GEOCHEMICAL EXPLORATION-1

Results of magnetic HGI and radiometric surveys in W. Canada

Leonard A. LeSchack
Topaz Energy Exploration Ltd. Calgary

This article presents four case histories in which ground-based magnetic horizontal gradient intensity (HGI) and radiometric surveys were used in Western Canada for cost-effective geochemical exploration for hydrocarbons.

The author has developed these two surface exploration techniques from published studies and adapted them for use on the prairies the past 7 years.

These surveys are used in conjunction with the usual geologic and seismic studies for:

1. Evaluating prospects and land;
2. Verifying seismic anomalies and inexpensively locat ing areas for conducting expensive 3D seismic surveys.

Occasionally, as in two of the case histories discussed, these surveys were used successfully as stand-alone exploration methods where seismic exploration is not effective.

The HGI and radiometric surveys measure, by geophysical methods, those effects associated with geochemical alterations due to vertical microseepage of hydrocarbons. The total cost, including permitting, data acquisition, data processing, and interpretation of the combination HGI and radiometric surveys is about 15% the total cost of a 3D seismic survey. Because of this, the author finds them an attractive and rapid survey adjunct to traditional exploration. They substantially reduce finding costs and significantly raise the probability of financial success.

HGI, radiometric theory

During the past decade and especially within the past few years, numerous discussions have appeared in the literature concerning hydrocarbon exploration based on measuring either the effects of vertical hydrocarbon migration from reservoirs, or the leaking hydrocarbons themselves.

The author's work is based on the theory presented by Saunders et al.1 and Thompson et al.2 that relates near-surface anomalies in uranium, potassium, topography, and magnetics to hydrocarbon deposits. This theory involves hydrocarbon microseepage by near vertical ascent of ultra-small (colloidal size) gas bubbles of light hydrocarbons (methane through the butanes) through a network of interconnected groundwater-filled joints, fractures, and bedding planes. Chemical and/or bacterial degradation of the hydrocarbons instigates diagenetic changes that enhance near-surface magnetic mineralization and alter the balance of uranium and potassium minerals.

Saunders et al.3 show that there are anomalously large amounts of diagenetic magnetic minerals in shallow-depth samples in 89% of the cases over 19 oil and gas fields. They suggest that nonmagnetic hematite is converted to magnetic iron minerals by chemically reducing effects of hydrogen sulfide produced as a by-product of microbial degradation of microseeping hydrocarbons and associated sulfate reduction. Machel and Burton4 say the most important magnetic minerals formed appear to be magnetite and pyrrhotite formed by dissolution of hematite.

R.S. Foote analyzed magnetic susceptibility in drill cuttings from hundreds of wells over, and surrounding, oil and gas fields, and his results correlated positively with low-altitude aeromagnetic data flown above these fields. The combination of aeromagnetic and magnetic susceptibility measurements of drill cuttings could be used to predict hydrocarbon discoveries 80-85% of the time.5

Foote6 has made powder X-ray diffraction analyses of drill cuttings from many boreholes in oil and gas fields over which aeromagnetic anomalies were observed, and reported that cuttings with high magnetic susceptibility values were mostly associated with maghemite (gamma Fe2O3). Lesser amounts of magnetite were observed with the maghemite, and greigite (Fe3S4) had been observed in cuttings from some wells. Foote's studies indicate most oil or gas wells have one major magnetic stratum, and may have two or more magnetic strata of varying thickness which may vary in depth from the surface to about 600 m. These strata are referred to by the present author as "Foote Horizons."

Saunders et al.7 suggested the following origin for observed low-potassium radiometric anomalies over subsurface hydrocarbons. Potassium in sediments is contained primarily in the clay mineral, illite, and occupies spaces between the layers in the clay structure. Degradation of microseeping hydrocarbons produces carbon dioxide which reacts with groundwaters to form carbonic acid. Carbonic acid ionizes to produce hydronium ions, H3O+, which are believed to replace potassium ions in the illite, and the potassium is removed in groundwater solution.

Saunders' proprietary studies of NURE (National Uranium Resource Evaluation) Program aerial gamma-ray spectral data covered approximately 19,608,000 acres in Texas, Oklahoma, Arkan- sas, Louisiana, Mississippi and Florida. The results showed characteristic thorium-normalized potassium and uranium anomalies above 72.7% of 706 oil or gas fields that were overflown.7 Saunders8 reported that ground-based gamma-ray spectral measurements have found characteristic anomalous low-potassium measurements over 12 out of 14 fields studied to date (86%).

The theory discussed above has yet to be conclusively validated; however, empirical evidence indicates effects of hydrocarbon microseepage can be measured over most oil and gas fields. Whatever the cause of the measured variations in radioactivity and soil or rock magnetic susceptibility over hydrocarbon emplacements, a sufficient number of independent studies of these associations exist to support conducting HGI and radiometric surveys for hydrocarbon exploration.

Survey methods

Our HGI survey is essentially a land-based version of the airborne survey discussed by Donovan et al.9 Using two Scintrex OMNI IV magnetometers accurate to 0.1 nT (one for surveying and the other as a base station), total field (and radiometric) data are recorded at 50 m intervals along linear profiles. Six to eight linear miles are surveyed per section, and fill-in data are recorded later as needed.

The diurnally corrected data are then gridded using a kriging algorithm as described by Olea.10 A map convolution filter,11 designed to focus on shallow anomalies in the "Foote Horizons" and designed to have the spatial period of hydrocarbon reservoirs expected in the survey area, then operates on the gridded data matrix to remove the regional magnetic effects of the Precambrian basement. The absolute values of the second horizontal derivative of the residual data are then mapped as the horizontal gradient intensity.

In the case histories at the Pierson, Man., fields and at the Rumsey Reef, a Scintrex GIS-4 gamma radiation spectrometer with 43 cc sodium iodide detector was either hand-carried or mounted forward of an all-terrain vehicle (ATV). A GPS navigation system was also hand-carried or mounted on the ATV.

For Waskada field, a Scintrex GAD-6 spectrometer with a 360 cc detector was used. This instrument and the GPS navigation system were mounted on an ATV. The magnetometer operator, on foot, trails 50 m behind the ATV to avoid any influence of the vehicle. He also avoids any other obvious cultural magnetic contamination.

Both the GIS-4 and the GAD-6 units measure total count, potassium (measured by K-40 radiation), uranium (measured by Bi-214 daughter radiation), and thorium (measured by Tl-208 radiation). A dwell time of 100 sec is used with either spectrometer and represents a compromise between counting statistics and survey logistical requirements.

Gamma radiation measurements in the uranium window must be corrected for diurnal variations in near-surface atmospheric radon concentrations and day to day variations due to changing atmospheric conditions. This is accomplished by making measurements at four separate control points at the beginning and end of each day's operations. Beginning and ending measurements in each channel for all control points are averaged and the averages are used to normalize every day's individual measurements from one day to another. Measurements are not made on rainy days to avoid erroneous counts due to Bi-214 radon daughter washout from the atmosphere.

As in the case of the HGI survey data, the corrected radiometric survey data are gridded for contouring using a kriging algorithm. The gridded data are expressed either directly as counts/100 sec, or as Z-scores (i.e., in terms of standard deviations about the mean). The difference between each grid value and the average of eight surrounding grid points located on a circle of 400 m diameter around that value is then computed. These difference values are mapped as the radiometric survey. This circular averaging process is an analytical attempt to attenuate effects of surface variations from clayey to sandy soils in the radiometric survey.

Thorium-normalized uranium and potassium measurements discussed by Saunders et al.1-7 are more stable than the total count measurements used for the Pierson, Man., surveys. The thorium-normalization pro- cess helps suppress variations in surface lithology, soil moisture content, vegetation shielding, and counting geometry which can interfere with accurate radioelement measurements. The author now routinely maps both the thorium-normalized potassium and "DRAD," the difference between thorium-normalized uranium and thorium-normalized potassium.

Although in a survey there are a variety of causes for either magnetic or radiometric anomalies to occur that have no relation to hydrocarbons, it is less likely for radiometric and HGI anomalies to occur at the same location without hydrocarbon leakage being the common causative factor. In a recent proprietary HGI/DRAD survey the author conducted over 39 contiguous sections of land, he observed that every significant HGI anomaly had a corresponding DRAD anomaly.

Manitoba case histories

In 1993 the author conducted an HGI and total count radiometric survey over what was called the North Pierson prospect. A downdip seismic "nose" contoured from old seismic data of the Mississippian erosion surface was being investigated.

At the time of the survey, there were no wells in the area covered. Also in 1993, a total count radiometric survey was conducted over a portion of South Pierson Lower Amaranth field. The intent was to see if the existing field could be extended or a new field discovered on offsetting land. Subsequent drilling on the mapped anomalies identified two new field discoveries in North Pierson and a new field discovery at South Pierson.

The Waskada study was motivated by the success of the HGI and radiometric surveys at the nearby Pierson fields, which have similar geology, and the fact that seismic surveys are ineffective in mapping the Lower Amaranth stratigraphic traps at Waskada.

Geology of the fields

Waskada field produces light gravity oil largely from the Lower/Middle Jurassic Lower Amaranth sand, which correlates with the Lower Watrous of Saskatchewan and Upper Spearfish of North Dakota.12 It lies unconformably over the erosionally truncated Mississippian Mission Canyon limestones, which are also oil producers and the presumed source for the Lower Amaranth oil.

Based on numerous wells in this field, it is clear to see that there are specific "channels" of higher productivity within the overall sand.

Fig. 1 [33074 bytes] shows the general location of Waskada field with respect to the underlying Mississippian oil production. The location of North Pierson (Mississippian) and South Pierson (Lower Amaranth) case histories is also shown. The sedimentary section occurs as a basinward thickening of Paleozoic and Mesozoic rocks. A significant angular unconformity at the Mississippian erosion surface separates the Paleozoic from the younger rocks. The lower member of the Amaranth formation is the basal Mesozoic unit. Fig. 2 [23074 bytes] shows its stratigraphic position and relation to underlying Mississippian rocks.

The major rocks of the Lower Amaranth are "red bed" dolomitic siltstones and sandstones interbedded with argillaceous siltstones and shales. It appears that deposition was probably related to intermittent shallow marine conditions.

The underlying Mississippian strata associated with Pierson and Waskada fields are described by Kent13 as inner shelf carbonates of burrowed, peloidal, and skeletal packstones and grainstones. Trapping of hydrocarbons in these Mississippian reservoirs is due to Mississippian erosion forming paleogeomorphic traps14 that are sealed by anhydritization immediately subjacent to the unconformity.

Predrilling surveys

Fig. 3 [51778 bytes], taken from LeSchack,15 shows an example of a GIS-4 total count radiometric survey over the South Pierson Lower Amaranth reservoir.

The productive sand channel in the lower left of the figure is clearly outlined by the typical potassium depletion over hydrocarbon emplacements. These channels cannot be easily seen with seismic surveys. There is little structural relief, and these stratigraphic traps are usually too thin to be clearly resolved by seismic data.

Based solely on that mapping, an exploratory well was drilled at "D" which made 1,000 bbl of oil the first month on production. Saunders8 similarly observed that the Paluxy formation sand channels in Texas (Lower Cretaceous) were also well defined by his radiometric surveys.

When the radiometric survey was being conducted, another company spudded the horizontal well 'E' in the section west of Fig. 3 and the 'heart' of the channel in the southwest of Fig. 3 was penetrated. The location of well 'E' was chosen on the basis of interpretation of subsurface geology alone. It produced 3,800 barrels of oil in its first month on production.

Fig. 4 [41007 bytes], also taken from LeSchack,15 shows an HGI study of North Pierson field. The reservoir is at the erosion surface of the Alida/Frobisher (MC-3) limestone of the Mission Canyon series. This formation is analogous to the underlying limestone at Waskada. The only difference is that at Waskada, which is closer to the northeastern edge of the Williston basin, the limestone at the erosion surface is the older Tilston formation (MC-1), the younger Alida formation having been eroded away.

This HGI survey was completed prior to any drilling. Well 'A' in Fig. 4 was subsequently drilled by a company that farmed in on the property. The well made 1,550 bbl of oil in its first month on production. The company drilling this well based its location on a seismic line and, at that time, ignored the author's data. As it turned out, the new field was stratigraphically, rather than structurally, controlled.

After drilling the successful well, the company sought the author's advice for location of well 'B' since it had no other seismic coverage in the area. Well 'B' was drilled, 168 bbl of oil were produced, but the well has been suspended. Following that, horizontal well 'C' was drilled and it made 1,680 bbl of oil in its first month on production. Wells 'F' and 'G' were subsequently drilled by the original lease holder with locations based solely on the HGI/radiometric survey. Both wells found oil. Well 'G,' as suggested by the HGI mapping, discovered a different pool from the others shown in Fig. 4. The well is now on production and the oil has a significantly different gravity from the earlier wells.

Fig. 5 [39363 bytes] is the concomitantly prepared total count radiometric map. Comparison of Figs. 4 and 5 show both apical and halo anomalies associated with the HGI anomalies.

In summary, five wells were subsequently drilled, and all found oil in two new pool discoveries. It is clear from Fig. 5 that the HGI anomalies take on shapes similar to those expected in this part of the Williston basin where limestone cuestas are eroded to form paleogeomorphic reservoirs.14

Of the wells drilled and shown on Fig. 4, the horizontal well 'C' provides the most compelling evidence of the accuracy of HGI surveying. Not only did the mapping predict encountering significant pay thickness beneath the surface of the lateral entry, it also predicted a breach or pinchout of effective reservoir which was, in fact, encountered during the horizontal drilling and is seen as the "erosional" embayment at the southern end of the lateral. This general congruence of HGI anomalies with the reservoir that caused them has been found in the current study, and has also been noted by Foote16 and Tompkins.17

Waskada follow-up

The motivation for the subsequent 320-acre Waskada study containing seven wells was to determine if the two economic producers shown in Fig. 6 [34968 bytes] are located in the same sand channel or "sweet spot."

The remaining wells include two dry holes and three marginal producers. As will be seen, and contrary to expectations, the two economic producers are located in two distinctly different, but clearly defined, channels.

Fig. 6 shows the location of the seven wells and their cumulative oil production. Superimposed on the area are the locations of HGI/DRAD survey points. Because all producers at the test site were drilled within 13 months of one another over 5 years ago, it is assumed that direct comparison of cumulative production totals is valid for this study.

Waskada results

The locations of the survey data points comprise an optimum data point spacing, far enough away from the wells to avoid magnetic contamination, yet close enough so that there is adequate coherence among the data as determined from the semivariograms used in the kriging process. There are no pipelines in this area to introduce other magnetic contamination. The oil is trucked away.

Fig. 7 [56552 bytes] shows the results of the DRAD survey. DRAD is the difference between thorium-normalized uranium and thorium-normalized potassium, as described above. The DRAD anomaly, as opposed to the potassium or total count anomaly, is expressed as a positive number, the more positive, the more anomalous. When the positive values are contoured, the coherent pattern of fluvial, possibly braided, channels is suggested.

Further qualitative comparison of the cumulative production of the well nearest each major anomaly within the channel system shows a positive correlation. For example, well 'S,' with the greatest production, clearly coincides with the most positive DRAD grid value. Well 'V,' with the second largest production, is just at the edge of the second largest anomaly on the map. Wells 'R' and 'T,' both dry holes, are associated with wide contiguous non-anomalous regions. Thus, it seems that DRAD mapping may be a valuable tool for predicting best drilling locations in the Lower Amaranth sand.

Fig. 8 [57567 bytes] shows the results of the concurrent HGI survey. The pattern and trend of the resulting contours is qualitatively different from the DRAD map, which was surprising, based on the observed similarity between maps produced by these two exploration techniques in the past, as is the case at North Pierson.

Examination of the sonic and resistivity logs for the wells in Fig. 6 suggests that Fig. 8 may be a map of the Mississippian reservoirs. Well 'W' actually produced 538 bbl of oil from a 2 m Tilston porosity zone, just beneath the tight anhydrite cap typical of Tilston reservoirs. Based on HGI contour morphology, wells 'W' and 'V' are too far down in the "subsequent valley" of the cuesta reservoir and are therefore wet. Based on the HGI data, it is expected that had well 'V' been drilled 200 m to the south, it would have been structurally near the top of the cuesta reservoir and productive. Well 'U,' based on contour morphology, is off the Tilston structure and is wet.

Well 'T' is located at the largest value HGI anomaly at which wells were drilled, and indeed has a 1 m zone of Mississippian porosity with a reasonably high resistivity, suggesting oil may be there. However, this well and well 'Q' are too far downdip the "consequent" slope of the cuesta structure and are therefore wet. Wells 'R' and 'S' are, following the contour pattern analogy, simply too far off the Mississippian structure and are therefore wet in their porosity zones.

END PART 1 OF 2

Publication
Geologic Investigations of the Kandik area, Alaska, and Adjacent Yukon Territory, Canada, published by Alaska Department of Natural Resources, Division of Geological & Geophysical Surveys, 794 University Ave., Ste. 200, Fairbanks, Alas. 99709-3645.

This set of maps and reports details the most extensive preserved stratigraphic section in Alaska, with rocks ranging in age from Precambrian to Tertiary.

The report includes new geologic mapping, stratigraphic descriptions, structural cross-sections, hydrocarbon source rock analyses, surface metal surveys, radiometric ages, detailed gravity survey and derivative maps, and subsurface seismic data.

Norway
Statoil plans to spud its first exploratory well on the Vema dome in the Norwegian Sea.

Statoil obtained blocks 6706/11 and 12, in 1,100-1,400 m of water, in the 1996 licensing round. The acreage is in the Voring basin (see map, OGJ, Nov. 7, 1994, p. 92).

British Petroleum spudded a well in April on the Nyk ridge feature about 25 km east of Vema dome.

Poland
Poland is developing one of its largest oil fields, discovered in 1993.

Buszewo-Mostno-Barnowko, or BEB, field has 14 wells, most of which have been production tested. The reservoir is the Main Dolomite member of Cycle II (Stassfurt) of Upper Permian Zechstein at 2,980-3,107 m.

The Zielona Gora Oil and Gas branch of Polish Oil & Gas Co. plans to start production in third quarter 1998 at the rate of 650 b/d of 41° gravity oil and 3.6 MMcfd of gas per well.

The development plan calls for drilling a sufficient number of wells to ensure gas delivery to a combined heat/power plant and build gas treatment and fractionation facilities.

POGC attributes to BEB field 488 million bbl of oil in place and 1.04 tcf of gas in place, not reserves as previously reported. The field is in central western Poland near the town of Gorzow Wiel- kopolski (see map, OGJ, May 5, 1997, p. 132).

Yemen
Canadian Occidental Petroleum Ltd. drilled three delineation wells in first quarter 1997, sharply expanding Tawila field areal extent and hiking production to more than 70,000 b/d.

The company is acquiring 2D and 3D seismic data to identify more locations on the block.

Gulf of Mexico
Stone Energy Corp., Lafayette, plans to drill two more wells this year on South Pelto Block 23.

The No. 23 well, drilled from the C platform, flowed on test of limited duration 220 b/d of oil and 2.3 MMcfd of gas on a 12/64 in. choke with 8,280 psi FTP from a 15,700 ft sand that is a new field pay. The well logged 100 ft of net pay in five sands. TD is 15,900 ft.

The D platform is being built, and production is to start in fourth quarter 1997.

Arizona
Ridgeway Petroleum Corp., Calgary, quickly controlled a carbon dioxide blowout at its sixth well in the Holbrook basin near St. John's (see map, OGJ, May 5, 1997, p. 127).

The well in 16-10n-31e, Apache County, was air drilled to TD 2,798 ft. The well blew out while being prepared to be logged.

Oklahoma
National Energy Group Inc., Dallas, said its 34-2 Powers, in Northeast Carpenter field of Custer County, is flowing 3.46 MMcfd of gas on an 8/64 in. choke with 4,200 psi FTP from unstimulated Clinton Lake Atoka sand perforations at 14,464-560 ft.

When Atoka bottomhole pressure becomes compatible with a Red Fork sand behind pipe, National plans to stimulate Atoka, complete Red Fork, and commingle the zones.

Texas

Gulf Coast Seneca Resources Corp., Houston, plans to commingle production from stacked laterals in Cretaceous Georgetown and Austin chalk at a well in Burleson County.

The 7H Wilkins, in Jerrys Quarters field near Clay, Tex., flowed 2.91 MMcfd of gas and 2 b/d of oil on a 28/64 in. choke with 1,550 psi FTP from Georgetown. This lateral was still recovering 1,090 b/d of drilling fluid.

The Austin chalk lateral flowed 7.21 MMcfd of gas and 302 b/d of oil with 703 b/d of drilling fluid on a 28/64 in. choke with 2,740 psi FTP. Seneca has a 93.96% working interest.

West

Conoco Inc. completed an exploratory test in the Val Verde basin.

The 1-9 Culbertson, in Terrell County, flowed 9.418 MMcfd of gas on a 26/64 in. choke from Pennsylvanian Strawn perforations at 11,686-794 ft.

It is about 1/2 mile south of Conoco's January 1995 South Park field discovery, which has produced 1.35 bcf of gas and 18,700 bbl of condensate from Pennsylvanian at 10,848-860 ft through January 1997, Petroleum Information/Dwights reported.

PI said Conoco has completed almost 20 successful Wolfcamp and Strawn wells on CCSD&RGNG survey Block 2 since late 1994, including the Broken Antler and Little Mesa discovery wells.