitude. The magnetic anomaly has its steepest gradient (dip) over the bounding faults of the body.
Nearly linear intrasedimentary faults or lateral termination of slightly magnetic sands produce correspondingly linear observed magnetic gradients that can be imaged in the same manner.
Figure 3 is a northeast sun shaded relief image of the horizontal derivative grid over the Rulison Gas Field. Townships are overlaid in blue for location and scale.
Actual subsurface fault/lineament locations are at linear trends, where lighter slopes facing the northeast sun turn to dark away from the sun toward the southwest. The line of change from light to dark is highlighted in pink, and is the fault/lineament location.
Figure 4 demonstrates a northwest sun angle for the same area, highlighting anomaly trends that are much weaker in a northeast orientation. Real time sun angle rotation on the computer screen is very helpful for detecting lineaments of a preferred orientation.
Highlighting All Dip Directions
A new method of lineament detection uses the full horizontal gradient, including both amplitude and azimuth.
Figure 5 shows a color SUNMAG image showing all dip directions as a unique color, with the color key superimposed in the top left corner of the figure. SUNMAG imaging delineates lineaments of all orientations simultaneously -- and, in fact, picks up more subtle faults than the gray single-direction sun shadowgraphs, because the horizontal gradient direction is a more sensitive indicator of fault anomalies than the gradient amplitude alone.
The computer automatically detects faults with an AUTOFAULT algorithm based upon a neural network technique for detecting discontinuities in the horizontal gradient azimuth and amplitude data.
The AUTOFAULT computer fault picks are superimposed as black lines on the figure 5 SUNMAG image to highlight dip compartments between faults.
Integrated Fault Interpretation
Integrating the magnetic fault picks (yellow) and 3-D seismic fault picks (red) with color contoured, estimated ultimate recovery from wells produces a good picture of why the better wells line up in a north-northwest direction (figure 6, GRI, 1997). The warmer colors correspond to wells with more than 3.5 Bcf per well estimated ultimate recovery.
The best wells in sections 17 and 20 lie between northwest trending sealing faults.
Within the 3-D seismic survey area, faults identified at the Mesa Verde level correlate locally with interpreted regional magnetic faults. The magnetically defined delineate pressure compartment boundaries within the field and extend beyond the 3-D seismic survey.
A comprehensive integration of well results, 3-D seismic and aeromagnetic imaging provides a model for locating lead areas in the greater Piceance Basin similar to the Rulison Field.
Although the aeromagnetic imaging has much coarser spatial resolution than 3-D seismic data, it is relatively inexpensive to use over a broad area of reconnaissance exploration prior to exploratory drilling and detailed 3-D seismic surveying.
Recognizing regional fault patterns in prospective basin-centered gas plays is a critical element based on known analog fields. Seismically interpreted faults and magnetically imaged faults in Wyoming's Green River Basin have confirmed that sealing faults bound the recently developed, 1 TCF Jonah gas field.
The similarity to the observed pattern of bounding faults around the better producing wells in the Rulison gas field point to a picture of compartmentalization of tight gas reservoirs.
An exploration strategy based upon this model can be applied to other gas and oil plays in the Rocky Mountains and Canada as well as to many basins around the world.