Table of Contents

Introduction

Part I

Appendix A

Appendix B

Appendix C

Part II A and B

Part II C and D

Part III

Part IV

References

Summary

Part III. Conceptualization of the Kansas High Plains Aquifer and its Recharge Characteristics, including Suggestions for Appropriate Recharge-quantification Techniques

1. Conceptualizing High Plains Aquifer Recharge

Understanding the sources of recharge and the spatial and temporal variability in recharge is basic to developing a conceptual model of recharge. Potential sources of recharge of the Kansas High Plains aquifer include precipitation, surface water (rivers, streams, ponds, playas, lakes, floods), return flow from irrigation, lateral ground-water flow into the aquifer from outside areas (for example, lateral inflows from the Colorado High Plains aquifer to Kansas), and cross-formational flow from adjacent aquifers (for example, the Dakota aquifer, as shown in section B4 of Part II).

Western Kansas has a semi-arid continental climate with moderate precipitation, low humidity, and high evaporation. Winters are relatively moderate and summers are often hot. The mean annual precipitation ranges from less than 17 inches in the west near the Colorado border to more than 30 inches in the easternmost extent of the High Plains aquifer (Equus Beds aquifer region). About three-fourths of the precipitation falls during the growing season (April through September). Average free water surface evaporation ranges from 52 inches in the easternmost extent of the High Plains aquifer to more than 68 inches in southwestern Kansas (Sophocleous, 1998b). The Kansas High Plains is characterized by flat to gently rolling terrain, which, in combination with the semiarid climate of western Kansas results in minimal surface runoff. The mean annual surface runoff in western Kansas ranges from less than 0.1 inch to about 1.1 inches (Sophocleous, 1998b).

Recharge generally increases with increased precipitation. The seasonal distribution in precipitation may be more important than the average annual precipitation because winter precipitation is more effective in recharging ground water than summer precipitation. As Scanlon, Dutton, and Sophocleous (2002) also pointed out, many think that if average annual potential evaporation is much greater than precipitation, there should be no ground-water recharge. However, the time scale of the calculations is important. Use of long time scales, such as yearly or monthly, can lead to an underestimation of recharge. Water-budget estimates should be conducted using data and time steps no larger than daily because precipitation at such smaller time scales can greatly exceed evapotranspiration and result in effective recharge. In addition to climatic factors, recharge is affected by soil texture/structure, and hydraulic conductivity (coarse-grained soils generally result in higher recharge rates than fine-grained soils), land cover (croplands generally result in higher recharge than grasslands and shrublands), and land use (irrigated lands result in higher recharge than drylands) (Sophocleous and McAllister, 1987; Sophocleous, 1992).

The High Plains aquifer of Kansas consists mainly of a heterogeneous sequence of unconsolidated deposits of sand, gravel, silt and clay of principally alluvial origin deposited during the Tertiary (the only stratigraphic unit of Tertiary age identified in western Kansas is the Ogallala Formation of Pliocene age—Gutentag, 1963) and Quaternary periods and unconformally overlies Permian, Jurassic–Triassic, and Cretaceous formations. The type and degree of cementation within the aquifer varies. Lime-cemented and silica-cemented beds of silty and sandy gravel (mortar beds) and sandy silt (caliche) occur throughout the aquifer and at the outcrop they form ledges or caprock.

The Ogallala Formation, which makes up the main part of the High Plains aquifer in western Kansas, was deposited primarily by easterly flowing aggrading streams carrying debris from the Rocky Mountains. A vast plain of braided streams and coalesced alluvial fans was formed. Ogallala sediments filled paleovalleys eroded into the pre-Ogallala surface. However, more recent studies in the southern High Plains (Gustavson, 1996) indicate that the Ogallala in Texas and eastern New Mexico consists of alluvial sediments that partly fill paleovalleys and widespread thick eolian sediments that cap both paleo-uplands and most fluvial sections. These strata, apparently deposited under mostly semiarid to subhumid climatic conditions, do not constitute coalescing or overtopping wet alluvial fans (Gustavson, 1996). Deposition of the High Plains aquifer in some areas was contemporaneous with dissolution of underlying Permian salt beds, resulting in additional ground-surface subsidence and increased accumulation of High Plains sediment. The lower part of the formation in paleovalley-fill alluvium tends to have more coarse-grained sediment and thus greater hydraulic conductivity than the upper part, although Breyer (1975) concluded that the distribution of sediment types within the Ogallala Formation is largely random. The major identifying feature of braided streams is the coarsening-upward as well as fining-upward sequences of alluvial deposits (Gutentag et al., 1984). This process of coarsening and fining of the alluvial deposits gives a random distribution of sediments in the High Plains aquifer, suggesting that the aquifer may behave as homogeneous on a regional scale. Test holes drilled within a 160-acre tract often show a predominance of clays and silts at one site and of sand and gravel nearby (Stullken et al., 1985).

The Quaternary deposits are of Pleistocene and Holocene age. Considerable thicknesses of both alluvial and eolian deposits occur at the surface of the High Plains in Kansas. Quaternary alluvium (stream-laid clay, silt, sand, and gravel) is the predominant type of Cenozoic deposit in most of western Kansas. Pleistocene loess mantles much of the upland areas in western Kansas, and Pleistocene and Holocene dune sands cover a significant portions of the High Plains area. Because of the similarity in composition, the contact between the Ogallala Formation and the overlying Pleistocene deposits is difficult to determine from drillers’ logs, gamma-ray logs, and some test-hole logs.

Figure III–1 shows an east-west cross section depicting the stratigraphy from western Stanton County to the Gray County line across township 28. The lithology of the wells and test holes has been simplified to show the aquifers (sand and gravel), aquitards (silt), and quasi-aquicludes (clay and caliche). Mixtures of materials were designated as to their major constituent for clarity of illustration. The slopes on the eroded bedrock and Ogallala surfaces in the eastern part of the area are moderate as opposed to steep slopes in the western part. The sediments are thickest where the slopes are moderate (Gutentag, 1963).

The configuration of the bedrock surface is a composite of subaerial erosional surfaces of several ages (Merriam and Frye, 1954). This surface also has been affected by structural movement and by subsidence associated with the solution of evaporites from Permian rocks (Gutentag et al., 1981). The pre-Ogallala surface south of the Arkansas River also has been modified by post-Ogallala erosion. The irregular bedrock surface in southwest Kansas between the Bear Creek and the Crooked Creek–Fowler faults (Gutentag et al., 1981) generally slopes at about 13.5 feet per mile—a gradient of 0.0026) to the east-southeast from 3,500 ft above sea level near the Colorado state line in southwest Stanton County to about 2,000 ft above sea level near the town of Meade in Meade County, Kansas. The Bear Creek and Crooked Creek–Fowler faults in southwest Kansas are attributed to dissolution of halite and gypsum from the Blaine Formation and Flower-pot Shale of the Lower Permian Nippewalla Group.

The High Plains aquifer ranges in saturated thickness from 0 to more than 550 ft (as of 2000), just south and west of Liberal in Seward and Stevens counties. Generally, the greatest saturated thickness is where the unconsolidated deposits overlie the deepest channels in the bedrock. In some areas the High Plains aquifer is hydraulically connected to the overlying alluvium, such as along the Arkansas and Cimarron River valleys. The Lower Cretaceous Dakota aquifer also is hydraulically connected to the High Plains aquifer in some locations—that is, the Ogallala Formation is not separated from the Dakota Formation by shale, clay, or other low-permeability units. General flow within the aquifer is eastward. Streams affect local flow patterns as they become discharge or recharge points for the aquifer. Based on average values of hydraulic gradient and aquifer characteristics, the velocity of water moving through the aquifer is about 1 ft/day (Gutentag et al., 1984), which is typical of sand and gravel aquifers.

Many areas of the aquifer have been irrigated since the 1940’s. Average annual withdrawal for irrigation was greatest during the 1980’s, but during the 1990’s the total rate of irrigation withdrawal decreased. Irrigation inefficiency probably was high during the 1940’s and 1950’s but decreased during the past few decades. Luckey and Becker (1999) estimated that irrigation inefficiency decreased from 24% during the 1940’s and 1950’s to less than 4% by the 1980’s.

Estimated recharge is much less than the water quantity extracted from the High Plains aquifer in western Kansas, resulting in significant long-term water-table declines as well as streamflow declines (Kromm and White, 1992; Sophocleous, 2000a,b; Schloss et al., 2000). Prior to heavy irrigation development, the Arkansas River received baseflow from the High Plains aquifer and the connected alluvial deposits. Under these conditions, ground water naturally flowed towards the river. At the present time, however, the water table has declined below the streambed so that the flowing river may be a recharge source for the underlying sediments.

Irrigation return flow may contribute significant amounts of recharge to the High Plains aquifer. The amount of return flow depends on irrigation rate, irrigation inefficiency, soil type, depth to water, and rate of downward movement (or velocity) of water from the root zone to the water table. Return flow may reach the water table much later than the year or the decade in which irrigation was applied, and the delay or lag time may increase as depth to water increases. The velocity of water moving downward through the unsaturated zone is an important, although poorly constrained, variable (Scanlon, Dutton, and Sophocleous, 2002). If the velocity is much greater than the rate of water-level decline, return flow quickly reaches the water table. If the downward velocity is similar to the rate of water-level decline, much of the return flow may be significantly delayed in reaching the water table, leaving more water in storage in the unsaturated zone. The magnitude and effect of return flow in different parts of the High Plains aquifer remain poorly understood (Scanlon, Dutton, and Sophocleous, 2002).

FIGURE III–1—East-west cross section through the Ogallala Formation and Quaternary deposits in Stanton, Grant, and Haskell counties, southwestern Kansas (cross section in Stanton and Grant counties from Gutentag, 1963).

2. Appropriate Techniques for Quantifying Recharge in the High Plains Aquifer

As we have seen in Part II, the main techniques that have been used for estimating recharge in the High Plains aquifer in Kansas are Darcy’s Law, annual water-table fluctuation analyses in combination with estimates of aquifer specific yield, ground-water modeling, soil-water budget modeling, and base-flow analyses. Although a number of techniques for quantifying recharge in the High Plains aquifer of Kansas have been used, it is apparent from the review of existing recharge estimates that additional recharge studies are required to better quantify recharge. As we also mentioned in Part I, section 9, one of the difficulties of determining appropriate techniques for quantifying recharge in the High Plains is that many techniques are restricted to measuring recharge rates within a certain range, which may not be known a priori before the recharge study is undertaken. Therefore, only different approaches that are likely to provide the most quantitative estimates of recharge can be suggested. Results provided by initial studies should be used as platforms for additional data to optimize the techniques and refine the recharge estimates. An iterative approach will be required to accurately quantify recharge rates, and a variety of approaches should be applied because of uncertainties in recharge estimates (see also section 9 of Part I). Results from the various techniques can be compared to determine uncertainties in recharge estimates.

Following similar recommendations on appropriate techniques for quantifying recharge for the major Texas aquifers (Scanlon, Dutton, and Sophocleous, 2002), we offer the following suggestions, especially in view of the general lack of tracer-based methodologies for recharge quantification in Kansas. Because of the generally thick unsaturated zone of the High Plains aquifer in western Kansas, many of the techniques for estimating recharge to the Ogallala aquifer may be based on the unsaturated zone. The absence of calcic soils or caliche may be used as a qualitative indicator of recharge. The absence of calcic soils or low levels of calcium carbonate suggest high recharge rates, such as beneath playas (Scanlon et al., 1997). Surface-water techniques may be appropriate for quantifying recharge from streams using channel-water budgets (differential streamflow measurements) or other techniques such as heat tracers and seepage meters.

Appropriate unsaturated-zone techniques may include the use of chloride concentrations in soil water. Low chloride concentration beneath playas in Texas suggest high recharge rates, whereas high chloride concentrations in interplaya settings suggest low recharge rates. Such studies have not yet been reported in Kansas. The chloride mass balance approach also may be used in sandy areas to quantify recharge rates; however, the accuracy of this approach decreases as recharge rates increase. However, using chloride to quantify recharge rates in irrigated regions would be difficult because of uncertainties in the chloride input to the system (Scanlon, Dutton, and Sophocleous, 2002). Bomb-pulse tritium may be appropriate for quantifying recharge in sandy areas where the bomb peak is expected to have moved beneath the root zone. The presence or absence of bomb-pulse tritium may also be used in irrigated regions to provide estimates of recharge; however, use of this technique is complicated because the irrigation water probably does not contain bomb tritium. Bomb-pulse ³⁶Cl/Cl ratios could also be used to quantify recharge in sandy areas where recharge is expected to be higher than in finer grained sediments. The ³⁶Cl/Cl bomb peak may be much more obvious than the ³H peak because of the long half-life of ³⁶Cl (301,000 yr) relative to that of ³H (12.43 yr). Unsaturated-zone modeling could be used to estimate recharge rates in irrigated and nonirrigated regions. However, such unsaturated-zone techniques have rarely been used in Kansas, especially beyond the plot-size scale.

Saturated-zone methods provide a more spatially averaged recharge rate than the point estimates provided by unsaturated-zone techniques. Water-table fluctuations may be used in areas of shallow water table (such as the Equus Beds and Great Bend Prairie portion of the High Plains aquifer) to quantify recharge rates. Wood and Sanford (1995) used chloride concentrations in ground water to estimate recharge in the north half of the southern Ogallala in Texas. Anthropogenic substances such as pesticides and CFC’s may provide qualitative indicators of high recharge rates. Tracers such as ³H, ³H/³He can be used to quantify recharge in sandy areas and irrigated areas. Inverse ground-water modeling may be combined with ground-water-age data on the basis of ³H, ³H/³He, and ¹⁴C to provide regional estimates of ground-water recharge.