Distribution of depositional complexes
The midcontinent of the United States straddled the Pennsylvanian paleoequator (Heckel, 1977). Warm equatorial seas were conducive to carbonate-sediment production through inorganic precipitation as well as by biochemical processes. Ibis setting persisted throughout the Carboniferous. However, siliciclastic sediments transported into an equatorial basin quickly overwhelmed the carbonate-producing agents. This occurred during regressions when fluvial systems drained nearby cratonic areas and more distant marginal orogenic belts.
The isopach and sandstone-isolith maps for the Banzet formation (figs. 28 and 29) outline the positions of preserved deltaic complexes. Isolated sandstone lenses on the isolith map probably represent the reworked remnants of relatively short-lived delta lobes. The well-log signatures that are associated with isopach thicks show that these thicks consist of blocky channel complexes and coarsening-upward sandstone bars (fig. 30). Using these interpretations as guides, two paleogeographic reconstructions were made showing the positions of delta complexes during the deposition of the upper and lower portions of the Banzet (figs. 56 and 57).
Figure 56--Paleogeographic reconstruction of study area at time that lower portion of Banzet Formation was being deposited.
Figure 57--Paleogeographic reconstruction of study area at time that upper portion of Banzet formation was being deposited.
An unpublished study made by Staton in Butler County, Kansas, indicates that there may be some thin limestone units in the lower portion of the Banzet (Brady, personal communication, 1987). This part of the study area is adjacent to the southern extension of the Nemaha uplift, which was apparently not a positive feature during this time. It is possible that some of the limestone-producing environments that existed at this time west of the uplift extended into the southwestern portion of the study area. If this were the case, then several feet of limestone may be present in the lower portion of the Banzet where I have placed some tightly cemented sandstone units. These possible limestone beds are not accounted for on the maps or cross sections in this report, but their presence would not significantly affect the interpretations made here.
Two mechanisms seem to control the distribution of sandstone bodies. First of all, well-documented eustatic fluctuations in sea level (e.g., Heckel, 1986) determine the positions of shorelines and siliciclastic point sources relative to the study area. Only during periods of eustatic regression were shorelines and deltaic systems located within the study area. During times of eustatic transgressions, shorelines and deltas were displaced to the north and east. Heckel (1986) has shown that at least two cycles of sea-level change are recorded within the Banzet.
The second mechanism that controlled sand-body distribution was delta-lobe shifting. As distributary channels overextended themselves seaward, they shifted laterally through the diversion of flow through crevasses formed during floods and storms. Abandoned portions of a delta system subsided as lobal sediments consisting mostly of mud became compact. Local transgressions occurred as marine waters reoccupied submergent lobes.
Eustatic changes and autocycles
Nature of transgressive and regressive sequences
Many transgressive and regressive sequences are found in the Banzet formation. Transgressive sequences commonly consist of sandstones overlain by marine shales. In some cases the sandstones fine upward, then coarsen upward, and are then capped with shale. This sequence represents delta-progradational sands, delta-destructive sands, and prodeltaic and shelf muds respectively. Similar sequences have been observed in the lower portion of the Cherokee Group and have been interpreted as being parts of deltaic complexes (e.g., Visher and others, 197 1). In the upper part of the Banzet, delta-front marine sandstones are capped by the Breezy Hill Limestone Member, representing a time when siliciclastic point sources were displaced northward and eastward. The end of Banzet deposition is marked by a widespread, black, phosphatic shale, the Excello shale. This shale, as well as the Oakley shale of the Verdigris Formation, represents deep-water anoxic conditions that covered the study area during times of widespread transgression (Heckel, 1977).
Regressive sequences commonly consist of delta-distributary sandstones that overlie marine shales and are sometimes capped with seatrocks and coal beds. These sequences generally are thicker than the transgressive sequence. Some transgressive sequences are extremely thin, such as the uppermost sequence that only consists of the Excello black shale that directly overlies the Mulky coal (fig. 58). Thin transgressive sequences are caused by low rates of sedimentation that accompanied rapid landward shifting of siliciclastic point sources.
Figure 58--Eustatic cycles and autocycles as they were recorded along cross section G-G'. Only initial transgressive events are labeled (T1-8). T1, T3, T6, and T7 seem to record eustatic transgressions, while others record abandonment and subsequent submergence of deltaic lobes.
Two widespread transgressive events are recorded by the Oakley shale, which marks the base of the Ardmore-Banzet interval, and the Excello shale, which marks the top of the interval. These units were observed on outcrops in nearly all wells and can be traced laterally throughout the midcontinent region and into the Appalachian basin to the east (e.g., see Ravn and others, 1984, for a summary of previous work). These two units represent transgressions that affected areas beyond the study area and were eustatic in nature.
The question arises as to whether other eustatic events can be observed within the Ardmore-Banzet interval, or whether the Banzet formation represents one eustatic regression between the two well-marked transgressions. When individual wells and cross sections made from closely spaced well logs are analyzed, several regressive-transgressive sequences can be delineated. The Bevier, Iron Post, and Mulky coal beds appear to mark the tops of regressive sequences that are each capped by transgressive marine shales (T3, T4, T7 on fig. 58). The transgressive-regressive cycle formed above the Iron Post coal and below the Breezy Hill Limestone Member by the marine shales of the Kinnison (T4 on fig. 58) and on the overlying generally coarsening-upward sequence of silty mudrocks and thin sandstones has not been previously described as an allocycle. Therefore, there are five regressive-transgressive cycles within the upper Cherokee interval that seem to be regional in extent (fig. 58): 1) between the top of the Oakley shale (T1) through the Ardmore limestone; 2) between the base of the mudrock-sandstone sequence that lies above the Ardmore limestone to the top of the marine shales (T3) that overlie the Bevier coal; 3) between the base of the coarsening-upward sequence that contains the Iron Post coal to the top of the Kinnison shale (T4); 4) between the base of the mudrocks and sandstones that overlie the Kinnison shale to the top of the Breezy Hill limestone (T6); and 5) the rocks above the Breezy Hill, including the Muk coal, to the base of the Excello shale (T7). With the exception of cycle 3, these cycles correspond to those observed by Heckel (1986) to be interregional.
Within the widespread eustatic regressive hemicycles, many local regressive-transgressive cycles can be observed. These cycles correspond to the progradation, abandonment, and submergence of deltaic lobes and associated shoreline features. Autocycles are restricted to eustatic regressive hemicycles because of the removal of siliciclastic point sources from the study area as shorelines were displaced landward during eustatic transgressions. Therefore, during transgressions, the only sediments deposited on the Cherokee shelf and adjacent areas consisted of thin layers of pelagic clays and carbonates--rock units that are observable on geophysical well logs. Eustatic regressions allowed siliciclastic depositional systems associated with shorelines and river mouths to reestablish themselves, and autocyclic processes resumed. The resulting siliciclastic sequences are dominantly regressive or upward-shoaling in character and are often capped by thin, destructional (reworked) sandstone sheets or pelagic shelf muds. Although their characteristics were determined by local fluvial and deltaic processes, stratigraphic distribution of autocycles was controlled by interregional mechanisms, most likely eustatic sea-level changes caused by climatic fluctuations.
Effective porosity and varying diagenetic styles
Many of the Banzet sandstone thicks illustrated on a sandstone-isolith map (fig. 29) are areas of petroleum production (fig. 59). However, not all the sandstone bodies within these sandstone thicks produce petroleum. In order for petroleum to accumulate in any rock, all the following conditions must be met: 1) the potential reservoir rock must have had communication with a hydrocarbon source rock at the time that petroleum was being generated, 2) the potential reservoir must have had effective porosity at the time that petroleum was migrating, and 3) a seal must be in place to prevent petroleum from migrating completely through the reservoir without accumulation.
Figure 59--Map of eastern Kansas showing distribution of oil fields that produce from pools in the Banzet formation. Dashed line is 10-ft (3-m) sandstone isolith from map on fig. 29. Note that some fields produce also from pools not in the Banzet. Data from Kansas Geological Survey oil production files updated through 1985.
The marine mudrock units of the Cherokee Group have been shown to be rich in organic carbon and have been interpreted as the most likely source rocks for the encased sandstone units on the Cherokee shelf (Baker, 1962). Organic geochemical data from Cherokee mudrocks provided by Hatch and others (1984) show that gray shales commonly have organic carbonate contents of 4-5%, and black shales often have more than 12% organic carbon. However, maturation modeling by time-temperature index calculations indicates that Pennsylvanian mudrocks in the Forest City basin should be immature, and that shales lower in the Paleozoic are more likely to be the source of oil in the Cherokee (Newell and others, 1986). It is possible that Cherokee oils have different sources depending upon whether they were generated in the more deeply buried portions of the Cherokee shelf or the shallower Forest City basin. Further analyses and petroleum-source-rock fingerprinting will ultimately allow the petroleum contribution of each Paleozoic mudrock interval to be evaluated across the study area.
Marine and delta-plain mudrocks seem to form the updip seals needed to form traps in the upper Cherokee. The main difference between petroleum-producing Banzet sandstone units and those that do not produce is the existence or nonexistence of effective porosity. In general, the thin delta-destructive sandstones have very little porosity. Most pore spaces are filled with carbonate cements that were precipitated early during the sediment's burial history. On the other hand, distributary-channel shoestring sandstone bodies are the most prolific petroleum producers in the study area. These units generally have greater than 15% porosity, a significant portion of which is secondary in origin.
While sands deposited in marine environments during the Pennsylvanian had interstitial waters saturated with respect to calcite, sands deposited in fluvial and delta-distributary channels were probably acidic and not saturated. As a result, channel sands were not pervasively calcite cemented during the early phases of their diagenetic histories. Instead, some grains were coated with chlorite. As burial continued and silica-saturated fluids entered channel sands from adjacent compacting mudrocks, silica overgrowths formed on detrital grains not coated with chlorite. Rocks with extensive chlorite coatings have very small mounts of silica overgrowth with resulting higher porosities and permeabilities (Woody, 1983; Lardner, 1984; Nelson, 1985).
Pore spaces that survived early stages of carbonate and silica cementation were subjected to at least one stage of carbonate cementation. Concomitant with, or just prior to late-stage carbonate cementation, silica grains and overgrowths were etched and significant volumes of silica were removed from the affected sandstone units. Secondary pores were formed prior to hydrocarbon migration as formation waters became acidic once more, dissolving much of the late-stage carbonates from some units.
Processes affecting reservoir properties
Depositional and early diagenetic conditions
Sandstone porosities and permeabilities are related to their depositional environments and diagenetic histories. Geochemical conditions at depositional sites seem to have controlled the nature and extent of early precompaction diagenetic processes. As a result, the well-sorted, reworked delta-destructional sandstones may have had high original porosities and permeabilities, but they also were more susceptible to early carbonate cementation. In this case, the free movement of carbonate-saturated waters under alkaline conditions led to the rapid, pervasive cementation of these rocks.
Another factor which may have contributed to early cementation was the interlayering of mud with thin sands. As muds began compacting, first carbonate-saturated and later silica-saturated fluids were flushed through the porous sand layers. Sandstones that were not pervasively cemented with carbonate were later subjected to silica cementation as diagenetic alterations of silt-sized and smaller feldspar grains and grains of unstable clay minerals released silica into waters that were being expelled during compaction.
Enhancement of reservoir properties
Channel-sandstone units have the highest porosities and probably the highest permeabilities within the Banzet formation. Significant portions of pore spaces in these units are secondary in origin (see appendix B). This type of porosity was recognized by noting the geometries of pores and by the presence of remnant grain particles. Secondary pores either have cross sectional shapes and sizes similar to those of grains or they form embayments into remnant grains. Because all sandstone samples were impregnated with blue-dyed polyester resin before thin sections were cut and ground, all naturally occurring pores appear blue in thin sections. Any materials that were removed during thin-section manufacture are colorless in thin sections. Therefore, secondary pores recognized using geometric criteria could not be misinterpreted as plucked particles or vice versa.
Secondary pores formed as a result of dissolution of feldspars and other unstable constituents and replacement of detrital and authigenic silica by late-phase carbonate cements, followed by still later dissolution of the carbonate cement. Both mechanisms of secondary porosity development seem to have taken place. Some specimens show secondary pores in rocks that have no evidence that carbonate cements were present (fig. 60). Other specimens have secondary pores that clearly were developed by dissolution of carbonate cements that had previously replaced parts of siliciclastic grains and siliceous overgrowths (fig. 61). The two phases of dissolution were mutually exclusive, since the chemistry required to dissolve carbonates (e.g., low pH or low temperature) is much different than that required to dissolve silicates (e.g., high pH or elevated temperatures). However, silicate dissolution possibly was concomitant with carbonate replacement of silicates.
Figure 60--Photomicrograph showing secondary pore spaces (P) formed as a result of feldspar (F) dissolution in a sample that does not show evidence of previous carbonate cementation. Sample 16.6, Bailey-Lohrengel 18 core (core #6 on fig. 2 1), Anderson County, Kansas. Plain polarized light; bar = 0.1 mm.
Figure 61--Photomicrograph of secondary pores (P) formed by dissolution of carbonate cements (C). Sample 1037.8, M. C. Colt-W. Lauber-101A well, Woodson County, Kansas. Plain polarized light; bar = 0.1 mm.
Keys to predicting quality of reservoir characteristics
The depositional styles depicted in this study show that potential reservoir rocks trend primarily northeast-southwest and, to a lesser extent, northwest-southeast. These are not new observations in that many authors, including Bass (1936) and Hulse (1979), have shown that these trends are common throughout the Cherokee Group. However, once sandstone trends are either delineated or predicted, high-quality reservoir characteristics are not always assured. Walton and others (in press) show that cross-bedded channel sandstones have the highest porosities and permeabilities among the Cherokee units they analyzed. Although this study does not include permeability measurements, point-counted porosities indicate that these relationships hold for the Banzet sandstone units analyzed. Reservoir characteristics vary within these units because of changes in the nature and intensities of diagenetic alterations. Therefore, the key to finding sandstone bodies with high porosities and permeabilities lies in the interpretations of the rock's diagenetic history in relation to depositional settings and subsidence history.
Because the tectonic setting for this portion of the midcontinent was one of relative stability during the time that the Banzet formation was being deposited, eustatic sea-level changes may have been the only significant cause of perturbations in sediment-accumulation rates. Therefore, the positions of sandstone bodies at the time of Desmoinesian and post-Desmoinesian eustatic transgressions and regressions may be an important key in determining which diagenetic alterations affected each sandstone body. This phase of the study is still in progress at this time and will be discussed more fully in a future paper.
Kansas Geological Survey, Geology
Placed on web Oct. 27, 2010; originally published 1989.
Comments to email@example.com
The URL for this page is http://www.kgs.ku.edu/Publications/Bulletins/GS3/06_geog.html