Normal faults occur in the Niobrara Chalk and Pierre Shale in numerous localities in South Dakota, Nebraska, and Kansas. Lakeshore outcrops provide a window into strata that otherwise outcrop poorly. Various criteria were used to distinguish mass-wasting features from faults that formed at the subsurface. Locally, these faults have variable orientations and throws that vary from centimeters to tens of meters. In the Niobrara Chalk, the fault zones typically display well-developed striae, occasional slickensides, limited to absent damage zones, and dilational jogs with coarse, microvein calcite fill. In places, the faults are associated with gentle monoclines and folds, but elsewhere they occur in horizontal strata. A weathering bias suggests these faults are more common than observed.

At any one locality, these faults have been or could be attributed to local tectonism, differential compaction, and/or glacial rebound. If due to tectonism, taken in aggregate, they indicate fairly widespread (albeit low-strain) activity, which is consistent with the idea of a critically stressed continental interior with migrating sites of faulting. An alternate and preferred hypothesis is that these are polygonal faults due to diagenetically driven deformation in fine-grained mud rocks. Subsurface polygonal fault systems have been described from the Niobrara Chalk and Pierre Shale in the Denver-Julesburg Basin area in eastern Colorado. Smectitic clays, associated with polygonal faulting elsewhere, are abundant in these units. The fault kinematics (extension in multiple directions), widespread character, and other traits are consistent with an origin by diagenetically driven deformation.


Darton (1905), in some of the earliest geologic exploration of the structure of the Great Plains, noted a proliferation of small faults in the Cretaceous Niobrara Chalk. Bass (1926), Diffendal et al. (2002), Shurr et al. (1994), and Nichols et al. (1994) described small normal faults in these strata from Kansas, Nebraska, and South Dakota, respectively. Considered locality by locality, these faults have been explained as due to local basement reactivation, underlying salt dissolution, or differential compaction. The history of basement reactivation–related faulting is of consequence in that it informs seismic risk assessment for the continental interior (e.g., Li et al., 2009). The distribution of faults in continental interiors also speaks to the rigidity of plate interiors (Sandiford, 2010).

The Niobrara Chalk and to a lesser degree the Pierre Shale stand out for the sheer number of small faults (minimally in the hundreds) they contain in comparison to other stratigraphic units in the Great Plains. Based on field work and the literature, this article addresses the character and significance of faulting in the Niobrara Chalk and Pierre Shale from a regional perspective. In particular, developments in the field of structural diagenesis (Laubach et al., 2010) provide new insight into possible fault-generation mechanisms. Localities discussed are depicted in Figure 1.


The Pierre Shale in particular is known for hosting slumps, and many of the outcrops studied are associated with steeper slopes produced by recent erosion, and so a critical endeavor in the field is to distinguish mass wasting–related slip surfaces from faults that formed in the subsurface and are unrelated to surface topography. In the field, the default assumption was that the fault was related to mass wasting, and it was characterized as being of deeper-seated origin if three or more of the following criteria were applicable. Other geologists have also considered faults in these areas as of subsurface origin and related to basement reactivation (e.g., Nichols et al., 1994; Shurr et al., 1994; Diffendal et al., 2002).

  • (1) Movement sense and/or orientation relative to topography: An offset sense on a solitary fault contrary to the slope direction was considered indicative of a fault of deeper-seated origin. If the fault strike was highly oblique to the slope, the fault was also considered as less likely to be related to mass wasting. Consideration was given to topographic complexities such as multiple slopes (e.g., truncated ridge) and slump curvature.

  • (2) Position in weathering profile and cliff: For exposures along reservoir shores (Francis Case, Lewis and Clark, and Harlan sites; Fig. 1), wave erosion has greatly hastened cliff retreat in many instances, so that a cross section shows the transition from weathered rocks on the edges to fresher rocks in the cliff core. For the Niobrara Chalk, cliff cores can expose relatively fresh medium- to dark-gray rock that still has significant organic content. This transitions to whiter chalk where the organics have been removed by weathering. Pierre Shale weathering penetrates much more deeply, and unweathered shale was not observed. A fault in the unweathered core of a cliff face was considered more likely to be of deep-seated origin.

  • (3) Fault geometry: Slumps are strongly dominated by curved fault surfaces (e.g., Fig. 2) that decrease in dip at lower slope positions. Planar faults that dip steeply throughout the exposure are considered much more likely to be of deeper-seated origin. This criterion is of less use for smaller outcrops.

  • (4) Associated coarse vein material in dilational jogs: Some faults were observed to have coarse vein fill (often green calcite spar) in dilational jogs and were considered to be of deeper origin based on an assumption that coarser vein fill is consistent with slow, stable growth conditions found at depth, and not with fluctuating near-surface conditions.

  • (5) Offset of weathering profiles versus overprinting: Offset weathering-related features indicated a young age and a mass wasting–related origin. An overprint of a weathering profile across the fault structure or deeper penetration along the fault due to increased fracture avenues for water is consistent with a preweathering age and subsurface formation.

  • (6) Fault truncation by younger deposits: Unfaulted overlying deposits of Miocene (primarily the Ogallala Group) age allow the timing of fault movement to be constrained to well before modern landscape development.

  • (7) Mass-wasting styles in the Niobrara Chalk versus the Pierre Shale: Extensive reservoir shoreline outcrops allow characterization of contrasting mass-wasting styles in the Niobrara Chalk versus the Pierre Shale. The former displays rock falls, topples, and distributed slope creep, often utilizing joint surfaces and weak ash layers, while the latter forms slumps of a variety of sizes and sometimes earth flows. In the Francis Case Lake area, the contact of the two stratigraphic units locally serves as a slump’s basal slip surface (Fig. 2). Despite cliff exposures kilometer long in some cases, slumps were unobserved in the Niobrara Chalk. Therefore, discrete normal faults observed in the Niobrara Chalk are considered significantly less likely to be related to mass wasting.


At least 45 normal faults are exposed in the cliff outcrops along the shores of Francis Case Lake (Figs. 1 and 3; Table 1). Shurr et al. (1994) described some of these faults and ascribed them to basement reactivation. Most of the faults were noted in the Niobrara Chalk, but some cut up into the basal Pierre Shale strata. Throws vary from centimeters up to 20 m. Individual fault surfaces are often corrugated with an approximately dip-line axis that is parallel to fault striae and slickensides (Fig. 3). Dilational jogs with green, sparry calcite vein fill are ubiquitous.

The faults found were fairly distributed, spanning the length of the lake (including some exposed in the dam spillway). However, there are ∼5-km-long cliff exposures where faults were not observed despite similar outcrop coverage to other places where they were observed. I estimate that <50% of the shoreline cliff lines were scanned for faults. Fault orientations were quite variable (Fig. 4), with a suggestion of a weak preferred orientation of fault strikes at 10°–30° azimuth.

To better gauge whether the orientation distribution differs significantly from a uniform/random one, I used a chi-squared test (Swan and Sandilands, 1995). With a null hypothesis of the sample population being drawn from a uniform strike distribution, a 6 bin chi-squared test returns a value of 0.11, which would lead to the acceptance of the null hypothesis given the standard 0.05 cutoff. My conclusion is that any preferred orientation is only weakly developed. Significantly, one of two common major joint directions in the Francis Case Lake area is subparallel to this direction. Often two to three subparallel faults occurred in relative proximity.


Alternating layers of darker and lighter beds in an upper portion of the Pierre Shale in the Big Bend area along the Missouri River are distinctly visible in relatively high-resolution imagery available from Google Earth (8/20/2012 imagery). This stratigraphic unit is distinctly less vegetated than other stratigraphic levels. It is not uncommon for significant Pleistocene glacial-related deposits to overlie it in this area.

In places, the Pierre Shale strata are offset by faults that, using some of the criteria discussed earlier, are interpreted as being subsurface in origin. Slumps are also common, and the associated slip surfaces and offsets have a distinctly different character, including: (1) shallowly dipping, consistently curved traces, (2) associated surface scarps and other topography attributable to slumping, or (3) offset of younger glacial deposits. In contrast, the faults interpreted as forming at depth appear steeply dipping and have orientations, positions, and kinematics inconsistent with mass wasting. Due to difficulty of access, the nature of the interpreted faults has not yet been verified in the field, except in one case.

Photointerpretation led to the identification of 50 faults of deep-seated origin in this area (Table 2). Scanning of Google Earth images of the Big Bend area was thorough, but elsewhere it was more sporadic. There were several-kilometer-long stretches where the outcrop coverage was similar to elsewhere where faults existed, but where no faults were found, suggesting that the faults occur in patches.

Estimates of strikes were made for all the faults, rounded off to the nearest 5°. Dip directions were often indeterminate. Rose diagrams of the strikes indicate that a weak preferred orientation in the 90°–120° bin may exist, along with a suborthogonal preferred orientation in the 0°–30° bin (Fig. 5). This is consistent with some exposures where multiple faults in proximity appeared suborthogonal. However, a chi-squared test similar to that described earlier returned a 0.09 value, leading to acceptance of the null hypothesis that the observed distribution (6 bins) is from a uniform population, given a <0.05 threshold. A 90% confidence value would lead to rejection of the null hypothesis and the conclusion that the sample reflects a nonuniform strike distribution.

In terms of which side went down, 15 had the northeast quadrant down, 12 the southeast, 6 the southwest, and 13 the northwest. A lack of consistent downthrow direction is supported by a chi-squared test with a null hypothesis of a uniform downthrow distribution amongst the four quadrants. The resulting probability value of 0.27 leads to acceptance of the null hypothesis, given the standard 0.05 threshold.


A structure class exercise and subsequent senior thesis (Williams, 2012) found an array of mostly normal faults (along with a few minor thrust faults) in the Niobrara Chalk exposed in the cliffs along the shores of Lewis and Clark Reservoir (Fig. 6). Perhaps 50% of the shoreline cliff outcrops were scanned. To my knowledge, these faults are undescribed in the literature, although Simpson (1960) did mention nine faults found in excavations for the Gavins Point Dam powerhouse, seven of which were normal faults with 10–50 cm of throw. Four strikes were ∼130° azimuth, and one was at ∼90° azimuth.

Observed fault throws varied from centimeters up to 3 m. Striae were more difficult to observe in the vertical cliff faces, but where found, they were dip-slip. Green calcite and gypsum vein material was associated with the faults. The gypsum may be a later replacement overprint, perhaps associated with weathering. Abundant ash layers in the sequence exhibit thickness changes indicating lateral flowage of the bentonitic ash associated with the faulting (Fig. 6). Stereonet plots indicate there is perhaps a weak preferred orientation of a north-south conjugate set, with the east-dipping set of the pair more common (Fig. 7), but multiple fault directions clearly exist. This strike direction is parallel to one of the well-developed joint preferred orientations in the host rocks (Williams, 2012). However, a strong sampling bias also exists in that the cliffs along the reservoir are strongly east-west oriented, following the prevailing joint direction, and so faults striking more north-south would be encountered more often. This is unlike along Francis Case Lake, where the dammed Missouri River meanders, and cliff exposures occur with multiple orientations.


Diffendal et al. (2002) described normal faults in the Niobrara Chalk and Pierre Shale along the shores of Harlan Reservoir in south-central Nebraska in some detail. While mass wasting–related slip surfaces also exist in the area (Machette et al., 1998), these normal faults have traits consistent with a deeper-seated origin. Diffendal et al. (2002) indicated that the faults in the Harlan County Lake area may continue down into the Precambrian basement and may have been reactivated several times in geologic history. Ogallala Group sediments and younger loess deposits are involved in local mass wasting (Machette et al., 1998), but for the faults discussed here, post-Cretaceous overlying deposits were unaffected, providing some constraint on their age, and consistent with a subsurface origin. The map by Diffendal et al. (2002, their fig. 50) shows the faults to be distributed along the south shore of the lake (where some of the larger cliff outcrops exist), and they argued that the faults may be much more common in the region, citing similar faults found to the south just over the border in Kansas. In addition, a U.S. Army Corps of Engineers (USACE) map of the dam foundation and spillway area clearly shows an array of distributed faults striking in multiple directions, but with one distinct set striking at 135° azimuth.

Most of the faults were observed in the Pierre Shale. Striae or slickensides were often not observed along these fault planes, although good dip-slip striae were seen in one fault in the Niobrara Chalk (Fig. 8). Gypsum veins along and in association with faults were common. Most of the outcrops are distinctly weathered, and some fibrous gypsum veining is interpreted as being weathering related, while other planar veins filled with selenite are associated with subsurface processes. Fault throws vary from centimeters to >5 m. Adjacent bedding has dips of up to 20° and forms a girdle pattern consistent with rotation about a subhorizontal north-south axis, although data are distinctly limited (Fig. 9).

As at other sites, multiple directions of faulting occur (Fig. 9). A roughly north-south preferred direction is evident in strike measurements, which is consistent with the map of Diffendal et al. (2002), but this pattern may be influenced by two factors. First, as with Lewis and Clark Lake, the roughly east-west preferred orientation of the shoreline outcrops produces a significant bias so that north-south–trending faults would be over-represented. Second, the faults recognized in the dam spillway show a different orientation pattern of strikes, and if included would change the plot substantially. The relatively low number of measurements (30 faults) is insufficient to capture a more complex distribution, which is hinted at by the data and does not allow a chi-squared test for uniformity. In aggregate, the data are consistent with multiple directions of faulting. Local veins and joints display substantially different strike preferred orientations (Fig. 10) than the apparent distribution in the fault plots. However, the preferred direction seen in the dam spillway map is subparallel to a preferred orientation displayed by the joint/vein strikes.


The literature documents numerous faults in the Niobrara Chalk of western Kansas. For northwestern Ellis County, Merriam (1963) described a suite of over 76 small faults in the Niobrara Chalk that had the following traits: steep dips, no preferential trend, slickensided calcite zones up to 20 cm thick, displacements difficult to determine but up to 25 m, and an apparent decrease and disappearance with depth. He attributed the faults to subsurface solution collapse and noted that such faults were common in the Cretaceous strata of western Kansas, especially in the Niobrara Chalk, and were also well exposed along the Smoky Hill River. Outcrops are not particularly extensive in this area, and thus it seems safe to assume that faulting is more common than described.

Near Russell Springs along the Smoky Hill River in Logan County, Carlton (1954) mapped an assemblage of normal faults in the Niobrara Chalk and underlying Pierre Shale (Fig. 11). Fifty-four approximate strike values were extracted from a portion of the map where faults were denser, and these show a complex distribution (Fig. 11). Two highly oblique preferred orientations may exist, but a chi-squared test similar to those described earlier herein for six 30° bins returns a value of 0.29 and leads to the acceptance of the null hypothesis of uniformity given the standard 0.05 threshold. In addition, the fault downthrow quadrant distribution was also tested for uniformity, and the returned value of 0.25 also leads to the acceptance of the null hypothesis of a uniform distribution, indicating there is also not a well-developed preferred facing direction for the faults. A regional joint set in the area is reported to trend approximately eastward. Carlton (1954) attributed jointing that is parallel and normal to the faults to local, and not regional, deformation.

Normal faults are both well exposed and easily accessible in Smoky Hill Member strata (upper part of Niobrara Chalk) in the Castle Rock area of central Kansas (Fig. 12). They display a complex strike distribution, with an apparent preferred orientation component striking ENE-WSW in the southern part of the exposures and are absent in the northern part. Overall, the area of outcrop is elongate north-south, providing an exposure bias. The faults are characterized by notably corrugated slip surfaces, abundant striae, some slickensides, and pervasive calcite mineralization. Cut slabs consistently show calcite microvein arrays that provide hard and dilational linkage between the corrugated slip surfaces (Fig. 13). The calcite is consistently sparry and not fibrous. Breccias and damage zones were not observed. The microvein textures suggest incremental development and fault creep. Throws vary from centimeters to tens of meters. While mostly subhorizontal, the bedding in proximity to several faults dips up to 15° in opposition to the adjacent fault dip, in a geometry that can be explained by hanging-wall rotation above a listric normal fault. Monument Rocks to the west of Castle Rock (Fig. 1) has a smaller set of exposures in the Niobrara Chalk, but is also easily accessible. Several similar normal faults can also be found here, and the other localities in western Kansas noted on Figure 1 also display normal faults.


While mostly normal faults strongly dominate, thrust faults also occur (Fig. 14). They are distinctly thin skinned in character, with offsets limited to a few meters and usually much less. They were found both in the Niobrara Chalk and the Pierre Shale, and in weathered and unweathered material. Some may be clay expansion and weathering related, but it is also known that Laramide compressive stresses penetrated well into the continental interior (van der Pluijm et al., 1997), and some of these thrusts could be due to far-field tectonic stresses. Not enough thrust fault orientations were available from an area to detect any preferred orientation, and their age relative to the normal faults could not be unequivocally determined.


Character and Significance of Fault Distribution?

A primary conclusion is that normal faults of subsurface origin, but limited in offset and associated strain, are widespread within these stratigraphic units in the central Great Plains (Fig. 1). These normal faults are found where better and more extensive outcrops occur. In addition to localities discussed here, isolated examples also occur elsewhere (e.g., Spencer dam in Nebraska; Fig. 15). Field relations clearly indicate that enhanced weathering penetrates deeper along these faults due to the associated fracture and vein system. Thus, the exposure bias is such that these faults would tend not to be exposed, suggesting they are likely even more common than observed. A similar suite of widely distributed normal faults is not known for older units in the Great Plains (such as the Dakota Sandstone, which also outcrops extensively along several reservoir shorelines).

Multiple working hypotheses for faulting in the craton include: deformation driven by far-field tectonic stresses and/or associated with basement reactivation (e.g., Shurr et al., 1994; Siguaw and Estes-Jackson, 2011), differential compaction or solution collapse, and/or glacial loading/unloading (e.g., Clark, 1982; Calais et al., 2010). Their occurrence in areas far from Pleistocene glaciation indicates these faults are not related to glacial rebound tectonics.

Townend and Zoback (2000), Zoback et al. (2002), Sandiford (2010), and Berglund et al. (2012) described how the interior of continents may be broadly critically stressed, causing distributed strain and migrating seismicity. In this case, the widely distributed faults in the Niobrara Chalk and Pierre Shale could represent the cumulative result of migrating intracontinental strain with time, reflecting reactivation of a myriad of basement structures since the Cretaceous. However, this is an unsatisfactory explanation for several reasons. The faults would be expected to be aligned in a pattern reflecting the kinematics of the deeper fault reactivation at that point and time. Instead, they show relatively complex strike distributions without preferred directions of downthrow. The weakly developed multiple preferred orientations that may exist could be the result of regional stresses organizing the fracture systems, the generation of which was driven by other loading mechanisms, and/or due to directional outcrop bias. The kinematic uniformity of the faults also contrasts with the diversity that might be expected with different basement features reactivated at different times. The lack of observations of similar faulting in overlying, younger Ogallala Group strata in Harlan County and in western Kansas suggests that fault formation was restricted in time. There is also the more general lack of a similar density of faults observed in older or younger rocks; if related to basement reactivation, the distributed faults in the Niobrara Chalk and Pierre Shale should be linked to regionally distributed faults in these older and deeper strata.

Considering the stratigraphic position of the fault arrays high in the Niobrara Chalk and Pierre Shale, there is a relative lack of possible underlying widely distributed less compactable features that could drive significant differential compaction. Collapse related to underlying salt dissolution is a possible driver in western Kansas, but there is a lack of evidence for such potential further north and east in Nebraska and South Dakota, where the normal faults are also seen. When considered as a group, the aforementioned multiple working hypotheses provide unsatisfactory explanations for these faults.

Possibility of a Diagenetic Origin?

Many of the attributes described here are consistent with those of polygonal fault systems, which are arrays of strata-bound normal faults that are relatively common in fine-grained marine sediments and that are thought to be due to diagenetically driven volume shrinkage associated with compaction, fluid expulsion, and associated changes in clay chemistry and colloidal state (e.g., Cartwright and Lonergan, 1996; Dewhurst et al., 1999; Cartwright, 2011; Sonnenberg and Underwood, 2013). They are part of an emerging field of structural diagenesis (Laubach et al., 2010) that focuses on the interplay between diagenetic and structural processes.

Distributed faults in the subsurface Niobrara Chalk of the Denver-Julesburg Basin in eastern Colorado and western Nebraska and Kansas are well known. Siguaw and Estes-Jackson (2011) attributed an array of such subsurface faults to reactivation of basement faults and dissolution of lower Paleozoic salt deposits. Allen (2010), in a detailed study of fracture evolution in the Niobrara Chalk of the basin posits a stage of post-Laramide extension. However, Sonnenberg and Underwood (2013) proposed that polygonal fault systems are well developed in this unit and in overlying Pierre Shale strata. They described the faults as layer-bound normal faults dipping 45° or less, with throws of 6–30 m, and horizontal lengths usually less than 1200 m. Two tiers occur, one in the top of the Niobrara Chalk and the other within the Pierre Shale. In addition, Kernan and Sonnenberg (2013) recently proposed that the faults in western Kansas are also due to polygonal faulting. Polygonal fault systems have also been recognized in other Cretaceous chalk deposits around the world, demonstrating the capacity, if not propensity, of chalks to form such fault arrays (Hibsch et al., 2003; Hansen et al., 2004; Sandrin et al., 2012; Tewksbury et al., 2014).

The polygonal fault system evident in Google Earth imagery from the Farafra region of Egypt (Tewksbury et al., 2014) is particularly instructive as to the complexity that can characterize these systems. The faults outcrop extensively as low, but visible, ridges due to the abundant calcite mineralization, and the fault pattern can be seen at a scale and resolution not available in seismic sections. The reader is strongly encouraged to view these images in Google Earth (long 27.154926°N, lat 28.048712°E). Polygons are 500–1000 m across, and the borders consist of dense and fairly wide arrays of subparallel faults. These polygon border fault arrays vary from rectilinear in some places to distinctly curved in others. Locally, suborthogonal patterns are common.

The faults observed in the Niobrara Chalk can came in local subparallel clusters, but on a larger scale of hundreds of meters to kilometers in dimension, they show a more complex strike and spatial distribution, traits consistent with the pattern seen at Farafra. A factor that can complicate interpretations is the phenomenon where tectonic stresses can serve to locally organize diagenetically driven fracture systems to different degrees (Maher and Shuster, 2012).

A parsimonious hypothesis is that the distributed faults described in this paper are the eastern continuation of polygonal faulting in the Niobrara Chalk and Pierre Shale seen in the Denver-Julesberg Basin (Sonnenberg and Underwood, 2013). Traits of the faulting seen in these units that can be explained by polygonal faulting include: their widespread character, the strong preponderance of normal faults as compared to other kinematics, the complex orientation distributions, the lack of a preferred facing direction, the limited throws in the same range as that documented for other polygonal fault systems, and microvein arrays in the fault zones that suggest incremental development by creep (as would be expected if driven by burial, dewatering, and diagenesis). Development by creep during shallow burial before the chalks were totally lithified may also explain the lack of fault damage zones associated with the chalks, which in turn has implications for the influence of these faults on subsequent fluid flow. Because polygonal faults can form at shallow depths during shallow burial (Cartwright et al., 2004), their widespread distribution may simply reflect where these chalks were initially deposited in the Western Interior Seaway.

Possible Polygonal Faulting Mechanisms?

The literature suggests polygonal faults are potentially driven by a variety of diagenetic mechanisms. Syneresis due to a volume-loss mineralogic change and/or colloidal changes is commonly invoked, and specifically transformation of smectitic clays has been identified as playing a primary role in the generation of polygonal fault systems (e.g., Cartwright and Dewhurst, 1998; Dewhurst et al., 1999). Changes in opal type (e.g., Davies, 2005), porosity collapse and compaction due to opal cement dissolution (Spinelli et al., 2007), and carbonate-related pressure solution (Hibsch et al., 2003) are also described as possibilities. Shin et al. (2008) experimentally and theoretically explored the role of mineral dissolution during diagenesis as a mechanism for polygonal fault system development. More speculative possibilities can be entertained, including changes linked to alteration of the organics, including biogenic methanogenesis, which may induce volume loss. Laubach et al. (1998) described how coal cleats are related to thermal maturation, providing one example of fracturing driven at least in part by alteration of the organic component. Organics also influence other diagenetic processes, such as inhibiting pressure solution in chalks (Hibsch et al., 2003). Both the Niobrara Chalk and Pierre Shale have significant smectitic clay content and are organic rich, and silicified horizons are known from the Niobrara Chalk, suggesting silica mobilization. Given their carbonate content, pressure solution is a possible mechanism. Laferriere (1992) discussed isotopic evidence for carbonate pressure solution in the Fort Hays Member of the Niobrara Chalk. In summary, multiple potential mechanisms that could drive syneresis exist for the Niobrara Chalk and Pierre Shale, and this should be an avenue of future research.

Goulty (2002) argued that syneresis is not necessary to produce polygonal faulting and that a low coefficient of friction during vertical compaction is a more parsimonious explanation. In this context, it is interesting to note that chalks in particular have a very low angle of internal friction. Cartwright et al. (2004) described a polygonal fault system from modern muds in Lake Superior, indicating that polygonal faulting can occur at quite shallow conditions, and so the lack of significant burial is demonstrably not an impediment to polygonal faulting. Indeed, Sonnenberg and Underwood (2013) associated polygonal fault development with the expulsion of excess pore water at depths less than 914–1524 m (3000–5000 ft). In the case of the sites discussed here (Fig. 1), burial depths varied and are not particularly well constrained, but for the top of the Niobrara and given stratigraphic thicknesses of the overlying Pierre Shale and Ogallala Group strata (e.g., Hattin and Siemers, 1978; Shurr et al., 1994), they were well less than a kilometer, and likely less than several hundred meters in the eastern part of South Dakota.

When did diagenesis and associated polygonal faulting occur? The known geologic history suggests two possibilities—the first during deposition of the Pierre Shale in the Late Cretaceous, and the second during deposition of the overlying Ogallala Group strata in the Miocene (providing renewed loading). The first possibility is considered the most likely given the lack of evidence of involvement of overlying Ogallala Group strata where they occur. The second develops from the idea that significant changes in pore-water chemistry as meteoric aquifers were slowly established could also influence diagenetic processes (e.g., changing pressure solution dynamics).

Significance to Intraplate Seismic Risk Assessment?

In areas where Quaternary sediments are absent, it can be very difficult to date the timing of faults in older strata, leaving open the possibility that the fault is recent and poses some seismic risk. For example, Shurr et al. (1994) considered such possibilities for faults in the Pierre Shale of South Dakota. In addition, as mentioned, models for critically stressed crust and associated migrating seismicity could explain a relatively widely distributed faulting pattern. However, for fine-grained mud rocks and especially chalks, polygonal faulting should be considered as an alternate hypothesis. The distinction can be complicated by a situation where an extant regional stress field or an existing anisotropy organizes the faults produced by diagenesis into two or more preferred directions, a pattern which can favor a reactivation interpretation. In distinguishing diagenetically driven from tectonic faults, the following criteria may be useful: (1) a strata-bound fault concentration in fine-grained mud rocks, (2) an approximately isotropic horizontal extension (normal faults in at least two directions), (3) slow, incremental development, and (4) formation during sedimentary loading or some other diagenetic trigger.


  • (1) Relatively small-scale normal faults are broadly distributed throughout exposures of the Niobrara Chalk and Pierre Shale of South Dakota, Nebraska, and Kansas.

  • (2) Their widespread distribution, complex orientation distribution, and kinematics are consistent with formation by diagenetically driven deformation as a polygonal fault system, similar to what has been described for fault systems in the more deeply buried, subsurface equivalent strata in the Denver-Julesburg Basin to the west. If so, the polygonal fault systems in the Niobrara Chalk may represent one of the larger such exposed examples, spanning at least four states.

  • (3) In addition to basement reactivation, diagenetically driven deformation should be entertained as a possibility for faults found in fine-grained stratigraphic units in the Great Plains.

  • (4) Abundant possibilities exist for the specific diagenetic processes driving polygonal fault development, but future research is needed to identify the drivers. The easily accessible Niobrara Chalk exposures provide an excellent opportunity to understand these systems at an outcrop scale.

Funding from the Petroleum Research Fund (#490405-UR8) and the GDL Foundation has helped support this research. A host of undergraduate students contributed to this research through the years, including: Jalot Al-Absy, Jake Anderson, Ben Bates, Daniel Bior, Joe Boro, Angie Burgett, Eric Bush, Matt Coan, Jace Cochran, Konal Dobson, Bronson Gerken, Alexandria Gilbert, Sarah Ferguson, John Glover, David Haase, Theresa Halligan, Kristyn Hill, Kyle Kloewer, Ryan Korth, Andy Lewis, Tony Maida, Aleece Nanfito, Sam Nath, Jordan Mertes, Laura Pickett, Sarah Pistillo, Phil Schiele, Nathan Schagel, Andrew Schwab, Chris Sautter, Erin Sherrill, Jenn Stilmock, Jermiah Taylor, Nick Valentour, Dave Vanosdall, Drew Williams, and Erin Young. Their efforts are much appreciated.