Arkansas Turbidite Studies May Aid Oklahoma Jackfork Exploration

Roger M. Slatt, Hisham Al-Siyabi, Eugene T. Williams
Colorado School of Mines
Golden

Charles G. Stone
Arkansas Geological Commission
Little Rock

Paul Weimer
University of Colorado
Boulder

Robert J. Davis
Schlumberger Wireline & Testing
Jakarta

Douglas W. Jordan
ARCO Venezuela Inc.
Caracas

During the past 2 years, more than a dozen successful Pennsylvanian Jackfork gas wells have been drilled in eastern Oklahoma,¹ and the play seems ready for expansion if drilling and completion costs can be contained.²

To date, wells are thought to produce from fractures in highly quartz-cemented, brittle sandstones. However, there is potential for stratigraphic accumulations of gas.^1-3

It is generally agreed that Jackfork Group rocks were deposited mainly as turbidites in the deepwater Ouachita basin.^{1, 4-7} A recent suggestion that significant amounts of Jackfork Group rocks are discontinuous and less predictable sandy debris flows⁸ has received considerable critical review in the recent literature^9-13 and at the 1996 American Association of Petroleum Geologists convention.^14-15 A shallower water origin has also been suggested for some outcrops in eastern Oklahoma.¹⁶

Montgomery² stated that "Jackfork depositional models will remain speculative, at least until stratigraphic and sedimentological analysis of Jackfork outcrops in eastern Oklahoma and western Arkansas can be carried out and integrated with well data from the current play." We concur with this statement and add that the excellent outcrop exposures in southwest-central Arkansas, which have been studied for many years by numerous geoscientists and engineers, can provide additional insights into potentially productive Jackfork strata in Oklahoma.

With this in mind, we provide below a summary of our observations, measurements, and interpretations of upper Jackfork outcrops in southwest-central Arkansas as a guide for interpreting the subsurface Jackfork in eastern Oklahoma.

GR log patterns

To minimize costs, most Jackfork wells are logged by conventional methods and not cored or logged with borehole imaging tools. Thus, facies interpretation is restricted to vertical well log patterns.

For example, logs of two wells through productive Oklahoma Jackfork sandstones exhibit both massive- and interbedded-appearing patterns (Fig. 1 [31,079 bytes]) .

Important stratigraphic questions concerning the logged intervals should include:

• Are they laterally continuous lobe deposits or discontinuous channel deposits?
• Are the thicker shales sufficiently continuous to vertically isolate overlying and underlying sandstones?
• Might the thinner sandstones be gas bearing, as is the case for the thicker, massive sandstones?

Similarly, well log patterns of individual sandstone beds or packages between laterally continuous shales are highly variable (Fig. 2 [47,516 bytes]) , making between-well correlations of sandstones difficult. Below we provide examples of outcrop gamma ray logs^17-18 and stratigraphic descriptions as an interpretation guide to individual Jackfork turbidite elements.

Turbidite elements and GR log response

As a basis for classification, we use Mutti and Normark's^19-20 five basic architectural elements that are common to both modern and ancient turbidite systems:

1. Channels and their fill;
2. Lobes;
3. Overbank deposits;
4. Channel-lobe transition deposits; and
5. Major erosional features (excluding channels).

Channels

A channel is the expression of negative relief that represents a long-term pathway for sediment transport.¹⁹

Topographic relief and features in modern channels and ancient channel-fills can be erosional, depositional, or of mixed origin.²¹ Either type of channel-fill usually contains the coarsest grained sediments at the base, which grade upward into mudstones or thin-bedded, finer-grained sandstones deposited during the waning stages of channel filling.

Upper Jackfork Group channel-fills are displayed in at least two areas in the vicinity of DeGray Lake, Ark., and one area near North Little Rock, Ark.

The first area is at Hollywood Quarry, near Hollywood, Ark.^6-13 The stratigraphic section at this quarry, with accompanying outcrop gamma ray log (Fig. 3, log A [20,251 bytes]) comprises the following intervals from the from the base upward:

a. shale that contains abundant sandstone, pebbly sandstone, and shale clasts;
b. 90 ft of mainly massive, medium-grained, quartzose, offset-stacked, lenticular sandstone and pebbly sandstone beds that are in erosional contact with interval 'a;'
c. 10 ft of black shale with thin, very fine-grained sandstone to siltstone beds, and
d. 30 ft of fine- to medium-grained turbidite and slurried sandstones interbedded with mudstones.

Vertical connectivity of these sandstones is good, as the shales are not continuous. Interval 'd' comprises interbedded, laterally continuous sandstones and shales that exhibit poor vertical connectivity.

The well log pattern illustrates this style of interbedding, however conventional well logs would preclude resolution of the thinner sandstone and shale beds seen in outcrop. The gamma ray log pattern of the entire succession (intervals 'b'-'d') is that of a thinning- and shalier-upward channel-fill (Fig. 3, log A).

The second area is at Pinnacle Mountain State Park, southwest of Little Rock.⁴ Here, an east-west trending linear chain of steep, discontinuous hills comprises thick, moderately- to steeply-dipping Jackfork sandstones and shales. The interval at the abandoned Fulk Mountain Quarry, and accompanying outcrop gamma ray log (Fig. 3, log B) comprises, from the base upward:

a. 50-60 ft of stratified, lenticular sandstone;
b. > 100 ft of sandstone and mudstone olistoliths within a contorted mudstone matrix;
c. > 100 ft of well-bedded, clean, quartzose, fine- to medium-grained sandstones that exhibit some channeling.

We interpret this section as large, discontinuous pods of high energy turbidite sandstones (interval 'c') that were transported through a linear channel on a relatively steep, unstable muddy slope and deposited within topographic lows between muddy slump blocks (interval 'b'). Modern turbidite sands have been found in this environmental setting.²²

Without the outcrop for observation, interpretation of the gamma ray log alone (Fig. 3, log B) would suggest interval 'c' is a clean, laterally continuous sandstone when in fact the sandstone occurs as discontinuous pods, and interval 'b' is mudstone with sandstone beds when in fact the interval contains sandstone olistoliths within mudstone.

The third area is in the DeGray Lake Spillway near Arkadelphia, Ark., which exposes a 1,000 ft thick stratgraphic section, much of which can be correlated laterally beyond the spillway.^{13, 26}

The section is comprised of thin- to thick-bedded, low- to high-density turbidite sandstones, thick pebbly sandstones to sandy conglomerates, thick debrite shales, and laminated shales and siltstones.^{4, 5, 23} The upper 230 ft of section is interpreted as a proximal lobe-channel-fill (Fig. 4 [35,804 bytes]) , which consists from the base upward of:

a. 20 ft of contorted debrite shale containing sandstone and mudstone olistoliths;
b. 130 ft of thick-bedded, erosional-based, massive-appearing to amalgamated, clean, quartz-cemented sandstones separated by thin, discontinuous mudstones;
c. 40 ft of laminated shale; and
d. 40 ft of clean, quartz-cemented, medium- to coarse-grained sandstones, pebbly sandstones, and sandy conglomerates. The gamma ray log pattern of this interval (Fig. 4, intervals 'a'-'d') corresponds to that of sharp-based, clean sandstones and sandy conglomerates with interbedded shales.

Lobes

Lobes on modern and ancient fans can extend areally for hundreds of square miles. Mutti and Normark¹⁹ differentiate between proximal, intermediate, and distal lobes. From the former to the latter, beds become thinner and finer grained, and transition from high to low density turbidites; scours are confined to proximal lobes.

We concur with Mutti's²⁴ interpretation that much if not all of the lower 770 ft at the DeGray Lake Spillway sequence comprises a series of lobe deposits (Fig. 4, intervals 'e'-'g'). An example of a series of relatively distal lobes is interval 'g' on Fig. 4. Within this interval, the thickening- and cleaning- upward lobe at 540-640 ft (Fig. 3, log D and Fig. 4) has been correlated with outcrops and subsurface core borings obtained about a mile away,²⁶ as well as in the Shell Oil Co. 1 Rex Timber well located six miles to the south.^4,7

Individual beds within this sequence have been measured and correlated laterally for distances up to 300 ft,^{4, 13, 25, 26} and calculations⁴ suggest they should extend for hundreds to thousands of feet before pinching out. These beds thicken and thin in alternate lateral directions, which is typical of compensation-style bedding of lobes.^{6, 13, 24}

Interval 'e' (Fig. 4) is interpreted as a series of proximal-distal lobes, many of which occur as thinning- and shalier-upward deposits.²⁴ A more proximal environment for much of this interval is indicated by the occurrence of thicker beds, slurried beds,⁴ slumped and resedimented beds, and scours of the type described below within a channel-lobe transition zone.

Overbank deposits

Both modern and ancient levee or overbank deposits consist of alternating beds of mudstone and current-laminated, fine-grained sandstone associated with relatively thin slumps, debris flows, channel sandstones, and crevasse-splay deposits.¹⁹ Beds can be laterally continuous or discontinuous depending upon the extent of local slumping or scouring.

Overbank deposits are difficult to unequivocally identify within the Jackfork Group owing to similarities with thin-bedded lobe deposits. An example of this difficulty is the fact that intervals 'e,' 'f,' and 'g' at DeGray Lake Spillway (Fig. 4) have been interpreted by others^{25, 27 28} as levee-overbank or possibly crevasse splay deposits.

Transition zone features

The channel-lobe transition zone within any turbidite system is the region that separates updip, well-defined channel-fill facies from downdip, well-defined lobe facies.¹⁹ As such, their identification in a well bore can be important for predicting updip channel deposits and downdip lobe deposits.

Features considered to be diagnostic of channel-lobe transition zones include the coarsest-grained sediments within the system, amalgamated beds, scoured surfaces, shallow cut-and-fill features, outsize mudstone rip-up clasts, localized disrupted and chaotic units, mud-draped scours, and cross-stratified, tractive-transport deposits.

This type of deposit is well exposed at quarries near Murfreesboro (Souter Quarry) and Arkadelphia, Ark. (Murray Quarry⁴), as well as at the DeGray Lake Spillway described above. At the Souter Quarry, 40 ft of strata near the base of a 200 ft measured stratigraphic succession (interval 'a' on Fig. 3, log E) comprises pebbly sandstone to sandy conglomerates.

The upper bedding plane surfaces of some beds are literally covered by outsize, platy shale clasts or their molds. Other bedding planes are very irregular, giving an almost worty appearance, and exhibit complex fluid flow features. Still other bedding planes contain an abundance of molds of wood fragments.

The most spectacular feature is a series of large, spoon-shaped erosional scours with mutually parallel long axes along the upper bedding plane surface of a few sandstone beds. Where this surface is overlain by another sandstone, the scours are amalgamated sand-on-sand flute molds.

Unfortunately, there is no diagnostic criterion from the outcrop gamma ray log (Fig. 3, log E) for identifying channel-lobe transition zone deposits. Interval 'b' at Souter Quarry (Fig. 3, log E) is tentatively interpreted as lobe deposits.

Scours

Although Mutti and Normark¹⁹ did not specifically include large-scale erosional features such as scours and slumps in their classification, they were described in earlier work,²⁹ so they are included here.

Small- to large-scale scours, slump scars, and other erosional features are well exposed at Big Rock Quarry near North Little Rock, Ark. This quarry exposes a 300 ft thick by 2,750 ft wide section of complexly interstratified, almost flat-lying sandstones and mudstones.

Quarry strata can be divided into three intervals (Fig. 3, log C):

a. a basal, contorted mudstone that represents the top of the lower Jackfork;
b. an overlying 150 ft thick interval of thick, lenticular sandstones and interbedded shales that is in unconformable contact with unit 'a;' and
c. a capping unit of thinner-bedded sandstone and shale accretionary beds that contain unusual shale-clast slurry breccias.⁴

The middle unit 'b' comprises massive, dark gray, fine- to very-fine grained sandstones interbedded with laminated shales and contorted and sometimes stretched mudstones.^4-5 Beds within unit 'b' are discontinuous at all scales owing to erosional scour, slumping, and infilling of sediment into scours and slump scars,³⁰ such that virtually no beds represented on three closely-spaced outcrop gamma ray logs across the quarry (Fig. 5) are correlative.

We interpret these strata as slope canyon fill, probably originally generated by retrogressive slope failure,²⁴ because sandstone unit 'b' sits in angular discordance with underlying contorted mudstone interval 'a,' and because the stratigraphic section in this area is broadly lenticular in the depositional strike direction over a distance of several miles.^{4, 31, 32}

Although unit 'b' appears on the outcrop gamma ray logs (Fig. 3, log C and Fig. 5) to contain thick sandstones interbedded with shales, none of these beds is laterally continuous between the logs (500-600 ft).

E&D interpretations

Log/core interpretation

We have presented examples of stratigraphic and sedimentologic features and associated gamma ray log patterns of the variety of turbidite elements present in outcrops of the Jackfork Group in southwest-central Arkansas, and probably also in eastern Oklahoma.

Different elements exhibit widely-varying geometries and vertical and lateral attributes, as well as variable gamma ray log motifs. In general, gamma ray log patterns characteristic of thick, sharp-based, clean sandstones, which may thin upward, may be interpreted as channel-fill deposits while patterns of thinner-bedded sandstones and interbedded shales may be interpreted as relatively more distal lobe deposits.

However, there can be considerable uncertainty in predicting geometries and attributes away from the well bore with conventional well logs. For example, the one gamma ray log from Big Rock Quarry (Fig. 3, log C) could be interpreted either as laterally discontinuous channel-fill or a series of laterally continuous lobes separated by shales if no outcrop information were available. Thus, we urge caution in attempting such predictions away from the well bore.

To improve interpretations and predictions away from the well bore, cores and/or borehole imaging tools should be considered when setting a drilling plan even though to obtain borehole images at this time requires drilling with water based mud.

The borehole tool delivers a high resolution electrical image of the borehole wall that can be used to interpret and orient structural and stratigraphic features encountered in the formation.

Although cost containment is essential for a successful Jackfork play,² additional up-front expenditures might prove fruitful for longer-term production because cores and borehole imaging logs can be integrated to provide information to aid interpretation of stratigraphic features away from the well bore.

To demonstrate this point, FMS (Schlumberger trademark) log patterns of Pliocene turbidites in the Gulf of Mexico have been matched with outcrop features of Jackfork strata at DeGray Lake Spillway and Big Rock Quarry.⁵ This comparison provides a basis for identifying FMS features that can allow prediction of laterally discontinuous canyon/ channel-fill and laterally continuous lobe elements.

Basically, strata that exhibit numerous stratigraphic discontinuities on borehole images are likely to exhibit such discontinuities at larger scale, while those that exhibit continuity of stratigraphic features on borehole images are likely to exhibit lateral continuity away from the well bore.

Seismic stratigraphic interpretation

The seismic stratigraphic expression of turbidite elements is expected to be variable because resolution depends upon frequency content of the seismic data, depth of the features to be imaged, thickness of the elements, structural and stratigraphic complexities, and acquisition and processing parameters.

The following salient seismic characteristics are for turbidite elements in a mud-dominated system such as the Jackfork Group.

If they are sufficiently large, channel-fill and large scale erosional features are probably the easiest to identify on multifold seismic data because of the truncation of individual seismic reflections (for an example in the Jackfork Group and overlying Johns Valley shale, see Pauli³).

Seismic stratigraphic expression of the fill of these features is highly variable, comprised of subparallel to chaotic to mounded reflections with variable amplitude and continuity. Buried leveed channels are best imaged using horizon slices from 3D seismic data.³³

If sufficiently thick (amalgamated), lobes (sheet sands) can be resolved on multifold seismic data, where they will consist of parallel to subparallel, generally high amplitude seismic reflections with good continuity. 3D horizon slices indicate high amplitude reflections can be traced over a few to hundreds of square miles.

Although on modern fans, overbank deposits adjacent to channels exhibit topographically elevated levees that are wedge-shaped in cross-section and thin away from the channel, buried leveed channels are generally thin and lack this "gull-wing" shape, probably due to compaction. Instead, the reflection character of levees is hummocky to slightly mounded to subparallel, with low to moderate amplitude and generally good continuity.

Channel-lobe transitional features cannot be resolved on 2D seismic data. If there is sufficient relief on erosional features associated with these deposits, 3D seismic horizon slices might be able to image them.

Reservoir factors

Gas production can occur out of any of the turbidite elements described above, as is the case in other turbidite deposits around the world.¹⁹ However, gas volumes and production characteristics might differ among elements.

For example, channel- and canyon-fill sandstones of the type decribed above probably will exhibit good vertical connectivity but will be elongate in character and contain internal discontinuities at a variety of scales so that gas migration will follow a tortuous path.

Lobe (and overbank) deposits contain thin sandstones that are more apt to be laterally continuous away from the well bore, but their vertical connectivity is typically poor. Lobe and overbank deposits often cover larger areas than channel deposits, but net-to-gross is lower in the former.

If Oklahoma Jackfork gas reserves are entirely within tight gas sandstones that are hydraulically fractured, then reservoir simulation probably is not applicable. However, simulation might become important if more porous stratigraphic accumulations are discovered.

In this instance, the design of numerical simulation models will be dependent on the nature of the interpreted depositional architecture. Different approaches towards numerical modeling are necessary for deposits where the interval is layered with good lateral continuity but poor vertical connectivity (lobes and overbank deposits), than would be the case for massive and unlayered intervals that have good vertical continuity but perhaps limited lateral extent (channels and canyon-fill).²⁶

Other factors that will affect simulation model design include the degree of natural fracturing, the probability of sand-on-sand contacts due to fault juxtaposition (as is the case at Hollywood Quarry²⁶), and the bed dip angles.

Conclusions

A variety of turbidite elements comprise upper Jackfork Group strata in southwest-central Arkansas. These elements are also thought to occur in potentially gas-productive Jackfork strata in eastern Oklahoma.

Based upon our studies of turbidite elements in southwest-central Arkansas, we conclude that important attributes such as geometry, lateral continuity, vertical connectivity, and areal extent of strata can be interpreted from conventional well logs, but with considerable uncertainty.

Cores and/or borehole imaging logs can provide important information about these attributes that cannot be obtained with conventional well logs, as can seismic reflection records if the different elements can be resolved.

Although gas production can occur from any of these elements, knowledge of their attributes can ultimately help in maximizing placement and orientation of exploration and development wells, and thus gas production.

A reference list is available upon request to Roger M. Slatt.

The Authors

Roger M. Slatt is professor and head, Department of Geology and Geological Engineering at Colorado School of Mines and director, Rocky Mountain Region Petroleum Technology Transfer Council. He has held technical and managerial positions at ARCO Research, ARCO International Oil & Gas Co., and Cities Service Co. He has written numerous technical publications and teaches courses for AAPG.

Hisham A. Al-Siyabi is pursuing a PhD degree in geology from CSM. His dissertation focus is sedimentology and stratigraphy of deepwater clastics. He received a BS in geology from the University of South Carolina in 1992 and an MS from CSM in 1994.

Charles G. Stone has been employed as a geologist with the Arkansas Geological Commission since 1957. He has worked extensively on the structure, stratigraphy, sedimentology, economic, and environmental geology in the Ouachita Mountains and Arkoma basin of Arkansas. Recent activities include mapping of 178 half-minute quadrangles for the COGEOMAP program.

Paul Weimer is an associate professor in the Department of Geological Sciences and director of the Energy and Minerals Applied Research Center at the University of Colorado at Boulder. He has published several articles and edited books on the petroleum geology of turbidite systems and 3D seismic applications. His research interests include sequence stratigraphy, basin analysis, and 3D seismic interpretation, with emphasis on the Gulf of Mexico.

Eugene T. Williams is an independent consulting engineer in Denver. He has an engineering degree from the University of Saskatchewan and an MBA from the University of Calgary. He is working towards a PhD in geology at CSM, where his thesis topic is the characterization and modeling of deepwater clastics.

Robert J. Davis is currently with Schlumberger Wireline and Testing, Jakarta, Indonesia. Prior experience includes several years with Schlumberger in New Orleans, where he specialized in borehole imaging interpretation. He has co-led courses for AAPG and internally for Schlumberger.

Douglas W. Jordan is with ARCO Venezuela Inc. and was previously with the Reservoir Evaluation Group of ARCO International as well as with Cities Service Co. He has co-authored numerous papers on applied reservoir geology, fluvial sedimentology, petroleum geology, trace fossils, and outcrop logging.

Copyright 1997 Oil & Gas Journal. All Rights Reserved.