Limited experiments on soil/sediment pH or lithologies have indicated the importance of these variables, but as yet have yielded no reliable method of correcting for these natural variations in the interpretation of adsorbed gas data. Research needs are suggested and discussed.

My objective today is to give you an overview of some exploration technology of which you might be unaware. I was employed at the Gulf Research & Development laboratory for nearly 10 years and at the Superior Oil laboratory for about 3 years. In both companies, I was involved in the development and application of geochemical pathfinder techniques in the exploration for buried mineral, oil and gas, and geothermal resources. For this workshop I will cover a number of examples regarding the detection of product leakage from petroleum storage facilities, important areas which demonstrate the application of some of the ultra-sensitive techniques that have been developed in leak detection technology.

I would also like to address the background concentration levels for light hydrocarbons, which have been established through more than 10 years of data gathering throughout the world. These natural near-surface hydrocarbon concentrations often exceed the levels that many state and other regulatory agencies are currently trying to impose on the industry. For example, the minimum detectable values that Florida and other states have set as normal background parameters are below natural ambient levels. These regulatory levels are being set lower and lower because of the increasingly sensitive measurement techniques, that now easily exceed these natural levels. It is important to be very cautious in establishing realistic background standards.

If one examines the database generated for petroleum exploration in many of the major oil companies today, you will find natural hydrocarbon seep levels that are far in excess of the regulatory values that are currently being set around these product storage reservoirs. These background, trace hydrocarbons, are measurable in the natural environment, for example in Glacier National Park and in other pristine places that have had no influence from man's activities. It is very important to integrate this exploration database with developing underground storage technology in order to avoid regulatory nightmares created by lower level limits which are uneconomic to maintain and obviously are impossible to control. This workshop discussion will also include underground coal gasification, an area where we have extensive published results available, in addition to considering data from underground storage operations such as salt domes, mined caverns, leaky well casings, pipelines, and petroleum product spills.

I will begin by describing the exploration technology, followed by storage leak detection technology, and finally I will discuss some of the sampling methods and their limitations. The Gulf Research and Development laboratory has supported a large geochemical research program since 1972 which was designed to develop surface and marine sniffer techniques for finding and mapping the ambient hydrocarbon levels in the natural environment, and then using this technology to define reservoirs and potential accumulation fairways. As shown in Figure 4.2.2.1, this research involved the sampling of free gases, from both near surface and marine environments. The onshore investigations include measurements of the free gases present in the atmosphere, and soil gases aspirated from a hole that is augered to varying depths. In addition to measuring the free gases, one can collect a sediment or water sample which can be brought back to a laboratory for analysis. Experience has shown that this laboratory sample analysis does not always reflect the leakage product. In many cases sediment samples were found to contain indigenous, naturally occurring hydrocarbons which were unrelated to the suspected leakage product.

The development of this free gas technology at GR&DC began in earnest in 1972 with the availability of the portable gas chromatograph, which allowed accurate measurement of the ambient background hydrocarbon levels associated with the natural environment. Comparison of these natural free gas measurements with those made on headspace gases from canned sediments clearly indicate that the headspace light hydrocarbon gases were noisy, fractionated toward oilier compositions, and generally less reliable for evaluation of buried products. Most of the data used for examples will involve only the light hydrocarbons -methane, ethane, propane, and butanes, and will not include gasoline range or aromatics, such as benzene, toluene, or xylene (BTX). There is, however, a lot of work currently in progress on measurement of these gases; perhaps we can discuss these BTX gases in the question and answer period. Experience suggests that the BTX gases are also often present in concentrations that are easily detectable, ranging up to the percent level in many of the natural seeps encountered in our light hydrocarbon gas database.

In trying various collection techniques to make the sampling easier and to maintain the samples integrity, we observed that headspace gases are less reliable than free gases collected in situ, involving both collection and absorption factors which are not well-defined. We tried analyzing headspace gases with fairly elaborate techniques, including circulating the headspace gases through a gas chromatographic sampling loop and periodically injecting into a GC column. Following careful measurement of the free headspace gases, we found that we could remove the soil from the cans and mix it in a blender and recover additional gas. If we took the soil sample after blending and put it into an acid bath, we would get still more gas. Experience with numerous samples from many diverse areas has indicated that these different gas fractions are often totally independent of one another, both in magnitude and in composition. The question then becomes - which gas fraction should be measured? Do we want to measure a specific fraction or should the total gas be analyzed?

Exploration experience indicates the presence of natural genetic gases present in the earth that have nothing to do with migrated products either from natural reservoirs or from underground storage reservoirs. Therefore, we must be very cautious to limit analyses to these migrated gases that are obviously the most important for evaluating losses from storage reservoirs.

John Hunt, who is currently one of the most noted and respected exploration geochemists in the United States today, made the statement at the 1981 national AAPG meeting in San Francisco that, "70% of the known hydrocarbon reserves in the world can be related to macroseeps." If we examine the microseep technology that is available today, you will note that we measure concentrations well below the elves that are being set by the regulatory agencies, and of course, this may yet be extended to levels below parts per billion. In defining these natural hydrocarbon levels, we have found that they are ubiquitous, they are present everywhere at levels that are measurable, and the first thing that we must do is recognize that fact. Gas chromatograms from over a known field and from a background area in the San Joaquin Basin, California (Figure 4.2.2.2) show natural hydrocarbon concentrations that are quite large, up to 1382 ppm in this case. Exploration studies have encountered many examples of natural seeps that are in the percent range.

The magnitude of these natural seeps observed in exploration geochemical programs are not a function of whether the deposits are economic or not, magnitudes are governed by tectonics, not economics. Geochemical prospecting is not a prospect tool per se. One cannot drill directly on the seep anomalies. The migration pathways may be tortuous and displaced laterally by fractures and other permeable zones. We have found hydrocarbons present in the near-surface soils in every productive hydrocarbon fairway that we have examined, recognizable at levels ranging from ppb up to percent. Natural soil gas compositions are real and repeatable, and have been used to predict the type of production, oil versus bas, that occurs in each basin surveyed. Extensive research at Gulf Research & Development Company has repeatedly demonstrated that free gas microseeps have compatible information to their associated reservoirs, Teplitz and Rogers (1935), Jones (1979), Janezic (1979), Mousseau & Williams (1979), Drozd et al., (1981), Richers (1984), and Price & Heatherington (1984), even including artificial underground gas generation and storage (Pirkle & Drozd (1984), Jones & Thune (1982) and Jones (1983).

A decidedly brief description of this soil gas geochemical technology follows. I would be pleased to recommend additional reading material for any interested parties. An illustration of the database generated and published on surface geochemical prospecting is shown in Figure 4.2.2.3, a U.S. basins map with black dots representing the surveyed areas. We have worked in most of the U.S. basins and have found that in the Sacramento Basin, for example, all the source rocks generate gas and all of the seeps are gas type. If we sample in the San Joaquin Basin, where there are oil and gas reservoirs, then we observe ethane, propane, and butane in much larger ratios relative to methane than in the Sacramento Basin. We can actually predict the source difference between the two basins based on soil gas compositions observed in shallow 3 meter (12 feet) deep augered holes. The GR&DC data has clearly demonstrated the usefulness of soil gas data for distinguishing compositional differences within and between basins. Predictions as to whether a potential basin contains gas or oil reservoirs is of obvious value in an exploration program.

The soil gas sampling technology developed at GR&DC is commercially available for gathering this type of exploration data and for evaluation of volatile product leakage from underground storage reservoirs. We chose to measure the most easily migrated (C1-C4) gases and are currently developing techniques for benzene, toluene and other gasoline range hydrocarbons. Exploration surveys involving methane, ethane, propane, and butane have indicted that the ratios of those gases change in response to the reservoirs, which obviously change in response to the source rocks which filled these reservoirs. The proof of this information requires conducting surveys that are large, basin wide, and statistically valid. Examination of such large surveys, involving at least 1000 to 3000 well distributed samples, indicates that the surface anomalies are not nearly as spotty and irreproducible as originally thought by those who have seen only limited data.

The gas chromatographs used at GR&DC to develop this database were internally built by Gulf technicians using a state-of-the-art technology which was never patented. This specialized gas chromatograph has a 3 foot alumna column which gives very good separation between methane, ethane, ethylene, propane, propylene, and iso and normal butane. These are all important gases for evaluation of product storage areas, as well as for surface prospecting. Two pictures illustrating a typical gas chromatograph installation, in the back of a Dodge power wagon on the bottom, and a Ford 4-wheel drive van are shown in Figure 4.2.2.4. Exploration Technologies, Inc. was formed in order to offer this technology to the oil industry and is currently composed of all ex-Gulf personnel.

We have conducted numerous studies using sediment samples collected in diverse geological situations, in addition to free gas techniques using in situ measurements. The free gas techniques appear to provide the most reliable data, which always matches the products known to be stored in the associated underground reservoirs. A large part of the success of these early studies can be attributed to the real time application of this technology which allows directing and confirming the significant anomalies in the field using the limited sampling time available rather than to depend on having only a few significant samples which may not cluster because of the limitations involved in collecting and returning samples to the laboratory for analysis.

Physical limitations to sample collection are generally imposed by the terrain, cultural activity, and the local environment. A major influence on any particular application comes in interacting, given the situation, and devising a method for taking a valid sample for measurement. This sampling problem is quite different depending on whether soils or waters must be collected.

The free soil gas techniques, originally developed at GR&DC requires drilling a hole to a specific depth, in most cases 12 foot, putting a packer in the ground to seal the hole from the atmospheric air, pumping the gases through a gas chromatograph, and analyzing the light hydrocarbons. This is shown diagrammatically in Figure 4.2.2.5. We have often expanded that analysis to include carbon monoxide, carbon dioxide, helium, hydrogen, mercury, radon, and a variety of other gases of interest in the natural environment.

The small power auger in the foreground in Figure 4.2.2.6 is the type of drill initially used and is actually drilling in the San Andreas fault zone. The metal covers in the background are tilt meter stations on Gold Hill in the Chalame Valley. This example provides the first case in which we measured helium associated with the San Andreas fault; helium is often observed as a potential earthquake prediction gas (Wikata et al., 1978). Other examples from the database illustrated on the U.S. basins map include sampling in hard rock terrain such as in the Copper Ridge dolomite in the folded belt in the Appalachian basin (Figure 4.2.2.7). In addition, we worked in foreign countries such as Oman (Figure 4.2.2.8), sampling in very sandy lithologies and arid environments.

We have found that the free soil gas technique yields the most reproducible results for defining the known source types in all of the areas surveyed, either in the U.S. or in foreign countries. Additional research is still ongoing for developing better collection methods for soil samples. Later in this paper I will describe some of the techniques that were developed for improving reproducibility for soil core analysis. The least reproducible technique which I will describe involves the headspace analysis from canned samples containing sediment cores. These headspace analyses can give inconsistent results, often exhibiting variability of over 150%. Loosely adsorbed and acid extraction techniques will also be discussed. Theoretically all three techniques should be equivalent. The problem lies in the fact that the desorbtion techniques do not always match the known storage products, as do the free gas measurements. The application of various extraction methods often yields populations of gases which are independent of one another. These different gases are independent because they were implaced in the rock matrix at different geological times. Therefore, it is logical to expect that the free gas method should analyze gases which were introduced into rocks in the most recent times, therefore, they are most likely to be directly related to current events.

Let us briefly consider the measurement techniques developed for gases dissolved in the offshore water column. GR&DC conducted marine geochemical surveys on a continuous basis over a number of years aboard several different research vessels. As shown in Figure 4.2.2.9 sampling was conducted simultaneously from three depths; water is pumped from a hull inlet, a mid-tow inlet from about 300 feet and a deep inlet from about 600 feet. The gases dissolved in water are purged through a stripper in which helium or nitrogen is sprayed into the chamber aiding the removal of gases from the solution. The gases are picked up by a carrier stream which delivers them to a gas chromatograph while the water is dumped through a drain. The three gas chromatographs are set up to analyze the gas stripped from the water every three minutes. Using three gas chromatographs to analyze the samples for each of three inlets reduces the risk of not having working deep tow system and greatly enhances the opportunity to produce the database shown on Figure 4.2.2.3 on the U.S. basins maps.

This technique works very well and has been discussed by Mousseau (1979) who related the gas values observed on all the continental shelves to those published for the open ocean values. The marine sniffer system is a sophisticated and continuous water headspace technique and provides an excellent way to characterize gases in water. In addition, we have put marine strippers like this on continuously pumped water wells, and analyzed the dissolved gas concentrations over extended time periods in order to measure the gas content of subsurface waters.

Differences between surface contamination from production platforms versus deep, natural microseeps are illustrated in Figure 4.2.2.10. The upper portion of Figure 4.2.2.10 exhibits the contamination from a producing area in the Gulf of Mexico, sampled through the hull inlet. At the bottom of Figure 4.2.2.10 can be seen three seeps emitted from the bottom. These are the sharp spikes which are termed localized anomalies. As shown, they obviously originate from the bottom and bear no relationship to the surface contamination.

Subsurface water samples must be collected under static conditions, just as they exist in the natural environment in order to properly analyze the dissolved gases. One should not draw a vacuum on the water before analyzing the light hydrocarbons because of the loss of volatile components from the solution. Very poor and nonreproducible results are obtained by using vacuum techniques to draw up the sample. It is very important to pump the samples under pressure to the surface, or even to bale them from the subsurface.

In contrast to sediments, headspace analysis of water is very straightforward. The methods published by McAuliffe et al., (1963, 1966, 1969) are perfectly adequate. There is no problem with analyzing gases in waters because all that one is dealing with is solubility and a matter of exchange, where temperature and salinity are the major factors, if you want to be quantitative. However, as soon as we attempt to analyze soils, for example from sea bottom cores, a problem develops because gas trapped in soils does not extract in the simple manner that waters allow. Solubility and salinity are no longer the major factors, as the trapping of gas in sediments does not appear to follow any easily described rules.

Large sub-bottom seeps often yield a geophysical anomaly which is referred to as a bright spot as shown in Figure 4.2.2.11. These bright spots are potential gas hazards. A drilling rig placed over a bright spot might be subject to explosion or a fire created by uncontrolled gas blowouts. Such blowouts could even cause the formation to breach, causing the loss of the drilling rig. A series of bright spots marching down a fault might indicate a gas migrated from a potential reservoir, along a fault or fracture, to a near-surface horizon where it then forms a micro-reservoir in the near-surface environment. It is this near-surface micro-reservoir which is the source of the surface geochemical anomaly. Such seismic bright spots are the result of the larger scale tectonic features and their associated macroseeps. Geochemistry is, however, capable of detecting much smaller hydrocarbon leaks, seeping along micro-fractures, which are not detectable by geophysics.

Because of gas dispersion in the water column, it is often desirable to collect point samples from the bottom in order to establish the exact location of the sea floor seep. As a substitute for free gas data, techniques have been developed for collecting and analyzing bottom sediments using dart or piston cores.

A classic exploration example is shown in Figure 4.2.2.12 as a profile of methane, propane, and butanes (the solid line is propanes plus butanes and the dashed line is methane). This data was collected over the Pineview field in the Wyoming Overthrust Belt in 1976 at a time when the field was fairly new and would not be expected to have much in the way of contamination problems from leaking well casings. Values as high as 80 ppm propanes plus butanes are found at the 12 foot depth in these holes.

The Pineview area also provides some initial tests of probe free gas sampling techniques. A table containing data collected with a probe driven one foot into the ground from three of the low areas and two of the high areas is shown in Figure 4.2.2.12. This experiment was conducted to ascertain whether or not a real flux of hydrocarbons was escaping into the atmosphere along this traverse. (The anomalies at sites P-2, P-6, and P-26 confirm the anomalies previously noted three years earlier in the 12 foot deep holes.) As shown, soil gas values at one foot depth are quite low in background areas; for example, propane and normal butane are about 28 to 52 ppb, respectively. However, the anomalous areas range upward to about 200 to 400 ppb at only one foot down in the soil.

We have often observed very large microseeps (some of which could be considered as macroseeps) within many of the large databases that we have gathered. These large microseeps can be used, not only for exploration, but also for evaluation of the presence of gas in the ambient environment. Samples from such areas could also be analyzed for such priority pollutants as benzene and toluene, and I am sure that we would find that they are often above the ppm level as well. A recent paper, Anderson et al., (1982), lists values in the percent range for aromatic hydrocarbons present in an oil seep in the Green Canyon area of the Gulf of Mexico. That is not an isolated case. We find these seeps within many areas of the Gulf of Mexico, the California offshore, and the Alaskan offshore. These examples provide only a part of the evidence from surface geochemical exploration programs that can be applied to the needs of the environmental industry.

Now, let us look at a few cases where we have tried to apply our seep technology to production leakage situations. These applications might be spills, casing leakage from wells (a very important aspect, leakage from underground storage reservoirs, and/or underground coal gasification reactors. Although most of the following examples were service projects, part of the information derived from performing this type of work was used to provide research information on environmental contamination as an aid to the exploration programs.

One of the important questions that the pipeline and natural gas storage industry would like answered is, just how important are natural seeps as compared to contamination from man-made sources? To provide the answer one must consider both the exploration and storage survey data. A depth profile of propane taken over two different crude oil spills is shown in Figure 4.2.2.14. As shown, ppm levels of propane are present at the very near-surface, however, the level decreases quite dramatically with depth because the source lies at the surface, rather than at depth. Many observations over well pads that have had minor oil spillage have exhibited profiles of this type. However, observations over a brand new well casing with a faulty cement job exhibited a very large seep on the well pad (Figure 4.2.2.15), which increases with depth to fairly substantial values. This well had a leak at about 7000 feet and was repressurized with oil from a hot oil tank. In this case, the seep was almost entirely propane and butane rather than ethane or methane, because of the not oil source, which would not be expected to contain significant light hydrocarbons.

Another application example shown in Figure 4.2.2.16 is taken from a survey over an underground propane storage reservoir, a mined cavern about 200 feet deep. This figure shows a log scale of propane collected over the cavern. An interesting secondary observation observed on this survey was an obvious color change noted on the soil cores. The soil was noted to change from a red-brown to a green-black directly over the top of the cavern where the largest seepage anomalies occur. These chemical changes are related to hydrocarbon seepage and might be used as an additional exploration tool to provide evidence of where the gas leakage has occurred around any type of storage cavern.

A storage cavern like this is also a good place to observe gas flux related to atmospheric phenomena. We have constructed plastic ground sheets, about 5 feet square, and have measured the gas flux related to meteorological and barometric conditions in order to answer the question - what are the effects of rainfall and barometric pressure? Figure 4.2.2.17 shows the variation with rainfall as vertical bars. A very large seepage anomaly is shown by the dashed line at the right edge of the first bar. The rain probably displaced the gas in the ground and caused it to come up underneath the ground sheet. However, as shown, we don't always see the same effect every time it rains. The monitoring of the barometer is shown as the very dark solid line at the top of Figure 4.2.2.17. In the central portion of this line, around the 19th, 20th, 21st, and 22nd days of the month very small barometric changes were observed. Note the small barometric lows that are expressed in the dark curve line and right underneath, in the light dotted curve line, are a couple of glass flux observations under ground sheets which have an absolutely direct relationship. Every time the barometer takes a little dip, some gas flux pops up under the ground sheet. Thus confirming that there is gas flux from seeps that escapes into the atmosphere.

A positive gas flux area such as a propane storage cavern provides an excellent opportunity to conduct some needed research on the relationship between the adsorbed and free propane in soil samples an soil gas. This experiment was suggested because we observed that canned samples which previously had headspace propane values of 20,000 to 40,000 ppm, had no detectable propane in the cans after only eight months. Although this is certainly longer than the 21 days we were talking about for holding VOA's, this still suggests a fairly rapid oxidation reaction for the propane, and provides that large microbiological colonies exist in the soils which are capable of completely consuming all the propane within the enclosed cans. After noting these microbiological oxidation results, we thought that this area would be a good choice for comparing free versus adsorbed gases from a known macroseep area, since perhaps only the free gases would be oxidized. These sample cans were sent to Horvitz Research Labs for acid-extractable total gas analysis. No adsorbed propane was found, suggesting that either microbes consumed both the free and adsorbed gases, or that the local soils do not adsorb propane under these specific EH/pH conditions.

To further test this theory, we went back and collected fresh canned samples from one seep area over the leaking reservoir and from a background site off the reservoir. Analyzing a fresh sample still hosed no adsorbed propane even when the sample was only two days old. These experiments were very interesting and suggested that we have microseeps with tremendous gas flux (large enough to attract and kill blow flies at the surface in this case); yet there is no adsorbed propane on the soil. If we had gone out and gathered cans of sample, then had them analyzed by some of the adsorbed gas techniques that are available in the industry today, we would not have found any evidence of the gas leakage. This makes one aware that all the potential measuring techniques are not equal, and may in fact yield spurious or invalid results. At this time, we don't understand all the factors which control gas uptake by soils.

Because of the obvious presence of microbes, we thought we would experiment with some of the bacteriological preservation techniques. Preserving samples for analysis is one of the techniques. Preserving samples for analysis is one of the industry's important problems because of our inability to analyze all the samples in real time. For example, if we need to collect samples in the Columbian jungle, it is very difficult to do the analysis on-site, and so the samples have to be transported to a laboratory for analysis. One microbiologist, Ron Ormland of the USGS that I talked with, told me there is no known guaranteed preservation technique. Microbiologists generally suggest one must steam the sample to kill all the organisms or freeze the samples. Steaming will clearly kill most of the microbes, but will also drive out all the volatile hydrocarbons, or perhaps even generate new products, both of which are undesirable results. Steam treating the samples is obviously inappropriate to our needs.

Both zerphiran chloride and methylene chloride have been suggested to be effective biocides. To test methylene chloride preservation techniques, we conducted another experiment in which we took 100 cans of sample, we ran one sample every week for 100 weeks, some at an outside lab, some at our own lab, and found larger quantities of biogenic methane in methylene chloride-preserved samples than in samples that were untreated. The generation of biogenic gas appears to be aided by the methylene chloride. Thus proving that one of the published sample preservation headspace techniques is inappropriate.

Additional experiments were conducted on sample collection techniques in an attempt to reduce analytical variability. For these experiments we collected over 100 cans using different collection techniques. We tried to can all the samples identically, so that the headspace conditions would be exactly the same in order to reduce variability from the headspace technique. The standard method used in the oil exploration industry for can headspace analysis is to volumetrically fill the cans with aqueous solution, followed by adding soil to displace a fixed volume of solution, leaving a fixed volume headspace. Even using this technique, we found that the results were still variable in this macroseep area, giving ± 150% variations in concentrations. Despite several additional efforts, we could not collect the cans such that their headspace analysis gave values repeatable as free gas (± 50%) even when dealing with a soil that was heavily gassed with propane. On the second experiment, instead of just filling the cans with individual aliquots of soil, as we did on the first experiment, we took a 5 gal bucket and homogenized a large soil sample and poured sample aliquots into each can. We figured that the soil slurry ought to provide more uniform results. This approach definitely improved the analytical results, however, the analytical variation was still ± 100%, Holding the volume constant was found to be very important in order to get the most reproducible results. However, that still did not remove all the variability. Perhaps the lithology of the soil is not uniform; we are not dealing with a homogeneous system which suggests that we had better be cautious in analyzing natural soil samples.

Another interesting factor in analyzing canned samples is the time lapse required for various gases such as helium to diffuse gases and has one of the smallest sticking coefficients, and thus should be fairly quick to get into the headspace. Reimer (1980) has reported a period of 18 days for helium to escape from dry soils in vacutainers. Some of our own experiments have indicated that the equilibrium time for canned headspace helium to escape from the soil at room temperature was nearly 24 hours without heating and shaking. Some form of mechanical disaggregation and heating are commonly practiced in order to get the volatiles into the headspace. Any time one deals with a soil matrix, there is the problem of never knowing if you have removed all of the component being measured. This can become a fairly severe problem if you are trying to do quantitative work.

A good caption for Figure 4.2.2.18 would be "preparation meets opportunity". Our capability in soil gas exploration was appropriate to the next example which really relates to the purpose of this conference. Let us consider the case of a leaky well in a hypothetical salt dome in which a product loss occurred from about 600 feet deep, through the well casing, and into the cap rock. The product migrated out into an adjacent town requiring the evacuation of people and creating both political and technical difficulties. Soil gas geochemistry was successfully used to evaluate the leakage situation and help locate the lost product, which could no longer be recovered from the storage well since it had migrated up through the cap rock and into the sands that overlie the salt dome.

A surface hydrocarbon survey was conducted using geochemical test holes drilled 30 feet deep to encounter the shallowest groundwater aquifer. Over 500 sample stations at 50 foot centers were installed throughout the town and over the general storage area. Figure 4.2.2.19 shows the location of the leakage well as a black dot. As shown in Figure 4.2.2.19, the lost product migrated updip toward the top of the dome and 3000 feet laterally along migration directions predictable from surface and photogeological studies. Major lateral leakage appeared to follow sand channel which may also reflect a fault related to the salt dome. The surface topographic contours on Figure 4.2.2.19 clearly show the top of the salt dome. The initial geochemical sampling was extended beyond the product leakage boundaries in order to allow a good estimate as to the location of the leakage product.

The geochemical sampling stations were installed using PVC pipe with water well screens on the bottom, and valves on the top, allowing them to be used for both leakage gas detection and later injected with nitrogen. These stations thus served a dual purpose. The leakage product apparently migrated into a sand layer at about 200 feet in depth, and from there, vertically pressured the shallow sands located only 30 feet below the surface. The first two relief wells were drilled to 100 foot depth along the line of this lineament and flared in order to release the subsurface pressure. More than thirty relief wells were drilled over the anomalous area in order to relieve the problem.

Another point of interest is the fact that this was a special product, an ethane-propane mix which has a unique signature as compared to the normal hydrocarbon products originally found in the reservoirs associated with such salt domes. The hook-shaped seep shown in Figure 4.2.2.19 turned out to be all propylene, and was traced to a previous spill, as documented from the historical records, which was previously thought to have no known surface expression. In this case, the company which lost the ethane-propane mix had to clean up the entire site.

The cleanup operation was facilitated by turning the 30 foot geochemical sample sites into eductor sites by installing a venturi tube on the top of the well casing. Nitrogen is run through the venturi tube to produce a small vacuum on the hole. One can then inject nitrogen into the ground in advance of the leakage area and allow that nitrogen to migrate through the ground toward the eductor sites where it escapes to the atmosphere. To aide in following this N2 flood process, we converted some of our helium/hydrogen chromatographs, used for exploration seep detection, to nitrogen/oxygen detectors. A typical response curve, observed, during this process is shown in Figure 4.2.2.20, along with type concentration of the product. This figure illustrates the rapidity with which the nitrogen front passed through the 30 foot aquifer in the area of this site. We conducted the nitrogen flood to remove and push the product out of the 30 foot sand, after it was relieved in the 200 foot sand, to prevent any recharge.

The time rate of cleanup response over the entire area, shown in Figure 4.2.2.21, burned out to be quite interesting for illustrating the variation in lithology and permeability, which was not uniform. In some cases the permeability was so low that 15 lbs of nitrogen pressure at 30 feet was not sufficient to push the gas 50 feet laterally over to the next soil gas stations; yet, if you poured a bucket of water on the ground near the injection site, the ground would froth and bubble from nitrogen that was escaping vertically through 30 feet of clay. This is in spite of the fact that the nitrogen would not travel laterally to the next eductor site. Some of these hot spots probably resulted from deep vertical migration rather than lateral migration through the 30 foot sand. These zones were the most difficult to cleanup.

To summarize, a soil gas survey was used to outline the leakage area, deep wells were drilled to relieve the product at depth, and a nitrogen flood was conducted for final surface clean-up. In addition the compositional data indicated whether the soil gas was from the product well or from a previous spill or pipeline. For example, we also detected several natural gas pipeline leaks in the city that had nothing to do with the storage product leakage. This latter application particularly points out the value of the gas chromatograph. Often consultants who are called in on these type jobs use small thermal conductivity meters which measure combustible gas. The soil gas is aspirated through the measuring coil in such instruments and is detected by a deflection on the meter. Because of the very poor sensitivity and selectivity of such instruments, a deflection of two divisions was observed to indicate a problem area. However, some predetermination is required to determine what the meter is responding to when small deflections are noted. For example, nitrogen used for the flood causes the same deflection on the meter as low levels of the leakage product. Many of the areas that were originally depicted as "hot" based on the LEL meter had no product in them whatsoever, but were pure nitrogen. If we had not used the gas chromatograph in this case, the cleanup operation would have lasted a lot longer and would have included attempts to clean up the nitrogen as well.

Following the clean-up operations, we installed an advance warning system to catch leakage from other storage wells before they can become a problem. As shown by the triangles in Figure 4.2.2.22, we installed at least two permanent stations as monitoring sites at every storage well casing. It would be better to have four geochemical sites per storage well, but economics always plays a role in such considerations. This type of monitoring system, if properly used, and sampled on a regular basis will allow early detection of well leakage before it appears at the adjacent boundary of the property. Application of such a system can be very beneficial for preventive maintenance and is highly recommended.

Another excellent example is taken from a study conducted over a 200 foot deep propane storage cavern. This particular site was a fairly small facility, as indicated by the scale on the right side of Figure 4.2.2.23. We put in 455 geochemical measuring stations to a depth of 20 feet using PVC pipe placed on 10 foot center.

This particular case provided an opportunity to determine the product leakage distribution and to conduct pressure pulse tests by injecting helium into the cavern as a tracer. It was fairly obvious that the main leak was coming from around the central shaft. Our immediate objective was to test the leakage rates and to determine whether or not the remedial efforts were successful.

Several opportunities to learn from this experience were encountered during this remedial period. The cavern pressure was decreased to ambient levels allowing us to measure the recharge leakage rate from the cavern. Upon recharging the cavern, we were able to measure the time that it took for the gas to appear at the surface. Following the recharge of the cavern with propane to its original pressure, we observed a value of over 90% of the original soil gas propane concentration. This occurred in a little over 15 days. As shown, most of the leakage came from around the central shaft. However, there is a large propane background in the soil, since the storage site had been there 20 years. This propane background makes it difficult to be sure the product we were now seeing at the surface came directly out of the reservoir during the sampling time period.

As a second more definitive test, we injected helium into the cavern, at a concentration of about 600 ppm, and found that in 15 days, not only had we moved the product, but as shown in Figure 4.2.2.24, we also detected helium at the surface. Helium showed us not only the leakage around the central cavern, but also exhibited a leak at the end of one of the drifts that we would have missed looking only at propane. This amount of helium is not enough to damage the product for sale, and yet still gives one adequate sample concentration for analysis.

This data suggests that migration is quite rapid, in other similar cases, for example from an underground coal gasification reactor, we have been able to establish migration times of from 2 to 15 days at depths of up to 1000 feet for changes of gas concentrations in the reservoir or cavern to be expressed at the surface. This means that migration does not follow a diffusion model, but is driven by pressure, along fault and fracture patterns and joints. The migration phenomenon is obviously pressure-driven, since diffusion would take several years to show up in most such situations.

Another very educational example is taken from an underground coal gasification reactor near the Rock Springs Uplift in the Rawlins, Wyoming area. In Figure 4.2.2.25 the shaded area shows the facility boundaries which lie over nearly vertical beds. These coal beds dip too steeply to do anything except gasify the coal in-place. The gasification pit lies about 600 feet vertically below the surface. We found during evaluation that bedding plane leakage and joint leakage along the dominant and subordinate joints are easily expressed in the resistant sandstones, and can be mapped from the surface. In this particular study, which was funded by the Department of Energy, the data is in the public file and has been published in part if anybody wants to look at it in more detail (Jones and Thune, 1982). Some of the plots constructed from this data show migrational chromatographic effects that were measured between wells a the gases moved through the earth. Figure 4.2.2.26 shows the hydrocarbon flux of methane plotted versus Julian days, with vertical bars separating sections representing different time periods.

We monitored this reactor every day for one month before the coal burn was started and monitored the site continuously until the initial gas flux from the reactor first appears at the surface. This required only two days. We continued to monitor the gas flux on a daily basis for about eight months, with a final measurement made over a year after the initial sampling. As a way to determine the flux of gas migrating to the surface, we installed permanent stations, such as shown in sites 22 and 27. These permanent monitoring sites consists of PVC pipe 20 feet deep which are about 50 feet apart. As shown in Figure 4.2.2.26, these two adjacent wells look very much alike in their leakage rates and patterns. Two other wells, sites 1 and 2 in Figure 27, show no sign of gas flux, yet obviously contain contamination levels of about 10,000 ppm which was apparently emplaced during an earlier retort that was 400 feet deep and located on strike a little to the southeast.

These data allow the determination of how long gas remains in the surface sediments, and on what magnitude of fluxes might occur under various pressure conditions. In monitoring the free gases, several permeability tests were conducted in which the gases were pumped down to ambient levels to see if they would recharge. The maximum gas values were observed at sites 22 and 27, which are directly updip of the outcrop of the coals being gasified. This maximum value occurred in the sandstone directly over the coal with 100,000 ppm methane showing up at the surface. Figure 4.2.2.28 illustrates the location and variety of seeps observed at the surface. Sites 85 and 86 had about 1200 ppm natural levels of hydrocarbon that had nothing to do with the retort. It was a natural seep which existed previously in the area. We pumped out the residual gases and did not contaminate this site during the retort process.

Using the observed values of propane flux from various areas at the site, one could construct a series of contour maps illustrating the changes every three days. If we had the aid of Walt Disney's illustrators, we could make a movie which demonstrates the movement of the gases at the surface as the retort pressures changed. The seep located to the northeast, right above, and along the baseline, provided one anomaly that was not influenced by the gasification process. In the central portion are the two vertical product wells and to the northeast, perpendicular to the strike, are the two other product wells. Also in the central part is a very small seep which follows the bedding plane along the direction of this seep. Note the vertical leak directly over the retort. As shown in Figure 4.2.2.28, the leakage patterns changed with time as did the pressure in the retort. The maximum pressure in the retort at 600 feet was 700 psi. This excess pressure caused an immediate change in the surface signature that occurred within two days.

Helium was occasionally injected during the stable part of the burn. This helium appeared in the seep gases at sites 22 and 27 within two days of its injection. As shown in Figure 4.2.2.28, there were at least three major leakage centers along the bedding plane, indicating that the leakage gas did not just migrate updip along the bedding plane and then migrate laterally along the bedding planes to fill the surface sediments with gas, but rather came up almost simultaneously in three different places. This suggests a fracture migration pattern with gases migrating along fracture avenues at depth, which were then communicated along vertical fractures to the surface, similar to what we noted in the case of the salt dome. In this case, the vertical migration zones were obvious during both the charging and discharging periods as the surface gas was depleted. I suggested that the Department of Energy should maintain this site for future research on soil gas analysis, since one could conduct various depth probe measurements to check the influence of joints and soil types on both vertical and lateral gas migration, and further investigate dissipation with time. This site is particularly useful for a research study because of the uniqueness of the gases; for example, carbon dioxide, carbon monoxide, hydrogen, and methane are the dominant gases. Because they are unique to the retort process, we know where and when they were generated. The maximum concentration of gas in the bedding plane leak is about 50,000 to 100,000 ppm at the peak generation of the retort, and falls off as the reactor pressure is reduced and finally filled with water. One can see the concentration decrease until they are well below 10,000 ppm a year later. This provides adequate information one can use in flux calculations for modeling, providing an excellent resource for further study. I have suggested that a graduate student at one of the coal research centers, such as at Laramie, should consider this modeling as a thesis project.

For comparison, Figure 4.2.2.29 shows an example of a natural seep at Glacier National Park that was encountered some years ago. The seep values ranged up to 108,500 ppm methane and over 1000 ppm helium, and was under positive pressure when encountered in a surface geochemical drill hole. Such large values are not very casually obtained from most surface samples. In this case, samples were taken every five feet along both the north and south sides of the highway. It was interesting to note that the anomalous gas values were displaced to the west on the north side of the highway. The offset occurred exactly along the strike of major faults associated with the disturbed belt of Montana which cuts through this area.

These natural leakage magnitudes were, in fact, larger than we found in our worst case situation with the underground coal gasification reactor, which had already decreased a factor of 10 below its initial values in less than a year. With improved casing and completion techniques, and with the knowledge of where the seeps are occurring, and the willingness to do a little remedial work, we could probably reduce the leakage to one-half or one-fifth of what it was in some of these early test cases. As you can see, some of the natural seeps exceed what you find in a pollution situation.

Research scientists at Arco, Cities Services, and Phillips Petroleum jointly conducted a research study which involved collecting samples of shallow 6 foot deep soil cores and soil gas in vacutainers using hand augered holes. Vacutainers are available from the medical area where they are used to take blood samples. Their biggest drawback is the fact that they cannot hold more than 7 to 8 psi positive pressure without leaking. In fact, at 5 to 8 psi positive pressure, the rubber seals will slip off if they are transported by air freight. They require constant atmospheric pressure to keep the lids intact. Serum bottles with crimp on lids provide much better seals for gas samples.

The initial study of Patrick Draw field was conducted as a GEOSAT test case and provided interesting data, but lacked adequate regional coverage. A second survey was performed by Richers, et al., (1986, in press AAPG Bulletin) which uses a better technique involving a pounder bar very similar to the KV type sampler that Dave Conway talked about in the first session. One hammers this bar into and out of the ground, inserts the probe into the ground in the pounder hole, and collects the sample in a small (150-200 cc) serum bottle by aspirating the free gas using a hand pump mounted on the soil gas probe. These serum bottles have crimp-on caps and can maintain up to 30 psi pressure. They are excellent gas containers; a sample of light hydrocarbons can be stored in them for several months without any serious degradation or loss problems. These bottles provide a simple way to bring back a subsurface sample to the laboratory for analysis, or over to a real time instrument truck, if the terrain does not permit access to the site.

In my estimation, free gas samples have always appeared to provide the very best quality for comparison to the original stored products. Collection and measurements of other types of samples have often proved to have serious problems. In an attempt to answer these collection problems, we have tried numerous experiments involving headspace gases, loosely adsorbed gases freed by disaggregation, and even tightly held gases freed by acid extraction.

The simplest disaggregation methods involve using a blender in which one places a core sample, disaggregates the sample, and purges the gas from the blender into a gas chromatograph. We found that if one keeps running the blender, one continues to extract gas. When do you stop? If we take the sediment material out of the blender and treat it with acid, we can get still more gas. To avoid this dilemma, we tried to develop a disaggregation technique that would be a little more reproducible.

We designed a small ball-mill system with a 35 cc volume having a septum side arm. We used ceramic balls in this mill to grind up the sample. The sample is collected as a core, which is put into a can, sealed and brought back to the laboratory, handling the sample as little as possible. To get an aliquot of sample for analysis, one must split the core, obviously exposing the sample. Volume displacement is probably the quickest and easiest method for measuring the sample which is then transferred into one of these ball mills. The ball mill is shaken on a Spex mixer that agitates the sample for about three minutes, before being placed in a hot water bath for three additional minutes. A constant temperature of 95oC gives reproducible results and buffers any heat-effects from grinding. One can then take the sample, remove an aliquot of the gas through the septum, and inject it into a gas chromatograph. This technique gives approximately 15 or 20% reproducibility. Thus, one can sample and resample the same core and get excellent repeatability. We tried to avoid using canned samples that were transferred by hand from the ground to the can. Even with the most diligent effort, the light gases will be lost. Although the ball mill method obviously loses some gas during collection, it is at least more reproducible than can headspace methods.

Some results comparing shallow probe to ball mill and auger drill (GR&DC soil gas) follow. The samples chosen to test these techniques were collected in the Rose Hill field in the Appalachian overthrust and from the Gulf Research laboratory. A comparison of data from Rose Hill using these three sampling methods is listed in Figure 4.2.2.30 (a 12 foot free gas analysis, disaggregated ball mill samples, and shallow 4 foot sample probe bottles) and shows a concentration of 241 ppm methane from the free gas analysis, 129 ppm from the disaggregated samples, and still lower values from the shallow 4 foot probe sample bottles.

In spite of the magnitude differences, these techniques suggest a common source. For example, percent gas wetness shows that the composition of the gas from these two different areas, using each of these techniques, was fairly uniform, but different in the two areas. As shown, it is about 30% wetness for the disaggregation samples and about 26% for the probe and 28% wetness for the free gas technique. Even though we are measuring different gas samples, taken from different depths at different times, we find that Rose Hill is a unique oil area. The Rose Hill oil field produces from a Trenton oil reservoir suggesting migration of light gases through Cambrian age rocks which are thrusted over the Trenton reservoir. This shows that the gases we observed at the surface are migrated gases, as opposed to syngenetic gases that did not migrate from a reservoir, but rather are related to near-surface source rocks. Data from numerous surface geochemical surveys published by Jones and Drozd (1983) have demonstrated that the soil gas hydrocarbon compositions and gas ratios are representative of the different source types of the different areas surveyed. For example, dry gas areas are about 95% methane or greater. In the Rose Hill area we compared three techniques. The adsorbed and 4 foot probe data was collected in 1983 while the 12 foot deep drill holes were collected in 1977. As shown, all three techniques found very similar values for the gas compositions, both as percentage ratios and the ratios of the gases methane to ethane and propane to ethane.

During the natural gas shortage several years ago, Gulf Research drilled three shallow (approximately 3000 foot) gas wells in order to avoid having to covert to coal for heating the laboratory. The wells were located by surface geochemical techniques and geologic fracture mapping, two out of these three wells produced gas from the Speechley formation. As shown in Figure 4.2.2.30, both the 12 foot deep augered holes and the adsorbed gas (by disaggregation) techniques agreed with one another, and defined the laboratory area as a gas area. In addition, both the chemical composition and the methane isotope values of the surface seeps observed before drilling were found to be in excellent agreement with the subsequent production.

Both the shallow probe and soil disaggregation techniques were designed to be performed by graduate students so that the cost of doing surveys could be reduced, allowing samples to be taken from a 4 foot depth instead of 12 foot, which requires drilling. Since Dr. Leo Horvitz has published numerous articles defining and explaining the acid extraction technique used in this study, I will not use any of my time to go into the details of the acid extraction technique.

In several comparisons between the mechanical disaggregation technique using the ball mill and the acid extraction using phosphoric acid, we found a fairly regular relationship; the high values are high, and the low values are low, and the compositions are similar. However, comparison of magnitudes of gas released by the various extraction techniques indicates that there is almost always more gas released by acid extraction. This suggests that there is gas trapped in calcareous sediments and iron oxide coatings within the soil environment which cannot be removed by disaggregation by mechanical means. The more calcareous the sediment sample, the larger this difference becomes between the two types of extractions. Samples from the Gulf of Mexico generally contain calcareous mud, yielding very large differences in gas volumes released between the two extraction techniques. Both techniques are very reproducible within themselves, but often yield quite different results. Compared to the free gas technique (both 4 foot probes and 12 foot drill holes), both adsorbed techniques preferentially lose some methane and ethane, and generally yield oilier compositions than the free gases measured in situ.

My final example in which all three of these techniques were applied is taken from one of the GEOSAT test case areas. The Patrick Draw survey area contains oil, gas, and oil-gas mixed reservoirs in the Cretaceous Almond Sands. The reservoirs are, for the most part, stratigraphic in nature, however, the Table Rock Field is anticlinal. These reservoirs are situated on the eastern flank of the Rock Springs uplift which has evidence of some late Cretaceous - early Tertiary (Laramid) deformation. The reservoirs are sealed by the Lewis Shale, and are underlain by the Erickson Sand. The Lewis is probably the source of the hydrocarbons in these reservoirs.

The propane concentrations measured at 4 foot depths in this study are shown in Figure 4.2.2.31 with the field outlines shown for reference. As shown, there is an obvious relationship between the northwestern fields and the high propane magnitudes.

In 1980 the University of Wyoming and the GEOSAT committee studies the foliage of the Patrick Draw area. In that study, an area with stunted sage and high overall albeito was detected on the Landsat, Landsat-D simulator, and high altitude photos. Subsequent study showed that the stress was recorded in the vegetation, and that this condition had existed for over 75 years. This suggest that cultural effects were at best minimal prior to field development.

The blight zone appears to be an area of high concentration of near-surface free hydrocarbon gases. The reason for this concentration (>30,000 ppm C1) is undoubtedly due to the presence of fractures as revealed by the lineament study and verified in part by seismic discontinuities. The blight zone is also located preferentially over the gas cap of the reservoir.

Although this field is under a pressure maintenance program by gas reinjection, the pressures have not exceeded the original formation pressure. Any leakage enhanced by artificial means should be restricted to the original leakage routes. The 4 foot probe data compared very favorably with the 12 foot drill free gas samples. In the blight zone, the 12 foot gas samples saturate the gas chromatograph.

In addition, we compared free and adsorbed gas values as shown in Figures 4.2.2.32 and 4.2.2.33. The correlation is poor. It is particularly interesting to note that there is no adsorbed gas in the blight zone where we found 3% methane gas with the probe. This observation appears to be similar to that previously noted over an underground propane storage reservoir where free gases were observed to be present and adsorbed gases were absent. This field has been extensively studied by the 70 companies involved in the GEOSAT program. The results are available in a large report published by the AAPG as a special publication, Abrams (1985). The GEOSAT study evaluated the lithology, the pH and even the plants, and still have not provided the answers to why these different cases exist. We don't know why adsorbed gas is not present in these macroseeps, but we clearly need to understand it. The adsorbed gas appears to be exactly 180o out of phase with the free gas that is observed in this area. Our experience has demonstrated that adsorbed gas does not necessarily reflect the free gas flux that is present in the field today and obviously does not reflect the product that is stored in the field today. We do not know if the adsorbed gas requires emplacement over geologic time, or is strongly affected by soil lithologies, pH, soil microbes, or other as yet unidentified factor.

I can provide additional details on the two soil extraction techniques that I have talked about if anyone is interested. They provide very reproducible results when run correctly. The important question is the significance of the results. Which technique do you use for evaluation of product that has leaked from a pipeline or disturbed area, or from an exploration target? As I hoped that I have demonstrated, there is a great need to do research in grab sample techniques involving soils. Based on fairly extensive experience, GR&DC scientist recommend the free gas techniques as the method of choice for mapping currently active gas migration.

REFERENCES

Abrmas, J.J., Conel, J.E., and Lang, H., 1985, the Joint NASA/GEOSAT Test Case Project, Final Report, Part 2, VII, Editor Helen N. Parley, AAPG, Tulsa, OK, p 11-6, 1177.

Anderson, R.K., Scanlan, R.S., Parker, P.L., and Behrens, E.W., 1983, "Seeps Oil and Gas in Gulf of Mexico Slope Sediment", Science, Nov. 11, pp. 619-621.

Drozd, R.J., Pazdersky, G. J., Jones, V.T., and Weismann, T.J., 1981, Use of Compositional Indicators in Prediction of Petroleum Production Potential, 1981 ACS National Meeting, March-April, Atlanta, GA.

Janezic, G.G., 1979, Role of Biogenic Light Hydrocarbon Generation in Near-Surface Prospecting, Presented at the AAPG Meeting, Houston, TX, April.

Jones, V.T., 1979, Predictions of OIl or Gas Potential by New Surface Geochemistry, Presented at AAPG Meeting, Houston, TX, April.

Jones, V.T., 1983, Surface Monitoring of Retort Gases From an Underground Coal Gasification Reactor: Time Dynamics, Presented at the 186th Annual American Chemical Society Meeting, Washington, DC, September.

Jones, V.T., and Drozd, R.J., 1983, Predictions of Oil or Gas Potential by Near-Surface Geochemistry; AAPG Bulletin, Vol 67, #6, pp. 932-952.

McAuliffe, C., 1963, Solubility in Water of C1-C9 Hydrocarbons, Nature Vol. 200, December 14, pp. 1092-1093.

McAuliffe, C., 1966, Solubility in Water of Paraffin, Cycloparaffin, Olefin, Acetylene, Cyco-Olefin, and Aromatic Hydrocarbons, Jour. Phys. Chemistry, Vol. 70, pp. 1267-1275.

McAuliffe, C., 1969, Solubility in Water of Normal C9 and C10 Alkane Hydrocarbond, Science, Vol. 163, pp. 478-479.QUESTION:

I have got a question for Vic. In the free soil gas analysis, where you stick a probe down in the soil to draw out a sample into a column, are you measuring a flux or are you measuring the equilibrium gas oncentration in that sample or are?

V. Jones:

We measure the equilibrium gas concentration of the soil gas aspirated from the sample hole, which is not necessarily a flux. Attempts to develop some of the other techniques and some of the various depth probe techniques were conducted to look at gradients, and hopefully to tell a live seep from a dead seep because of the indigenous hydrocarbons contained in the soil. The proble is that any technique you use, mixes gas with free and adsorbed states because of the interaction which the soil created by the sampling tools. The auger itself, in making the hole, is a disaggregation method. A fair percentage of the gas that is free is, in fact, librated from the soil by a disaggregation method. his is one reason why we tried to set up a core sampling technique that would then disaggregate in the same that the auger does. However, as shwon, the different techniques do no always give the same result. Sometime there is more free gas than there are adsorbed sites, and we really do not know how to measure how much soil we extract in one of those augered holes. Using the ball mill technique, we work with a core, so that we can get a true concentration (g/g) by weight in the soil, although the adsorbed concentrations will not necessarily match the free gas concentrations.

Comment:

We have done some work on these methods with contractors, and I would say that the reactions to the methods and the data generated have been, at best, mixed. They are only very midly successful.

I would like to ask Victor where he finds the hydrocarbons in the free zone? Very frequently, as a geologic feature, the surface area where, even over aggregate and fairly porous soils, there is a hard band caliche which forms a farily effective vapor barrier to the soils underneath the surface. Do you find that you have to penetrate such a barrier before you get good soil measurements?

V. Jones:

Not necessarily, but you may get much larger magnitudes beneath such a barrier. In our experience, the best quality samples come from holes drilled into the water table. However, although the probe technique is shallow, it is adequate in the majority of cases. A slight problem with theprobe techniques is that sampling at the 4 foot level is influenced by barometric and meteorological effects, which primarily operate in the weathering zone near the surface -- an interface zone that can influence the gas measurements. If you have a shallow probe sample that indicates a large anomaly which intersects a small stream channel containing a lot of sand and very little clay, you may find

*Presented at the American Petroleum Institute Workshop for Sampling And Analytical Methods For Determining Petroleum Hydrocarbons In Groundwater And Soil, November 27-29, 1984, Denver, Colorado.