U.S. Geological Survey
Water-Resources Investigations Report 02-4021

Prepared in cooperation with the
CITY OF WICHITA, KANSAS

Download this report as a PDF file (633 KB)

Download a free copy of Acrobat Reader

Historical Contributions of Phosphorus From Natural and Agricultural Sources and Implications for Stream Water Quality, Cheney Reservoir Watershed, South-Central Kansas

By Larry M. Pope, Chad R. Milligan, and David P. Mau

CONTENTS

Abstract
Introduction

Setting
Previous Investigations
Purpose and Scope

Methods

Soil-Core Collection, Processing, and Analysis
Quality Control

Sequential Replicate Samples
Split Replicate Samples

Historical Contributions of Phosphorus

Natural Soil Concentrations
Agricultural Enrichment

Particle-Size Weighted Concentrations
Phosphorus Mass

Natural and Agricultural Sources

Phosphorus and Sediment Retention Efficiencies of Cheney Reservoir
Implications for Stream Water Quality
Summary and Conclusions
References

FIGURES

Figure 1. Maps showing location of Cheney Reservoir watershed and nonagricultural soil-coring sites, south-central Kansas
Figure 2. Photograph showing soil-coring site 27, cemetery east of Langdon, Kansas
Figure 3. Graph showing comparison of total phosphorus concentrations and particle sizes less than 0.062 millimeter in diameter between 2-inch and 8-inch soil-core samples collected from 43 nonagricultural sites in and near Cheney Reservoir watershed, September 1999
Figure 4. Map showing distribution of total phosphorus concentrations in 2-inch and 8-inch soil-core samples at 43 nonagricultural coring sites in and near Cheney Reservoir watershed, September 1999
Figure 5. Map showing distribution of soil particle size in 2-inch and 8-inch soil-core samples at 43 nonagricultural coring sites in and near Cheney Reservoir watershed, September 1999
6-9. Graphs showing:

Figure 6. Relation between total phosphorus concentrations in bottom-sediment core samples and normalized depth of samples from selected coring sites in Cheney Reservoir, 1965-97
Figure 7. Relation between mean percentage of bottom-sediment particles less than 0.062 millimeter in diameter and mean concentrations of total phosphorus in bottom-sediment samples from 10 coring sites in Cheney Reservoir, August 1997
Figure 8. Comparison of particle-size weighted mean total phosphorus concentrations in soil-core samples from 43 nonagricultural sites in and near Cheney Reservoir watershed, September 1999, and in bottom-sediment samples from Cheney Reservoir, August 1997
Figure 9. Comparison of phosphorus transported to Cheney Reservoir from natural and agricultural sources, 1965-98

Figure 10. Map showing location of streamflow sampling sites in Cheney Reservoir watershed
Figure 11. Graph showing comparison of measured mean total phosphorus concentrations in water samples from five streamflow sampling sites in Cheney Reservoir watershed and mean concentrations estimated on basis of agricultural-enrichment factor of 2.9

TABLES

Table 1. Location of soil-coring sites in and near Cheney Reservoir watershed, south-central Kansas
Table 2. Results of analyses of total phosphorus concentrations and particle-size determinations for selected soil cores in and near Cheney Reservoir watershed, south-central Kansas, September 1999
Table 3. Description of streamflow sampling sites in Cheney Reservoir watershed, south-central Kansas
Table 4. Number of samples analyzed and mean concentrations of total phosphorus in streamflow during base-flow, runoff, and long-term streamflow conditions at sampling sites 1-5 in Cheney Reservoir watershed, south-central Kansas, 1997-2000

CONVERSION FACTORS AND ABBREVIATIONS

Historical Contributions of Phosphorus From Natural and Agricultural Sources and Implications for Stream Water Quality, Cheney Reservoir Watershed, South-Central Kansas

Abstract

INTRODUCTION

The effects of eutrophication can be most pronounced on lakes and reservoirs that serve as water-quality integrators (summation of upstream constituent characteristics) for their watersheds. Cheney Reservoir, on the North Fork Ninnescah River in south-central Kansas (fig. 1), may be a reservoir with the potential for detrimental effects from nutrient enrichment and, in fact, has a history of algal blooms and taste-and-odor problems in treated drinking water (Cheney Reservoir Watershed Task Force Committee, written commun., 1996).

To address the eutrophication issues in Cheney Reservoir and nutrient (nitrogen and phosphorus species) transport from the Cheney Reservoir watershed, the U.S. Geological Survey (USGS) entered into a cooperative study with the city of Wichita in 1996. The study was designed to assess the occurrence and transport of selected water-quality constituents, including phosphorus, within the watershed and to describe their potential effect on Cheney Reservoir, a major source of public-water supply for Wichita and surrounding communities.

Results from the study will be useful to watershed managers in evaluating the sources and transport of phosphorus in the Cheney Reservoir watershed and the potential for meeting in-stream total maximum daily loads (TMDLs) for phosphorus under existing or proposed land-use and land-management practices. A TMDL is a calculation of the maximum amount of a contaminant that a water body can receive without violating State water-quality standards. Results also may have transferability to other watersheds in Kansas with similar land-use characteristics and phosphorus distribution patterns. The rather unique sampling methods used for this study could have transferability throughout the Nation where a comparison of the historical transport of water-quality constituents to natural conditions (unaffected by human activity) is required. This comparison would be particularly suited to watersheds with downstream impoundments where a record of historical transport may be contained in bottom sediment.

Setting

In south-central Kansas, precipitation and temperature vary substantially throughout the year. The mean annual (1961-90) precipitation is about 27 in. but ranges in the vicinity of the Cheney Reservoir watershed from about 25.8 in. at Pratt (fig. 1) to about 29.2 in. at Hutchinson (National Oceanic and Atmospheric Administration, 1992). The mean annual temperature is 56.6 ÷F, with a mean monthly range of 30.7 ÷F in January to 81.3 ÷F in July (Kansas Department of Agriculture and U.S. Department of Agriculture, 1997, p. 8).

Topography in the Cheney Reservoir watershed ranges from flat to gently sloping hills. Total topographic relief is about 550 ft, with maximum local relief (within 1 mi) of about 50 ft.

The North Fork Ninnescah River Valley and the surrounding plains are underlain by consolidated rocks (mainly with thin layers of limestone, dolomite, siltstone, gypsum, and salt) of Permian age (230 to 280 million years old) covered by unconsolidated stream-related (mainly sand and gravel) and windblown (mainly silt and clay) deposits of Pleistocene age (less than 1.5 million years old) (Zeller, 1968). The unconsolidated deposits are a source of water for domestic supply, irrigation, and livestock and range in thickness from little or no saturated thickness east of Red Rock Creek (fig. 1) to more than 160 ft in the northern part of Pratt County (Hansen, 1991).

Soils in the Cheney Reservoir watershed generally are classified as clayey loam on the uplands to sand or sandy loam on low-lying areas or where slopes are less than about 3 percent. Many of the soils in the watershed are subject to erosion by wind and rainfall runoff (Rockers and others, 1966).

Previous Investigations

Purpose and Scope

METHODS

Soil-core samples were collected at 43 coring sites (fig. 1, table 1) in and near Cheney Reservoir watershed on September 1-9, 1999. The sites were selected to represent soils with natural concentrations of total phosphorus and to provide as wide a spatial distribution as possible in and near Cheney Reservoir watershed.

Soil-Core Collection, Processing, and Analysis

Forty of the 43 locations selected for this study were in cemeteries (an example is shown in figure 2), two were on abandoned farmsteads, and one was on Kansas State University (KSU) land. Although several of the selected cemeteries were not located in the Cheney Reservoir watershed, it was believed that soil conditions at these nearby cemeteries would be similar to those soils within the watershed and that the positive benefits from a larger, spatially diverse data set would offset possible negative effects of dissimilar soils. Discussions with local residents and cemetery caretakers indicated that landscape fertilizers are not commonly, if ever, applied to cemetery soils in this watershed. However, to avoid sampling soils that may have been fertilized, cemeteries with unusually viable, well-maintained lawns were not selected for this study. The two abandoned farmsteads were included to increase spatial definition in the coring-site distribution. Each farmstead had been abandoned for more than 50 years, and the KSU land had no history of fertilizer applications.

Within each cemetery, sampling locations were selected in the oldest part, as indicated by headstone dates, or in areas as yet to receive burials. At each sampling location (fig. 1), soil cores of two lengths (2 and 8 in.) were collected using a 2-in. diameter soil auger. Two soil-core lengths were collected to determine vertical variability in soil phosphorus concentrations and to represent the parts of the soil profile most likely eroded and transported in runoff. The soil auger was cleaned between sampling locations in a sequence of nonphosphorus soap/water mix, tapwater rinse, and deionized-water rinse. Each soil sample was sieved through a 2-mm opening sieve (U.S. Standard No. 10) to remove plant debris, mechanically homogenized, and subsampled for analytical determinations. All soil samples were analyzed for total phosphorus at the USGS National Water-Quality Laboratory in Denver, Colorado, according to methods described in Fishman (1993). Soil particle size was determined for the percentage of sand and silt and (or) clay (percentage greater than or less than 0.062 mm in diameter, respectively) at the USGS sediment analysis laboratory in Iowa City, Iowa, according to methods described in Guy (1969).

Quality Control

Sequential Replicate Samples

Within-site variability was evaluated through the collection and analysis of primary and two sequential replicate soil-core samples (table 2) collected at coring sites 13, 20, 26, 27, and 40 (fig. 1). Sequential replicate soil-core samples were collected from different locations within the coring sites (cemeteries). All coring locations in cemeteries were either in the oldest developed parts or in unused areas (areas as yet to receive burials). Variation among the three resulting soil phosphorus concentrations and percentage of particle size less than 0.062 mm (silt and clay) at each coring location and depth was evaluated with a technique used by Dufour and others (1981) that expressed variation as a percentage of the mean with the equation:

(1) where
where s is the standard deviation,
where n is the number of replicate analyses, and
where _
where x is the mean replicate values from the analyses.

Within-site variability in soil phosphorus concentrations ranged from 2.2 percent of the mean in the 2-in. core samples from coring site 20 (fig. 1) to 31 percent of the mean in the 2-in. core samples from coring site 13. The average within-site variability was 16 percent of the mean for the 2-in. core samples and 11 percent of the mean for the 8-in. core samples from all five coring sites. The average within-site variability for both coring depths (2 and 8 in.) at all five coring sites with sequential replicate core samples (table 2) was 13 percent of the mean.

Within-site variability in percentage of soil core as silt and (or) clay (particle size less than 0.062 mm in diameter) was not as large as that for soil phosphorus concentrations. Variability in percentage of silt and (or) clay ranged from 0.2 percent of the mean in the 8 in. core samples from coring site 26 to 25 percent of the mean in the 2-in. core samples from coring site 27. The average within-site variability was 9.8 percent of the mean for the 2-in. core samples and 5.4 percent of the mean for the 8-in. core samples from all five coring sites. The average within-site variability for both coring depths at all five coring sites with sequential replicate core samples was 7.6 percent of the mean.

Evaluation of within-site variability for soil phosphorus and percentage of silt- and (or) clay-size particles produced two conclusions. First, the 2-in. core samples had greater within-site variability than the 8-in. core samples. This greater variability in the shorter core samples may be related to natural variability in organically bound phosphorus near the surface of the soil profile. This near-surface variability, in turn, probably is the result of variation in vegetative cover (organic production) within the coring site (cemetery). The greater particle-size variability in the 2-in. core samples probably is the result of preferential sorting and distribution of certain soil particles sizes that result from water and windblown processes. Closely associated with these erosional processes are variations in microenvironments that either may promote or mitigate the movement, distribution, and segregation of certain soil particle sizes within a coring site. Microenvironmental characteristics include type and percentage of vegetative cover, land slope, precipitation, orientation to prevailing wind direction, and proximity to structures, obstacles, or natural features that affect the potential for water or windblown erosion. The variability in these surface characteristics has less effect (smaller variability) on longer (8-in.) core samples because of the larger mass of these samples and their ability to dilute these near-surface effects. Second, soil phosphorus concentrations and percentage of silt- and (or) clay-size particles from the analysis of sequential replicate samples indicate that, although within-site variability may be large at some coring sites, the average soil phosphorus concentrations and percentage of silt- and (or) clay-size particles from the entire network of coring sites would be reasonably represented by single sampling locations within sampling sites.

Split Replicate Samples

Variability in total phosphorus concentrations in soil cores resulting from a combination of sample collection, processing, and analytical procedures was evaluated with split replicate core samples (table 2) collected at coring sites 9, 18, 28, 29, and 34 (fig. 1). Split replicate samples of both 2- and 8-in. soil-core samples were prepared by mechanically homogenizing each sample at the collection site and splitting the homogenized mass into two separate samples each of which was independently analyzed for total phosphorus. Variability between split replicate analyses was calculated as a percentage of the mean using equation 1.

Variability in total phosphorus concentrations in split replicate soil-core samples ranged from 0.7 percent of the mean in the 2-in. core samples from coring site 18 to 46 percent of the mean in the 2-in. core samples from coring site 9. The average variability was 14 percent of the mean for the 2-in. split replicate samples and 7.6 percent of the mean for the 8-in. split replicate samples. The average variability for all split replicate samples from both coring depths at all five coring sites was 11 percent. With the exception of the unusually large 46 percent of the mean from coring site 9, variability between all other split replicate samples was less than 17 percent of the mean. Excluding the 46 percent of the mean from the calculation of average variability reduced average variability for both coring depths at all five coring sites from 11 to 6.9 percent of the mean. Generally, therefore, results from the analyses of split replicate soil-core samples presented in table 2 indicate that sample collection, processing, and analysis of soil-core samples for total phos-phorus were adequate and not a substantial source of variability.

HISTORICAL CONTRIBUTIONS OF PHOSPHORUS

Estimates of natural concentrations of total phosphorus in soil were made from the analyses of soil-core samples collected from 43 nonagricultural sites in and near Cheney Reservoir watershed. These natural concentrations were compared to total phosphorus concentrations in Cheney Reservoir bottom sediment to estimate the extent to which the agricultural distribution of phosphorus has enriched soil phosphorus concentrations in the watershed. A phosphorus-enrichment factor was calculated and used to estimate mean total phosphorus concentrations in water at selected streamflow sampling sites in the Cheney Reservoir watershed that may have occurred under conditions of natural concentrations of soil phosphorus. Before results from coring sites with quality-control samples (table 2) were used in calculations, statistical analyses, or graphical presentations, primary and replicate analyses were averaged.

Natural Soil Concentrations

Natural concentrations of total phosphorus in soil in and near Cheney Reservoir watershed were estimated from soil-core samples collected from 43 nonagricultural sites (fig. 1) and analyzed for total phosphorus (table 2). Generally, total phosphorus concentrations were larger and had a wider range in the 2-in. soil-core samples than in the 8-in. core samples (fig. 3). Total phosphorus concentrations ranged from 74 to 539 mg/kg with a median concentration of 245 mg/kg in 2-in. soil-core samples compared to a range of 50 to 409 mg/kg with a median concentration of 166 mg/kg in 8-in. soil-core samples. In contrast, there was little difference in soil particle-size composition (percentage of particle size less than 0.062 mm) in either ranges or median values between the 2-in. and 8-in. soil-core samples. Soil particle size less than 0.062 mm ranged from 0.4 to 88 percent for both 2-in. and 8-in. soil-core samples with median values of 44 and 41 percent, respectively. Some of the potential reasons for differences in total phosphorus concentrations between the 2-in. and 8-in. soil-core samples were discussed previously in the "Quality Control" section of this report.

A previous investigation of total phosphorus transport in streams of the Cheney Reservoir watershed in 1997-98 (Pope and Milligan, 2000) identified an increasing trend in total phosphorus yields in a west-to-east direction. Phosphorus yields ranged from about 0.13 lb/acre in the western part of the watershed to about 0.37 lb/acre in the eastern part. To determine if this trend is partly the result of natural variability in total phosphorus concentrations in soil, a plot of the spatial distribution of total phosphorus concentrations (fig. 4) in core samples from the 43 nonagricultural coring sites (fig. 1) sampled during the September 1999 investigation was examined. Although not definitely evident in figure 4, it appeared that the eastern part of the watershed had the larger natural concentrations of total phosphorus in soil. To verify this observation, a statistical comparison was made between coring sites located in the western half of the watershed (coring sites 1-19, 22, 23, 25, fig. 1) with those in the eastern half (coring sites 20, 21, 24, 26-43, fig. 1)

Median total phosphorus concentrations in 2-in. and 8-in core samples were 36 and 47 percent larger, respectively, in samples collected from the eastern coring sites than in samples collected from the western coring sites. These percentages were statistically significant at least at the 0.07 level of significance (p-value 0.07 or less) as determined by a Wilcoxon rank-sum test. Part of the west-to-east increase in phosphorus concentrations may be associated with a corresponding increase in the percentage of silt- and (or) clay-size particles (particles less than 0.062 mm in diameter) in samples from the eastern coring sites (fig. 5). Median percentages of silt- and (or) clay-size particles were 17 and 34 percent larger in 2-in. and 8-in. soil-core samples, respectively, from the eastern coring sites than in samples from the western coring sites. The determination that soils are sandier in the western part of the watershed is substantiated by Rockers and others (1966), Horsch and others (1968), and Dodge and others (1978). Although the west-to-east variability in natural soil phosphorus concentrations in the Cheney Reservoir watershed has a part in explaining the previously determined west-to-east trend in total phosphorus yield, other factors such as variability in land-use and land-management characteristics, precipitation and water yield, and topography may contribute substantially more to the 185-percent increase in total phosphorus yields from west to east in the Cheney Reservoir watershed.

Agricultural Enrichment

To evaluate the effects of agriculture on soil phosphorus concentrations in the Cheney Reservoir watershed, concentrations of total phosphorus in soils from nonagricultural areas were compared to total phosphorus concentrations in Cheney Reservoir bottom sediment. A previous examination (Pope, 1998) of Cheney Reservoir bottom sediment indicated that phosphorus transport from the watershed has been increasing since at least construction of the reservoir, probably as a result of agricultural activities, as evident in bottom-sediment cores from Cheney Reservoir (fig. 6) and that a direct relationship exists between total phosphorus concentrations and percentage of silt- and (or) clay-size particles in bottom sediment (fig. 7). This relation of particle size to total phosphorus also has been identified by other researchers (Keulder, 1982; Dong and others, 1983; Viner, 1984). Therefore, variability in total phosphorus concentrations among samples of soil or sediment with varying particle-size composition may reflect only the variability with particle size and not that associated with agriculture. To make comparisons of this type, all samples first must be normalized relative to particle-size composition. This normalization process also compensates for the preferential transport of silt- and (or) clay-size particles in runoff from agricultural fields (Rhoton and others, 1979; Dean, 1983). The preferential transport of silt and clay particles with their potentially larger phosphorus concentrations relative to the source soil naturally would tend to enrich deposited material in Cheney Reservoir.

Particle-Size Weighted Concentrations

Soil samples collected from nonagricultural locations (fig. 1) for this study and previously collected bottom sediment from Cheney Reservoir (Pope, 1998) were normalized (weighted) relative to the percentage of the silt and (or) clay fraction of the particle-size composition. Particle-size weighted (hereinafter referred to as weighted) mean total phosphorus concentrations for both the 2-in. and 8-in. soil-core samples from nonagricultural locations in the Cheney Reservoir watershed were calculated with the equation:

(2) where
where C is the concentration of total phosphorus in 2- or 8-in. soil-core samples, and
where P is the percentage of the silt and (or) clay particle-size fraction in 2- or 8-in. soil-core samples.

A similar technique was used to calculate weighted mean total phosphorus concentrations in Cheney Reservoir sediment from results of bottom-sediment core analysis as presented in Pope (1998).

The weighted mean total phosphorus concentration in Cheney Reservoir bottom sediment was substantially larger than weighted mean total phosphorus concentrations in either the 2-in. (1.8 times larger) or 8-in. (2.6 times larger) soil-core samples from nonagricultural areas in and near Cheney Reservoir watershed (fig. 8). The sediment in Cheney Reservoir was derived almost exclusively from lands used for agricultural purposes (crops and livestock); therefore, the agricultural enrichment of phosphorus in this sediment as evident from figure 8 is probably the result of agricultural use of commercial phosphorus fertilizers and the surface application of livestock manure.

The differences between weighted mean total phosphorus concentrations in the two soil-core depths (fig. 8) probably reflect a natural buildup near the surface of the core profile of organic phosphorus from plant or microbial origins or possibly windblown deposits from nearby agricultural land. Because of a much smaller sample mass associated with the shorter soil core, this buildup would have a greater effect on mean total phosphorus concentrations (produce a larger mean concentration) in a short (2-in.) core sample than it would on the larger mass in a longer (8-in.) core sample.

Wind-transport processes also might be a reason for the larger weighted mean total phosphorus concentration in the shorter soil cores. Nearly all coring sites shown in figure 1 are surrounded by or in close proximity to agricultural lands where phosphorus fertilizers are used. South-central Kansas is subject to strong winds during times when cropland is relatively bare of vegetation. These winds have the potential for erosion of cropland with chemical transport to and deposition on nonagricultural areas. If phosphorus enrichment resulting from wind transport is substantial, then the differences between Cheney Reservoir bottom sediment and the nonagricultural soil samples could be even greater than that indicated in figure 8. Unfortunately, the extent, if any, of phosphorus enrichment resulting from wind transport is not known and probably cannot be quantified.

The 2.6 enrichment factor calculated using 8-in. soil-core samples collected at the 43 nonagricultural sites probably is the more representative factor for the Cheney Reservoir watershed because (1) results from 8-in. soil cores would minimize the effects (relative to results from 2-in. soil cores) of phosphorus enrichment by wind transport from nearby agricultural fields, and (2) results from 8-in. soil cores probably are more indicative of total phosphorus availability from land tilled for crop production, which is a substantial land use in the Cheney Reservoir watershed. Most tilled (plowed or cultivated) land probably is worked to at least an 8-in. depth and is the land most susceptible to erosion and transport of phosphorus.

Phosphorus Mass

An agricultural-enrichment factor for total phosphorus was calculated using the mean total phosphorus concentration in Cheney Reservoir bottom sediment (500 mg/kg) and the mean total phosphorus concentration (184 mg/kg) in 8-in. soil-core samples collected from 43 nonagricultural coring sites (fig. 1, table 1) used in the study described in this report. The mean total phosphorus concentration in 8-in. soil-core samples was calculated from data presented in table 2 after first averaging, by coring site, results of primary and replicate sample analyses. The agricultural-enrichment factor for total phosphorus in the Cheney Reservoir watershed was calculated by dividing the mean concentration in Cheney Reservoir bottom sediment (500 mg/kg) by the mean concentration (184 mg/kg) in 8-in. soil-core samples. The result of this calculation (2.7) produced an agricultural-enrichment factor for total phosphorus that was essentially the same as the 2.6 factor calculated from the results of the study described in this report. In summary, the information presented by Mau (2001) and the agricultural-enrichment factor calculated from that information provide supporting evidence for the validity of the particle-size weighted approach and methodology used in the study described in this report.

Both the 2.6 and 2.7 calculated enrichment factors for total phosphorus are conservative estimates of the effect of agricultural inputs of phosphorus in the Cheney Reservoir watershed because the factors do not account for (1) the phosphorus mass dissolved or suspended in the water at the time of collection of bottom-sediment samples, and (2) the mass of phosphorus historically discharged from the reservoir. These masses, however, can be estimated, added to the total mass of phosphorus in bottom sediment in Cheney Reservoir as determined by Mau (2001), and subsequently, used in the calculation of a historical mean agricultural-enrichment factor.

The phosphorus mass (in pounds) dissolved or suspended in the water at the time of collection of bottom-sediment samples (1997-98) was estimated from a calculation of the storage capacity of Cheney Reservoir at the conservation-pool elevation in 1998, 154,000 acre-ft (Putnam and others, 2000; Mau, 2001), multiplied by the mean concentration of total phosphorus, 0.10 mg/L (Milligan and Pope, 2001), of water in storage and by a unit conversion factor (2.720). The mean concentration of total phosphorus in water in storage was estimated as the mean concentration in water discharged from Cheney Reservoir during the course of the Cheney Reservoir watershed study (1997-2000). Estimated total phosphorus mass in Cheney Reservoir water at the time of collection of bottom-sediment samples was 42,000 lb.

The phosphorus mass historically discharged from Cheney Reservoir was estimated as a combination of mass actually discharged (into the North Fork Ninnescah River) and mass as part of the volume of water withdrawn by the city of Wichita for municipal supply. The phosphorus mass (in pounds) actually discharged from the reservoir was estimated from a calculation of volume of water discharged from the reservoir October 1, 1965, through September 30, 1998 (2.96 million acre-ft; Mau, 2001), multiplied by an estimated historical mean concentration of total phosphorus (0.07 mg/L) and by a unit conversion factor (2.720). The historical mean concentration of total phosphorus was estimated on the basis of the fact that the mean concentration of total phosphorus in reservoir discharge during the course of the Cheney Reservoir watershed study (1997-2000) was 0.10 mg/L (Milligan and Pope, 2001) and that total phosphorus transport from the watershed has been increasing with time (Pope, 1998; Mau, 2001). Therefore, the 0.10-mg/L mean concentration calculated for reservoir discharge (1997-2000) probably reflects a historical maximum mean concentration. No comprehensive historical (1965-98) database exists for concentrations of total phosphorus in water discharged or withdrawn from Cheney Reservoir. Consequently, the estimated 0.07-mg/L concentration of total phosphorus represents a compromise between what is believed to be an upper limit (0.10 mg/L) in mean total phosphorus concentration and an estimated lower limit (0.05 mg/L) determined on the basis of best professional judgement. Estimated total phosphorus mass actually discharged from Cheney Reservoir (1965-98) was calculated at 564,000 lb.

The phosphorus mass (in pounds) as part of the volume of water withdrawn by the city of Wichita for municipal supply was estimated from a calculation of volume of water withdrawn from the reservoir October 1, 1965, through September 30, 1998 (694,000 acre-ft; Mau, 2001), multiplied by an estimated historical mean concentration of total phosphorus (0.07 mg/L) and by a unit conversion factor (2.720). Estimated total phosphorus mass as part of the volume of water withdrawn for municipal supply was calculated at 132,000 lb.

Mau (2001) calculated the total mass of phosphorus in bottom sediment of Cheney Reservoir at 7.7 million lb. The total mass of phosphorus historically transported into Cheney Reservoir from the watershed is a summation of phosphorus mass in bottom sediment (7.7 million lb), in the reservoir water at the time of bottom-sediment sample collection (42,000 lb), in water discharged from the reservoir (564,000 lb), and in water withdrawn from the reservoir (132,000 lb). Total phosphorus mass historically transported (1965-98) into Cheney Reservoir was 8.4 million lb.

A historical mean agricultural-enrichment factor for total phosphorus can be calculated as the ratio of the historical phosphorus mass, expressed as a mean concentration of total phosphorus in sediment transported into Cheney Reservoir, to (divided by) the mean concentration of total phosphorus in soils from nonagricultural areas of the Cheney Reservoir watershed. The historical mean total phosphorus concentration in sediment transported into Cheney Reservoir (542 mg/kg) was calculated as total phosphorus mass divided by total sediment mass. Total phosphorus mass (8.4 million lb) was previously calculated. Mau (2001) calculated the total sediment mass deposited in Cheney Reservoir during 1965-98 (15.4 billion lb); however, this mass did not include an estimate for sediment discharged from the reservoir.

Sediment mass discharged from the reservoir (actually discharged and withdrawn) was calculated using the same method and water volumes as used in the calculation of phosphorus mass except a historical mean concentration of sediment in water discharged from Cheney Reservoir was estimated on the basis of a mean concentration of total suspended solids in water (9.8 mg/L) discharged from the reservoir during 1997-2000 (Milligan and Pope, 2001). Total sediment mass discharged from the reservoir during 1965-98 was calculated at 97.0 million lb. Total sediment in reservoir water at the time of bottom-sediment sample collection (1998) was believed to be insignificant and was not calculated. Therefore, total sediment mass historically (1965-98) transported into Cheney Reservoir was 15.5 billion lb, and the historical mean agricultural-enrichment factor was 2.9 (542 mg/kg divided by 184 mg/kg).

It should be emphasized that the 2.9 enrichment factor is a historical mean and that a current (2001) enrichment factor may be greater because of an apparent increasing trend in phosphorus transport from the watershed (fig. 6). In summary, agricultural activities during 1965-98 have increased the transport of phosphorus into Cheney Reservoir 2.9 times greater than that expected under natural concentrations of phosphorus in watershed soil.

Natural and Agricultural Sources

Historical contributions of phosphorus to Cheney Reservoir from natural and agricultural sources could not be measured directly, but instead, were estimated from the previously calculated 2.9 historical mean agricultural-enrichment factor for total phosphorus. The enrichment factor, in effect, indicates that the mass of phosphorus historically transported into Cheney Reservoir is 2.9 times greater than the mass expected under conditions where the watershed soil contains an average natural concentration of soil phosphorus. In other words, the total mass of phosphorus historically transported into Cheney Reservoir (8.4 million lb) is about 2.9 times the mass from natural sources; thus, the mass from natural sources is about 8.4 million lb divided by 2.9, or 2.9 million lb (35 percent of the total mass historically transported). The percentage of this total mass of phosphorus transported into Cheney Reservoir that originated from agricultural sources, therefore, was about 65 percent (5.5 million lb) (fig. 9).

The percentages shown in figure 9 represent historical averages and may not accurately describe current (2001) conditions. However, these percentages do indicate that, because of the large percentage of phosphorus from agricultural sources, changes to agricultural or land-management practices that would mitigate the movement of sediment and associated phosphorus or changes in phosphorus application and distribution in the watershed have a large potential for reducing the mass of phosphorus transported to Cheney Reservoir.

PHOSPHORUS AND SEDIMENT RETENTION EFFICIENCIES OF CHENEY RESERVOIR

The potential for surface-water bodies to retain phosphorus varies, but many studies have documented large retention efficiencies. For example, phosphorus retention efficiencies for wetland and swamp areas in Florida were calculated at 72 percent (Moustafa and others, 1998) and 82 percent (Moustafa, 1999), respectively; an agriculturally affected lake in southwest Finland had a retention efficiency of greater than 80 percent (Ekholm and others, 1997); and a multipond system in southeastern China had a retention efficiency of about 98 percent (Weijin and others, 1998). More locally, Juracek (1998) calculated a historical phosphorus budget for Hillsdale Lake in northeast Kansas (fig. 1) that produced a phosphorus retention efficiency of 89 percent. The 92-percent phosphorus retention efficiency calculated for Cheney Reservoir, therefore, seems reasonable.

In a process similar to that for phosphorus, reservoirs also have the potential to retain large percentages of the inflow sediment load. Jansson and Erlingsson (2000) calculated a sediment retention efficiency of about 80 percent for Cachi Reservoir in Costa Rica. Tobin and Hollowed (1991) calculated a sediment retention efficiency for Kenney Reservoir in northwestern Colorado that ranged from 91 to about 98 percent; the latter percentage corresponds well to the 99-percent sediment retention efficiency calculated for Cheney Reservoir in the study described in this report.

Large retention efficiencies of water-quality constituents such as phosphorus or sediment may have water-quality implications for, and may reduce the useful life of, a reservoir. Phosphorus accumulation in reservoir bottom sediment may provide a source of internal loading of phosphorus (release of bioavailable phosphorus from sediment) (Morris and Fan, 1998), particularly under anaerobic (absence of oxygen) conditions at the sediment-water interface (Alaoui Mhamdi and others, 1994; Istranovics, 1994). This released bioavailable phosphorus may increase the rate of algal production (Horne and Goldman, 1994) and, subsequently, the occurrence of taste-and-odor problems in treated drinking water (Juttner, 1984; Hoson, 1992). However, under aerobic (oxygenated) conditions at the sediment-water interface, phosphorus release from sediment may be greatly suppressed (Erickson and Auer, 1998).

On the basis of limited depth profiling of dissolved oxygen concentrations collected during the Cheney Reservoir watershed study (1997-2000), anaerobic conditions in Cheney Reservoir have not occurred. Additionally, Gibson (1997) provided evidence that indicated there is a phosphorus-concentration threshold at approximately 1,000 mg/kg dry sediment below which sediment tended not to release phosphorus back into the water column. Phosphorus concentrations in Cheney Reservoir bottom sediment did not exceed this threshold (Pope, 1998; Mau, 2001).

Although Cheney Reservoir has a large retention efficiency for phosphorus (92 percent), other factors such as well-oxygenated water and relatively small concentrations of phosphorus in bottom sediment (less than 1,000 mg/kg) may limit algal proliferation and subsequent occurrences of taste-and-odor problems in treated drinking water from the reservoir by inhibiting the internal loading of phosphorus. In fact, taste-and-odor problems in treated drinking water occurred during several summers in the early 1990s (Cheney Reservoir Watershed Task Force Committee, written commun., 1996) but were not observed and reported during the Cheney Reservoir watershed water-quality study (1996-2001) until the summer of 2001.

One of the principal concerns with the transport of sediment into Cheney Reservoir is a loss of reservoir storage capacity (Cheney Reservoir Watershed Task Force Committee, written commun., 1996). Decreases in reservoir storage capacity can affect reservoir allocations used for flood control, drinking-water supplies, recreation, and wildlife habitat. The part of a reservoir designed for sediment storage is called the inactive conservation storage pool and commonly is designed to provide storage for 100 years of sediment deposition (Morris and Fan, 1998). As of 1998, 34 years of sediment deposition had occurred in Cheney Reservoir, which equates to 34 percent of the design life of the reservoir. A previous investigation of Cheney Reservoir (Mau, 2001) determined that during the first 34 years in the design life of Cheney Reservoir, 27 percent of the allocated sediment storage capacity had been used. Although the sediment retention efficiency at Cheney Reservoir was large (99 percent), the design life of the reservoir is being met provided that the common design-life specifications were used in the design of the reservoir.

IMPLICATIONS FOR STREAM WATER QUALITY

In 1996, the USGS entered into a cooperative study with the city of Wichita to assess the occurrence and transport of selected water-quality constituents, including phosphorus, within the Cheney Reservoir watershed. As part of that study, six streamflow sampling sites were established in the watershed (fig. 10, table 3). Streamflow sampling sites 1-5 were located on the main stem of the North Fork Ninnescah River or its tributary streams. Sampling site 6 was located at the outflow of Cheney Reservoir. A continuous record of streamflow was collected at all six sampling sites. Samples of streamflow were collected during base flow (low-flow conditions) and during runoff (high-flow conditions) ( table 4).

Automatic samplers were installed at streamflow sampling sites 1-5. The analysis of samples collected from these automatic samplers (automatic samples) and additional manually collected samples (manual samples) comprised the water-quality database for the Cheney Reservoir watershed assessment. Most runoff samples were collected with automatic samplers. Manual samples were collected to provide depth- and width-integrated composite samples representative of the average chemical composition of the stream's cross-sectional area. Automatic samples were collected from a single point in the stream cross-sectional area and, therefore, may not be representative of the average chemical composition of the stream's cross-sectional area at the time of sample collection. Samples were collected from January 1997 through December 2000. Samples were analyzed for total phosphorus concentrations at the city of Wichita laboratory using U.S. Environmental Protection Agency method 365.2 (U.S. Environmental Protection Agency, 1999).

Concentrations of total phosphorus determined from automatically collected point samples were adjusted to approximate the average cross-sectional phosphorus concentration provided by depth- and width-integrated manually collected samples. This approximation was accomplished through the development and application of sampling-site-specific linear-regression relations between total phosphorus concentrations determined for pairs of concurrently collected automatic and manual samples from streamflow sampling sites 1-5 (fig. 10). Details of this adjustment procedure are presented in Milligan and Pope (2001).

More samples of streamflow for analysis of total phosphorus were collected during runoff than during base flow (table 4) because the majority of phosphorus from the Cheney Reservoir watershed is transported during runoff conditions. For example, at sampling site 4 (fig. 10), the main inflow site to Cheney Reservoir, it was estimated that 72 percent of the phosphorus in streamflow at this site (1997-2000) was transported during runoff. This percentage mirrors the percentage of samples collected during runoff at sampling site 4 (73 percent, (table 4).

The 72-percent transport estimate for phosphorus during runoff was calculated using a regression equation that related phosphorus load (PL, dependent variable), in grams per second, to streamflow (Q, independent variable), in cubic feet per second. This equation [PL = 0.00063Q^1.423; R² (coefficient of determination) equals 0.84, p-value of less than 0.0001] was an updated version of an equation used in a previous investigation of phosphorus transport in the Cheney Reservoir watershed during 1997-98 (Pope and Milligan, 2000). It was updated to include total phosphorus and streamflow data through 2000. Daily mean streamflow (Putnam and others, 1998-2001) for periods of base flow and runoff were used as input variables for the regression equation. Daily loads and, ultimately, percentage transport for base flow and runoff were calculated.

On average, total phosphorus concentration in samples of streamflow collected during runoff were about 2.0 (sampling site 2) to 4.4 (sampling sites 3 and 5) times larger than concentrations in samples of streamflow collected during base flow (table 4). Mean concentrations of total phosphorus in samples of streamflow also were calculated as a long-term (1997-2000) average (table 4) using all available (base-flow and runoff) adjusted automatic and manual sample concentrations (fig. 11).

One of the purposes of determining natural concentrations of total phosphorus in soil is the estimation of the potential to achieve stream-water-quality goals under natural conditions, all other watershed factors being historically constant. This information may have substantial watershed-management implications.

Mean total phosphorus concentrations at streamflow sampling sites 1-5 that have occurred under natural concentrations of total phosphorus in soil were estimated by dividing the previously determined mean total phosphorus concentrations calculated from streamflow samples (1997-2000) by an agricultural-enrichment factor of 2.9. A comparison of measured mean and estimated mean total phosphorus concentrations (on the basis of a historical mean agricultural-enrichment factor of 2.9) in water from streamflow sampling sites 1-5 is presented in figure 11. An examination of results presented in figure 11 indicates that under natural concentrations of total phosphorus in soil, mean total phosphorus concentrations in water from two of the five streamflow sampling sites upstream from Cheney Reservoir (sampling sites 3 and 5) would not meet the water-quality goal of 0.10 mg/L established by the Cheney Reservoir Watershed Task Force Committee. The conclusion, therefore, is that even if soils in the Cheney Reservoir watershed were maintained with natural concentrations of total phosphorus, some streams or stream segments would not meet water-quality goals assuming all the other historical watershed conditions remained the same.

The implication of the information presented in figure 11 is that the water-quality goal for total phosphorus in some streams in the Cheney Reservoir watershed may not be met simply by reducing phosphorus concentrations in soil (reducing the amount of phosphorus applied). Instead, meeting the goal could involve a combination of approaches-for example, reducing the agricultural distribution of phosphorus and implementing changes in land-use, land-management, and agricultural practices that reduce the loss of phosphorus from soil to water or the movement of phosphorus-laden soil in runoff (Devlin and others, 1998). Previous investigations have shown a direct relation between phosphorus concentrations and percentages of the fine silt- and (or) clay-size (less than 0.062 mm) particles (Galeone, 1996; Pope, 1998). Therefore, changes in the watershed that would mitigate the transport of these fine particles also would reduce the movement of phosphorus to surface water and, potentially, the eutrophication of Cheney Reservoir.

The possibility of substantial reductions in estimated long-term mean concentrations of total phosphorus in water from several sampling sites (for example, sites 1, 2, and 4; figs. 10 and 11) probably is not great. An analysis of total phosphorus concentrations in water from Kings Creek (USGS gaging station 06879650) located on the Konza Prairie Biological Station (KPBS) (fig. 1) in northeast Kansas was conducted to compare results from the agriculturally dominated Cheney Reservoir watershed to a "natural" (predevelopment) prairie-grass environment. The KPBS is representative of an e astern Kansas native tallgrass prairie ecosystem (Kansas State University, 2001).

Between 1980 and 1996, the USGS collected and analyzed 76 samples of streamflow from Kings Creek for total phosphorus. The mean concentration of total phosphorus in these samples was 0.06 mg/L (data on file at USGS office in Lawrence, Kansas), not too dissimilar to the estimated long-term mean concentrations determined under natural soil phosphorus conditions at sampling sites 1, 2, and 4 in the Cheney Reservoir watershed (fig. 11). This 0.06-mg/L concentration may provide insight as to a lower threshold limit of total phosphorus in eastern Kansas, tallgrass prairie streams under "natural" conditions. Therefore, it is probably unrealistic to expect that a long-term mean concentration of total phosphorus in water from a managed agricultural watershed would be less than 0.06 mg/L regardless of whether the soil phosphorus concentrations in the watershed were natural concentrations or not. However, this supposition is qualified with the fact that in 1987 the KPBS had a permanent herd of bison established in the Kings Creek watershed and that in 1992 an experimental herd of cattle was introduced (Kansas State University, 2001). It is unknown to what extent these animal populations have affected the long-term mean concentration of total phosphorus in water from Kings Creek. The available data set is not adequate for that type of analysis.

SUMMARY AND CONCLUSIONS

REFERENCES

Larry M. Pope
U.S. Geological Survey
4821 Quail Crest Place
Lawrence, KS 66049-3839
Telephone: (785) 832-3548
Fax: (785) 832-3500
Email: mpope@usgs.gov