by
R.F. Mercer

Continental Laboratories Inc.
Calgary, Alberta, Canada.

Sufficient evidence exists to suggest that misinterpretation of wellsite gas detection data is quite common. The question is often asked, "How big a gas show should I expect to get from a zone that will make a well?" Such misunderstanding may often be traced to a lack of familiarity with the fundamental principles of gas detection and interpretation.

To illustrate these fundamental principles, a drilling model is presented to demonstrate the effects of bit penetration. The model is analyzed to explain theoretical gas detection response to the penetration of a hydrocarbon bearing zone. Gas show characteristics, as transmitted to the surface by the drilling fluid, are specifically related to bit penetration. A careful analysis for the drilling model derives four

classifications of gas present in the drilling fluid. These are:

1. Liberated gas
2. Produced gas
3. Recycled gas
4. Contamination gas

A strong case is presented to show that all drilling fluid hydrocarbons may be classified into one of the four categories.

Definitions are provided for each type of gas.

1. Liberated gas is defined as gas mechanically liberated by the bit into the drilling fluid as the bit penetrates the formation.
2. Produced gas is defined as gas produced into the drilling fluid from a specific zone in response to a formation pressure which exceeds the opposing effective hydrostatic pressure.
3. Recycled gas is defined as gas which has been pumped back down the hole to appear a second time at the surface.
4. Contamination gas is defined as gas artificially introduced to the drilling fluid system from a source other than the rock formations.

Geological and drilling engineering implications of each category are discussed. Abnormal pressure detection and surveillance receive special consideration.

It is concluded that a working knowledge of basic principles of interpretation is absolutely requisite to an effective use of wellsite gas detection data.

The author has noted over the last few years that considerable misunderstanding exists with regard to techniques of interpretation of hydrocarbons found in the drilling fluid during penetration. This paper will seek to provide a basic understanding of these principles to assist concerned persons in the efficient use of gas detection data. These fundamentals should provide the basis for a creative understanding of the mechanics of gas liberation and detection as well as assisting the responsible person to render accurate and timely decisions more efficiently.

Such information should prove beneficial as applied to a number of different phases of the total drilling and completion operation. It will be shown that mud gas data is of great value in delineating the precise thickness of potential reservoirs as well as providing assistance in identifying the hydrocarbon-water interface.

Generally speaking, down hole logs accurately measure physical characteristics of the rock in place but do not specifically measure distinct physical characteristics of hydrocarbons. The presence of hydrocarbons is projected on the basis of relative changes in tool response for one measurement compared to other measurements covering the same interval.

In contrast, gas detection data measures a specific physical characteristic of hydrocarbons and consequently responds directly to their presence. The gas detector is not capable, however, of giving definitive data regarding important physical characteristics of the rock. The interdependence of gas detection data and down hole logs should therefore be readily apparent.

It will also be shown that intervals of hydrocarbon potential can, in many instances, be identified with greater resolution than is available from other sources. Such information has proven to be of valuable assistance to operators requiring a very selective perforation program.

The presence of hydrocarbons in the mud system is shown to be an important indicator with respect to abnormal pressure detection. Continuous surveillance is essential to the maintenance of high standards of blow out protection and general rig safety. In addition to basic safety consideration, gas detection data has proven useful to drilling engineers in determining the precise mud density requirement to ensure adequate blow out protection while maintaining maximum bit penetration.

Before a detailed consideration of gas detection interpretation is presented, a short background may prove helpful to those not directly familiar with gas logging techniques.

The basic function of wellsite hydrocarbon detection has been defined in a previous paper as follows: "to detect and position with respect to depth all hydrocarbon accumulations." (1)

This definition relates specifically to techniques employed at the wellsite to determine potential hydrocarbon bearing zones. Both drilling fluid and cuttings detection are included. A significant increase in the concentration of hydrocarbons in the mud system is referred to as a "gas show".

In contrast, wellsite gas analysis is performed by chromatography to define the composition of "detected" hydrocarbons as an aid in pre-determining the character of the hydrocarbons in a reservoir as well as assisting in the proper classification of those hydrocarbons contained in the mud system. The primary function of wellsite gas analysis has also been defined in the earlier paper as follows: "to analyze each hydrocarbon accumulation to include the identity and the relative proportion of each component". (2)

Typical gas detection and analysis equipment in use throughout the world generally incorporates one of two standard detectors. By far the most popular total gas detector is the catalytic filament detector (CFD). It operates on the principle of catalytic combustion of hydrocarbons in the presence of a heated platinum wire at concentration levels below the lower explosive limit. The increasing heat due to combustion causes a corresponding increase in the resistance of the platinum wire filament. This resistance increase is measured through the use of a wheatstone bridge circuit and recorded as, "units of gas".

The second detection technique is provided by the flame ionization detector (FID). This detector functions on the principle of hydrocarbon molecule ionization in the presence of a very hot hydrogen flame. These ions are subjected to a strong electrical field resulting in a measurable current flow which is then amplified and recorded. A thorough discussion of this unique technique as applied to wellsite gas analysis is available in the literature. (3)

The total gas detector and the chromatograph are normally installed as companion instruments at the wellsite and provide a continuous monitoring of the drilling fluid and well cuttings for the presence and composition of hydrocarbons.

Although cuttings gas detection and analysis is of considerable value in its own right, it will not be discussed in this paper.

Unfortunately the term "gas detection" has often proven to be misleading because it appears to suggest that gas detection equipment is only of service in locating gas reservoirs. This is not the case. As shown by Evans, Rogers and Bailey, (4) mature liquid hydrocarbon reservoirs are characterized by rich compositions of all components in the gasoline range C4 through C7 with a good distribution of gases in the range C2 through C4 plus reasonable quantities of C1. These facts demonstrate that gas detection equipment should be more properly called hydrocarbon detection equipment since it is effective in locating both gas and liquid hydrocarbon reservoirs.

Requisite to a clear understanding of the interpretation of mud-gas data is consideration of the source of hydrocarbons as they occur in the drilling mud. To assist in this consideration, a simple drilling model is proposed which illustrates the impact of bit penetration through hydrocarbon accumulations. A series of cases is presented where variations in the configuration of the mud-gas data indicate specific differences in the response of the hydrocarbon bearing zone to bit penetration and subsequent rig operations.

The model will show that the geometry of the gas show recorded by the instrumentation and plotted with respect to time is directly related to significant characteristics of the hydrocarbon zone as well as the impact of concurrent drilling operations. It will become apparent that the configuration of the gas show as recorded directly from the drilling mud is of greater interpretive significance than the magnitude of the gas show. When instrument chart data recorded versus time is digitized and plotted in graph format versus depth, the magnitude of the gas show may be faithfully reproduced but the configuration of the show is usually lost.

Thus it becomes obvious that basic and vital interpretation must be derived from a detailed analysis of the instrument charts themselves and not solely from a plotted graph. The basic function of the plotted graph should be to collate, according to depth, pertinent data produced from various sources. This graph then provides a broader understanding of the hydrocarbon accumulation and a convenient means for future reference.

To illustrate these concepts, a diagrammatic technique has been employed which graphically relates the gas detector response plotted versus time to the actual penetration of the rock by the drilling bit through the penetration rate curve plotted versus depth. This technique allows direct

comparison of the geometry of the gas response to actual rock penetration.

A. LIBERATED GAS 1. Full Hole Drilling

Figure one illustrates a typical situation where a bore hole is created through a hydrocarbon bearing zone and the total bottom hole pressure (TBP) is greater than the formation pressure (FP). During penetration, the bit continuously introduces to the mud system components of the rock contained in the cylinder defined by the hole size and the thickness of the interval. As the bit penetrates, it mechanically creates pseudo-permeability and allows material contained in the absolute pore volume of the cylinder to enter the mud system and be transported to the surface. The term pseudo-permeability is suggested because the liberating action of the bit is purely mechanical and not directly related to the inherent rock permeability.

If the pore volume contains hydrocarbons, it is evident that the hydrocarbons contained in the cylinder of rock will be transported to the surface in various proportions of two possible forms. First, liberated directly into the mud or produced into the mud from the cuttings as they are subjected to ever decreasing hydrostatic pressure. Second, retained by the cuttings chips themselves. It therefore follows that the primary source of hydrocarbons available to the gas detection equipment under these conditions derives from the cylinder of rock mechanically liberated into the mud system by bit action. This is a type one gas response.

In figure one the penetration rate curve corresponding to the porous interval shows a characteristic drilling break as the bit drills though the sandstone. Such drilling breaks are often invaluable in determining the thickness of porous intervals. The hypothetical gas detector response shows a typical record of the concentration of hydrocarbons in the mud versus time.

The concentration of liberated hydrocarbons in the mud is primarily a function of the following factors:

1. Penetration rate
2. Absolute pore volume
3. Formation pressure

Other factors are also of concern, such as oil and gas saturations, mud return flow rates and hole size. It is assumed for purposes of our example that these additional factors are not pertinent to our discussion of basic principles but should be considered when evaluating the significance of a particular gas show.

Substantial increases in any of the three named factors will normally have a visible effect on the gas detector response. In the normal case, the rate of penetration is the most important single factor in determining the magnitude of the gas show. The effect of penetration rates will be discussed in greater detail later in the paper.

If the bit were penetrating the cylinder of rock at a constant rate, and if the porosity and the formation pressure were exactly constant throughout the interval, it is reasonable to assume that an equilibrium would be established between the volume of mud circulating through the bit and the volume of gas mechanically liberated to the mud system. This supposition is shown diagramatically in figure one but of course rarely exists in reality. The lag time is shown as the circulating time commencing with bit contact of the porous interval and terminating with commencement of the gas increase. If the gas response comprises only liberated gas, it is reasonable to conclude that the gas response would begin to end one lag time after the bit ceases to penetrate hydrocarbon bearing porosity.

Since the formation pressure is normally constant throughout a single porous interval, it is reasonable to conclude that variations in the magnitude of the liberated gas reading are related to the remaining two parameters; penetration rate and porosity. Should the penetration rate be relatively constant, show magnitude variations can often be related directly to rock porosity with resolution capability less than one meter.

Figure one, a type one gas response, is the normal case because a margin of safety is always desired when penetrating possible blow out zones. Figure one shows the typical situation where mud filtrate has invaded the porous formation while wall cake was being deposited on the surface of the hole. Since the pressure differential across the hole - rock interface is positive, it is evident that no additional hydrocarbons beyond those contained in the rock cylinder contribute to the liberated gas response.

In unusual circumstances of high formation permeability, low formation pressure and exceedingly high total bottom hole pressure, it is possible that mechanically liberated hydrocarbons may be pumped directly into the formation and not return to the surface. A further variation of this possibility may occur if filtrate invasion immediately preceded the bit. Resident hydrocarbons may be flushed by the filtrate so that the bit would mechanically liberate only mud filtrate and not pre-existing hydrocarbons. While these two possibilities occur

with extreme rarity, they should be considered in instances where gas shows were normally expected but did not occur.

1. In a type one gas response (ie. where the formation pressure is less than the total bottom hole pressure) only mechanically liberated gas forms the show.
2. The configuration of the gas show is an early indication of the thickness of the liberating interval and possibly the quality of the porosity as well as the depth and thickness of the most porous interval.
3. The presence of a type one gas response gives no direct indication of the presence or absence of

2. Coring

The principle of mechanical liberation through a type one gas response zone should have obvious implications for coring because only a small portion of the normally drilled cylinder of rock is being exposed to mechanical liberation by the core bit. It logically follows that a much smaller quantity of liberated gas is introduced to the mud system per meter of penetration. Often, coring rates are considerably slower than full hole bit penetration rates which would further decrease the liberated gas quantity per mud volume. These factors usually combine to result in the common phenomenon of considerably lower gas readings while coring.

In the event that mud gas is not completely volatilized in the settling pit but is pumped back down the hole, the gas detector may record a second appearance of a pre-existing show. This phenomenon is diagrammed in figure one where the liberated gas show has recycled to the surface for the second time and is designated R.

An analysis of the usefulness of recycled gas is available in the literature. (5)

Recycled gas may be identified by the application of certain tests. The recycle should be no larger than the original response but should be similar in shape. The composition of the recycled response may be misleading in that the more volatile hydrocarbons are often liberated to the atmosphere in the pits and under the influence of a degasser. The result is the analysis of the recycled response shows a larger proportion of heavy ends.

From the beginning of the primary gas response to the beginning of the recycled gas response in circulating time is a good indication of the total circulating time of the mud system. Such direct information may often be helpful in assuring the accuracy of an estimated lag time.

Figure two demonstrates possible alternative explanations for instances where the duration of the gas show does not seem to extend throughout the entire period of probable liberation as projected from other indicators such as the penetration rate.

In a type one gas response (TBP>FP) only liberated gas would comprise the gas response.

If the geometry of the show is solved in a manner consistent with the principles derived in figure one, a significant variation within the interval of the drilling break becomes apparent. Two alternative

explanations are suggested in (A) or (B) as shown in figure 2.

1. Since gas was mechanically liberated only from the top portion of the drilling break, it is probable to assume that the best porosity occurs through that interval. The absolute pore volume is probably diminished or absent through the bottom of the section resulting in no liberation. If liberation should occur from the bottom of the zone and not from the top, this explanation would be favored over (B) because gas does not naturally occur under water in a contiguous reservoir. In case (A) the constant penetration rate throughout the drilling break would probably reflect better bit performance in sandstone than shale. The distinction between drilling porous and nonporous sandstone appeared to be of little consequence by comparison.
2. If indications suggest that the porosity does in fact continue throughout the interval as delineated by the drilling break, it is probable that the absolute pore volume in the upper section contains only hydrocarbons while the lower pore volume is filled with water. This principle can be exceptionally helpful in conjunction with log saturation indications in determining the gas-water interface or transition zone.

4. EFFECT OF PENETRATION RATE CHANGE ON SHOW MAGNITUDE Figure three again considers a type one gas response where only liberated gas is present.

If the porosity, formation pressure and mud pump volume remain constant throughout the drilling of the zone, the magnitude of the gas show becomes a direct function of the rate of penetration. A substantial decrease in the rate of penetration (designated (A) in figure 3) should result in a corresponding decrease in the liberated gas - versus mud volume equilibrium designated (B).

Of course a decrease in penetration rate through a zone of given thickness would require a greater total period of penetration. This longer drilling interval would result in a response of decreased magnitude but of longer duration.

Consequently, during extremely slow penetration of hydrocarbon accumulations where no particular drilling break is evident and where rock porosity is especially low it is entirely possible that no recognizable gas response as such will occur. The presence of liberated gas may be very difficult to ascertain in addition to the carried background gas in the mud system at that time. Such a situation would require careful interpretation in conjunction with all other qualifying information before one might conclude that significant liberation did not occur.

Experience has shown that the penetration rate is the most important single factor governing magnitude in a type one gas response.

This principle clearly suggests the error of using show magnitude as the single or primary criteria in judging significance.

Of particular concern is the dangerous combination of effects from various liberated gas zones may exist up hole which are effectively contained by the existing hydrostatic pressure. Extreme care must be taken not to liberate large quantities of gas from a thick downhole reservoir by drilling through it too rapidly. The result may be to decrease the effective hydrostatic on upper zones due to gas cutting of the mud thus allowing it to blow out. Penetration rate should be reduced to minimize the quantity of liberated gas in the mud system thereby maintaining sufficient effective hydrostatic pressure uphole.

5. PRODUCED GAS

Figure four illustrates the abnormal case where the total bottom hole pressure (TBP) is less than the formation pressure (FP). The gas response resulting from such a situation is characterized by significant differences from those previously discussed and is designated a type two gas response.

Figure four shows the usual situation where the hole does not begin to make fluid immediately upon penetration of the zone but the gas response commences at one normal lag time. Such responses are characterized by exceptional initial magnitude and the continuation of the response

beyond the time normally anticipated for the termination of the liberated show.

If the source zone is clearly defined by the penetration rate and other available geological data, it becomes apparent that the formation is contributing additional hydrocarbons to the mud system beyond those mechanically liberated.

Significant contrasts in interpretation result from a type two gas response.

1. There is now no direct relationship between mechanical liberation and mud circulation, therefore definitive analysis of the source zone thickness and quality becomes extremely difficult. The magnitude of the gas response can no longer be related to the general significance of the source zone in comparison with other type one gas responses.
2. The presence of produced gas demonstrates conclusively that at least some degree of effective permeability is present. This direct evidence of permeability is in contrast to the absence of any definitive evidence in a type one response where only mechanical liberation occurs.
3. Since produced gas is generally independent of mechanical liberation and its attendant controlling factors, it is reasonable to expect that the configuration and magnitude of type two gas responses encountered while coring would be generally independent of the mechanical characteristics of the coring operation.

Occasionally drilling operations require the introduction of oil in various forms to provide additional pipe lubrication, etc. Oil based muds are often used to minimize formation damage through elimination of excessive water loss. Diesel is the normal oil phase used in inverted oil emulsion muds. Diesel in its natural state does not contain volatile hydrocarbons and therefore is not disruptive to gas detection equipment. However, diesel is often transported in containers which have previously carried volatile crudes and may therefore retain some volatile gases. Hydrogen gas is often detected in pipe iron or associated with the setting action of cement. Occasionally mud additives or various chemical reactions in the mud will provide other hydrocarbons or combustible gases which may be detectable by wellsite total gas detectors. All of these examples comprise combustible gas sources which are not indigenous to the rock formations and must be identified accordingly when detailed interpretation is desired.

After working around gas detection equipment for some time, rig personnel become aware of what gas sources can be added to the mud to influence gas detector readings. One must of course establish that the gas source was not deliberately introduced by a member of the drilling crew.

At certain times mud conditions are such that the introduction of large volumes of air into the mud system cause "pseudo gas responses". These responses do not reflect increased gas concentration in the mud but rather greater gas trap efficiency when the air-rich mud reaches the surface. This phenomenon may occur after trips when a float is used or from kelly air introduced during connections. Such pseudo responses are often called "kelly responses", or have a distinct effect on "trip gas responses". Trip gas will be considered in detail later in the paper.

Figure five portrays a type two gas response and suggests subsequent rig operations which can be used to deal with an abnormally pressured interval with due regard to safety and optimized penetration.

1. CONNECTION GAS

In this hypothetical situation the rig experienced a type two gas response. After continuing circulation for some time, the magnitude of the readings continued to increase. At that point a decision was made to increase the mud density which eliminated the produced gas and returned the mud system to the pre-existing background. The time scale is of course very compressed in this example and does not accurately portray the time span often necessary to eliminate large quantities of produced and recycled produced gas. Subsequent to the elimination of produced gas, a connection was made. Evidence of the connection appears on the gas detector chart as a decrease in the carried background reading where no mud was circulated during the connection. At approximately one lag time after circulation was resumed, a response occurred. This response was deemed produced gas as it was related to the connection and was not liberated from the formation being penetrated one lag time before the response.

Because there is no evidence of produced gas in the system while circulating, it is apparent that the mud density plus the annular pressure drop (APD) are sufficient to create a total bottom hole pressure greater than the formation pressure. Therefore, the connection gas peak experienced after the first connection subsequent to penetrating the gas zone may be related predominantly to swabbing of the zone rather than to insufficient hydrostatic pressure. Swabbing may occur when the kelly is raised for a connection. Because the annular pressure drop is lost during periods of no circulation, the bottom hole pressure is equal to the hydrostatic pressure for static mud systems. This decrease of bottom hole pressure may be a factor in the magnitude of the connection peak. Since the swabbing effect is not measurable , it would be difficult to ascertain with any degree of accuracy the significance of connection gas peaks which result when bit movement on connections extends above the gas zone.

On the next connection, however, when it was certain that no bit swabbing occurred, no connection gas peak resulted. This fact suggests that the mud may be too heavy since no produced gas resulted from loss of the annular pressure drop.

The mud density was subsequently reduced until a moderate connection gas occurred with no increase in background values while circulating. The formation pressure of the producing zone is bracketed as follows: The formation pressure is approximately equal to or greater than the hydrostatic pressure, however, the formation pressure is less than the hydrostatic pressure plus the annular pressure drop. Such circumstances represent the optimum mud density for containing the zone yet providing positive evidence that the mud density is not excessively high.

Subsequent reduction in mud density resulted in measurable quantities of produced gas becoming apparent in the mud system during circulation. This fact suggested that the formation pressure was now greater than the total bottom hole pressure and that the mud density had been reduced too much. The mud density was then increased to restore the ideal condition of moderate connection gas peaks with no evidence of produced gas while circulating.

2. TRIP GAS

Trip gas is the general term applied to produced gas which characteristically occurs within one lag time after a trip is completed and circulation has been resumed. Three basic factors influence the presence, location and magnitude of the trip gas .

1. The effect of bit swabbing the entire hole. This effect is influenced to a considerable degree by such factors as the speed at which the pipe is tripped out of the hole, variations in hole size, the configuration of a packed hole assembly, and tripping out with a full hole core barrel.
2. The time over which these factors influence the static mud system.

The basic principles previously discussed with regard to connection gas of course also apply to trips. The most significant difference between trips and connections is the extreme accentuation of these influences during a round trip as compared to the relatively minor influence of a connection.

This accentuation of effect should immediately suggest the seriousness of ensuring absolute control over any previously drilled zone exhibiting abnormal pressure characteristics before a trip is attempted. It would be extremely foolish to suspend circulation and commence a trip in the midst of a formation gas response without first ascertaining whether it was a type one or type two response.

The rig is never more vulnerable than during a trip out of the hole especially when the hole is not kept full. Mr. A.S. Murray reported to the Canadian Oil Scouts Association that "80% of all blow outs occur during tripping." He noted that "80% of blow out problems occur while drilling with insufficient mud density and failure to fill the hole while tripping." He also observed that "80% of blow outs come from normal pressure zones and occur in wells less than 2500 meters deep." He concluded that, "practically all these blow outs could have been prevented because 80% of all blow outs are the result of human failure." (6)

It is apparent from the principles of gas detection as previously discussed, that the gas detector is of limited assistance in providing early warning for zones which blow out immediately upon contact. Proven methods of well control have been developed and should be employed in these cases. The constant drill pipe method is especially helpful in establishing control over such a zone until proper mud density can take effect. Statistics have shown, however, that this type of blow out zone is the exception rather than the rule. Therefore, careful surveillance of all produced gas indications, especially resulting from tripping, becomes a very important rig safety indicator.

Figure six presents a hypothetical situation where a type two gas response has occurred followed by the elimination of produced gas through the effect of increased mud density. Subsequently, circulation was terminated for a trip. Upon completion of the trip, circulation was resumed and a typical trip gas response was received at approximately on lag time. This response suggested by its configuration that the mud density was sufficient to keep the well under control and no dangerous conditions were experienced during the trip.

If the trip gas response had occurred in its entirety at a lag time somewhat less than one total lag time, this would indicate the possibility of a produced gas source at some shallower depth in the hole. Such information derives from the fact that the first circulation after a trip gives some indication of the response of various hydrocarbon accumulations present throughout the length of the uncased hole.

If the trip gas response indicates a very early onset which cannot be attributed to a shallower source, this may indicate the extent of gas swabbing up the hole following the bit on the trip out.

In the instance of a normal trip gas response we know that the formation pressure is greater than the hydrostatic pressure minus the swabbing effect or no response would have occurred. But other indications suggest that the formation pressure is less than the hydrostatic pressure alone. Exorbitant trip gas peaks and unexplained variations in magnitude should cause the seriously interested person to ascertain the explanation for this condition. An extended trip gas response after the main peak suggests that the formation has continued to produce after circulation was resumed. This circumstance would suggest the hole is barely in balance and should be treated with great care during subsequent trips. The formation pressure in this case may be greater than the total bottom hole pressure until the swabbed trip gas is circulated out of the system.

On occasion, factors will combine to render the gas detection data received when a zone is drilled apparently less significant than it is, especially if no indications of produced gas are seen. The first trip gas response after such a zone may be significant indication of the presence of permeability and shed greater light on the original interpretation.

Occasionally abnormal pressure intervals may be drilled where no indications are received from other wellsite sources, such as penetration rate and sample examination. The most probable explanation is fracturing. If such a zone is penetrated in the absence of total gas detection equipment, and therefore there is no evidence regarding the presence of produced gas, the rig crew may be unaware of the danger awaiting them on the next trip. In many instances, a gas response with large proportions of produced gas will occur while drilling. Subsequent attempts to kill the zone, thus eliminating the produced gas by raising the mud density may prove unsuccessful. Even after a considerable increase in mud density, the produced gas seemingly continues. One must first be certain that the gas is not a continuing build up of recycled gas. If the produced gas does in fact exist and persist, the facts would suggest an unusually high pressure interval of very low permeability which has apparently established a flow equilibrium with the circulating mud system as a function of deliverability rather than formation pressure. Had the deliverability been greater, the zone would probably have blown out immediately upon contact.

If such evidence exists, efforts to completely eliminate all traces of produced gas may be unwarranted.

On occasion, evidence of produced gas, such as connection gas and trip gas may diminish over extended periods of drilling with no apparent change in mud characteristics. Such evidence may suggest that the source zone was of limited permeability and has now depleted to a sub-hydrostatic formation pressure.

All of these factors indicate the importance of bringing to bear all available gas detection data on the interpretation of specific hydrocarbon accumulations including the more subtle indications which are observed over long periods of mud surveillance especially on connections and trips.

This paper has shown that all hydrocarbon indications in the drilling mud system may be classified in one of the four derived categories. An understanding of basic principles discussed in the paper is necessary for proper classification. Accurate classification of hydrocarbons appearing in the mud system from time to time is a necessary prerequisite for dependable data interpretation.

Further benefits accrued from an understanding of these principles include the following;

1. The basis is formed for better communication between the logging crew and the wellsite geologist together with the drilling personnel.
2. Unnecessary or improper drilling activities may be prevented.
3. Early decisions with respect to drilled hydrocarbon accumulations may be rendered with due regard for safety. Some examples are as follows:

1. Do not stop circulating for a trip in the middle of a gas show until it has been clearly ascertained if the show is type one or type two.
2. Mud density increase may be initiated immediately if necessary without losing the annular pressure drop during periods of non-circulation.
3. Instructions regarding special care during tripping may be given to the drilling crew immediately after penetration of hydrocarbon accumulations which show characteristics conducive to a possible blow out.

1. A careful application of these principles should provide the basis for valuable co-interpretation with the down hole logs and should also assist in planning specific perforating schedules.

The paper has attempted to establish four new terms with specific meanings as they apply to the interpretation of drilling mud gas detection data. Therefore, remember when you are next confronted with a gas response from a gas detector, the question is not "How many units was the increase?" but rather, "Was it liberated, produced, recycled or contamination?"

1, 2, 3 "The Use of Flame Ionization Detection in Oil Exploration". Mercer,

Mr. Mercer was granted the Bachelor of Arts degree in Geology from the Johns Hopkins University, Baltimore, Maryland. He completed one further year of graduate studies at Hopkins in Sedimentation and Sedimentary Petrography and Petrology.

After serving with Shell Oil Company as a Junior Geologist on a field mapping party in Montana, he completed his military requirement including graduation from the Field Artillery, Officer Basic Course conducted at Ft. Sill, Oklahoma.

In 1959, he moved to Montana where he accepted a position with Continental Laboratories Inc., of Billings, a geological wellsite service company. The next three years comprised field assignments at wellsites in the Northern Rocky Mountain States and the Dakotas. During this period, he became familiar with all phases of geological wellsite supervision, including the use of gas detection equipment.

In 1962, Mr. Mercer was transferred to Calgary, Alberta, and was promoted to Canadian Manager. In 1967 he was elected Vice President with responsibility for Canadian Operations.

He is a member of the Canadian Society of Petroleum Geologists, the Canadian Well Logging Society, and the Association of Professional Engineers, Geologists and Geophysicists of Alberta.