Cycles and Cycle Hierarchies
Wanless and Weller (1932) introduced the term "cyclothem" to describe upper Carboniferous (Pennsylvanian) cyclicity of the Illinois basin. Jewett (1933) applied the term to the cyclicity he observed in the Permian rocks in Kansas. This description was modified and elaborated by Elias (1937), who placed all the major facies encountered within Permian cycles into an idealized depth-related sequence (see also Moore et al., 1934). A number of detailed sedimentary and paleontological studies of individual Wolfcampian cyclothems and their member-scale lithologic units followed (e.g., Imbrie, 1955; Hattin, 1957; Lane, 1958; Laporte, 1962; McCrone, 1963; Imbrie et al., 1964). Moore (1936) introduced the concept of "megacyclothems" based on his work in the Virgilian of Kansas, and subsequently (Moore, 1964) divided the lower Permian section into megacyclothems, which he correlated across the Kansas outcrop belt.
The facies sequence of Wolfcampian cyclothems typically begins with a thin marine limestone overlain by a gray fossiliferous shale/mudrock. One or more additional limestone-shale/mudrock alternations may follow. An interval of variegated red and green mudrocks with extensive paleosol development lies above these shallow marine facies. This general pattern persists to the top of the Chase Group. Thin dolomite and dolomitic mudrock units occasionally interrupt the thick interval of finer-grained red siliciclastics and evaporites in the overlying Sumner Group. Cyclicity developed in the Hutchinson Salt Member of the Wellington Formation appears to be similar to the Wolfcampian cycles. Also, cyclicity persists in the siliciclastic-dominated intervals of the Sumner and Nippewalla Groups, but detailed stratigraphic classification and nomenclature remain unresolved due to limitations of surface exposures and efforts to distinguish and correlate individual cycles.
The Wolfcampian cyclothems were originally defined using repetitive facies patterns rather than discontinuity surfaces. Elias (1937) constructed smooth depth curves for these cyclothems, even though actual cycles often did not closely match his ideal facies sequence (Mudge and Yochelson, 1962). Nonetheless, Elias (1937) constructed sea-level curves in which the environmental change from deepest to shallowest water were centered on those facies. In particular, he relied on a depth-controlled biofacies model to identify the points of maximum transgression for his cycle model. This resulted in very different curves being drawn for cycles of similar lithologic complexity. The absolute magnitude of sea-level fluctuation and the stratigraphic position of maximum transgression have been subsequently disputed.
Until recently, the focus has remained on the bathymetric interpretation of specific facies rather than on the character of the surfaces that bound them. Although discontinuities are very abundant within the Wolfcampian and are associated in many cases with sharp lithologic contacts, they have received comparatively little attention. Understanding the meaning and temporal duration of these surfaces is critical for defining cycle patterns and periodicities, and for interpreting their regional significance (Miller and West, 1993, 1998). Recent efforts have been made to reinterpret the cyclothems of the Council Grove and Chase Groups within a sequence stratigraphic context (Mazzullo et al., 1995, 1997; Mazzullo, 1998; Miller and West, 1998; Boardman and Nestell, 2000; Olszewski and Patzkowsky, 2003; Boardman et al., 2009). Furthermore, Mazzullo (1998) has attempted to define systems tracts within the Chase Group cycles. Such efforts mark an important future direction for cyclostratigraphic research in the midcontinent. In other recent work, the occurrence of certain trace-fossil associations have been found to correlate with hiatuses in deposition associated with transgressive and regressive events (Chaplin, 1996). Thus, ichnology may provide another tool for identifying these important surfaces in the subsurface.
Sequence boundaries are defined as unconformities resulting from erosional truncation or subaerial exposure and their correlative conformities (van Wagoner et al., 1988). Although not necessarily associated with maximum eustatic sea-level lowstand, sequence boundaries do represent the maximum seaward limit of terrestrial sedimentation and the time of maximum subaerial exposure of the shelf (Posamentier et al., 1988). However, on the shelf, depositional sequence boundaries are difficult to recognize beyond the limits of individual valley-fill systems (van Wagoner et al., 1990; Aitken and Flint, 1995). This problem is accentuated in the Permian of Kansas where incised valleys and valley fills are virtually absent. Furthermore, the red and green mudrock intervals of the Permian cyclothems are commonly composed of multiple subaerial exposure surfaces ranging from desiccation cracks to mature paleosols (Joeckel, 1991; Miller and West, 1993; Miller et al., 1996). Stacked paleosol profiles are ubiquitous features. The problem of identifying sequence boundaries thus becomes one of the determining criteria for selecting among several exposure surfaces in a stacked series. Although defining precise boundaries is difficult, general sequence boundary intervals can nonetheless be recognized. When rocks of marine origin overlie paleosols, sequence boundaries are more easily recognized as shown by Olszewski and Patzkowsky (2003).
Another important surface in depositional sequences is the marine transgressive surface. Correlation of midcontinent Wolfcampian cyclothems, on the outcrop and in the subsurface, has historically been based on marine limestones that can be recognized from Nebraska to Oklahoma. The base of the stratigraphically lowest, fossiliferous, fully marine limestone occurring above a paleosol-bearing interval is equivalent to the transgressive surface (van Wagoner et al., 1988). Commonly, these marine limestones directly overlie and partially truncate the uppermost paleosol profiles. The contacts often appear to be erosive, although little or no relief is evident at an outcrop scale, and are typically overlain by intraclastic beds up to 20 cm (0.7 ft) thick (Miller and West, 1998). These transgressive surfaces can be identified with relative confidence and also can be used as informal boundaries for the cyclothems (figs. 16A-C, 28A-D).
Meter-scale cycles are both ubiquitous and prominent within the Wolfcampian cyclothems of eastern Kansas (Archer et al., 1995; Miller and West, 1993, 1998; Olszewski and Patzkowsky, 2003). Flooding surfaces overlie paleosol profiles and other indicators of subaerial exposure, or mark sharp changes in depth as indicated by lithology and fossil content. Thin (<2-cm-thick) skeletal and/or intraclastic lags mark these cycle-bounding flooding surfaces. A variety of carbonates ranging from marine bioclastic limestones to laminated intertidal limestones and calcareous mudrocks overlie the flooding surfaces. These carbonate facies are followed by siliciclastic units that include gray fossiliferous mudrocks and variegated red and green mudrocks. Although rocks containing evidence of subaerial exposure cap many of these small-scale cycles, they do not show basinward shifts of facies.
At the other end of the temporal hierarchy from meter-scale sequences, Ross and Ross (1988) have argued that the upper Carboniferous (Pennsylvanian) and Permian cyclothems can be grouped into larger-scale transgressive-regressive cycles that can be recognized and correlated among the world's cratonic shelves using fusulinids, bryozoans, and other taxa. They recognized four such globally correlated cycles within the Council Grove and Chase Groups. The first begins at the Roca Shale and extends upward to the Grenola Limestone, the second extends from the Eskridge Shale to the Wreford Limestone, the third from the Matfield Shale to the Barneston Limestone, and the fourth from the Doyle Shale to the Nolans Limestone. Ross and Ross (1988) also identified four global cycles within the overlying Sumner and Nippewalla Groups. The Wellington Shale, Ninnescah Shale to Stone Corral Formation, Harper Sandstone to Salt Plain Formation, and Cedar Hills Sandstone to Dog Creek Formation represent these cycles. However, the lack of fossils, faunal provinciality, and the common occurrence of restricted environments in stable cratonic areas prohibited them from recognizing and correlating cycles in the upper Permian. Interestingly, the thickest and most widespread limestone units cap the lower Permian (Wolfcampian) cycles of Ross and Ross (1988). Similarly, the variegated mudrock units at the base of these cycles all contain evidence of significant times of subaerial exposure with some of the most mature paleosol profiles in this stratigraphic interval. This suggests that the cycles delineated by Ross and Ross (1988) may indeed record the transgressive and regressive extremes of a higher-level cyclicity.
Cycle durations for the upper Carboniferous (Pennsylvanian) cyclothems of the midcontinent have generally been estimated at 200,000 to 400,000 years (Heckel, 1986). This periodicity provides a first approximation for cyclothem duration in the Permian. Within the Permian, generally 4--6-m-scale cycles occur within each cyclothem (Miller and West, 1993) and thus probably record time-periods of 40,000 to 100,000 years. These values are in broad agreement with the common occurrence of mature paleosols capping these meter-scale cycles that likely required tens of thousands of years to form. The large-scale, globally correlated cycles identified by Ross and Ross (1988) are estimated to range from 1.2 to 4 million years in duration, with an average of 2 million years.
Higher-level cycle sets were also recognized in the upper Carboniferous (Pennsylvanian, Missourian) strata (Watney et al., 1995). These cycle sets are based on stacking patterns of lower-level cycles along the shelf margin bordering the northern Anadarko basin and parallel changes in the development of cycle components, particularly "core" shales and paleosols. Trend analysis of Th/U ratios obtained by gamma-ray spectral logs show distinctive and coupled patterns of redox conditions in these strata at different positions on the Kansas shelf (Watney et al., 1992, 1995).
Controls on Cyclicity and Sedimentation
Like the upper Carboniferous (Pennsylvanian) cyclothems of the midcontinent, sedimentary cycles of the Wolfcampian probably record glacio-eustatic fluctuations in sea level as well as paleoclimatic and tectonic influences on clastic-sediment supply. Wanless and Shepard (1936) were the first to propose a glacio-eustatic model for cyclothem formation in the upper Carboniferous (Pennsylvanian). This model was followed by Elias (1937) in his interpretation of the lower Permian cycles. The widely accepted cyclothem model of Heckel (1977), based on the Missourian of Kansas, also assumes primary eustatic control. Recent models for Permian cyclothems have incorporated climatic as well as eustatic controls on the observed lithologic patterns (Miller and West, 1993; Miller et al., 1996; West et al., 1997; Olszewski and Patzkowsky, 2003).
Because upper Carboniferous (Pennsylvanian) to Permian sedimentary cycles coincide temporally with Gondwanan glaciation (Crowell, 1978; Veevers and Powell, 1987; Crowley and Baum, 1991), glacio-eustasy is recognized by most workers as the predominant forcing mechanism for cycle formation. Climatically forced eustatic models have been supported by comparisons of the estimated periodicities of upper Carboniferous (Pennsylvanian) cyclothems with Milankovitch orbital modulations (Heckel, 1986, 1994; Boardman and Heckel, 1989; Boardman and Nestell, 1993). The estimated cycle duration and the observed nested hierarchy of cycles are broadly consistent with glacio-eustatic curves reconstructed from the Pleistocene isotopic record (Denton and Hughes, 1983; Crowley and North, 1991). It also is significant that the glacially driven Pleistocene cycles are asymmetric with rapid sea-level rises followed by slow falls interrupted by minor deepening. This pattern seems to mimic that seen in the lower Permian (and upper Carboniferous) cycles with abrupt transgressive surfaces at the bases of cyclothems followed by a series of flooding surface-bounded meter-scale cycles (Watney et al., 1991; Miller et al., 1996).
Changes in the global extent of glacial ice cover would be expected to affect both the potential amplitude of glacio-eustatic sea-level fluctuations and the amount of exposure of the midcontinent "shelf" during lowstand. A recent summary of the age distribution of glacial deposits by Frakes et al. (1994) suggested two peaks in continental glaciation: one in the Westphalian (Desmoinesian) and one in the Asselian-Sakmarian (Wolfcampian). Although both Desmoinesian and Wolfcampian cyclothems were formed during times of extensive Gondwana glaciation, they are very different in their lithologic compositions (West et al., 1997). Generally, Desmoinesian cycles are siliciclastic with associated coal beds, Missourian cycles are largely carbonate, siliciclastics are more conspicuous in Virgilian cycles, and well-developed paleosols typify the Wolfcampian. These differences probably reflect the increasing emergence of continents during the Carboniferous and Permian as well as long-term climatic change. Emergence of the shelf and shallowing of seaways that connected Kansas with the Permian ocean are supported by the evaporitic cycles that dominated Leonardian deposition. The rapid waning of glaciation at the end of the Wolfcampian explains the absence of well-developed cyclothems in post-Leonardian rocks.
Several observations at both the meter and cyclothem scale indicate that climate, in addition to sea-level change, must be considered to adequately explain the lithologic expression of the cyclicity. First, a consistent carbonate-to-siliciclastic pattern of meter-scale sequences exists within both the open-marine facies and paleosol-bearing intervals of cyclothems. Both carbonate and clastic units display a range of facies recording environments from open marine to subaerial exposure (Miller and West, 1993). A simple bathymetric facies model does not work for these cycles. Furthermore, the shallow marine and paralic facies of the carbonate units typically contain dry-climate indicators such as replaced-gypsum nodules, evaporite-crystal molds, laminated dolomitic mudrocks, and tepee structures. When present in the subsurface, bedded gypsum commonly is closely associated with carbonate units. For example, within the Council Grove and Chase interval in Riley County, subsurface gypsum beds occur above the Burr limestone, above the Middleburg limestone, immediately below the Kinney limestone, and at the top of the Fort Riley limestone (Twiss, 1991a). Thus, relatively arid conditions seem to be associated with times of carbonate deposition. These observations are consistent with a model for climatic control over facies development proposed by Cecil (1990). In this model, clastic sediment transport is predicted to be highest in seasonal wet-dry climates and lower in both arid and tropical wet climates (see also Wilson, 1973; Perlmutter and Matthews, 1989). Carbonates and evaporites accumulate during arid and semiarid conditions, and mappable coal beds form during relatively wet climates where clastic influx is low. This model of climatic control over chemical and clastic deposition can be combined with paleosol evidence and sea-level curves to yield integrated models for cyclothem formation (Miller and West, 1993).
Halite, in cycles dominated by halite, represents an extension of the dry-climate carbonate phase. The thick mudrock that caps the halite-dominated cycles was deposited late during the regression, and the associated dissolution of the underlying halite can be explained by a change from dry to a seasonal wet-dry climate, analogous to the carbonate-siliciclastic cycle. Alternatively, inundation during halite accumulation may have flooded the basin margin and inhibited clastic influx.
A persistent pattern has been recognized in the characteristics of the paleosols and exposure surfaces of successive meter-scale cycles within a given cyclothem (Miller et al., 1996; McCahon and Miller, 1997). For most of the thicker variegated mudrock units within the Council Grove and lower Chase Groups (i.e., Roca Shale, Eskridge Shale, Blue Rapids Shale, Speiser Shale, and Matfield Shale), a very similar vertical succession of paleosol profiles has been observed. Paleosols from the lower part of these units have reddish-brown calcic profiles with carbonate nodules and rhizocretions indicative of semi-arid to subhumid conditions. By contrast, pseudoanticlines (mukkara structures) and other features of vertic paleosols that form under highly seasonal monsoonal conditions characterize the uppermost greenish-gray paleosols within these units. When present, salt-influenced natric paleosols occur near the base or top of variegated mudrock intervals (i.e., Roca Shale, Hooser Shale, Easly Creek Shale) or within gray-mudrock units sandwiched between marine carbonates (i.e., Legion shale and Salem Point shale) (McCahon and Miller, 1997). These lithologic and paleosol patterns suggest that climate changed together with relative sea level during the course of cyclothem deposition. In order of decreasing aridity, a spectrum of features can be recognized as follows: evaporites, tepee structures, laminated dolomitic mudrocks, natric paleosols, calcic paleosols, vertic paleosols, and organic-rich shales. The consistent vertical pattern of these climatically sensitive sedimentary features within Wolfcampian cycles indicates that climate changed from arid to progressively wetter and more seasonal conditions during the formation of each cyclothem (Miller et al., 1996; McCahon and Miller, 1997).
A spectrum of lithologic and paleosol climate indicators provides a basis for reconstructing climate change within the early Permian (Wolfcampian) of Kansas. Patterns of change in these climate indicators reveal a hierarchy of cyclic climate changes from those occurring during the formation of single paleosols (recording tens of thousands of years) to long-term climate trends encompassing the entire late Paleozoic.
At the highest level of stratigraphic resolution, climate change can be recognized within meter-scale cycles (Miller et al., 1996). The consistent carbonate-to-clastic pattern of these meter-scale cycles, regardless of their position within the cyclothem, suggests some climatic control (fig. 71). The paleosols capping these cycles are commonly composite or polygenetic profiles and provide additional climate information. In nearly all cases, the latest stage of pedogenesis seems to have occurred under the driest conditions. The reddish-brown calcic paleosol profiles, for example, have the highest carbonate concentrations above horizons showing well-developed clay cutans, suggesting carbonate precipitation following clay illuviation.
Figure 71--Interpreted relative sea-level and climate curves for the Roca Shale and Grenola Limestone. Arrows mark flooding surfaces that bound meter-scale parasequences. The vertical line for the sea-level curve marks the sediment surface, a position to the left represents subaerial exposure and paleosol development. The vertical line for the climate curve indicates the threshold between carbonate deposition (drier climate) and siliclastic deposition (wetter climate) (modified from McCahon and Miller, 1997).
The lithofacies and paleosols of successive meter-scale cycles within cyclothems of the Council Grove and Chase Groups indicate climate trends toward wetter and more seasonal climates (fig. 71). The carbonates and mudrocks of meter-scale cycles immediately above transgressive surfaces usually do not contain diagnostic climatic indicators. However, the shallow-marine and paralic facies of the carbonate units of subsequent meter-scale cycles typically contain dry-climate indicators. Within the overlying variegated-mudrock facies of cyclothems, meter-scale cycles with salt-influenced natric paleosols occur below those with calcic paleosols, thus recording semiarid conditions that, in turn, occur below those with vertic paleosols, indicating monsoonal climates. Based on the interpreted maturity of paleosol development, the boundary intervals of the cyclothems tend to be located near the middle of the variegated mudrock facies (Miller and West, 1998). A straightforward reading of this pattern would result in the conclusion that the driest climates were associated with early highstand (i.e., occurring a few meter-scale cycles above the transgressive surface), and the wettest climates with lowstand (i.e., within meter-scale cycles between the sequence boundary and transgressive surface) (McCahon and Miller, 1997; Miller and West, 1998).
Long-term Climate Changes During the Permian
Parrish (1998) has addressed pre-Quaternary climate changes based on the geologic record. Based on the distribution of coals, red beds, eolian sandstones, and evaporites, a consistent trend toward increased aridity has been recognized in North America from the late Carboniferous into the Triassic (Parrish, 1993; Parrish and Peterson, 1988; Kutzbach and Ziegler, 1993; Golonka and Ford, 2000; Gibbs et al., 2002). This drying trend also is recorded in the paleobotanical record (Phillips and Peppers, 1984; Phillips et al., 1985; Cross and Phillips, 1990; DiMichele and Aronson, 1992; DiMichele et al., 2001; DiMichele et al., 2009). Within the midcontinent, a long-term climate trend toward increased aridity is recorded by the increase of red terrigenous clastics and evaporites and the virtual absence of coal beds and channel sandstones within the Permian relative to the upper Carboniferous (Pennsylvanian) (West et al., 1997). In particular, the Wolfcampian seems to have been a time of major climatic transition from generally wetter conditions in the Virgilian to significantly drier conditions in the Leonardian and Guadalupian. This trend is reflected in the changing lithologic composition of cyclothems (West et al., 1997). In particular, carbonates become progressively dominated more by sabkha-like dolomitic and peritidal facies with nodular evaporites (Mazzullo et al., 1995, 1997), and the clastic intervals become increasingly silty. Calcic and vertic paleosols characteristic of the Council Grove Group give way to more poorly developed paleosols and evidence of evaporites. No well-developed vertic paleosols, indicating monsoonal conditions, appear above the Wymore Shale Member of the Matfield Shale.
Two global climate models attempt to explain the climate trends within Wolfcampian glacial/interglacial cycles, as well as the longer-term trend toward drier climates throughout the Permian. One of these models assumes a zonally organized global climate system, and the other emphasizes the establishment of a non-zonal circulation pattern caused by an intensifying Pangean "megamonsoon."
Using a zonal model of global circulation, Matthews and Perlmutter (1994) attempted to predict the general direction and extent of climatic change during glacial-interglacial cycles. The humid and arid climatic zones of the earth today are largely controlled by the position of global atmospheric-circulation cells. Latitudinal shifts in these circulation cells associated with Milankovitch-driven cyclicity would have resulted in the latitudinal migration of climatic zones. During glacial periods the mid-latitude dry high-pressure atmospheric-circulation cells would shift toward the equator, and during interglacials the humid equatorial low-pressure cell (Inter-Tropical Convergence Zone) would expand into higher latitudes (Matthews and Perlmutter, 1994). These shifts of circulation cells during glacial-interglacial cycles could generate climate alternations between humid tropical during interglacials to arid temperate during glacials, but only for latitudes between 15 and 20 degrees. However, more recent paleogeographic reconstructions (Witzke, 1990; Scotese and McKerrow, 1990; Scotese and Golonka, 1992) do not place the midcontinent above 10 degrees north by Wolfcampian time (compare the position of Kansas in figs. 51, 52).
The zonal model also may be relevant to understanding the long-term climate trend toward increasing aridity from the late Carboniferous through the Permian. One problem in applying this zonal global climate model to the Permian of the midcontinent is the uncertainty in determinations of paleolatitude. Was the progressive northward movement of Pangea sufficient to move the region of the midcontinent into higher and drier latitudes? The paleogeographic reconstructions of Witzke (1990) show a northward shift of Kansas from near the equator to about 20 degrees north from the middle Morrowan (Westphalian) to the Guadalupian, a change in latitude that might well have had a major impact on long-term climate change. However, alternative reconstructions by Scotese and McKerrow (1990) show only a shift from 3 degrees south to 12 degrees north for the same time interval.
As an alternative to the zonal model, global circulation models for Pangea (Parrish et al., 1982; Kutzbach and Gallimore, 1989; Crowley et al., 1989; Patzkowsky et al., 1991) have strongly indicated the presence of a monsoonal circulation pattern that intensified from the early Permian to a maximum in the Triassic. Development of this Permian "megamonsoon" was a consequence of the assembly of the Pangean supercontinent and the increasing symmetry of the landmass across the equator (Parrish, 1993). Establishment of a Permian monsoon would have disrupted zonal circulation and diverted the moisture-laden equatorial easterlies flowing from the Tethys, resulting in a drying of equatorial Pangea. With the development of a monsoonal circulation pattern, the drying of the midcontinent in the Permian can be explained even if the Kansas area remained within low paleolatitudes.
Miller et al. (1996) have proposed a cycle model assuming that the climate of the midcontinent was strongly affected by a Pangean monsoon, and fluctuations in the intensity of the monsoon produced oscillations between wetter and drier conditions. Both climate models (Kutzbach and Guetter, 1984) and Pleistocene paleoclimate data (Fairbridge, 1986; Crowley and North, 1991; McKenzie, 1993) indicate that monsoons are strengthened during early interglacial periods and significantly weakened during glacial periods due to albedo effects. Thus, during interglacial periods when the monsoon was strong, the wet equatorial air would have been diverted to the north or south, resulting in a dry midcontinent. However, the weakening of the monsoon during glacial periods would have permitted the equatorial easterlies to penetrate into the continental interior. This model thus predicts that strong monsoons during interglacial highstands would have been associated with more arid conditions in the midcontinent, and weakened monsoons during glacial lowstands would have been associated with wetter, although still very seasonal, conditions.
The timing and location of orogenic activity also may have had a significant impact on midcontinent climate. Gondwana and Laurasia collided in the Namurian (Chesterian) to produce the Appalachian and Mauritanide mountains. As discussed by Rowley et al. (1985), the rise of this high mountain range should have acted as a high-altitude heat source producing an area of low pressure, in a manner similar to the modern Himalayas. Being located at the equator, however, these mountains would have intensified the normal equatorial low pressure, thus inhibiting the development of fully monsoonal conditions and producing high rainfall in the mountains (Rowley et al., 1985; Patzkowsky et al., 1991; Otto-Bliesner, 1993). However, with the end of orogenic activity and the subsequent erosion of the mountains, their climatic influence would have declined, permitting the development of fully monsoonal conditions during the Permian (Rowley et al., 1985). This scenario is supported by changes in the coal-swamp vegetation of the Appalachian basin during the upper Carboniferous (Pennsylvanian) (Phillips et al., 1985). The delay in establishment of a Pangean monsoon caused by the Alleghenian Orogeny may have made the onset of monsoonal conditions more abrupt. Perhaps this accounts for the rapid climatic shift in the Wolfcampian suggested by the cyclothems in Kansas (West et al., 1997).
Kansas Geological Survey, Geology
Placed on web April 27, 2010; originally published April 2010.
Comments to email@example.com
The URL for this page is http://www.kgs.ku.edu/Publications/Bulletins/257/06_hist.html