--> Aeromagnetic Imaging of Faults in a Basin-Centered Gas Play, by William C. Pearson, #40078 (2003)
[First Hit]

Datapages, Inc.Print this page

  Click to view article in PDF format.

 

Aeromagnetic Imaging of Previous HitFaultsNext Hit in a Basin-Centered Gas Play*

By
William C. Pearson1

Search and Discovery Article #40078 (2003)

 

*Adapted for online presentation from the Geophysical Corner column in AAPG Explorer, May, 2001, entitled “Finding Previous HitFaultsNext Hit in a Gas Play” and prepared by the author. Appreciation is expressed to the author, to R. Randy Ray, Chairman of the AAPG Geophysical Integration Committee, and to Larry Nation, AAPG Communications Director, for their support of this online version.

1Geophysical consultant, Lakewood, CO ([email protected])

General Statement

Recent technologic advances in aeromagnetic acquisition, processing, imaging, and interpretation provide effective methods for locating Previous HitfaultsNext Hit in sedimentary basins. Previous HitMappingNext Hit 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 Previous HitfaultsNext Hit 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 Previous HitfaultsNext Hit seen on both 3-D seismic and aeromagnetic data.

 

uGeneral statement
uFigure captions
uRulison field
uImproved techniques
uMagnetic block model
uDip directions
uIntegrated fault interpretation
uConclusion

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

uGeneral statement
uFigure captions
uRulison field
uImproved techniques
uMagnetic block model
uDip directions
uIntegrated fault interpretation
uConclusion

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

uGeneral statement
uFigure captions
uRulison field
uImproved techniques
uMagnetic block model
uDip directions
uIntegrated fault interpretation
uConclusion

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

uGeneral statement
uFigure captions
uRulison field
uImproved techniques
uMagnetic block model
uDip directions
uIntegrated fault interpretation
uConclusion

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure Captions

Figure 1. The Rulison Gas Field, located in Colorado's Piceance Basin.

 

 

Figure 2. 3-D magnetic block model with computed field and shaded image of first horizontal derivative amplitude. The peaks of the horizontal derivative amplitude overlie the linear block boundaries (Previous HitfaultsNext Hit).

 

 

 

Figure 3.Shaded-relief image of horizontal derivative amplitude of the 5,000-6,000-foot depth-sliced, reduced-to-pole magnetic intensity. The sun in the northeast illuminates several northwest fault trends. Orange Previous HitlinesNext Hit are AUTOFAULT picks of Previous HitfaultsNext Hit within Rulison Gas Field.

 

Figure 4. Another view of the same shaded-relief image of the same horizontal derivative amplitude, this time with the sun in the northwest illuminating weak northeast trends. Again, orange Previous HitlinesNext Hit are AUTOFAULT picks.

 

 

Figure 5. SUNMAG image with block AUTOFAULT Previous HitlinesNext Hit shows Previous HitfaultsNext Hit identified by 5,000-6,000-foot depth-sliced, filtered reduced-to-pole magnetic field. Color key in upper left corner shows direction of dip for each color. Rectangular window shows location of Figure 6.

 

Click here for sequence of Figures 3, 4, and 5.

 

Figure 6. A contoured estimated ultimate recovery from wells in the Rulison Gas Field (GRI, 1997). Interpreted Previous HitfaultsNext Hit from 3-D seismic (red) and from aeromagnetic survey (yellow) produce lateral seals to the tight gas sand compartments within the field. The footprint of the full-fold 3-D seismic survey is outlined in black.

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 Mesaverde 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 Previous HitfaultsNext Hit 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 Previous HitlinesNext Hit 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 Previous HitlinesNext Hit -- in this case, a 250m by 1,000m array of Previous HitlinesNext Hit.

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 Previous HitgridNext Hit. The total magnetic intensity Previous HitgridNext Hit 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 Previous HitfaultsNext Hit of the body.  Nearly linear intrasedimentary Previous HitfaultsNext Hit 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 Previous HitgridNext Hit 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 Previous HitfaultsNext Hit 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 Previous HitfaultsNext Hit 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 Previous HitlinesNext Hit on the Figure 5 SUNMAG image to highlight dip compartments between Previous HitfaultsNext Hit.

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 Previous HitfaultsNext Hit.

Within the 3-D seismic survey area, Previous HitfaultsNext Hit identified at the Mesaverde level correlate locally with interpreted regional magnetic Previous HitfaultsNext Hit. 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.

Conclusion

Recognizing regional fault patterns in prospective basin-centered gas plays is a critical element based on known analog fields. Seismically interpreted Previous HitfaultsNext Hit and magnetically imaged Previous HitfaultsNext Hit in Wyoming's Green River Basin have confirmed that sealing Previous HitfaultsNext Hit bound the recently developed, 1 TCF Jonah gas field.

The similarity to the observed pattern of bounding Previous HitfaultsTop 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.

Return to top.