Kansas Geological Survey, Open-File Rept. 90-27
Annual Report, FY89--Page 7 of 10
The factor controlling fluid flow through rock is hydraulic conductivity, which is a function of permeability, porosity, grain size, sorting, cementation, composition, sedimentary structures and stratification assuming a fluid of uniform density and viscosity (Freeze and Cherry, 1979). These factors determine the degree to which strata transmit water and thus determine whether a hydrostratigraphic unit is considered to be an aquifer, aquitard or aquiclude. The definition of unconformity-bounded depositional sequences and their internal progradational events allows one to determine the positions of depositional environments through time and space. The depositional energies within each environment control the grain size, sorting and composition of sediment deposited into lithofacies. This in turn controls the hydraulic conductivities of the lithofacies. The movement of environments through time controls the stratigraphic architecture within sequences. Therefore, the spatial arrangement of aquifers and confining beds can be estimated from the sequence stratigraphy and can be used to determine bulk hydrogeologic properties.
The Dakota aquifer framework in Kansas is composed of strata belonging to the Dakota Formation, Kiowa Formation and Cheyenne Sandstone. In western Kansas the arrangement of these stratigraphic units is straightforward. However, architectural changes in lithologic composition of these units are more pronounced in a depositional dip direction than along the depositional strike (Figure 7.4, 7.5 and 7.6). From west to east, strata are increasingly nonmarine and lithologically complex, and there is an increase in stratigraphic truncation along unconformities. In general, the proportion of more permeable lithofacies in the stratigraphic section increases from west to east.
Within the lower Cheyenne Sequence, the Kiowa Formation marine shale facies decreases in thickness due to lateral changes eastward into more permeable shoreface and coastal plain facies of the Longford Member and Cheyenne Sandstone (lower Dakota aquifer). Fluvial Cheyenne Sandstone thins depositionally to the east. Where present, the thick sequence of marine shale in the Kiowa Formation is an aquitard, the "Kiowa Shale" aquitard, separating the Dakota Formation (upper Dakota aquifer) from the underlying Longford Member and Cheyenne Sandstone aquifer units (lower Dakota aquifer). Across the basal unconformity, the Cheyenne Sandstone is in contact with underlying Jurassic Morrison Formation and Permian Cedar Hills Sandstone and provides a hydraulic connection between the Longford Member and the Cedar Hills.
Within both the J and D sequences, which make up the Dakota Formation, nonmarine facies become dominant toward the east. Fluvial sandstones are coarser grained and better sorted than nearshore sandstones in the Dakota Formation and probably more conducive of fluids. Fluvial channel sandstones have varying degrees of lateral and vertical connections. The J and D sequences are separated to the west by shoreline deposits which consist of interbedded sandstones and shales. Shales are more likely to be laterally continuous in this depositional environment than in the coastal-plain deposits and therefore are probably an aquitard between the sequences. Eastward amalgamation of coastal-plain sandstones of the sequences would cause them to act as one hydrostratigraphic unit. The D unconformity is usually associated with an overlying sandstone. In the outcrop it is coarse grained and laterally extensive. This sandstone may be an important conduit between separate sandstone bodies. The total thickness of the Dakota aquifer including the "Kiowa Shale" aquitard ranges up to more than 750 ft in parts of western Kansas.
Along the south to north cross section N-N' (Figure 7.6), Kiowa Formation shales and the Longford Member separate the Dakota Formation from Permian strata. The Kiowa consists of shale and is an aquitard. However, the Longford contains interbedded sandstones and shales and probably behaves heterogeneously with respect to vertical fluid flow. Inadepuate data preclude the determination of continuous or interconnected sandstone bodies within the member that would act as continuous aquifers. Fractures may provide a hydraulic connection between aquifers in the Permian and the upper Dakota aquifer (Dakota Formation).
Taken as a whole, the Dakota aquifer framework can be thought of as being both heterogeneous and anisotropic due to the "patchiness" of the lithologies represented. Perhaps more appropriate to this discussion, the aquifer framework can be subdivided into two components, (1) permeable sandstones with high porosity (20-30%) and (2) less permeable mudstones (siltstones and shales) with high storage due to the relatively high porosity (30% or more) of these lithologies. Some preliminary studies by Butler and Liu (1989) of the behavior of heterogeneous aquifers like the Dakota during pumping indicate that the storage in the mudstones may be the primary source of extracted water. Furthermore, porosity and acoustic logs of boreholes penetrating the Dakota suggest that the mudstones surrounding the sandstone aquifers are not well-compacted and may be slightly permeable with relatively high porosities (30-35%). This implies that on a local scale (macroscale) the fluid-flow hydraulics will be directly related to the geometry and distribution of lithologies in the aquifer (the sedimentary architecture). The sedimentary architecture of these clastic sediments is a function of the environment of depositon and any diagenetic changes that have occurred since deposition. Thus, it should be possible to relate trends in the hydrologic properties distribution to regional changes in the sedimentary architecture of the aquifer framework.
Figure 8.1 is a histogram of hydraulic-conductivity data computed from pump tests of the Dakota aquifer in Kansas. Applying a "goodness-of-fit" chi square test, the log hydraulic-conductivity data are not quite normally distributed at the 5% level of significance, suggesting that the distribution of hydraulic conductivity is not random but is controlled by some structure or process. Figure 8.2 is a map of the distribution of the data points and hydraulic conductivities arranged by class. Based on the 33 observations from southwest, south-central and north-central parts of the state, the geometric mean is approximately 19 ft/day with a standard deviation of almost 5 ft/day. Figure 8.3 shows the distribution of hydraulic conductivity in standard deviations from the mean in Kansas. Both Figures 8.2 and 8.3 show that hydraulic conductivity appears to increase in a northeasterly direction from southwest into central Kansas.
Figure 8.1. Distribution of hydraulic-conductivity data computed from pump and drill-stem tests of the Dakota aquifer in Kansas.
Figure 8.2. Geographic distribution of hydraulic-conductivity data for the Dakota aquifer in Kansas.
Figure 8.3. Geographic distribution of hydraulic-conductivity data in standard-deviation units from the mean for the Dakota aquifer in Kansas.
Figure 8.4 is a plot of hydraulic conductivity vs. depth to the midpoint of the water-producing zone of the pumping well. The data are further classified by lithofacies represented in the water-producing zone. Ignoring the differences in lithofacies of the tested interval, the hydraulic conductivity is negatively correlated with depth (or overburden thickness) which confirms the findings of Belitz and Bredehoeft (1988) in the Denver basin. They attributed the negative correlation to decreased porosity from the weight of the overburden and diagenetic effects resulting from deep burial in the basin. Looking at where the Dakota aquifer is present in the deeper part of the Kansas subsurface, drill-stem test data and core analyses indicate a less permeable aquifer than when the aquifer is in the shallow subsurface. Figure 8.5 shows the distribution of all of the hydraulic conductivities of the Dakota aquifer vs. depth to midpoint of the measured interval (overburden thickness), including core data from Merriam et al. (1959). The data show a grouping of the hydraulic conductivities according to the testing method used, which is probably related to the volume of aquifer framework tested (Halderson, 1986). The data also suggest that as the volume of the aquifer tested increases, the hydraulic conductivity increases. This correlation could be related to the influence of interconnections between individual sandstone aquifers (Fogg, 1986), fracturing or other heterogeneities. A closer look at the data of Merriam et al. (1959) shows that the hydraulic-conductivity geometric mean of sandstones deposited in fluvial channels is slightly higher than that of sandstones deposited nearer the shoreline. This is discussed in a later section of this chapter.
Figure 8.4. Hydraulic conductivity vs. depth (overburden thickness) to the midpoint of the water producing zone in the Dakota aquifer.
Figure 8.5. Hydraulic conductivity vs. depth and as a fuction of testing method for the Dakota aquifer of Kansas.
From this discussion, it seems that the distribution of hydraulic conductivity is at least in part related to the sedimentary architecture. In central Kansas, the thicker, more permeable sandstones are more likely to be associated with nonmarine sedimentation eastward of the Cretaceous interior seaway. Landward of the shoreline stacked sequences of highly permeable channel sandstones are commonly found on geophysical logs. As a result, the bulk hydraulic conductivity of the aquifer is higher in the central part of the state than in the western part. Using all of the data, there appears to be a slight decrease in hydraulic conductivity that correlates with increased thickness of overburden above the tested interval. However, this conclusion may be misleading due to the differences in scale of the tested interval.
Some insight into the flow of water on a small scale (approaching a microscale) can be gained by looking at the distribution of intrinsic permeability and porosity in the sandstones of the Dakota aquifer. This information can be useful if these hydrogeologic properties can be related to specific depositional processes, sandstone texture or style of bedding. Relating these properties back to some characteristic of the sandstones makes it possible to derive qualitative generalizations concerning the effect of the framework on water flow at small scales. It is important to recognize, however, that the degree to which these generalizations can be extended to larger scales, encompassing larger volumes of the aquifer, is limited due to the large proportion of mudstone in the framework.
The purpose of this section is to demonstrate the effect of sandstone textures and sedimentary structures on the distribution of intrinsic permeability and porosity in the sandstones of the Dakota aquifer, using core analysis data from Merriam et al. (1959). The intrinsic permeability is a function of the properties and characteristics of the porous medium, including the porosity. The hydraulic conductivity is a function of both the intrinsic permeability and the properties of the transmitted fluid. Merriam et al. (1959) reported two sets of effective porosity and permeability data from oriented samples of a core collected from the #1 Beaumeister, drilled in Cheyenne County. Effective porosity and the horizontal component of intrinsic permeability, kh, were measured on 80 core plugs collected along the portion of the core penetrating the Dakota aquifer framework. Merriam et al. also reported 47 measurements of effective porosity and the horizontal and vertical components of intrinsic permeability, kh and kv, for different intervals of the Dakota aquifer framework penetrated by the core. These data were provided by the Guy F. Atkinson Company. Merriam et al. used an air method to measure permeability and the data are uncorrected for Klinkenberg effects. The method used to obtain the other data set is not specified in their report.
Figure 8.6 contains the stratigraphic interpretation of the Dakota aquifer framework penetrated by the core at the #1 Beaumeister well location. Core plugs were obtained from and hydrogeologic-properties measurements made on a fluvial channel sandstone near the base of the Dakota Formation in the J sandstone and shoreface sandstones in the D sandstone interval of the Dakota Formation and the Longford Member of the Kiowa Formation (Hamilton, 1989). The channel sandstone lithofacies consists of a multi-storied sequence of massive, fine- to medium-grained, subrounded to rounded, well-sorted sandstone in the lower part and crossbedded and ripple-laminated, calcareous to noncalcareous, poorly sorted micaceous sandstone in the upper part. This facies was deposited by unidirectional currents of varying strength and competency in stream valleys during the J sandstone trangression of the Western Interior sea. The clay mineralogy of this lithofacies is primarily illite and kaolinite with minor chlorite (Merriam et al., 1959). The shoreface sandstones consist of bioturbated ripple-laminated and horizontal- to wave-ripple laminated, medium- to fine-grained sandstones with secondary silt, mica and glauconite. The clay mineralogy is variable but consists primarily of illite and kaolinite with minor montmorillonite (Merriam et al., 1959). The horizontal- to wave-ripple laminated sandstone in the core is well sorted and fine grained. Ripple-laminated sandstones contain asymmetric current ripples with silt and carbonaceous laminae. The horizontal- to wave-ripple sandstones were deposited by bimodal currents alternating with slack water conditions in a subtidal marine setting. The ripple-laminated sandstone was deposited in shallow marine water by unidirectional currents that alternated with periods of slack water, allowing for deposition of carbonaceous material by settling.
Figure 8.6. Stratigraphy of the Dakota and Kiowa Formations interpreted from the #1 Beaumeister core and the distribution of porosity and horizontal permeability data from Merriam et al. (1959).
Table 8.1 summarizes the variation of intrinsic permeability and porosity in the channel and shoreface sandstones for the two data sets. Both data sets show that the geometric means of kh and kv are considerably higher in the channel than in the shoreface sandstones. However, the kv/kh ratios are approximately 1 in the channel sandstones but are nearly 2 in the shoreface sandstones. These ratios suggest that for the most part the channel sandstones are more homegeneous than the shoreface sandstones. However, the large variance of the shoreface kv/kh data set suggests that these ratios are highly variable. The porosity data show that the channel sandstones are slightly more porous than those deposited in shoreface environments. However the large variance in the shoreface data suggests a high degree of variability. Interestingly, the data show that even though the sandstones have been loaded by more than 2000 ft of overburden, the porosity of these sandstones is relatively high. The mean porosity is 30.6% and 24.4% for the Guy F. Atkinson Co. supplied data and for the Merriam et al. data set, respectively.
Table 8.1. The variation of intrinsic permeability and porosity in the channel and shoreface sandstones of the Dakota and Kiowa formations penetrated by the #1 Beaumeister.
Figures 8.6, 8.7 and 8.8 show the vertical variation of kh and porosity for the measurements made by Merriam et al. and the kv, kh and kv/kh ratios for the data supplied by the Guy F. Atkinson Company. In the channel sandstone at the base of the J sandstone interval, the intrinsic permeabilities and their ratios are highly variable ranging from 0.25 to 2.26 in the upper part of the lithofacies. However, in the lower part, the intrinsic-permeability variation is subdued and, below 2290 ft, the ratios tend towards a value of one. This change in behavior of the intrinsic permeability coincides with the transition from massively bedded medium-grained sandstones to crossbedded and ripple-laminated fine-grained sandstones in the channel sandstone lithofacies. A similar pattern can be seen in the porosity data except that some fluctuation of the data occurs near the bottom of the channel sandstones near the 2300-ft and 2270-ft depths. In the shoreface sandstone data, all of the hydrogeologic-properties data exhibit a high degree of variability with depth in the core, including the kv/kh ratio which varies from 0.03 to 4.89. Figure 8.9 shows that kv and kh are better correlated in the channel sandstone (r = 0.74) than in the shoreface sandstone (r = 0.52). These observations suggest that sorting and bedding characteristics have a significant effect on the vertical and horizontal components of intrinsic permeability in both types of sandstone bodies. The channel sandstone lithofacies consists of well-sorted sands with little fines, whereas in the shoreface sandstones the degree of sorting is highly variable. Additionally, the intrinsic permeabilities are generally much higher in the massively bedded channel sandstone than in the crossbedded and ripple-laminated sandstones associated with the channel sandstone lithofacies.
Figure 8.7. The variation of the vertical and horizontal components of intrinsic permeability (kv, kh) and their ratio with depth in the channel sandstone penetrated by the #1 Beaumeister.
Figure 8.8. The variation of the vertical and horizontal components of intrinsic permeability (kv, kh) and their ratio with depth in the shoreface sandstone penetrated by the #1 Beaumeister.
Figure 8.9. Plot of vertical (kv) and horizontal (kh) data for channel and shoreface sandstones in the #1 Beaumeister Core. Data supplied by the Guy F. Atkinson Company as reported by Merriam et al. (1959).
Further confirmation of the differences between the channel and shoreface sandstone lithofacies can be seen by examining the relationship between the porosity and kh. Because the intrinsic permeability is directly related to the porosity in granular porous media (Mitchell, 1976), then differences in the porosity-intrinsic permeability relationship can be used to distinguish between lithofacies as there are other factors which influence kh, such as the degree of sorting, grain shape, packing, cementation or sedimentary structures. Figure 8.10 shows differences in the trend of porosity-intrinsic permeability data between the lithofacies, suggesting that these other factors are influencing the intrinsic permeability in one or both lithofacies.
Figure 8.10. Plot of horizontal permeabilities (kh) vs. porosity for channel and shoreface sandstones in the #1 Beaumeister Core. Data from Merriam et al. (1959).
The Dakota aquifer framework can be subdivided into three regionally significant hydrostratigraphic units on the basis of lateral continuity of lithofacies and sequence boundaries. A lower unit, an aquifer, consists of the Cheyenne Sandstone and Longford Member of the Kiowa Formation. The middle unit, an aquitard, consists of a shale facies of the Kiowa Formation that is laterally continuous in western Kansas. The upper unit, an aquifer, consists of the Dakota Formation. The upper and the lower hydrostratigraphic units are hydraulically connected in central Kansas where the aquitard is not present.
On a macro-scale, the upper and lower aquifer units of the Dakota aquifer can be thought of as consisting of two components: (1) permeable sandstones with relatively high porosities and (2) less permeable mudstones with high storage. The hydraulic conductivity of the Dakota aquifer framework at this scale ranges from less than 10 ft/day in southwest Kansas to more than 50 ft/day in central Kansas (the geometric mean of the data is 19 ft/day). The variation of hydraulic conductivity is related to the testing method used and the variation in lithology of the aquifer and may be related to the thickness of overburden above the tested interval. Reliable storativity data are meager but the average is estimated to be in the range of 1 x 10-5.
On a micro-scale, the porosity and intrinsic permeability of the sandstones in the Dakota aquifer are controlled by texture and sedimentary structures. Data published in Merriam et al. (1959) show that channel sandstones are more porous and permeable in the vertical and horizontal directions and show less variability in these properties as a group than do shoreface sandstones. Additionally, the massively bedded sandstones are more permeable and exhibit less anisotropy than do the shoreface sandstones.
These results indicate that the porosity and transmissive properties of the Dakota aquifer can be directly tied to the nature of the lithologies comprising the framework at scales ranging from mega (regional) to micro (an individual sandstone body). The data also underscore the need to understand the depositional framework of the strata that comprise the Dakota aquifer prior to using the data predictively.