ACKNOWLEDGMENTS ii

EXECUTIVE SUMMARY iii

INTRODUCTION 1

Problem 1

Study Objectives 1

Purpose of Report 2

Selection of the Well Sites 2

Well Installation, Monitoring, and Sample Analysis Procedures 4

DEERFIELD SITE 5

GARDEN CITY SITE 15

DODGE CITY SITE 25

IMPLICATIONS OF THE RESULTS TO SALINITY MIGRATION 34

REFERENCES 37

APPENDIX A: Well Construction Information Forms 38

APPENDIX B: Geophysical Logs: Gamma and Conductivity 53
 
 

The contract for the Upper Arkansas River Corridor Study is funded by the State Water Plan Fund of the Kansas Water Office. Joe Anderson, Exploration Services Section, drilled the wells installed by the Kansas Geological Survey. Lawrence Hathaway and Truman Waugh in the analytical services group of the Geohydrology Section analyzed the water samples collected from the wells. Jeff Lanterman of the Division of Water Resources, Kansas Department of Agriculture, assisted in some of the logistics of site preparation before drilling. Jeff Lanterman and Eric Hargett, also in the DWR, measured some of the water levels in the wells. Mark Schoneweis of the KGS drafted some of the map and well schematics figures. The reviews of Susan Stover of the Kansas Water Office and Eric Hargett are appreciated and were used to edit the report.
 
 

The Upper Arkansas River Corridor Study, supported by the State Water Plan Fund of the Kansas Water Office, was established to document the fate and effects of contaminated Arkansas River flows on the ground water in the river valley, and to clearly determine the links among flows in river, increased levels of water contamination in the aquifers, and lowered ground-water tables. Saline water from the Arkansas River has been infiltrating into the High Plains aquifer underlying the river valley and areas irrigated by diverted river water. The study determined that there was little information on the depth distribution of the salinity in the aquifer. The study selected three sites along the river corridor for installation of multi-level observation wells sampled and analyzed ground waters, measured water levels, and conducted aquifer tests in the wells. The information from the sites has greatly increased understanding of the subsurface fate of the saline water along the river valley.

The first well site is located within Deerfield, east-central Kearny County, about one mile north of the Arkansas River. The second site is in Garden City, Finney County, approximately ¼ mile south of the Arkansas River. The third site is about 3 miles west of Dodge City, Ford County, about ¼ mile south of the Arkansas River. The shallow aquifer at all sites contains water with a sulfate concentration between 1,500 and 1,700 mg/L, which represents a mixture of Arkansas River water with ground water. Each site has a markedly different salinity distribution with depth below the shallow aquifer, from a gradual decrease to freshwater at Deerfield to penetration of salinity throughout the aquifer at Garden City to a sharp decrease to freshwater at Dodge City.

Substantial amounts of clays in the subsurface appreciably retard the downward movement of salinity at the Deerfield site. The source of the saline river water affecting the aquifer at Deerfield is not directly from the river channel, but from the river water diverted to the west and north of the site. The downward hydraulic gradient at the site continues to provide a drive for downward migration of salinity. Thus, the deepest portion of the aquifer, which still contains freshwater, can be expected to slowly increase in salinity in the future.

The High Plains aquifer at the Garden City site contains sediments that are generally more permeable than those at the Deerfield site. U.S. Geological Survey test-hole data for a location very near the site showed that only the shallow alluvial aquifer contained slightly saline water in 1961; the main aquifer yielded very fresh water with a sulfate content of 24-52 mg/L. Today, the water in the alluvial aquifer is much more saline than in 1961 and the shallow salinity has penetrated to the bottom of the aquifer. The proximity of the Garden City site to the river water source, the downward head gradient, and the permeable sediments in the stratigraphic section allowed the saline river water to penetrate into the aquifer at a faster rate than at Deerfield.

Although the location of the Dodge City site relative to the Arkansas River is similar to that of the Garden City site, there is a sharp gradient from saline water in the alluvium to freshwater throughout the underlying main portion of the High Plains aquifer. The coarse nature of the alluvium, the abundance of low permeability sediments underlying the alluvium, the small downward head gradient in the underlying aquifer, and the discontinuous presence of water in the river channel are major factors explaining why the saline water has not appreciably penetrated below the alluvial aquifer. Substantial declines in the water levels of the main aquifer might possibly lead to migration of salinity below the alluvial aquifer. Management of ground-water withdrawals in the valley, especially during periods of high flows in the Arkansas River, will be important for protecting water quality in the main aquifer.
 
 

INTRODUCTION

Problem

The Arkansas River in southeastern Colorado and westernmost Kansas is one of the most saline rivers in the United States. Most of the river flow entering Kansas from Colorado is lost in the river stretch from the state line to Dodge City. The loss is due to infiltration from the river channel and consumptive use and seepage of water diverted from the river for irrigation. Saline water can infiltrate from the river and penetrate deeper in the aquifer below fields irrigated with diverted river water because water levels have declined in the High Plains aquifer from consumptive pumping of ground water. The major water-producing portion of the High Plains aquifer is also known as the Ogallala aquifer. The alluvial aquifer of the Arkansas River valley overlies the Ogallala aquifer along most of the river course in southwest Kansas. The amount of saline water entering the subsurface below the river and irrigated fields is large enough and the dissolved solids concentration high enough that it could contaminate all of the ground water in over a 500 square mile area of the river corridor to a sulfate concentration of over 1,000 mg/L within 50 years, if the saline river water could completely mix with the subsurface water. The ground waters that have been and could become contaminated include municipal supplies for Syracuse, Lakin, Garden City, Cimarron, and Dodge City.

The areal and vertical distribution and temporal variation of the saline waters, and the future movement of these contaminated waters have not been well known in the Arkansas River corridor. Factors such as the distribution of lower permeability silt and clay layers in the High Plains aquifer and the slow subsurface flow retard the movement of the saline water. Flows down the gravel packs of unsealed irrigation wells, either during periods between use or after abandonment, can contribute to the amount of saline water reaching the deep aquifer. An assessment of the extent of the ground water contamination and the mechanisms controlling the salinity transport is critical for developing plans for minimizing or mitigating water-quality problems in the aquifers.

Study Objectives

The Upper Arkansas River Corridor Study is a Kansas Water Plan project that is providing data and research for the Upper Arkansas Basin section and 2010 Water Quality Objectives in the Water Plan. The basic objectives comprise major parts of the objectives listed under the water-quality and ground-water decline issues in the subsection on the Arkansas River Corridor Subbasin in the Upper Arkansas Basin section of the Kansas Water Plan:

• Water-Quality Issue: Document the fate and effects of contaminated Arkansas River flows on the alluvial, Ogallala, and Dakota aquifers in the river valley.
• Ground-Water Decline Issue: Clearly establish the links among decreased flow in the Arkansas River, increased levels of water contamination in the alluvial and Ogallala aquifers, and lowered ground-water tables.

• Characterization of the water quality and documentation of the fate of Arkansas River water within the Arkansas River corridor;
• Determination of the distribution and sources of salinity and nitrate within the river corridor;
• Evaluation of the migration of poor quality river water into the freshwater aquifers underlying the river corridor; and
• Evaluation of the implications of saline water migration on possible management strategies within the river corridor.

In order to determine the distribution and migration of salinity in the High Plains and alluvial aquifers, we examined existing ground-water information. This information included lithology, water-levels, aquifer tests, water quality, hydrologic models, and other relevant data. Information on the depth distribution of water levels, water quality, and hydrologic properties of the aquifers was particularly limited. Therefore, we selected three locations for the installation of multi-level observation wells along the corridor. Water samples and field measurements and tests at the sites provided data that greatly increased the understanding of the fate of saline Arkansas River water in the valley. This report describes the multi-level well sites, including well construction information, water levels, and water-quality data. Another report, "Performance and Analysis of June 1998 and October 1998 Slug Tests in Kearny, Finney, and Ford Counties" by Butler and Healey (1999) presents the results of hydraulic tests performed on the wells.

Selection of the Well Sites

We selected the location of all the sites to be within the area in which ground waters of the High Plains aquifer have been impacted by infiltration of saline Arkansas River water. We determined the impact area from water-quality data that existed or that that the study obtained in the river corridor. Each of the sites represents different conditions along the river corridor. The sites are located in Deerfield in eastern Kearny County, in Garden City in Finney County, and just to the west of Dodge City in Ford County (Figure 1).

The first well site selected (within the City of Deerfield) represents an area where seepage of ditch irrigation water diverted from the Arkansas River, as well as water in the river valley, were thought to be potential sources of ground-water contamination. In addition, the location is near public water supply wells that yield water showing a substantial salinity increase since the late 1970’s. The legal location of the site is 24S-35W-11CBB (NW/4 of NW/4 of SW/4 Sec. 11, T. 24 S., R. 35 W.).

We selected the second site (in Garden City) to be as close as possible to the only location where prior data existed for water quality at more than two depths within the river valley. The site represents an area next to Garden City where seepage of Arkansas River water from the river channel is the source of ground-water contamination. In 1961, the U.S. Geological Survey drilled two test holes near Highway 83 about ¼ mile south of the Arkansas
  River. The USGS collected water samples from two depths in one hole and from three depths in the other hole for a total of 5 observations, all at different depths. The previous data provided an excellent opportunity for examining the change in salinity of the ground water with time at the depths for which the USGS obtained data. The legal location of the site is 24S-32W-19CABA (NE/4 of NW/4 of NE/4 of SW/4 Sec. 19, T. 24 S., R. 32 W.).

The third site (Dodge City) represents a more eastern location within the upper Arkansas River valley. River flows at the location have been sporadic over the last few decades in the Dodge City area. Water in the river valley is currently used for the public supply of Dodge City as well as for other purposes. The site provided a location where the effect of substantial river flows over the last several years, in comparison with previous dry periods, could be examined on the water quality of the High Plains aquifer underlying the river valley. The legal location of the site is 26S-25W-32BBBB (NW/4 corner Sec. 32, T. 26 S., R. 25 W.).

Well Installation, Monitoring, and Sample Analysis Procedures

The Kansas Geological Survey (KGS) drilled the four shallowest wells at the Deerfield site and the three shallowest wells at the Garden City site. Clarke Well & Equipment, Inc. drilled the deepest well at the Deerfield site, the two deepest wells at the Garden City site, and all four wells at the Dodge City site. KGS and Clarke Well staff logged the cuttings obtained during the drilling of the holes. The KGS followed the drilling of all wells installed by Clarke Well. The deepest hole at each site went to bedrock and we selected the screened interval depth to be in permeable sediments as close to the bottom of the hole as practical. We selected the screened intervals for the other wells based on the cuttings and geophysical logs of the first deepest well, and, for the Garden City site, the depths of the former test holes drilled by the USGS. All the wells and screens are composed of PVC. The two shallowest wells at the Deerfield and Garden City sites are 4 inch in diameter, whereas the deeper wells are 2.5 inch diameter. All of the wells at the Dodge City site are 2.5 inch diameter. The annular space is sealed with cement or bentonite from several feet above the screened intervals up to the surface. The boreholes are generally a few feet deeper than the planned depth of the screened interval bottom. In some cases, blank casing sumps of a couple feet in length are at the end of the screen. Construction and lithologic log information for the wells is in Appendix A. Well depths measured with a tape soon after well completion and later before slug tests differ from those on the construction information forms for a few wells by up to a few feet. We used the depths measured from the top of the casing minus the sump length and minus two feet for the casing length above the land surface for the graphs in this report.

The wells are numbered from 1 to 4 or 5 from deepest to shallowest at each site. The Deerfield and Dodge City wells are in a line from deepest to shallowest, with about 15 ft between each well. The Garden City wells are in a rough circle, each also separated by approximately 15 ft. Concrete pads protect the base of the well at the surface. The Deerfield site wells are locked and protected from vehicles by steel cables strung along posts anchored in concrete. The Garden City and Dodge City sites include steel protectors with locks over each well. The Dodge City site also has protection posts along the edge of the field road.

The KGS or Clarke Well & Equipment developed the wells soon after installation. We further developed all the wells before sampling and aquifer tests. The KGS and Clarke Well & Equipment logged the cuttings from the wells. We ran geophysical logs (gamma and inductive conductivity) for the wells soon after installation and again several months later. The logs for the deepest wells are in Appendix B. The natural gamma log responds primarily to fine-grained sediments such as clays that generally contain more radioactivity than coarser-grained deposits. The gamma logs contained some high readings that did not correlate well with the conductivity logs. Some of these units appear to occur in some of the coarser sediments and may represent high radioactivity associated with feldspars in the deposits. The focussed electromagnetic induction log measures formation conductivity, including the conductivity of the formation fluid. Therefore, the conductivity is influenced by clay content, salinity, and moisture content. The metallic parts of centralizers used above the screened interval appear as small spikes, and the small screws used to anchor the glued casing joints result in tiny spikes at approximately 20 ft intervals on the conductivity logs.

The Division of Water Resources (DWR) in the Kansas Department of Agriculture and the KGS periodically measured water levels in the wells. The KGS determined aquifer parameters such as hydraulic conductivity for screened intervals using proven methods of slug-test analysis (Butler and Healey, 1999).

We sampled the wells at the end of or soon after development and then at a later time to examine changes in the water quality. Water extraction from the well was by air lifting or submersible pumping at the end of development and later by submersible pumping. We removed more than several casing volumes and monitored the specific conductance and water temperature to make sure the readings were stable before sample collection. We collected the samples in polyethylene bottles and perserved them in an ice chest for transport to the KGS laboratories. During the sampling of the wells by pumping a year or more after installation, we filtered the water through a 0.45 m m membrane filter before collection. We collected the field filtered sample for nitrate determination in a separate bottle and preserved with hydrochloric acid.

The KGS analytical laboratories in Lawrence, Kansas determined chemical properties and the concentrations of major and minor dissolved constituents in the water samples. The laboratories maintain quality assurance/quality control procedures that include calculation of charge balance errors and analysis of USGS standard reference waters. The estimated maximum errors in the constituent concentrations given in this report are typically a few percent or less. The KGS analyzed the last set of samples for all wells for ammonium ion; the concentration in all samples was <0.05 mg/L as NH₄-N.
 
 

DEERFIELD SITE

The observation well site at Deerfield is located in the southwest part of the city to the west of the athletic fields of the public school land (Figure 2). It is approximately one mile north of the Arkansas River. There are five wells screened at different depths in the alluvial and main portions of the High Plains aquifer (Table 1 and Figure 3). The Kansas Geological Survey (KGS) installed the four shallowest wells in June 1997. Clarke Well & Equipment installed the deepest well in February 1998.

The site is located at the edge of the alluvial valley of the Arkansas River. The shallowest well is in alluvial terrace deposits. The lithology of the High Plains aquifer at the site is characterized by a substantial amount of fine-grained material, including many thick clay units, in comparison with other areas in the Arkansas River corridor. Figure 4 displays the general lithology of the well boreholes by colored intervals based on the geologic logs. Note the difference in the logging of the cuttings by the KGS in comparison with Clarke Well & Equipment. The gamma and conductivity logs tend to show a relatively large amount of variability in the sediments with depth (Appendix B). The bedrock is at a depth of 345 ft below the land surface.
 
 

• ¹ Units for information in this and remaining columns are ft.
• ² Depths are from top of casing in this and remaining columns. The top of the casing is approximately 2 ft above land surface.
• ³ Neat cement grout
• ⁴ Depth includes a sump at bottom of screen

The total dissolved solids (TDS) concentrations for the well waters range from a little over 2,700 mg/L in the shallowest wells to 460 mg/L in the deepest well (Table 2). Except for the freshest, deepest water, sulfate is the dissolved constituent that contributes the greatest percentage of the TDS content. Concentrations of the major dissolved constituents calcium, magnesium, sodium, sulfate, and chloride are relatively similar in waters from the two shallowest wells and also decrease substantially with depth in the other well waters. Figure 5 shows the change in TDS, sulfate, and chloride concentrations with depth. The depths used in this and other figures in this report represent the middle of the screened interval. The nitrate concentration also decreases substantially with depth, whereas the fluoride content of the well waters approximately doubles from the shallowest wells to the deepest well (Figure 6).
 
 

Sodium concentrations are somewhat greater than calcium contents in waters from the shallowest two wells. The sodium concentration decreases with depth at a greater rate for well numbers 2 and 3 than for calcium. Bicarbonate contents decrease at a smaller rate with depth than the other major constituents to a level of about 190 mg/L in the two deepest well waters. Boron concentrations also decrease in the 3 shallowest well waters and then reach essentially the same contents in the two deepest wells.

The sulfate and TDS concentrations exceed the recommended level for drinking water of 250 mg/L and 500 mg/L, respectively, in waters from all the wells except the deepest one. The sulfate concentration in the shallowest two wells also exceeds the maximum recommended for stock use (1,000 mg/L). None of the other constituents exceeds standards for drinking, stock, or irrigation water use, although the nitrate-nitrogen content of the shallowest well is near the maximum contaminant level for drinking-water of 10 mg/L, and the fluoride concentration of the deepest well is essentially the same as the maximum level of 1 mg/L recommended for irrigation water use.

The KGS and the DWR have periodically measured water levels in the wells since installation (Table 3). Figure 7 displays the change in water levels with depth and time. In general, the deeper the well, the greater the depth to water and the greater the temporal variation in level. The shallowest two wells have nearly the same water level, although the deeper well consistently has a level the same or deeper than the shallower well. The middle well (#3) has a level that is always between those of the two shallower and two deeper wells. The relative depth to water levels in the two deepest wells changes with time. The relative direction of temporal change (up or down) in water level is essentially the same for all wells. The large water-level changes in the summer for the deepest two wells reflect the greater rate of pumping for irrigation and municipal use during that season in the area.
 
 

Table 3. Water-level data for the Deerfield observation well site. The values are depth to water below land surface.

The change in hydraulic head with depth indicates that there is a potential for downward flow of water at the site. The water levels in the shallowest three wells are always at a higher elevation than for the river and the alluvial aquifer adjacent to the river to the south. Except for parts of the summer, the water levels in the deepest wells are also at a higher elevation than for the river and in the adjacent alluvium to the south. The deepest observation well is screened within a permeable aquifer unit that lies within the same depth range as for the screened interval of the highest capacity well (and the one most frequently pumped) of the three municipal wells in Deerfield. This Deerfield well is located ¼ mile to the north of the observation well site and yields water with a sulfate concentration that is about two or more times that of water in the deepest observation well. The hydraulic head and water-quality data indicate that the primary source of salinity affecting the aquifer at the observation well site in Deerfield is from Arkansas River diverted from the river to areas to the west and north of the site, rather than water that infiltrated from the river channel into the aquifer to the south. The canal of the Great Eastern Ditch passes through Deerfield just south of the largest capacity municipal well of the city and to the north of the observation well site.

The salinity decrease with depth for the aquifer waters at the observation well site is almost linear. The decrease fits a process of slow downward migration of saline water and mixing with the deeper, fresher water. The migration is retarded by the substantial amount of clay in the aquifer in the area. Salinity that had already penetrated into the aquifer in nearby areas could also reach the site by lateral migration through the most permeable zones. The subregional direction of ground-water flow is generally to the east in the Deerfield area. The salinity of water from the deepest observation well is somewhat greater than that of the water supply of Deerfield before it began to become impacted in the mid-1970’s by migration of the diverted river water. The deepest observation well has a sulfate content of about 160 mg/L in comparison with the 95-120 mg/L range for the municipal well water of Deerfield during 1949-1973. The different shape of the change in nitrate concentration with depth (Figure 6) in comparison with that of the sulfate and TDS change (Figure 5), and the generally low concentration of nitrate in Arkansas River water (usually less than 3 mg/L), suggest that the source of higher shallow nitrate is local and at or near the surface.

Changes in the concentration of TDS and the major constituents during monitoring generally have not been great enough to be of significance for the wells. Small decreases in the salinity of the waters from the two shallowest wells might represent such effects as increased rainwater recharge during the period on water quality, the effect of ground-water flow in response to hydraulic head relationships at the site, and leakage of less saline water in the irrigation canal during the last few years of higher river flows from Colorado in comparison with several years earlier.
 
 

GARDEN CITY SITE

The observation well site at Garden City is located in the southern part of the city approximately ¼ mile to the south of the Arkansas River and about 100 yds to the east of Highway 83 (Figure 8). There are five wells screened at different depths in the High Plains aquifer (Table 4 and Figure 9). The Kansas Geological Survey (KGS) installed the three shallowest wells in June 1997. Clarke Well & Equipment installed the two deepest wells in February 1998.

The TDS concentrations for the well waters range from near 3,000 mg/L in the middle three wells (in depth) to about 2,200 mg/L in the deepest well (Table 5). Sulfate is the dissolved constituent that contributes the greatest percentage of the TDS content in all of the well waters; its depth distribution is similar to that of TDS (Figure 10). Although the depth distribution of sodium content is also relatively similar to that of sulfate and TDS, the patterns of calcium and magnesium are slightly different. Sodium concentrations are close to the calcium contents in waters from the shallowest well. The sodium concentration is greater than the calcium in the next two deeper wells but is smaller than the calcium content of waters from the two deepest wells. The data suggest that cation exchange altered the concentrations as saline waters migrated through the sediments of the aquifer.
 
 

Nitrate and fluoride concentrations do not change substantially with depth (Figure 11). The nitrate-nitrogen content of the shallowest well water is the highest for the site (in the range 2.5-3.3 mg/L during the monitoring), but is not much greater than the concentrations for the deeper wells (1.9-2.2 mg/L). Bicarbonate contents are relatively similar in the three shallowest wells and decrease to about 200 mg/L in waters from the deepest two wells. Boron concentrations first increase then decrease in depth in a pattern similar to that for sodium and sulfate.

The sulfate and TDS concentrations substantially exceed the recommended levels for drinking water of 250 mg/L and 500 mg/L, respectively, in waters from all the wells. The sulfate concentration in all the well waters also exceeds the maximum recommended for stock use (1,000 mg/L). None of the other constituents exceeds or is close to standards for drinking, stock, or irrigation water use.

The KGS and the DWR have periodically measured water levels in the wells since installation (Table 6). Figure 12 displays the change in water levels with depth and time. For all measurements, the deeper the well, the greater the depth to water. The average difference between the water levels in the shallowest and deepest wells is about the same as that for most of the measurement period for the Deerfield site. Water levels in the three shallowest wells at Garden City vary by about the same amount with time. The temporal variations in the water levels of the deepest two wells are very similar and are substantially greater than the water-level variations in the shallower wells. The water levels in the middle well (#3) are consistently between the levels of all the other wells. The middle zone appears to represent a transition to the substantially deeper water levels of the lowermost 100 ft of the High Plains aquifer at the site. The relative direction of temporal change (up or down) in water level is the same for all wells, indicating a great enough hydraulic connection throughout the aquifer to produce responses to changes at all levels. The largest water-level changes are in the summer and reflect the greater rate of pumping for irrigation and municipal use during that season in the area.