Gerry G. Calhoun
Independent Geologist
Midland, Tex.
Geochemical Evaluation Research Team (GERT) is an experiment conceived to identify surface geochemical exploration methods that are effective in finding oil.
Many of the methods use technology developed in the last 5 years. In fact, one of the challenging jobs for the geochemical explorationist is keeping up with new techniques and improvements.
The way to minimize dry holes is to screen prospects with carefully selected geochemical techniques prior to acreage acquisition and seismic surveys.
HOW GERT WORKS
The format of GERT was to recruit operators that had a conventionally generated prospect they intended to drill regardless of the outcome of geochemical surveys.
GERT then contracted the various methods listed below to be run on each well during 1988 or 1989. Each contractor predicted the outcome of the wildcat based solely on his evidence.
The companies that supported this effort are: Enserch Exploration Inc., Fina Oil & Chemical Co., Hunt Oil Co., and Maxus Energy Corp., all of Dallas; Gunn Oil Co., Wichita Falls; IP Petroleum Inc. and Mobil Oil Corp., Midland; Meridian Oil Inc., Primary Fuels Inc., and Union Texas Petroleum Corp., all of Houston; Sandia Oil & Gas Co., San Antonio; Yates Petroleum Corp., Artesia, N.M.; Robert Landreth, and Tom Hillin.
Fourteen prospects were involved in the study (Fig. 1). The New Mexico projects were Silurian Fusselman seismic-closure traps, as were those in Glasscock County, Tex.
The Pecos County project had Ouachita thrust-Ordovician Ellenburger objectives. The Concho-Menard County wildcats had stratigraphic traps in the Pennsylvanian King sand as their target. The north Midland basin prospects were seismically controlled Pennsylvanian carbonate buildup plays.
WELL CLASSES
In GERT, a wildcat was classified as a success if it found economic volumes of oil, and a marginal producer was classified as a failure because it did not generate positive cash flow.
The successful wildcats were the Hunt discovery and subsequent offset in Garza County and Hillin's Fisher County Pennsylvanian Canyon (Palo Pinto) producer, a one well field.
The marginal producers are Landreth's Permian San Andres well in Dawson County, Meridian's well in Borden County, and Primary Fuels' well in southern Glasscock County.
To date 18 wells have been drilled on 12 of the prospects. Two prospects have not been drilled: Fina's in Howard County, Tex., and Yates' in Chaves County, N.M.
Two of the 18 wells were chosen to illustrate the results of the study so far.
GERT PROGRAM
GERT examined 13 separate methods on each of 14 prospects in the Permian Basin.
The following discussion will list and describe each technique.
The radar was run by Geochemical Exploration Services (GES), Dallas, with an instrument modified to specifications by Luke Gournay, a consultant to GES. This is the only radar instrument the author can recommend based on his own field experience.
The micromagnetics data were acquired with a ground based cesium vapor instrument by GES and interpreted by them for their prediction.
The author applied a Simple trend surface-residual program to GES data to remove basement effects and emphasize the high frequency anomalies. This was one of the more effective methods.
The soil gas phase of the study was hampered by a lack of data. Due to financial restrictions, GERT could acquire only about 40 points per prospect, which proved to be too few.
The contractor asked that the soil gas method be dropped from the statistical study. This does not condemn the method.
LANDSAT, ELEMENTAL ANALYSIS
Randall Anderson, Satellite Exploration Consultants Inc., Midland, conducted the Landsat experiments, which were among the most accurate of the methods.
The key elements of this phase were lineaments, tonal anomalies, and a Band 5 to Band 4 ratio image that emphasizes the iron content in the soils.
Barringer Laboratories Inc., Golden, Colo., ran the elemental analysis with inductively coupled plasma (ICP) equipment. The author interpreted the data.
Barringer analyzed potassium, calcium, strontium, and manganese. Seven samples were taken near the wildcat site, and three were located around a nearby deep dry hole.
Obviously this was a minimum effort, and better results could be obtained with a better financed project.
Iron and sulfur concentrations vary with proximity to oil; however, sulfur was not available from Barringer Laboratories. Iron concentrations did not vary due to the solvent used in the experiments. The strontium to calcium ratio was analyzed and yielded negative results.
The manganese/potassium (Mn/K) ratio proved effective by contrasting the concentrations near the wildcat with samples taken around a nearby dry hole. By normalizing this ratio, it is possible to minimize the effects of soil changes between the two sites.
FLUORESCENCE, MICROBIAL DATA
Barringer also ran the fluorescence data, using the 10 sample set per prospect format and obtaining 73% predictive accuracy.
This is a measure of the vertically migrating aromatics. These compounds are the most water soluble of all hydrocarbons and therefore the most mobile. This is an excellent method to identify a live seep.
The microbial experiment, also a Barringer project, involved culturing each sample for 1 month. The time lag is a distinct disadvantage. The fluorescence and microbial results were combined to make a single prediction.
RADIOMETRICS
Energy Exploration Inc., Houston, ran the focused or windowed radiometrics with a 1,000 cu in. sodium iodide crystal.
Shielding and observing window are proprietary; however, the technique proved to be one of the most effective of the 13 predictors.
Robert Foote, an independent contractor operating as Geoscience & Technology Inc., Dallas, tested the magnetic susceptibility of drill cuttings. He examined the first 2,000 ft of cuttings from the wildcat for the presence of authigenic iron minerals.
A problem developed because half the members forgot to save shallow samples. Nevertheless, this method has proven quite effective in differentiating between producers and dry holes in studies unrelated to GERT.
TELLURICS, RADIUM SULFATE
Don Chambers, Decco Operating Inc., Abilene, ran the most controversial of the methods, Geoprobe (audio magneto tellurics).
This site specific, depth specific too[ proved fairly accurate in picking formation and porosity tops. This proves nothing about the merits of a similar tool, the Petrosonde by Geophysical international, Dallas.
Chambers also used the standard radiometrics device. Crystal size is 5 cu in., and soil change corrections were made visually.
Target Exploration Co., Columbia, Md., directed the radium sulfate experiment. Its data was derived from a single profile across each prospect totaling 20-30 stations from augured holes 18 in. deep. Target and Energy Exploration generously participated in the project without pay.
HUNT DISCOVERY
The Hunt discovery is naturally of interest because of its success.
The ground based radar map (Fig. 2) shows a medium intensity anomaly associated with Happy field to the north and broken anomalies confined to topographic lows around the prospect. A dry hole was predicted.
The micromagnetic data (Fig. 3) are definitely a departure from regional gradient, and a discovery was predicted.
The residual micromagnetic survey (Fig. 4) shows a small positive anomaly at the drillsite, and a discovery was predicted. The 10 gamma contour represents data beyond one standard deviation. Shallow samples were not collected, so magnetic susceptibility could not be run.
The Landsat image (Fig. 5) contains strong curvalinears and tonal anomalies. A discovery was predicted.
Focused or windowed radiometrics (Fig. 6) outlined the producing feature and predicted a discovery.
The Geoprobe results (Fig. 7) are posted on logs subsequently run on the Hunt well. Geoprobe picked the "C" zone porosity 21 ft high and the basal Pennsylvanian shaly lime 18 ft high.
The survey was terminated at 8,250 ft above the Ellenburger by the Hunt field representative. A dry hole was predicted, and the well was barren in the Strawn.
The Target data (Fig. 8) show anomalous concentrations of radium sulfate surrounding the drillsite; however, the feature was not considered strong enough to merit a discovery prediction.
Soil fluorescence (Fig. 9) on the Hunt area shows concentration of benzene-xylene-toluene and the multiringed aromatics similar to a generic oil. A discovery was predicted.
Compare this with a spectrograph result from the Enserch dry hole (Fig. 10). The naphthalene spike is the result of contamination because it is used in herbicides and pesticides.
Likewise, the Mn/K ratio (Table 1) shows a 3:1 contrast between the wildcat and dry hole. Dry wildcats were characterized in both of these methods by ratios at or below unity.
Fig. 11 is the after-drilling structure map submitted by the operator.
UNION TEXAS WELL
One of the more interesting dry holes is the Union Texas well in Glasscock County.
The radar experiment (Fig. 12) reveals a medium to strong anomaly centered over the Fusselman field to the north, while the wildcat location is haloed with the propane response depicted by the radar image. The linear features parallel faulting on seismic. A dry hole was predicted.
Despite a closed anomaly on the micromagnetics map (Fig. 13), GES predicted a dry hole.
The residual micromagnetic picture (Fig. 14) shows a strong positive feature flanking the drillsite--a discovery was predicted.
The Target Exploration profile (Fig. 17) shows an altered anomaly west of the drillsite, and a dry hole was predicted.
Sample magnetic susceptibility on the Union Texas well shows a negative pattern below 100 units, and a dry hole was predicted (Fig. 19).
Shallow drill cuttings in a producer usually carry authigenic iron minerals that cause susceptibilities in the 300-400 range.
The Landsat image illustrates converging lineaments and positive drainage anomalies, giving a positive prediction (Fig. 15).
Energy Exploration outlined a sizable producing feature (shaded area) with the windowed radiometrics data (Fig. 16).
The Geoprobe picks are posted on the density logs (Fig. 18). The base of Woodford was missed by 75 ft, and the top of Fusselman was off 4 ft. Oil production was predicted.
The Mn/K ratio is about one, indicating a wildcat result similar to the dry hole used in the experiment (Table 1).
This well was notable for the number of positive predictions it drew. After integration of the velocity survey from the wildcat and a subsequent north-south seismic line, the post mortem map shows opportunities for Fusselman production.
Perhaps this area merits a second wildcat.
ASSESSING RESULTS
It has been commented that a contractor could predict all wildcats to be dry and be assured of a very high rating.
Thus it would require a program resulting in at least 26 discoveries before one could truly evaluate these tools.
When a technique shows pessimistic evidence and a dry hole results, the author considers this to be valuable data, for it is the job of the explorer to minimize the number of dry holes his employer drills. The author emphasizes only those methods that accurately predicted at least one of the discoveries.
The scoreboard that resulted from the 179 experiments is shown (Table 2). The more successful techniques will be discussed in some detail. Since the various methods are used to solve different problems in exploration, there are no "best" methods.
The most common mistake in applying geochemical techniques is to get gun-barrel vision and use only one method. The author recommends using no fewer than three, each measuring a different physical or chemical parameter. One of the three should be fluorescence, soil gas, absorbed gas, or radar so that live seeps can be established.
If a regional study as large as 1,000 sq miles has been done using conventional methods, the first geochemical screen should be Landsat.
In GERT, Landsat identified the Hunt discovery and offset and 11 of the 15 dry holes for an overall predictive accuracy of 76%.
The windowed radiometric device employed by Energy Exploration also predicted the Hunt discovery and offset and 14 of the 15 dry holes for an 88% accuracy.
The fluorescence-microbial analysis by Barringer correctly picked both discoveries and seven of 11 dry holes for an overall accuracy of 73%. This appears to be the more accurate of the live seep indicators.
The micromagnetic and residual methods have a solid basis in physics and chemistry.
They identified the Hunt discovery and nine of the 14 dry holes for an overall accuracy of about 62%.
The Mn/K ratio identified both discoveries and achieved an 83% predictive accuracy. This method is not well documented in the literature.
By combining Landsat, Energy Exploration, fluorescence, and residual micromagnetics and applying Fisher's Exact Test statistics to their combined predictions results in the following conclusions:
The combination of the four techniques accurately predicted eight oil producers and inaccurately predicted 15 oil wells.
The four techniques accurately predicted 31 dry holes and were inaccurate at predicting only three dry holes.
Applying Fisher's Exact Test to these results yields numeric odds of 0.186. This statistic indicates that the odds are greater than 50:1 that the predictions are systematic and related to the results.
LESSONS LEARNED
The lessons learned in GERT include:
- The most effective methods to date are Landsat, Energy Exploration radiometrics, residual micromagnetics, fluorescence-microbial, and the Mn/K ratio.
- Even a suite of geochemical methods should not be used as a stand alone exploration program.
- Soil sampling methods should involve no less than 70 samples, equally distributed between "prospective" and "barren" areas.
- GERT results should not be extended to other geochemical contractors using similar equipment or analysis methods, nor should they be projected beyond climatic or geologic conditions similar to the Permian Basin, nor does this report constitute endorsement by the author of the contractors that participated in GERT.
- As it stands, GERT is not finished. At least 12 more iterations must be added before it can be considered a valid statistical study. The pioneer companies that supported the initial phase have helped lay a firm groundwork, but a large amount of geochemical research lies ahead.
If contractors and the industry rise to meet the worldwide competition successfully, it will be with the help of disciplined geochemical surface exploration.
New ideas have always opened up new eras of exploration, and the data developed by GERT can provide a new way to look at mature basins.
BIBLIOGRAPHY
Barringer, A.R., and Lowell, J.S., Multiple Correlation Geochemical Prospecting, in Unconventional Methods in Exploration, Vol. 4, 1986, pp. 201-219.
Beghtal, F.W., and Hitzman, D.O., Microbial Oil Survey Technique Evaluation of New Field Wildcat Wells in Kansas, APGE Bull., Vol. 3, No. 1, 1987, pp. 1-39.
Curry, William H., Evaluation of Surface Gammas Radiation Surveys, in Unconventional Methods in Exploration, ISEM, Vol. 3, 1983, pp. 25-40.
Foote, Robert S., Correlations of Oil & Gas Producing Areas with Magnetic Properties of the Upper Rock Column, APGE Bull., Vol. 4, No. 1, 1988, pp. 47-61.
Morris, Richard V., Spectral and Other Physics-Chemical Properties of Hematite, Journal Geoph. Res., Vol. 90, No. B4, 1985, pp. 3,126-44.
Pirson, Sylvan J., Significant Advances in Magneto Electric Exploration, in Unconventional Methods in Exploration, ISEM, Vol. 2, 1981, pp. 169-196.
Price, Leigh C., Aqueous Solubility of Petroleum as Applied to its Origin, AAPG Bull., Vol. 60, No. 2, 1976, pp. 213-244.
Rice, Gary K., Near-Surface Hydrocarbon Gas Measurement of Vertical Migration, included in Unconventional Methods in Exploration, ISEM, Vol. 4, 1986, pp. 183-200.
Riecker, Robert E., Hydrocarbon Fluorescence and Migration of Petroleum, AAPG, Vol. 46, No. 1, 1962, pp. 60-75.
Townsend, Timothy E., Discrimination of Iron Alteration Minerals in Visible and Near-infrared Reflectance Data, Journal Geoph. Res., Vol. 92, No. B2, 1987, pp. 1,441-54.
Copyright 1991 Oil & Gas Journal. All Rights Reserved.
