Relations of Streamflow and Specific- Condu9tance Trends to Reservoir Operations in the Lower Arkansas River, Southeastern Colorado

By Michael E. Lewis and Daniel L. Brendle

U.S. GEOLOGICAL SURVEY

Water-Resources Investigations Report 97-4239

Prepared in cooperation with the COLORADO SPRINGS UTILITIES; PUEBLO BOARD OF WATER WORKS; SOUTHEASTERN COLORADO WATER CONSERVANCY DISTRICT; PUEBLO COUNTY, DEPARTMENT OF PLANNING AND DEVELOPMENT; CITY OF AURORA, DEPARTMENT OF UTILITIES; ST. CHARLES MESA WATER DISTRICT; UPPER ARKANSAS AREA COUNCIL OF GOVERNMENTS; UPPER ARKANSAS WATER CONSERVANCY DISTRICT; CITY OF PUEBLO, DEPARTMENT OF UTILITIES; PUEBLO WEST METROPOLITAN DISTRICT; FREMONT SANITATION DISTRICT; CITY OF ROCKY FORD; CITY OF LAS ANIMAS; AND CITY OF LAMAR

Denver, Colorado 1998 U.S. DEPARTMENT OF THE INTERIOR BRUCE BABBITT, Secretary

U.S. GEOLOGICAL SURVEY Thomas J. Casadevall, Acting Director

The use of firm, trade, and brand names in this report is for identification purposes only and does not constitute endorsement by the U.S. Geological Survey.

For additional information write to: Copies of this report can be purchased from:

District Chief U.S. Geological Survey U.S. Geological Survey Information Services Box 25046, Mail Stop 415 Box 25286 Denver Federal Center Federal Center Denver, CO 80225-0046 Denver, CO 80225 CONTENTS

Abstract...... 1 Introduction...... ^^ 2 Purpose and Scope...... 4 Description of Study Area...... 5 Methods of Trend Analysis ...... 6 Water Administration and Reservoir Operations ...... 8 Pueblo Reservoir...... 8 John Martin Reservoir ...... 9 General Discussion of Streamflow and Specific Conductance in the Arkansas River...... 10 Relations of Streamflow and Specific-Conductance Trends to Reservoir Operations in the Arkansas River...... 11 At Canon City...... _^ 11 Above Pueblo...... 15 Near Avondale...... 23 At Las Animas...... 28 Below John Martin Reservoir...... 35 AtLamar...... 41 Summary...... 45 References Cited...... 47

FIGURES 1. Map showing location of study area and streamflow-gaging stations...... 3 2-27. Graphs showing: 2. Example of a boxplot...... 7 3. Daily mean Streamflow at selected Arkansas River gaging stations, 1982...... 10 4. Median specific conductance at selected sites on the Arkansas River, 1990-93...... 12 5. Daily mean Streamflow at station 07096000 (Arkansas River at Canon City), 1964-74 and 1975-94...... 13 6. Total annual volume of water imported into the Arkansas River from the Colorado River Basin, 1964-94...... 14 7. Specific conductance at station 07096000 (Arkansas River at Canon City), 1964-94 ...... 15 8. Monthly specific conductance at station 07096000 (Arkansas River at Canon City), 1964-74 and 1975-94...... 16 9. Daily mean Streamflow at station 07099400 (Arkansas River above Pueblo), 1966-74 and 1975-94...... 18 10. Specific conductance at station 07099400 (Arkansas River above Pueblo), 1966-94...... 19 11. Monthly specific conductance at station 07099400 (Arkansas River above Pueblo), 1966-74 and 1975-94...... 20 12. Duration frequency of specific conductance at station 07099400 (Arkansas River above Pueblo); Arkansas River at diversion for the Pueblo Board of Water Works; and Arkansas River at diversion for the St. Charles Mesa Water District, 1966-74 and 1975-94...... 22 13. Daily mean Streamflow at station 07109500 (Arkansas River near Avondale), 1969-74 and 1975-94...... 24 14. Specific conductance at station 07109500 (Arkansas River near Avondale), 1969-94 ...... 26 15. Monthly specific conductance at station 07109500 (Arkansas River near Avondale), 1969-74 and 1975-94...... 27

CONTENTS III 16. Duration frequency of specific conductance at station 07109500 (Arkansas River near Avondale), 1969-74 and 1975-94...... 29 17. Daily mean streamflow at station 07124000 (Arkansas River at Las Animas), 1961-74 and 1975-94...... 30 18. Relation of cumulative annual streamflow at station 07109500 (Arkansas River near Avondale) to cumulative annual streamflow at four downstream sites, 1966-94...... 32 19. Specific conductance at station 07124000 (Arkansas River at Las Animas), 1961-94 ...... 33 20. Monthly specific conductance at station 07124000 (Arkansas River at Las Animas), 1961-74 and 1975-94...... 34 21. Median end-of-month contents of John Martin Reservoir, 1955-79 and 1980-94...... 36 22. Daily mean streamflow at station 07130500 (Arkansas River below John Martin Reservoir), 1955-79 and 1980-94...... 37 23. Specific conductance at station 07130500 (Arkansas River below John Martin Reservoir), 1955-94...... 39 24. Monthly specific conductance at station 07130500 (Arkansas River below John Martin Reservoir), 1955-79 and 1980-94...... 40 25. Daily mean streamflow at station 07133000 (Arkansas River at Lamar), 1964-79 and 1980-94...... 42 26. Specific conductance at station 07133000 (Arkansas River at Lamar), 1964-94...... 43 27. Monthly specific conductance at station 07133000 (Arkansas River at Lamar), 1964-79 and 1980-94...... 44

TABLES 1. Main-stem stations for which streamflow and specific-conductance trends were analyzed ...... 5 2. Step-trend results on the daily mean streamflow at station 07096000 (Arkansas River at Canon City) between 1964-74 and 1975-94...... 14 3. Step-trend results on specific conductance at station 07096000 (Arkansas River at Canon City) between 1964-74 and 1975-94...... 17 4. Step-trend results on the daily mean streamflow at station 07099400 (Arkansas River above Pueblo) between 1966-74 and 1975-94...... 19 5. Step-trend results on specific conductance at station 07099400 (Arkansas River above Pueblo) between 1966-74 and 1975-94...... 21 6. Step-trend results on the daily mean streamflow at station 07109500 (Arkansas River near Avondale) between 1969-74 and 1975-94...... 25 7. Step-trend results on specific conductance at station 07109500 (Arkansas River near Avondale) between 1969-74 and 1975-94...... 28 8. Step-trend results on the daily mean streamflow at station 07124000 (Arkansas River at Las Animas) between 1961-74 and 1975-94...... 31 9. Step-trend results on specific conductance at station 07124000 (Arkansas River at Las Animas) between 1961-74 and 1975-94...... 35 10. Step-trend results on the daily mean streamflow at station 07130500 (Arkansas River below John Martin Reservoir) between 1955-79 and 1980-94...... 38 11. Step-trend results on specific conductance at station 07130500 (Arkansas River below John Martin Reservoir) between 1955-79 and 1980-94...... 41 12. Step-trend results on the daily mean streamflow at station 07133000 (Arkansas River at Lamar) between 1964-79 and 1980-94...... 43 13. Step-trend results on specific conductance at station 07133000 (Arkansas River at Lamar) between 1964-79 and 1980-94...... 45

IV CONTENTS CONVERSION FACTORS

Multiply By To obtain

acre 0.4047 hectare acre-foot (acre- ft) 1,233 cubic meter acre-foot per year (acre-ft/yr) 1,233 cubic meter per year cubic foot per second (ft3/s) 0.02832 cubic meter per second foot (ft) 0.3048 meter inch (in.) 25.4 millimeter mile (mi) 1.609 kilometer square mile (mi ) 2.59 kilometer ton per year (ton/yr) 0.9072 meter ton per year

The following terms and abbreviations also are used in this report: microsiemens per centimeter at 25 degrees Celsius (|iS/cm) milligram per liter (mg/L)

CONTENTS Relations of Streamflow and Specific-Conductance Trends to Reservoir Operations in the Lower Arkansas River, Southeastern Colorado

By Michael E. Lewis and Daniel L. Brendle

Abstract stations located between Pueblo Reservoir and John Martin Reservoir were analyzed for trends To provide for the better management that may have occurred after 1974, which corre­ of Streamflow in the lower Arkansas River, sponds to the construction of Pueblo Reservoir. two main-stem reservoirs were constructed. Data from the two stations located downstream John Martin Reservoir, constructed near from John Martin Reservoir were analyzed for Las Animas in 1948, and Pueblo Reservoir, trends that may have occurred after the imple­ constructed near Pueblo in 1975, provide for mentation of a new reservoir operating plan in flood control, irrigation, municipal water supply, 1980. and recreation. Both reservoirs have the potential At the station in the upper basin, stream- to alter specific conductance in the Arkansas flow increased significantly and specific conduc­ River because of Streamflow management. A tance decreased significantly after 1974 during change in specific conductance could affect the the low-flow months, January, February, and intended use of the water as an agricultural or March. These trends apparently were caused domestic water supply. Step-trend analysis of by the increased importation of low-specific- Streamflow and specific-conductance data for conductance water after 1974 from the Colorado the Arkansas River was used for determining if River Basin into the Arkansas River. At the three the operation of Pueblo Reservoir or John Martin stations located between Pueblo Reservoir and Reservoir had affected Streamflow or specific John Martin Reservoir, Streamflow and specific conductance in the lower Arkansas River. The conductance primarily were affected by Pueblo nonparametric Mann-Whitney-Wilcoxon rank- Reservoir operations. After 1974, at the two sum test was used for trend analysis. stations located 0.4 and 24 miles downstream Streamflow and specific-conductance data from Pueblo Reservoir, Streamflow generally collected at five streamflow-gaging stations on increased during most months of the growing the lower Arkansas River and at one station on season and decreased during November through the upper Arkansas River were analyzed for February. The Streamflow trends at these two trends. The station in the upper basin was stations largely were attributed to the storage of included in the analysis to differentiate between water in Pueblo Reservoir during winter and to trends in the lower basin that were caused by the release of that stored water during the growing differences in the quantity or quality of inflow season in order to meet downstream irrigation from the upper basin or were caused by reservoir needs. At the station 0.4 mi downstream from operations in the lower basin. Data from the Pueblo Reservoir, specific conductance decreased station in the upper basin and from the three during most months between September and

Abstract 1 April and increased during the high-flow months, water supply for most of the 165,580 people who live May and June. This trend was caused by the in the five counties that compose the lower Arkansas mixing of seasonally low-specific-conductance River Valley and the primary agricultural irrigation water and seasonally high-specific-conductance supply for about 300,000 acres of irrigated land in water in Pueblo Reservoir, thus narrowing the lower basin. Because of this dependence on the the annual range in specific conductance in Arkansas River as a municipal and an agricultural water supply, the quality of water in the Arkansas the reservoir outflow. Few trends, except for River is very important. increased specific conductance in June, August, The quality of water in the Arkansas River is and December, were detected in specific conduc­ markedly different throughout the study area. Specific tance at the station located 24 miles downstream conductance, which is directly related to dissolved- from Pueblo Reservoir. The increase in specific solids concentration, increases downstream from a conductance probably was caused by the median of about 500 jiS/cm near Pueblo to about combined effects of water storage and mixing 3,900 )J,S/cm at Lamar (fig. 1); this range in specific in Pueblo Reservoir and the increased inflow conductance corresponds to a range in dissolved- of relatively high-specific-conductance water solids concentration of about 340 to 3,600 mg/L from Fountain Creek. At Las Animas, located (Cain, 1987). The downstream increase in specific 120 miles downstream from Pueblo Reservoir, conductance and, hence, in dissolved solids, largely streamflow increased significantly during all is due to the consumptive use of surface water months after 1974. Specific conductance tended and ground water for agricultural irrigation (Miles, 1977). High specific conductance is indicative of high to decrease during all months, but the decreases dissolved-solids concentration; dissolved solids can generally were not statistically significant at the affect the suitability of water for domestic, industrial, 95-percent confidence level. and agricultural uses. The secondary maximum At the two stations located downstream contaminant level (SMCL) for dissolved solids in from John Martin Reservoir, specific conductance drinking water is 500 mg/L (U.S. Environmental was affected by changes in John Martin Reservoir Protection Agency, 1986). In the lower Arkansas operations, increases in the reservoir inflow, River, 500 mg/L of dissolved solids is equivalent and decreases in the specific conductance of to a specific conductance of about 700 to 800 uS/cm the reservoir inflow. Specific conductance (Cain, 1987). At higher levels, drinking water may decreased during September through April and have an unpleasant taste or odor or even cause did not change substantially during May through gastrointestinal distress. Additionally, high dissolved- August. These trends were very similar to trends solids concentrations can cause increased deterioration observed immediately downstream from Pueblo of plumbing fixtures and appliances. Relatively expensive, advanced water-treatment processes, such Reservoir and were attributed largely to increased as reverse osmosis, are needed to remove excessive storage and increased mixing of seasonally low- dissolved solids from water. and seasonally high-specific-conductance water Agriculture also can be adversely affected in John Martin Reservoir. These factors tended by high-specific-conductance water. Depending to increase the minimum specific conductance on the crop, agricultural losses might occur when and decrease the maximum specific conductance dissolved-solids concentrations reach 700 to in the reservoir outflow. 850 mg/L (U.S. Department of the Interior, 1994), which is equivalent to a specific conductance of about 950 to 1,200 uS/cm in the Arkansas River INTRODUCTION (Cain, 1987). With increasing specific conductance, special agricultural management practices may be The lower Arkansas River in southeastern needed and crops having a substantial salinity toler­ Colorado extends about 200 mi downstream from ance may need to be grown. Generally, crops with a Pueblo Reservoir to the Colorado-Kansas State line higher salinity tolerance have a lower market value (fig. 1). The Arkansas River is the primary municipal than salt-sensitive crops (Miles, 1977).

Relations of Streamflow and Specific-Conductance Trends to Reservoir Operations in the Lower Arkansas River, Southeastern Colorado O) c

CO O)

T3 CO CO 0)

T3 3 to

3 D>

INTRODUCTION Streamflow in the lower Arkansas River Additionally, Streamflow and specific-conductance is regulated by the storage and release operations trends were evaluated at station 07096000 (Arkansas of two main-stem reservoirs, Pueblo Reservoir River at Canon City), which is in the upper basin about and John Martin Reservoir. Because of Streamflow 28 mi upstream from Pueblo Reservoir (fig. 1). This manipulation, both reservoirs can cause increases station was included in the analysis to help differen­ or decreases in the dilution potential in the river and, tiate whether trends in the lower basin were caused thus, can affect specific conductance. Although the by differences in the quantity or quality of inflow from specific-conductance conditions and the relations the upper basin or were caused by reservoir operations between specific conductance and Streamflow in the in the lower basin. Although specific-conductance Arkansas River have been well documented (Miles, trends were the main focus of this study, Streamflow 1977; Cain, 1985, 1987), the effects of main-stem trends also were analyzed because Streamflow and reservoir operations on specific conductance specific conductance generally are correlated; there­ have not been systematically studied and fore, changes or trends in specific conductance often reported. can be explained in terms of the associated change or Beginning in 1988, the U.S. Geological trend in Streamflow. Survey initiated a basinwide study of water All Streamflow and specific-conductance quality in the Arkansas River and of the effects of data used for trend analysis in this study are available certain water-supply operations on water quality, from the U.S. Geological Survey National Water Data including the relations of Streamflow and specific- Storage and Retrieval System (WATSTORE). Addi­ conductance trends to reservoir operations in the tionally, all Streamflow and specific-conductance data lower Arkansas River. The study was conducted used for trend analysis in this study were published in in cooperation with the Colorado Springs Utilities; annual data reports (U.S. Geological Survey, 1959, Pueblo Board of Water Works; Southeastern 1960, 1961, 1962-65, 1963a, 1963b, 1964, 1965a, Colorado Water Conservancy District; Pueblo 1965b, 1966-75a, 1966-75b, 1976-95). County, Department of Planning and Development; At individual stations, the record length for city of Aurora, Department of Utilities; St. Charles specific-conductance data generally was shorter than Mesa Water District; Upper Arkansas Area Council the record for Streamflow. Therefore, the streamflow- of Governments; Upper Arkansas Water Conservancy record length for individual stations was shortened District; city of Pueblo, Department of Utilities; to match the specific-conductance record for that Pueblo West Metropolitan District; Fremont station. This matching of record lengths facilitated Sanitation District; and cities of Rocky Ford, the direct comparison of Streamflow and specific- Las Animas, and Lamar. conductance trends at individual stations. Because record lengths differed between stations (table 1), Purpose and Scope trend results were not quantitatively compared between stations. At most stations, the effects of tribu­ This report presents: (1) Step-trend analysis tary Streamflow and specific conductance were not results for Streamflow and specific conductance considered in the trend analysis because of a lack of at three main-stem Arkansas River streamflow- tributary Streamflow and specific-conductance data. gaging stations located between Pueblo Reservoir The exception was station 07106500 (Fountain Creek and John Martin Reservoir and at two main-stem at Pueblo), which is tributary to the Arkansas River stations located between John Martin Reservoir near Pueblo (fig. 1) and for which there exists long- and the Colorado-Kansas State line, and (2) a determi­ term Streamflow and specific-conductance data. nation of whether significant Streamflow and specific- Streamflow and specific-conductance data from conductance trends are related to the operations of the the three stations located between Pueblo Reservoir reservoirs. Daily mean Streamflow data and discrete and John Martin Reservoir were evaluated for trends specific-conductance data were used in the trend that might have occurred after the construction of analyses. The five main-stem stations were selected Pueblo Reservoir in 1975. Data from the two stations for trend analysis because of the availability of located downstream from John Martin Reservoir long-term Streamflow and specific-conductance data. were evaluated for trends that might have occurred

Relations of Streamflow and Specific-Conductance Trends to Reservoir Operations in the Lower Arkansas River, Southeastern Colorado following the implementation of the 1980 John Martin The semiarid climate of the study area is charac­ Reservoir operating plan (Arkansas River Compact terized by low to moderate precipitation, substantial Administration, 1980). Similarly, data collected in evaporation, low humidity, moderate to intense winds, the upper basin at station 07096000 were evaluated and a large daily range in temperature. Mean annual for trends that might have occurred after 1975 because precipitation ranges from 12 in. at Pueblo to 15 in. this station was used in the study to evaluate trends at Lamar. About 75 to 80 percent of the annual in the quantity and quality of inflow into Pueblo precipitation falls as rain during the growing season. Reservoir and the lower basin. Throughout the area, potential evapotranspiration greatly exceeds precipitation. Land use along the Arkansas River in the Table 1 . Main-stem stations for which streamf low and specific-conductance trends were analyzed lower Arkansas River Basin is predominantly agricultural. Major crops are alfalfa, corn, wheat, U.S. Geological Period of and sorghum; about 300,000 acres are irrigated. Most Station Survey station trend name of the irrigated acreage is located in the alluvial valley number analysis of the Arkansas River, along the major tributaries, and Arkansas River at 07096000 1964-94 near off-channel reservoirs. Crop types grown in the Canon City valley generally vary downstream by their salinity Arkansas River above 07099400 1966-94 tolerance. Vegetables and other salt-sensitive crops are Pueblo grown on proportionally more acreage upstream from Arkansas River near 07109500 1969-94 La Junta where salinity is lower; alfalfa, which is rela­ Avondale tively salt tolerant, is grown on proportionally more Arkansas River at 07124000 1961-94 Las Animas acreage downstream from La Junta where salinity is much higher (Miles, 1977). Arkansas River below 07130500 1955-94 John Martin Reservoir The Arkansas River is a partially penetrating Arkansas River at 07133000 1964-94 stream that is incised into the alluvial deposits that Lamar form the valley-fill aquifer of the Arkansas River Valley. The valley-fill aquifer, which extends from Pueblo to the downstream end of the study area, Description of Study Area is an unconfined system that directly underlies and is in hydraulic connection with the Arkansas River. The The study area includes the Arkansas River aquifer width varies from 1 to 14 mi and averages Basin in southeastern Colorado from the foothills of 3 to 5 mi. The thickness of the alluvium varies from the Rocky Mountains near Canon City to Lamar, a 0 to about 250 ft. The alluvium consists of fairly distance of about 200 mi (fig. 1). The Arkansas River well-sorted sand and gravel with minor amounts of headwaters are located to the northwest of the study clay. Depth to water varies from 0 ft in wetlands in the study area to about 40 ft in eastern Colorado, area near Leadville. The river flows south and east and the saturated thickness varies from less than through mountainous terrain before emerging from 10 ft to about 210 ft. Ground-water flow in the the mountains near Canon City, at an altitude of about alluvial aquifer generally is from west to east (Hurr 5,350 ft. At Pueblo, the river is impounded to form and Moore, 1972; Taylor and Luckey, 1974; Nelson Pueblo Reservoir. Downstream from Pueblo, the and others, 1989a, b, c). river flows eastward across flat terraces and almost Snowmelt from the mountainous upper basin level flood plains, an area commonly referred to is the primary source of streamflow in the Arkansas as the lower Arkansas River Valley. Immediately River. Snowmelt runoff usually begins in late April downstream from Las Animas, the river is impounded or early May and peaks in June. In addition to native by John Martin Reservoir. About 58 mi downstream snowmelt runoff, streamflow in the Arkansas River is from John Martin Reservoir, the river flows into supplemented by the transmountain diversion of water Kansas. The Arkansas River drains an area of about from the Colorado River Basin. Transmountain water 25,400 mi O in Colorado, including 4,669 mi 0 upstream is diverted into the Arkansas River Basin at locations from Pueblo Reservoir. that are more than 150 mi upstream from Pueblo

INTRODUCTION Reservoir. The transmountain water is imported location on the Arkansas River changed after the during the summer and can be stored in several off- construction of Pueblo Reservoir (1975) or after channel reservoirs in the upper basin or in Pueblo the implementation of the John Martin Reservoir Reservoir. The imported water may be released from 1980 operating plan. storage to meet downstream municipal or irrigation Step-trend analysis was done using the non- water-supply demands. Rainfall runoff, ground-water parametric Mann-Whitney-Wilcoxon rank-sum test inflow, and irrigation-return flow also contribute to (Bradley, 1968). This nonparametric procedure was flow in the Arkansas River. A substantial amount of selected because the data (streamflow and specific the water in the Arkansas River is diverted and conductance) generally were not normally distributed, consumptively used for irrigation in the study area. based on graphical data analysis. Nonparametric During 1955-94, the median annual streamflow procedures have more power (or efficiency) than para­ in the Arkansas River decreased downstream by metric procedures in cases where there is a substantial about 88 percent from Pueblo (448,000 acre-ft/yr) departure from normality (Helsel and Hirsch, 1988). to Lamar (53,700 acre-ft/yr), largely because of In addition to doing trend analysis on daily irrigation diversions. Irrigation-return flow from mean streamflow data and specific-conductance data tributary streams, drainage ditches, and the alluvial that were grouped by month, specific-conductance aquifer supplements flow in the Arkansas River, and data were analyzed for trends with data grouped by much of the streamflow in the river downstream from season. One problem with using monthly specific- La Junta consists of irrigation-return flow during parts conductance data is that the sample sizes are smaller of many years (Cain, 1987). compared to data sets consisting of several months of data. The p value for hypothesis testing is affected by sample size. For a given trend magnitude and vari­ METHODS OF TREND ANALYSIS ance, p values tend to increase as the sample size decreases (Helsel and Hirsch, 1992); therefore, it Trend analysis can be used to determine if becomes more difficult to reject the null hypothesis streamflow or water quality has changed over time. of no trend as the sample size decreases. Grouping In this study, step-trend analysis was used to deter­ specific-conductance data by season increased the mine streamflow and specific-conductance trends sample sizes of the data sets being analyzed. The in the Arkansas River. In a step-trend analysis, data seasonal grouping was based on the timing of water- collected before a specific time are assumed to be storage and release operations for Pueblo Reservoir from a distinctly different data population than data and John Martin Reservoir. For the streamflow- collected after that time. The difference between the gaging stations located between Pueblo Reservoir data populations is assumed to be one of location (for and John Martin Reservoir, data were grouped by example, mean or median), but not necessarily of scale growing season (March 16-November 14) or winter- (for example, variance or interquartile range). The storage season (November 15-March 15). For the step-trend analysis is much more specific than other streamflow-gaging stations located downstream from trend analyses (for example, monotonic trend analysis) John Martin Reservoir, data were grouped by growing because step-trend analysis requires that a particular season (April 1-October 31) or winter-storage season fact, the time of the change, is known before any (November 1-March 31). A thorough description examination of the data. It is imperative that the deci­ of the factors affecting these seasonal groupings sion to use step-trend analysis not be based on prior is provided in the following "Water Administration examination of the data because prior examination and Reservoir Operations" section. would bias the significance level of the test. Signifi­ Qualitative comparisons of streamflow and cance levels, as represented by the individual p values specific-conductance data were made using boxplots. of each test, were for the two-sided trend test because Boxplots are useful because variability between data no prior determination of the direction of trends was sets and unusually large or small values in a data set made. For this study, a significant test was defined can easily be seen. For this report, boxplots were at a 95-percent confidence level (p<0.05). The step- constructed to compare monthly differences in daily trend analysis is particularly well suited to this study mean streamflow and specific-conductance data at because the purpose of the study was to determine particular main-stem sites before and after implemen­ if streamflow or specific conductance at a particular tation of reservoir operations. Boxplots contain the

Relations of Streamflow and Specific-Conductance Trends to Reservoir Operations in the Lower Arkansas River, Southeastern Colorado following information (fig. 2). The horizontal line the 25th and 75th percentiles. The bottom of the inside the box represents the median value (50 percent vertical line on the boxplot is the smallest value within of the data are greater than this value and 50 percent of 1.5 times the IQR of the box. The top of the vertical the data are less than this value). The lower line of the line is the largest value within 1.5 times the IQR of box is the 25th or lower quartile (25 percent of the data the box. Outside values, shown as *, are greater than are less than this value). The upper line of the box is 1.5 times and less than 3 times the IQR from the box. the 75th percentile or upper quartile (75 percent of The far outside values, shown as O, are greater than the data are less than this value). The interquartile three times the IQR from the box. The number of data range (IQR) contains the values between the 25th values used to construct each boxplot is presented at and 75th percentiles and is the difference between the top of the boxplot.

17 13 Number of values

Far outside value greater than 3.0 times 16 0 the interquartile range 0 3.0 times the interquartile range 15

Outside value between 1.5 and 3.0 times 14 greater than the interquartile range

Q 13 1.5 times the interquartile range Z O O 12 LLJ CO Between the 75th percentile and 1.5 times DC greater than the interquartile range

LJJ W m U_ 1U 75th percentile O 00 => 9 O Median ^ Interquartile range

25th percentile

HI Between the 25th percentile and 1.5 times DC the interquartile range CO

1.5 times the interquartile range

Outside value between 1.5 and 3.0 times less than the interquartile range

3.0 times the interquartile range

Far outside value less than 3.0 times 0 less than the interquartile range

Figure 2. Example of a boxplot.

METHODS OF TREND ANALYSIS Statistical analyses were made of daily mean Pueblo Reservoir streamflow values. The daily mean streamflow is the mean streamflow for a given date and site. The median Pueblo Reservoir is used for the storage daily mean streamflow for a given month is the and regulation of water that is imported into the median value of all daily mean streamflows for that Arkansas River Basin from the Colorado River month during a specified period of time. Basin as part of the Fryingpan-Arkansas Project (hereafter referred to as the Project). The Project is a multipurpose water development constructed WATER ADMINISTRATION by the Bureau of Reclamation. The main purpose AND RESERVOIR OPERATIONS of the Project is to divert unappropriated water from the western slope of Colorado for use on In Colorado, water law is based on the the more populated, water-limited eastern slope. doctrine of prior appropriation. The prior appropria­ The Project began importing water in 1972. From tion doctrine holds that the water in a State is the prop­ 1972 through 1994, the Project imported an annual erty of the public, which has a vested right to the use median volume of about 47,300 acre-ft into the of this water. Specifically, the doctrine states that the Arkansas River Basin (U.S. Department of the first in time to use the water is first in right to receive Interior, 1996). During summer, the Project diverts that water in subsequent years. Prioritized direct-flow water from the western slope to the eastern slope, water rights for the Arkansas River were established where the water is held in storage until municipal as long ago as 1859. As irrigated agriculture spread or irrigation water-supply demands need to be throughout the basin, the list of prioritized water satisfied. Imported water may be stored in the rights grew rapidly, and the Arkansas River and upper basin or in Pueblo Reservoir, the farthest its tributaries were fully appropriated for normal downstream facility of the Project. Imported or average years by the mid-1880's (Abbott, 1985). water generally is stored in the upper basin as In most areas, water rights with priorities dated long as possible to minimize evaporative losses, after 1887 are little more than flood rights, which which are lower in the upper basin than in Pueblo allow diversion of water only in periods of higher Reservoir. During winter, water stored in the than average streamflow (Abbott, 1985). Flood upper basin may be released to the river for down­ rights do not provide a dependable supply of water stream storage in Pueblo Reservoir in order to because these flows generally occur at times inconve­ create upper basin storage space for the importation nient to farming operations or at rates in excess of of western slope water during the upcoming snow- canal capacities. Water-storage rights were developed melt runoff. and reservoirs were constructed to take advantage of the flow not available to direct diversions, which Storage began in Pueblo Reservoir in 1974, includes streamflow in excess of direct-flow water and the darn was completed in 1975. The reservoir rights (flood rights) and streamflow during the had an initial storage capacity of 357,678 acre-ft nonirrigation season (winter water) from November (Lewis and Edelmann, 1994). Since water was first through March. During 1880 through 1910, storage impounded in Pueblo Reservoir, reservoir storage rights were established that allowed for the yearly has fluctuated because of variations in the inflow diversion and storage of almost 500,000 acre-ft of and in the demand for stored water. Most of the Arkansas River water in off-channel reservoirs. Addi­ storage space in Pueblo Reservoir is reserved for tionally, two large main-stem reservoirs were built in Project water, although storage of some non-Project the lower Arkansas River Valley to manage Arkansas water is granted under a limited number of storage River streamflow. In 1948, the U.S. Army Corps contracts. Most of the annual inflow to the reservoir of Engineers completed construction of John Martin usually occurs during May through July. Reservoir Reservoir, a main-stem reservoir east of Las Animas storage generally decreases substantially by the end (fig. 1). In 1975, the Bureau of Reclamation com­ of the growing season because of decreased inflow pleted construction of Pueblo Reservoir, a main-stem and large downstream demands for irrigation reservoir west of Pueblo (fig. 1). water.

Relations of Streamflow and Specific-Conductance Trends to Reservoir Operations in the Lower Arkansas River, Southeastern Colorado The operation of Pueblo Reservoir, in particular storage period. During the winter-storage period from the Winter Water Storage Program (WWSP), has a November 1 to March 31, all inflow to the reservoir notable effect on the historic streamflow regime of the was required to be stored, except that as much as /5 lower Arkansas River. In winter (November-March), 100 ft /s of water could be requested by Colorado prior to the implementation of the WWSP in 1975, water users downstream from the dam (Abbott, 1985). irrigators in the lower Arkansas River Valley gener­ During the remainder of the year, river flow was ally diverted appropriated Arkansas River water onto stored, although Colorado could demand the a fallow fields to increase soil moisture for later use by release of as much as 500 ft /s of the water entering crops during the growing season. Alternatively, this the reservoir and Kansas could demand releases of that a water could have been stored during the winter and part of the inflow between 500 and 750 ft /s. Provi­ then released to the river for the downstream irrigators sions were made for the rate of release of stored water, to use during times when streamflow was insufficient without reference to the volume of stored water to meet irrigation needs. However, under Colorado assigned to each State. To ensure that each State water law, storage of water that is diverted with direct- received its share of stored water, release demands had flow water rights is not permitted. Therefore, the to be made concurrently. Although the Arkansas River WWSP was created to allow several irrigation canal Compact was developed to ensure that Colorado and companies downstream from Pueblo Reservoir to Kansas irrigators received their legal shares of store their direct-flow water in Pueblo Reservoir, Arkansas River water, the plan had several problems John Martin Reservoir, and in several private off- and generally was unsatisfactory to both States channel reservoirs during the winter and to use this (Abbott, 1985). Historically, following the winter- water during the crop-growing season. Under the storage period, reservoir storage was usually drawn WWSP, winter water storage is allowed from down to empty or almost empty very early in the irri­ November 15 to March 15. Generally, WWSP water is gation season, often by the middle of April. Because released from storage at times when streamflow is not of the unsatisfactory nature of this operation, a resolu­ large enough to meet irrigation demands. This situa­ tion was adopted by the Arkansas River Compact tion usually occurs in early spring or late summer and Administration in 1980. This resolution commonly is autumn. Winter water was stored every year from referred to as the 1980 operating plan (Arkansas River 1975 to 1994, except during the 1977-78 winter- Compact Administration, 1980). Under the new plan, storage season. During 1975-94, the median annual any water not immediately called for and released to volume of water that was stored in Pueblo Reservoir downstream irrigators is stored in separate storage as part of the WWSP was about 42,200 acre-ft accounts for the States of Colorado and Kansas. Either (Colorado Division of Water Resources, written State can call for the release of its stored water inde­ commun., 1995). pendently of the other. Two other recent changes have been made in the operation of John Martin Reservoir John Martin Reservoir that affect reservoir storage and streamflow down­ stream from John Martin Reservoir. A 10,000-acre-ft John Martin Reservoir is a 608,200-acre-ft permanent recreation pool was established in 1976, main-stem reservoir located 58 mi upstream from the and three irrigation canal companies have been Colorado-Kansas State line (fig. 1). The reservoir is allowed to store their approved WWSP water in used for flood control, irrigation-water storage, and John Martin Reservoir. Thirty-five percent of the recreation. Storage of irrigation water in John Martin winter water that the three canal companies store in Reservoir is by agreement between the States of John Martin Reservoir is shifted to Arkansas River Colorado and Kansas, under the terms of the Arkansas Compact use and is subject to downstream release. River Compact. The Arkansas River Compact is an These two changes, in conjunction with the 1980 oper­ agreement between Colorado and Kansas, signed in ating plan, have substantially increased the long-term 1948, which ensures both States will receive their storage of water in John Martin Reservoir and have percentage share of Arkansas River flows. The altered the flow regime in the Arkansas River down­ Compact agreement dictated a winter- and a summer- stream from the reservoir.

WATER ADMINISTRATION AND RESERVOIR OPERATIONS 9 GENERAL DISCUSSION OF main-stem reservoirs. The general spatial and temporal STREAMFLOW AND SPECIFIC patterns of Streamflow in the Arkansas River are illus­ CONDUCTANCE IN THE trated in the hydrographs for six main-stem Arkansas ARKANSAS RIVER River stations for an average runoff year (1982) (fig. 3). At station 07096000 (Arkansas River at Canon Streamflow in the Arkansas River varies spatially City), which represents inflow from the upper basin to and temporally because of the regular timing of snow- Pueblo Reservoir, the predominant effects on stream- melt runoff in the upper basin, the irregular timing flow are due to snowmelt runoff and the release of of rainfall runoff in the lower basin, the large magni­ stored water from off-channel reservoirs. Streamflow tude of irrigation diversions, and the operations of generally increases with snowmelt runoff in April or

8,000 \ \ \ r i i r

Arkansas River at Canon City (07096000) Arkansas River above Pueblo (07099400) 6,000

4,000

Q 2,000 LU CO DC LU Q_ 0 i i i i i i \- LU 8,000 \ \ i i i i i i i i r LL O CD Arkansas River near Avondale (07109500) Arkansas River at Las Animas Q 6,000 (07124000)

£ O 4,000

LU OC 2,000 CO

i i i i i i i i i i i I o ^ 8,000 i i i ir^ i i \ i i i i i i i r

Arkansas River below John Martin Reservoir (07130500) Arkansas River at Lamar 6,000 (07133000)

4,000

2,000

i i i I ._ I____L Afbu. JFMAMJJASOND JFMAMJJASOND 1982 1982 Figure 3. Daily mean Streamflow at selected Arkansas River gaging stations, 1982.

10 Relations of Streamflow and Specific-Conductance Trends to Reservoir Operations in the Lower Arkansas River, Southeastern Colorado May and peaks in the middle or latter part of June. is a function of the evaporative concentration of The recession of snowmelt runoff usually is supple­ dissolved solids. In 1990-93, during a period of inten­ mented by off-channel reservoir releases in the upper sive data collection, the median specific conductance basin in July and August. Natural base flow and increased downstream from 276 |0,S/cm at Canon reservoir releases maintain streamflow during the City to 3,855 \iS/cm at Lamar (fig. 4). Miles (1977) low-flow period from October through March. At reported that the dissolution of soluble sedimentary station 07099400 (Arkansas River above Pueblo), materials between Canon City and Pueblo Reservoir located 0.4 mi downstream from Pueblo Dam, stream- increases specific conductance in that reach. From flow is regulated by the operation of Pueblo Reservoir. Pueblo to Lamar, specific conductance primarily The hydrograph for station 07099400 is similar to the increases because of the consumptive use of irrigation hydrograph for station 07096000, except that stream- water and the concomitant increase in the concentra­ flow is considerably smaller at station 07099400 tion of dissolved solids. The rate of increase in specific during November through March. At station 07109500 conductance is larger downstream from Catlin Dam (Arkansas River near Avondale), located about 24 mi (fig. 4) because irrigation-return flow composes a downstream from Pueblo Reservoir, streamflow is larger percentage of streamflow in this reach than it affected by Pueblo Reservoir operations and by does upstream from Catlin Dam (Cain, 1987). substantial tributary inflow from Fountain Creek and the St. Charles River. The shape of the hydrograph for station 07109500 is similar to the hydrograph for RELATIONS OF STREAMFLOW AND station 07099400, except that tributary inflow from SPECIFIC-CONDUCTANCE TRENDS occasional rainfall runoff produces distinctly larger TO RESERVOIR OPERATIONS IN peaks, and the streamflow is substantially larger at THE ARKANSAS RIVER station 07109500 during winter, owing to tributary inflow from Fountain Creek. At station 07124000 (Arkansas River at Las Animas), which is located At Canon City 110 mi downstream from Pueblo Reservoir, the magni­ tude of streamflow is substantially smaller than at the Station 07096000 (Arkansas River at Canon upstream stations. The decrease in streamflow is attrib­ City) is located in the upper Arkansas River Basin utable to irrigation diversions. As previously discussed about 28 mi upstream from Pueblo Reservoir (fig. 1). in the "Description of Study Area" section, much of Streamflow and specific-conductance trends were the streamflow in the river downstream from La Junta analyzed at this site to determine if the quantity and (fig. 1) consists of irrigation-return flow during parts of quality of water flowing from the upper basin into the most years. Downstream from John Martin Reservoir at lower basin were different during the period before the stations 07130500 (Arkansas River below John Martin completion of Pueblo Reservoir (pre-1975) compared Reservoir) and 07133000 (Arkansas River at Lamar), to the period after the completion of Pueblo Reservoir the annual hydrographs are regulated by the storage (post-1974). Streamflow and specific-conductance and release operations in John Martin Reservoir. The data were available for 1964-94. Streamflow and reservoir gates typically are closed during November specific-conductance data from 1964 through 1974 through March, and all inflow is stored; winter stream- were compared to data from 1975 through 1994. flow at stations 07130500 and 07133000 is main­ The median annual streamflow at station tained by ground-water discharge. Requests for reser­ 07096000 increased from about 520,800 acre-ft/yr voir releases of stored water by downstream irrigators in 1964-74 to about 538,600 acre-ft/yr in 1975-94; usually begin in the first 2 weeks of April and continue however, this difference in streamflow was not statisti­ episodically through October. Rainfall runoff that cally significant (p=0.92). Although there was no is generated upstream from the reservoir is attenuated significant change in the median annual streamflow, by storage in the reservoir. Streamflow between the daily mean streamflow changed substantially in John Martin Reservoir and Lamar is substantially several months of the two periods (fig. 5). Daily mean decreased by irrigation diversions. streamflow increased significantly (p<0.05) between Specific conductance in the Arkansas River 1964-74 and 1975-94 in January, February, March, markedly increases downstream from Pueblo (fig. 4). April, June, October, November, and December The downstream increase in specific conductance (table 2). Most of the increases occurred during low

RELATIONS OF STREAMFLOW AND SPECIFIC-CONDUCTANCE TRENDS TO RESERVOIR OPERATIONS 11 IN THE ARKANSAS RIVER Q. 4,500 LJJ LJJ ^ 4,000

LJJ o CC 3,500 LJJ CL

3,000

2,500

2,000 " Q LU n z < 1,500 O D Q Z o 1 ,000 o o i 500 O LJJ Q_ CO 0 300 250 200 150 100 50 RIVER MILES UPSTREAM FROM THE COLORADO-KANSAS STATE LINE Figure 4. Median specific conductance at selected sites on the Arkansas River, 1990-93.

flow when natural base flow was supplemented by increase in transmountain water is attributable to reservoir releases in the upper basin. Daily mean the Fryingpan-Arkansas Project importations, which streamflow decreased significantly in July, August, began in 1972. The increased importation of western- and September (table 2). slope water and the release of this water from storage The monthly streamflow trends probably were during otherwise low-flow months probably accounts caused by differences in the quantity of imported for the increase in streamflow at Canon City during western-slope water for the two periods of analysis. October-April. Water is released from upper basin The transmountain importation of Colorado River storage in the winter to create storage space for addi­ Basin water into the Arkansas River Basin has tional transmountain imports during the coming snow- occurred since the late 1800's. The imported water has melt runoff season. There were insufficient data for a been used to meet mining, agricultural, municipal, and historical analysis of monthly trends in the amount of industrial water needs. Some of the imported water transmountain water that has been released to the is diverted directly into the main stem of the upper Arkansas River. Arkansas River, and some of the water is diverted out Although there is a 13-year gap (1977-89) in of the upper basin via closed conduit flow. During the specific-conductance record, the data do provide 1964-74, prior to the completion of Pueblo Reservoir, some important information about the quality of the median annual volume of water that was imported water that entered the lower basin before and after into the Arkansas River Basin and released to the river the construction of Pueblo Reservoir. A visual assess­ was about 62,900 acre-ft (fig. 6). The median annual ment of the data indicates that the range of specific- volume of imported water increased significantly conductance values (130-380 fiS/cm) was relatively (p=0.01) to about 103,000 acre-ft during 1975-94 constant during 1964-94 (fig. 7). The median (fig. 6). These values do not include uie water that was specific conductance decreased about 19 percent diverted out of the basin via closed conduit flow. The from 307 ^iS/cm in 1964-74 to 250 |iS/cm in 1975-94.

12 Relations of Streamflow and Specific-Conductance Trends to Reservoir Operations in the Lower Arkansas River, Southeastern Colorado 10,000 600

7,000 620 341 330 5,000 * 341 4,000 * 620 3,000 it

O 600 600 O 2,000 330 six LU CO DC LU 620 620 Q_ -r- 330 I- 1,000 620 560 600 LLI 620 LU _ 341 330 LL 700 308 O CQ 500 341 J_ O 400

300

200

LU DC CO 100

< LU 70 o m 50 1964-74 3J < 40 m Q z o 30 1975-94 o 20 z= m3J H (/> i m m 3) 10 January February March April May June July August September October November December s Figure 5. Daily mean streamflow at station 07096000 (Arkansas River at Canon City), 1964-74 and 1975-94.

< O m z 3) (/> Table 2. Step-trend results on the daily mean streamflow at station 07096000 (Arkansas River at Canon City) between 1964-74 and 1975-94

[ft /s, cubic feet per second; N, number of values; p value is the significance level of the test; <, less than; NS, trend not statistically significant; I, statistically significant increasing trend; D, statistically significant decreasing trend]

1964-74 1975-94 Month Median Median p value Significance1 streamflow N streamflow N (ft3/s) (ft3/s) January 304 341 378 620 <0.01 I February 284 308 351 560 <01 I March 268 341 385 620 <.01 I April 268 330 369 600 <.01 I May 874 341 792 620 .16 NS June 1,800 330 2,140 600 .04 I July 1,420 341 1,150 620 <.01 D August 1,020 341 703 620 <.01 D September 380 330 369 600 .04 D October 232 341 329 620 <.01 I November 344 330 425 600 <01 I December 344 341 412 620 <01 I A statistically significant trend was defined as having a p value less than or equal to 0.05.

180,000

LLJ 160,000 LLJ DC O Median annual volume 140,000 of imported water for DC 1975-94 LLJ

120,000 Q LU DC O 100,000

LL O Median annual volume LLJ of imported water for 80,000 1964-74 ID O

60,000

< _l < 40,000

I 20,000 I I 1963 1965 1970 1975 1980 1985 1990 1995 Figure 6. Total annual volume of water imported into the Arkansas River from the Colorado River Basin, 1964-94.

14 Relations of Streamflow and Specific-Conductance Trends to Reservoir Operations in the Lower Arkansas River, Southeastern Colorado 600 I I \ \ I I I T \ I I 1 I I 1 T I I I I I 1 I I I ITT

QC LU 550 LU

500 LU O QC LJLJ 450 Q_ CO Z CO LJLJ D 400 2(0 LJLJ _l toUJ O ° 350 QC CO

300 Z O LJLJ LJj" Q O LO 250 Z

200 Q Z O 150 O O - 100 LJLJ D_ " 50

I I i i I I I I I I i I I I I I I i i I I I I I I I I I I I 1964 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 Figure 7. Specific conductance at station 07096000 (Arkansas River at Canon City), 1964-94.

Boxplots of monthly specific-conductance data for has a larger dissolved-solids concentration and 1964-74 and 1975-94 indicate that, after 1974, specific conductance than the imported water, the median specific conductance tended to decrease historically represented a substantial part of the during most low-flow months and increase during streamflow. late summer (fig. 8), thus correlating well with the monthly streamflow trends. Specific conductance decreased significantly in January, February, Above Pueblo and March (table 3), which are months having low streamflow. Although specific conductance The quality of water in the Arkansas River tended to increase during July through September above Pueblo (station 07099400) is important from a (fig. 8), the differences were not statistically drinking-water perspective because the river is the significant (table 3). municipal water supply for the Pueblo Board of Water The significant increase in streamflow Works and for the St. Charles Mesa Water District. and the significant decrease in specific conductance The diversion point for the Pueblo water supply is for January, February, and March probably were about 4 mi downstream from station 07099400 caused by an increase in the amount of stored (fig. 1), and the diversion point for the St. Charles imported water that was released to the river, Mesa water supply is about 8.5 mi downstream from upstream from Canon City, during low flow. The the station. Streamflow and specific-conductance data larger flow volume increased the dilution potential were available at station 07099400 for 29 years from of the river during a period when base flow, which 1966 through 1994.

RELATIONS OF STREAMFLOW AND SPECIFIC-CONDUCTANCE TRENDS TO RESERVOIR OPERATIONS 15 IN THE ARKANSAS RIVER o>

-o

03

b c o 1C 03 O .> 03 I

CD O> 1^O o c ,g -, "os < t*

CD O> c o o o 'o CD Q. W

00 S S 8

SniS130 9Z IV O) d3d Nl '30NV10nQNOO Oldl03dS

16 Relations of Streamflow and Specific-Conductance Trends to Reservoir Operations in the Lower Arkansas River, Southeastern Colorado Table 3. Step-trend results on specific conductance at station 07096000 (Arkansas River at Canon City) between 1964-74 and 1975-94

[uS/cm, microsiemens per centimeter at 25 degrees Celsius; N, number of values; p value is the significance level of the test; <, less than; NS, trend not statistically significant; D, statistically significant decreasing trend]

1964-74 1975-94 Median Median Month specific specific p value Significance1 N N conductance conductance (uS/cm) (uS/cm) January 312 11 255 9 0.03 D February 321 7 273 4 .02 D March 334 12 276 10 <01 D April 333 9 326 10 .33 NS May 208 8 188 14 .52 NS June 173 9 158 13 .87 NS July 161 10 208 10 .17 NS August 226 9 250 14 .06 NS September 278 8 330 6 .11 NS October 322 12 322 9 .97 NS November 311 11 280 5 .19 NS December 318 7 266 6 .06 NS A statistically significant trend was defined as having a p value less than or equal to 0.05.

The median annual streamflow at station spring and the latter parts of the growing season 07099400 increased from about 445,200 acre-ft/yr in when streamflow generally is insufficient to meet 1966-74 to about 518,400 acre-ft/yr in 1975-94; the irrigation needs. The net effect of this operation difference in the median annual streamflow was not downstream from Pueblo Reservoir is decreased statistically significant (p=0.23). Although there was streamflow in November through February and no significant change in the median annual stream- increased streamflow during early spring and late flow, there were differences in the daily mean stream- summer or autumn. flow that were segregated by month in each of the two Specific conductance at station 07099400 periods (fig. 9). Following the construction of Pueblo has changed markedly since the construction of Reservoir, the daily mean streamflow increased Pueblo Reservoir (fig. 10). The median specific significantly in March, April, June, July, September, conductance decreased significantly (p<0.01) from and October (table 4). Most of the increases occurred 625 |iS/cm in 1966-74 to 496 |iS/cm in 1975-94. during the growing season when demands for irriga­ The most obvious change in specific conductance was a narrowing of the range in specific conductance that tion water were largest. Significant decreases in occurred after 1974. Since 1974, the annual maximum streamflow occurred in January, February, May, specific conductance has tended to decrease, and November, and December (table 4). This basic the annual minimum specific conductance has tended pattern of increased spring, summer, and fall stream- to increase slightly. The annual maximum specific flow and decreased winter streamflow is largely attrib­ conductance decreased from a range of about utable to the WWSP. As discussed in the "Pueblo 800 to more than 1,000 |iS/cm in 1966-74 to a Reservoir" section, the WWSP allows irrigators to range of about 500 to 750 |iS/cm in 1975-94. store water in Pueblo Reservoir from November 15 to Specific conductance tended to decrease during March 15. This stored water usually is released to the low flow, September through April, and increase river to meet downstream irrigation needs during early during high flow, May through August (fig. 11).

RELATIONS OF STREAMFLOW AND SPECIFIC-CONDUCTANCE TRENDS TO RESERVOIR OPERATIONS 17 IN THE ARKANSAS RIVER 10,000 30 3J 9,000 < 2. 8,000 (D Q) 7,000 CO | 600 620 ° w 6,000 620 S.S- 5,000 600 270 4,000 * 279 279 3,000 2,500 600 * 620 aQ) a.3 Z 2,000 0 w O * UJ 1,500 270 600 CO DC 560 270 9 620 o 1,000 620 279 o 900 620 a. LU 800 LU 700 LL 270 279 600 i O _ 279 CD 500 ID 252 O 400 279

300 O 250 200

LU 150 DC CO 100 90 LU 80 70 1 * 60 < Q 50 40

30 25 1966-74 20 1975-94 15

10 January February March April May June July August September October November December Figure 9. Daily mean Streamflow at station 07099400 (Arkansas River above Pueblo), 1966-74 and 1975-94. Table 4. Step-trend results on the daily mean streamflow at station 07099400 (Arkansas River above Pueblo) between 1966-74 and 1975-94

[ft3/s, cubic feet per second; N, number of values; p value is the significance level of the test; <, less than; NS, trend not statistically significant; I, statistically significant increasing trend; D, statistically significant decreasing trend]

1966-74 1975-94 Month Median Median p value Significance1 streamflow N streamflow N (ft3/s) (ft3/s) January 300 279 112 620 <0.01 D February 245 252 175 560 <.01 D March 164 279 271 620 <.01 I April 237 270 418 600 <.01 I May 824 279 806 620 .03 D June 1,625 270 2,120 600 <.01 I July 1,160 279 1,460 620 <.01 I August 867 279 866 620 .16 NS September 326 270 400 600 <.01 I October 202 279 320 620 <.01 I November 283 270 201 600 <.01 D December 315 279 107 620 <.01 D 1A. statisti< ;ally significant trend was defined as having a p value less than or equal to 0.05.

1,100 1 1 i 1 > 1 « 1 1 i 1 i 1 i 1 i 1 i 1 1 " 1 ' 1 i 1 .

2] 1,000 COMPLETION OF PUEBLO RESERVOIR ^^ LJJ ! , _ ^ 900 1 . - LJJ - 0 . m rr *.. 9 _ - i§ * * ** . m -" 700 !dd LJJ CO O ^ Oco ! *% * «. DC LJJ . % O LJJ 600 * ** * ~ i o . ? . iL ~Z QLJJ 1 s * - Oft 500 - * ^ - Z ,_ s -

O - =) 400 * : ' / Q ^ Z * O O ; O 300 LT O LJJ g) 200

mn ~ i I i I i I i I i i i i i i i i i i i i i i i i i i 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 Figure 10. Specific conductance at station 07099400 (Arkansas River above Pueblo), 1966-94.

RELATIONS OF STREAMFLOW AND SPECIFIC-CONDUCTANCE TRENDS TO RESERVOIR OPERATIONS 19 IN THE ARKANSAS RIVER - I S 10

O T~ O

, § H I I g- W O)

JD

o o "t CD CD O

c o cc «

]H

£ 3 O) S33d03Q 92 IV a3I3l/MIIN30 d3d Nl '30NVIOnQNOO Oldl03dS

20 Relations of Streamflow and Specific-Conductance Trends to Reservoir Operations in the Lower Arkansas River, Southeastern Colorado These differences in specific conductance were statis­ high flow probably because of mixing of low-specific- tically significant for January, March-June, and conductance inflow from snowmelt runoff with higher September-December (table 5). On a seasonal specific-conductance water in the reservoir (fig. 11). basis, the median specific conductance increased Estimates were made of the frequency of from 460 jiS/cm in 1966-74 to 465 jiS/cm in occurrence of specific-conductance values at 1975-94; the increase was not statistically significant station 07099400 and at the diversion points for (p=0.51) during the growing season. In the winter- the domestic water supplies for the Pueblo Board storage season, the median specific conductance of Water Works and the St. Charles Mesa Water decreased significantly (p<0.01) from 720 |iS/cm in District. This information and the observed relations 1966-74 to 575 |iS/cm in 1975-94. Historically, between specific conductance and dissolved-solids before the construction of Pueblo Reservoir, specific concentration were used to estimate the percentage conductance was highest during low streamflow in the of time the dissolved-solids concentration exceeded winter. With completion of the dam, the high-specific- the SMCL for drinking water (500 mg/L). A compar­ conductance inflow (October-April) began to mix ison of specific-conductance data that have been with and be diluted in the reservoir by lower specific- conductance water derived from snowmelt runoff, collected at station 07099400 with data that have been resulting in decreased specific-conductance water collected at the diversion points for the Pueblo Board flowing out of the reservoir during low flow. A of Water Works and the St. Charles Mesa Water decrease in the specific conductance of the reservoir District indicates that specific conductance increases inflow, as indicated by the 19-percent decrease in about 3 percent per mile in that reach. An estimate the median specific conductance at Canon City, also was made, based on this relation, of the specific contributed to the decreased specific conductance at conductance at the diversion points for the two station 07099400. Additionally, before the completion domestic water supplies (fig. 12). These estimates of Pueblo Dam, the annual minimum specific conduc­ were made by increasing the specific conductance at tance occurred during high flow, May through August. station 07099400 by 12 percent for the diversion for the Following the closure of Pueblo Dam, the specific Pueblo Board of Water Works and by 27 percent for the conductance at station 07099400 increased during diversion for the St. Charles Mesa Water District.

Table 5. Step-trend results on specific conductance at station 07099400 (Arkansas River above Pueblo) between 1966-74 and 1975-94

[uS/cm, microsiemens per centimeter at 25 degrees Celsius; N, number of values; p value is the significance level of the test; <, less than; NS, trend not statistically significant; I, statistically significant increasing trend; D, statistically significant decreasing trend]

1966-74 1975-94 Median Median Month specific specific p value Significance1 conductance N conductance N (uS/cm) (uS/cm) January 673 9 574 4 0.04 D February 718 9 604 3 .14 NS March 790 8 559 11 <.01 D April 625 13 572 12 .05 D May 404 12 557 13 <01 I June 258 12 375 13 <01 I July 380 9 377 15 .86 NS August 340 9 408 12 .34 NS September 593 9 451 12 .05 D October 748 9 514 12 <.01 D November 716 10 557 4 .03 D December 710 8 549 6 .02 D A statistically significant trend was defined as having a p value less than or equal to 0.05.

RELATIONS OF STREAMFLOW AND SPECIFIC-CONDUCTANCE TRENDS TO RESERVOIR OPERATIONS 21 IN THE ARKANSAS RIVER 1,400

1966-74 1,200

1,000

^ 800

600 EQUIVALENT TO THE SECONDARY MAXIMUM CONTAMINANT LEVEL FOR DISSOLVED-SOLIDS CONCENTRATION 400 OF 500 MILLIGRAMS PER LITER (U.S. Environmental Protection Agency, 1986; Cain, 1987)

200

10 20 30 40 50 70 100

1,400

1975-94 1,200

EQUIVALENT TO THE SECONDARY MAXIMUM CONTAMINANT LEVEL 1,000 FOR DISSOLVED-SOLIDS CONCENTRATION OF 500 MILLIGRAMS PER LITER (U.S. Environmental Protection Agency, 1986; Cain, 1987)

800

600

EXAMPLE

The specific conductance of the 400 Arkansas River at the diversion for the St. Charles Mesa Water District was equal to or exceeded 718 microsiemens per centimeter about 27 percent of the time during 200 1975-94.

2 3 4 5 7 10 20 30 40 50 70 100 PERCENTAGE OF TIME INDICATED SPECIFIC CONDUCTANCE WAS EQUALED OR EXCEEDED

STATION 07099400 (ARKANSAS RIVER ABOVE PUEBLO)

____ ARKANSAS RIVER AT DIVERSION FOR PUEBLO BOARD OF WATER WORKS

.... ARKANSAS RIVER AT DIVERSION FOR ST. CHARLES MESA WATER DISTRICT Figure 12. Duration frequency of specific conductance at station 07099400 (Arkansas River above Pueblo); Arkansas River at diversion for the Pueblo Board of Water Works; and Arkansas River at diversion for the St. Charles Mesa Water District, 1966-74 and 1975-94.

22 Relations of Streamflow and Specific-Conductance Trends to Reservoir Operations in the Lower Arkansas River, Southeastern Colorado Cain (1987) determined, using regression analysis, most of the concern and focus at this station are related that a specific conductance of 718 |iS/cm at to specific conductance and the suitability of the river station 07099400 was equivalent to a dissolved- as an irrigation supply. Streamflow and specific- solids concentration of 500 mg/L. The conversion conductance data were available at station 07109500 also was assumed to be applicable at the two water- for 1969-94. supply diversion points. Therefore, the dissolved- The median annual streamflow in 1969-74 solids concentration at station 07099400 and at (623,000 acre-ft/yr) was not significantly different the two water-supply diversion points was estimated (p=0.65) from the median annual streamflow in to exceed 500 mg/L when specific conductance 1975-94 (625,200 acre-ft/yr). Similarly, the median exceeded 718 |iS/cm. During 1966-74, the estimated annual streamflow from the upper basin in 1969-74 exceedance of a specific conductance of 718 |iS/cm (535,700 acre-ft/yr) and 1975-94 (538,600 acre-ft/yr), was 28 percent at station 07099400, 45 percent at as indicated by the record at station 07096000, was the Pueblo Board of Water Works diversion, and not significantly different (p=0.69). The temporal 55 percent at the St. Charles Mesa Water District diver­ nature of streamflow, however, changed appreciably sion point. During 1975-94, after the construction of during the two periods (fig. 13). After the completion Pueblo Reservoir, the exceedance of a specific conduc­ of Pueblo Reservoir (1975), streamflow generally tance of 718 |iS/cm decreased to about 1 percent at increased during March through October and station 07099400, 4.3 percent at the Pueblo Board decreased during November through February of Water Works diversion, and 27 percent at the (fig. 13). The decreases in daily mean streamflow St. Charles Mesa Water District diversion point during November through February were all statisti­ (fig. 12). These results indicate that the chemical cally significant; streamflow increased significantly quality of the Arkansas River, in terms of specific during March, April, June, August, and October conductance and dissolved-solids concentration, (table 6). The decreased winter streamflow was has improved in the 8.5-mi reach between Pueblo caused by the storage of water in Pueblo Reservoir Reservoir and the St. Charles Mesa Water District as part of the WWSP. As previously noted in the diversion since 1975, when Pueblo Reservoir "Pueblo Reservoir" section, the median annual volume was completed. The improved quality of water is of water stored in Pueblo Reservoir during 1975-94 as attributable to two factors: (1) Decreased specific part of the WWSP was 42,200 acre-ft. The significant conductance in the upper Arkansas River, probably increase in streamflow during March, April, June, because of the increased importation of Colorado August, and October is attributable to the combined River Basin water; and (2) dilution of reservoir effects of the release of stored WWSP water from inflow having elevated specific conductance during Pueblo Reservoir and increased inflow from Fountain low flow by low-specific-conductance water in Creek. The median annual streamflow at the tributary Pueblo Reservoir. station 07106500 (Fountain Creek at Pueblo) increased from about 37,000 acre-ft/yr in 1969-74 to about 67,000 acre-ft/yr in 1975-94. The median Near Avondale daily streamflow of Fountain Creek at Pueblo during March through October increased 50 percent from The largest main-stem streamflow in the about 48 ft3/s in 1969-74 to about 72 ft3/s in 1975-94. Arkansas River occurs at station 07109500 (Arkansas This increase in streamflow likely is partly attributable River near Avondale) (fig. 1) because of substantial to increased unit runoff and increased municipal tributary inflow from the St. Charles River and wastewater discharge from the Colorado Springs Fountain Creek and because the station is upstream area. Increased unit runoff probably has resulted from from most of the large irrigation canals that divert the substantial growth and the associated paving most of the flow from the lower Arkansas River. of permeable surfaces in the greater Colorado Streamflow at station 07109500 is strongly affected by Springs area. Much of the flow in Fountain Creek Pueblo Reservoir operations because of the proximity is derived from sewered (treated wastewater effluent) of the station to Pueblo Reservoir and because of the and unsewered (lawn irrigation) wastewater from absence of substantial streamflow diversions between Colorado Springs and several smaller municipalities the reservoir and the station. In terms of water quality, in El Paso County (Edelmann and Cain, 1985).

RELATIONS OF STREAMFLOW AND SPECIFIC-CONDUCTANCE TRENDS TO RESERVOIR OPERATIONS 23 IN THE ARKANSAS RIVER on mil-

§***[ E ? a.a> r--m $ 2 T3 03

CO o>

* «D C O

if CO 03 CO c 03

o o _ o < r~

O CO 03 *- S 03 O H= 5 oT "* CO o> I £ u- sc E ? -"5 3 03 \L- -Jxxlx ^T"T^-T- ~3ra O CO oo o o o oo o o oo o o o oo o o o LO CO

24 Relations of Streamflow and Specific-Conductance Trends to Reservoir Operations in the Lower Arkansas River, Southeastern Colorado Table 6. Step-trend results on the daily mean streamflow at station 07109500 (Arkansas River near Avondale) between 1969-74 and 1975-94

[ft3/s, cubic feet per second; N, number of values; p value is the significance level of the test; <, less than; NS, trend not statistically significant; I, statistically significant increasing trend; D, statistically significant decreasing trend]

1969-74 1975-94 Month Median Median p value Significance1 streamflow N streamflow N (ft3/s) (ft3/s) January 500 186 332 620 <0.01 D February 462 168 365 560 <.01 D March 414 186 480 620 <.01 I April 452 180 636 600 <.01 I May 1,220 186 1,050 620 .73 NS June 1,950 180 2,410 600 <.01 I July 1,560 186 1,650 620 .09 NS August 993 186 1,090 620 <.01 I September 497 180 540 600 .40 NS October 386 186 472 620 .04 I November 560 180 403 600 <.01 D December 540 186 298 620 <.01 D 1 A statistically significant trend was defined as having a p value less than or equal to 0.05.

From 1969 to 1995, the discharge from the Colorado a range of about 1,000 to 1,100 |aS/cm in 1969-74 Springs wastewater-treatment plant increased from to a range of about 1,000 to 1,450 |aS/cm in 1975-94 about 25 ft3/s to about 50 ft3/s (V.L. Card, Colorado (fig. 14). Monthly specific conductance differed Springs Utilities, written commun., 1996; Edelmann between 1969-74 and 1975-94 (fig. 15), but most and Cain, 1985). Although the median daily stream- of the differences were not statistically significant flow of Fountain Creek at Pueblo during November (table 7). The small number of significant differences through February increased from about 63 ft ^/s in partly may be due to the small amount of specific- 1969-74 to about 87 ft3/s in 1975-94, the increase conductance data in 1969-74 (table 7). On a seasonal was not large enough to offset the seasonal decrease in basis, the median specific conductance in the growing streamflow of the Arkansas River at Avondale, which season increased from 580 |j,S/cm in 1969-74 to resulted from the storage of WWSP water in Pueblo 700 |j,S/cm in 1975-94; the increase was not statisti­ Reservoir. cally significant (p=0.07). In the winter-storage Specific conductance at station 07109500 had season, the median specific conductance increased less variability between the pre-Pueblo Reservoir significantly (p<0.01) from 900 |aS/cm in 1969-74 (1969-74) and post-Pueblo Reservoir (1975-94) to 1,050 |aS/cm in 1975-94. periods (fig. 14) than specific conductance at the Significant increases in specific conductance upstream station 07099400. Although the annual occurred during June, August, and December (table 7). minimum and maximum specific conductance tended The increase in the median specific conductance in to increase after 1974, the median specific conduc­ December (900 to 1,130 |J,S/cm) coincided with a tance for 1969-74 (778 |J,S/cm) was not significantly significant decrease in streamflow (table 6). The different (p=0.16) from the median specific conduc­ increased median specific conductance in June tance for 1975-94 (812 |J,S/cm). The annual minimum (338 to 455 jiS/cm) and in August (470 to 550 jiS/cm) specific conductance increased from a range of about coincided with significant increases in streamflow 200 to 400 |aS/cm in 1969-74 to a range of about for those months (table 6). This relation between 250 to 600 |aS/cm in 1975-94 (fig. 14). The annual streamflow and specific-conductance trends is unusual maximum specific conductance increased from because increased streamflow generally is expected to

RELATIONS OF STREAMFLOW AND SPECIFIC-CONDUCTANCE TRENDS TO RESERVOIR OPERATIONS 25 IN THE ARKANSAS RIVER ' I I I I I I I I I I I I I I I I I I I I I I I I I

LU 1,800 " I- ^ COMPLETION OF PUEBLO RESERVOIR LU ^^

Z 1,600 _ - LU O . DC 9 £ 1,400 * - LU ^CO - LU|JU1 1 1 J 1 ,200 . . \ COO ^ ** * * '* * gco \« * * y.LU 1,000 % - ^DC . * * * 9 » ZLU±0 ^ * -D / ** /* * Win 800 * .** ' Ooj " ..... ' . . *. & " ". ' " O ^3 600 « . Q z * * * O * * . O ' . * * * H o ^ . * ^ LL * . . * O . * % ' LU 0- 200 t - CO

0 I I I I I I I i I i I i i I i I i I I i I i I I I 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 Figure 14. Specific conductance at station 07109500 (Arkansas River near Avondale), 1969-94.

result in increased dilution and decreased specific at Pueblo) decreased from 1,500 |iS/cm in 1969-74 conductance. There are two likely causes of the to 1,380 jiS/cm in 1975-94. Although specific increased specific conductance in June and August: conductance decreased in Fountain Creek at Pueblo, (1) Mixing of low-specific-conductance and high- it remained substantially larger than specific conduc­ specific-conductance water in Pueblo Reservoir tance in the Arkansas River near Avondale. As previ­ resulted in increased specific conductance in the reser­ ously mentioned in this section, the median annual voir outflow during months that historically had the streamflow at station 07106500 increased from about lowest specific conductance, and (2) dissolved-solids 37,000 acre-ft/yr in 1969-74 to about 67,000 acre-ft/yr loads from major tributaries increased. Apparently, in 1975-94, with streamflow increasing during all both situations occurred and together caused increases seasons. The median specific conductance in Fountain in specific conductance during June and August. A Creek was converted to an equivalent dissolved-solids lack of data prevented an analysis of streamflow and concentration based on relations described by Cain specific-conductance trends in the St. Charles River (1987). The estimated median dissolved-solids (fig. 1), which is tributary to the Arkansas River concentrations were multiplied by the median annual about 4.5 mi upstream from station 07109500. streamflow to obtain an estimate of the median annual Adequate data did exist for general trend analysis dissolved-solids load contributed by Fountain Creek at Fountain Creek (fig. 1), the other major tributary to the Arkansas River in 1969-74 and 1975-94. in the reach, although there were not enough data Based on these estimates, dissolved-solids loading to do monthly trend analysis. The median specific increased about 58 percent from 53,000 to conductance at station 07106500 (Fountain Creek 84,000 tons/yr.

26 Relations of Streamflow and Specific-Conductance Trends to Reservoir Operations in the Lower Arkansas River, Southeastern Colorado 00 -fc *

. CD CO

CO O)

(A 03 (A

1

oO 10 O) o

= O

< 0 Is V)

* CD C Ic3 13 ID T3 CO r- O) O) § O

Q. CO

in T 0)

SniS133 8330930 SZ IV D> d3d SN3l/\l3ISOa3ll/\l Nl '30NV10nQNOO Oldl03dS E

RELATIONS OF STREAMFLOW AND SPECIFIC-CONDUCTANCE TRENDS TO RESERVOIR OPERATIONS 27 IN THE ARKANSAS RIVER Table 7. Step-trend results on specific conductance at station 07109500 (Arkansas River near Avondale) between 1969-74 and 1975-94 [uS/cm, microsiemens per centimeter at 25 degrees Celsius; N, number of values; p value is the significance level of the test; <, less than; NS, trend not statistically significant; I, statistically significant increasing trend]

1969-74 1975-94 Median Median Month specific specific p value Significance1 N N conductance conductance (uS/cm) (uS/cm) January 970 6 1,090 16 0.17 NS February 950 5 960 13 .66 NS March 900 4 844 20 .56 NS April 925 6 852 27 .66 NS May 600 7 734 29 .38 NS June 338 10 455 25 <.01 I July 320 5 454 21 .15 NS August 470 7 550 27 .01 I September 850 7 718 24 .60 NS October 950 7 827 17 .37 NS November 840 7 890 17 .22 NS December 900 7 1,130 18 <.01 I A statistically significant trend was defined as having a p value less than or equal to 0.05.

The duration frequencies of specific conduc­ at this station is substantially smaller than at tance at station 07109500 for 1969-74 and 1975-94 station 07109500 (fig. 3) because several large irriga­ were compared to the salinity-hazard classifications tion canals divert most of the streamflow in the for irrigated crops (fig. 16). The salinity hazard is a 96-mi reach between the stations. Irrigation-return relation developed by the U.S. Salinity Laboratory flow composes a substantial fraction of the stream- (Richards, 1954) that describes the qualitative flow at station 07124000 (Cain, 1987); therefore, effect of saline water on irrigated crops. The hazard the specific conductance is considerably higher is based on the specific conductance of the water than at upstream sites (fig. 4). Streamflow and and is divided into four classes of salinity hazard specific-conductance data were available at ranging from low (Class Cl) to very high (Class C4). station 07124000 for 1961-94. Water at station 07109500 generally was Class C2 The median annual streamflow at station or C3. Class C2 water (250-750 fAS/cm) is defined 07124000 increased significantly (p=0.01) from as having a moderate salinity hazard and can be used 77,200 acre-ft/yr in 1961-74 to 149,400 acre-ft/yr on crops having a moderate salt tolerance without in 1975-94. This difference represents, on average, needing special irrigation practices for salinity control. an increase in the daily mean streamflow of about o Class C3 (750-2,250 fiS/cm) is defined as having a 100 ft /s. The daily mean streamflow at station high salinity hazard. The changes in specific conduc­ 07124000 increased in every month after 1974 (fig. 17); tance that occurred at station 07109500 after 1974 did the differences in streamflow were significant for all not result in a substantial change in the salinity-hazard 12 months (table 8). The increased streamflow prob­ classification of the water (fig. 16). ably is attributable to a combination of factors, including the WWSP and associated changes in growing-season and nongrowing-season irrigation At Las An imas practices and the increased importation of water from the western slope for irrigation. The effects of these Station 07124000 (Arkansas River at factors are greatest at Las Animas because it is the Las Animas) is located about 120 mi downstream farthest downstream station between Pueblo Reservoir from Pueblo Reservoir (fig. 1). Streamflow and John Martin Reservoir.

28 Relations of Streamflow and Specific-Conductance Trends to Reservoir Operations in the Lower Arkansas River, Southeastern Colorado 1,500

100 1 2 3 4 5 7 10 20 30 40 50 70 100 PERCENTAGE OF TIME INDICATED SPECIFIC CONDUCTANCE WAS EQUALED OR EXCEEDED Figure 16. Duration frequency of specific conductance at station 07109500 (Arkansas River near Avondale), 1969-74 and 1975-94.

RELATIONS OF STREAMFLOW AND SPECIFIC-CONDUCTANCE TRENDS TO RESERVOIR OPERATIONS 29 IN THE ARKANSAS RIVER 3J 3J 30,000 20,000 r 420 434 10,000 620 7,000 600 620 II(D {/} o> s 5,000 600 434 4,000 434 420 3,000 620 2,000 620 " 0) 620 560 600 600 0) 3 1,000 Q. Q. 0 W O 700 392 434 O LU 500 L 434 C/) 400 420 rr 300 434 LU Oo a. 200 t) LJJ 100 LL 70 O m 50 ID 40 O 30 ± 20

33 O 10 (0 at Ql 7 9 ^ 5 4 I LJJ rr 3 1 * 2

LJJ 1 0.7 0.5 0.4 0.3 0.2

0.1 0.07 0.05 0.04 0.03 0.02

0.01 January February March April May June July August September October November December Figure 17. Daily mean Streamflow at station 07124000 (Arkansas River at Las Animas), 1961-74 and 1975-94. Table 8. Step-trend results on the daily mean streamflow at station 07124000 (Arkansas River at Las Animas) between 1961-74 and 1975-94

[ft3/s, cubic feet per second; N, number of values; p value is the significance level of the test; <, less than; I, statistically significant increasing trend]

1961-74 1975-94 Month Median Median p value Significance1 streamflow N streamflow N (ft3/s) (ft3/s) January 36 434 135 620 <0.01 I February 36 392 130 560 <.01 I March 24 434 79 620 <.01 I April 20 420 33 600 <01 I May 28 434 108 620 <.01 I June 250 420 484 600 <.01 I July 168 434 272 620 <.01 I August 63 434 126 620 <.01 I September 32 420 59 600 <.01 I October 29 434 74 620 <.01 I November 28 420 69 600 <.01 I December 30 434 130 620 <.01 I JA statistically significant trend was defined as having a p value less than or equal to 0.05.

Prior to the WWSP, many irrigators diverted 30 mi upstream from Las Animas, receives a substan­ winter streamflow and applied the water to barren tial percentage of the WWSP water and the imported fields in order to maintain soil moisture. Because of Project water during the irrigation season. Because of this practice and the already seasonally low stream- its proximity to the Fort Lyon Canal, the Las Animas flow, most winter streamflow was diverted and site probably benefits from increased irrigation-return consumed upstream from Las Animas. After the flow to the river from land irrigated by the Fort Lyon WWSP began, winter streamflow was stored in Pueblo Canal. On average, the Fort Lyon Canal received Reservoir, John Martin Reservoir, and in several small about 57,400 acre-ft/yr of WWSP water and about off-channel reservoirs. Releases of stored WWSP 10,400 acre-ft/yr of Project water since 1975 water to downstream irrigators generally were made at (Thomas C. Simpson, Southeastern Colorado Water high rates during the irrigation season in order to mini­ Conservancy District, written commun., 1997). mize transit losses. Therefore, a larger percentage of The relation of annual streamflow at Avondale, winter water, which was released from reservoir where streamflow in the lower Arkansas River Basin storage, probably flows to downstream irrigation is largest, to annual streamflow at four downstream canals, thereby increasing the volume of applied water sites (fig. 18) illustrates the effects of the WWSP and the associated volume of irrigation-return flow to and transmountain imports on streamflow. Double- the river. Additionally, as part of the WWSP and the mass curves (fig. 18) are plots of cumulative values 1980 operating plan for John Martin Reservoir, three of one variable compared to cumulative values of large irrigation-canal companies have been allowed to another variable. The theory of the double-mass store winter water in John Martin Reservoir as an curve is that a graph of the cumulation of one quantity alternative to storage in Pueblo Reservoir or in private compared to the cumulation of another quantity off-channel reservoirs. This water now flows by during the same time period will plot as a straight Las Animas, whereas prior to the WWSP, it may have line as long as the data are proportional; the slope of been diverted and consumptively used. The Fort Lyon the line will represent the constant of proportionality Canal, which diverts streamflow from the river about between the quantities (Searcy and Hardman, 1960).

RELATIONS OF STREAMFLOW AND SPECIFIC-CONDUCTANCE TRENDS TO RESERVOIR OPERATIONS 31 IN THE ARKANSAS RIVER 20,000,000

18,000,000

16,000,000

14,000,000

12,000,000 LU

cc. LU 10,000,000 W LL

3 o 8,000,000 z <

LU 6,000,000

=> 4,000,000

O 2,000,000

0 10,000,000 20,000,000 0 10,000,000 20,000,000 CUMULATIVE ANNUAL STREAMFLOW AT NEPESTA, CUMULATIVE ANNUAL STREAMFLOW AT CATLIN DAM, I NACRE FEET PER YEAR IN ACRE-FEET PER YEAR

20,000,000

5,000,000 10,000,000 5,000,000 10,000,000 CUMULATIVE ANNUAL STREAMFLOW AT LA JUNTA, CUMULATIVE ANNUAL STREAMFLOW AT LAS ANIMAS, IN ACRE-FEET PER YEAR IN ACRE-FEET PER YEAR EXPLANATION _ _ RELATION BETWEEN CUMULATIVE ANNUAL RELATION BETWEEN CUMULATIVE ANNUAL STREAMFLOW AT UPSTREAM AND STREAMFLOW AT UPSTREAM AND DOWNSTREAM STATIONS FROM DOWNSTREAM STATIONS FROM 1966 THROUGH 1978 1979 THROUGH 1994

Figure 18. Relation of cumulative annual streamflow at station 07109500 (Arkansas River near Avondale) to cumulative annual streamflow at four downstream sites, 1966-94.

32 Relations of Streamflow and Specific-Conductance Trends to Reservoir Operations in the Lower Arkansas River, Southeastern Colorado A break in the slope of the double-mass curve means that flowed from Avondale past Nepesta, Catlin that a change has occurred in the constant of propor­ Dam, La Junta, and Las Animas was 75 percent, tionality. The double-mass curves in this analysis 72 percent, 18 percent, and 14 percent, respectively. (fig. 18) indicate that the accumulative rate of stream- From 1979-94, the percentage of annual streamflow flow past the gaging stations downstream from that flowed from Avondale past Nepesta, Catlin Dam, Avondale changed substantially after 1978 and that La Junta, and Las Animas was 78 percent, 78 percent, the degree of change increased downstream. A larger 27 percent, and 26 percent, respectively. percentage of annual streamflow at Avondale flowed Although the range in specific-conductance past downstream sites after 1978 than before. The values at station 07124000 changed little after 1974 (fig. 19), the median specific conductance decreased change probably occurred after 1978 because the significantly (p<0.01) from 3,000 LiS/cm in 1961-74 WWSP only had operated for 2 years prior to 1977, to 2,500 LiS/cm in 1975-94. Specific-conductance when its operation was curtailed for 1 year until 1978. values were smaller in all months during 1975-94 The effects of transmountain imports also probably (fig. 20), but the only statistically significant differ­ had some effect on the change in the constants of ences in specific conductance occurred in March, proportionality that occurred in 1979. On average April, and June (table 9), despite the significant from 1966-78, the percentage of annual streamflow increases in streamflow during each month (table 8).

6,000

5,500 Completion of Pueblo Reservoir DC LJJ LJJ 5,000

LJJ O 4,500 DC LJJ Q_ ».* 4,000 CO LJJ ^CO LJJ _J CO.LU 3,500 0° DC CO v O LJJ 3,000 ZO LJJ LJJ Q O 10 2,500 Z ^ < h- I- < O 2,000

O O 1,500 O LL o LJJ Q_ 1,000 CO

500

1960 1962 1964 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 Figure 19. Specific conductance at station 07124000 (Arkansas River at Las Animas), 1961-94.

RELATIONS OF STREAMFLOW AND SPECIFIC-CONDUCTANCE TRENDS TO RESERVOIR OPERATIONS 33 IN THE ARKANSAS RIVER 3D 3D 5,000 < 2. (D CB

12 17 26 30 13 14 4,500 17 #

20 13 30 rr 16 19 18 16 25 16 16 LU 4,000 18 UJ 20 3 13 Q. Q. 13 r1-! 19 25 0

LU O 3,500 DC LU 29 Q. §« ^ 7H 3,000

§ffi °- LU 2,500

i= LU *

Z £! 2,000 £< O=) T Q O 1,500 O O LL o 1 1 a! 1,000 I CO

1961-74 I 500 1975-94

January February March April May June July August September October November December Figure 20. Monthly specific conductance at station 07124000 (Arkansas River at Las Animas), 1961-74 and 1975-94. On a seasonal basis, the median specific conduc­ an irrigated-agriculture perspective, but water quality tance during the growing season decreased signifi­ also is important from a domestic water-supply cantly (p<0.01) from 2,820 ^iS/cm in 1961-74 to perspective because the city of Lamar diverts water 2,250 ^iS/cm in 1975-94. Similarly, during the winter- from the river to provide artificial recharge to the storage season, the median specific conductance in the alluvial aquifer in the vicinity of its municipal well growing season decreased significantly (p=0.02) from field. Streamflow and specific-conductance data were 3,300 ^iS/cm in 1961-74 to 2,730 ^iS/cm in 1975-94. available at station 07130500 for 1955-94. The decrease in specific conductance at Las Animas The median annual streamflow at did not change the salinity-hazard classification of the station 07130500 increased significantly (p<0.01) water. The water retained the same C3 (high salinity from about 142,000 acre-ft/yr during 1955-79 to hazard) to C4 (very high salinity hazard) classifica­ about 231,400 acre-ft/yr during 1980-94. The tions that were common to the period prior to the median annual streamflow that entered the lower construction of Pueblo Reservoir (Richards, 1954). basin from the upper basin, as indicated by the record at station 07096000, increased an insignificant amount Below John Martin Reservoir (p=0.20) from 481,000 acre-ft/yr in 1955-79 to about 574,400 acre-ft/yr in 1980-94. The increased median Station 07130500 (Arkansas River below annual streamflow at station 07130500 probably John Martin Reservoir) is located 0.2 mi downstream is attributable to the combined effects of the 1980 from John Martin Reservoir (fig. 1). A substantial reservoir operating plan, the decreased diversion change in reservoir operations occurred with the adop­ and consumptive use of winter streamflow upstream tion of the 1980 operating plan, as discussed in the from the reservoir, increased irrigation-return "John Martin Reservoir" section; therefore, stream- flows resulting from irrigation with WWSP and flow and specific-conductance trends were evaluated Project water, and the storage of winter water in for changes that might have occurred after 1979. John Martin Reservoir by three canal companies. The Water quality at station 07130500 is important from effect of these factors is evidenced by a large increase

Table 9. Step-trend results on specific conductance at station 07124000 (Arkansas River at Las Animas) between 1961-74 and 1975-94

[uS/cm, microsiemens per centimeter at 25 degrees Celsius; N, number of values; p value is the significance level of the test; NS, trend not statistically significant; D, statistically significant decreasing trend]

1961-74 1975-94 Median Median Month specific specific p value Significance1 N N conductance conductance (uS/cm) (uS/cm) January 3,000 13 2,240 16 0.51 NS February 3,300 13 2,790 18 .26 NS March 3,650 12 3,100 19 .01 D April 3,750 17 3,410 30 .04 D May 3,420 20 2,850 26 .26 NS June 1,500 19 1,200 29 .05 D July 1,440 20 1,260 30 .95 NS August 2,300 18 1,630 25 .22 NS September 2,920 16 2,460 25 .18 NS October 2,560 13 2,430 17 .80 NS November 3,500 13 3,110 16 .21 NS December 3,280 14 2,655 16 .22 NS 1A statistically significant trend was defined as having a p value less than or equal to 0.05.

RELATIONS OF STREAMFLOW AND SPECIFIC-CONDUCTANCE TRENDS TO RESERVOIR OPERATIONS 35 IN THE ARKANSAS RIVER in the total annual streamflow in the Arkansas River March tended to decrease or remain constant (fig. 22). about 3 mi upstream from John Martin Reservoir All increases in daily mean streamflow, except for at station 07124000. The median annual streamflow April, were statistically significant (table 10). The at station 07124000 increased 142 percent from increase in streamflow during the growing season was 76,400 acre-ft/yr in 1955-79 to 185,200 acre-ft/yr a function of the increased availability of water in the in 1980-94. Additionally, as discussed in the reservoir and in the changes in reservoir-operating "John Martin Reservoir" section, 35 percent of the practices. Reservoir storage that was previously winter water stored in the reservoir by three canal released in the spring was released throughout the companies was shifted to Arkansas River Compact use summer. Although streamflow tended to decrease and was subject to downstream release. Prior to 1980, significantly during winter, the winter streamflow after the winter-storage period, reservoir storage generally was so small that the decreases were rela­ usually was drawn down to empty or almost empty tively inconsequential (table 10). very early in the irrigation season, often by the middle Specific conductance at station 07130500 of April (fig. 21). From 1955 through 1979, reservoir changed markedly after the implementation of storage .was completely depleted by April 30 in 15 of the John Martin Reservoir 1980 operating plan the 25 years. Reservoir storage increased substantially (fig. 23). The median specific conductance in all months after 1979 (fig. 21). decreased significantly (p<0.01) from 2,700 )J,S/cm Daily mean streamflow at station 07130500 in 1955-79 to 2,260 |iS/cm in 1980-94. The most tended to increase during April-October after 1979, obvious change in specific conductance was a whereas daily mean streamflow in November through narrowing of the range in values after 1979.

140,000 n i i r i i i i r 1771

120,000 1955-79 LJJ LJJ Li. 1980-94 ^j 100,000 EC O v\ & 80,000 H LJJ Z O 60,000 EC O EC LJJ 40,000 CO LJJ EC

20,000

JAN FEB MAR APR MAY JUNE JULY AUG SEPT OCT NOV DEC Figure 21. Median end-of-month contents of John Martin Reservoir, 1955-79 and 1980-94.

36 Relations of Streamflow and Specific-Conductance Trends to Reservoir Operations in the Lower Arkansas River, Southeastern Colorado 10,000 7,000

5,000 775 4,000 465 450 750 450 3,000 775 2,000 465 465 465 750 775 750 450 I -i- 775 1,000 700 I 700 O 775 o o 500 r 775 775 Tl 111 CO 400 750 DC 300 450 111 O. 200 I- ULJ 111 LL 100 O CD 70 ^ O 50 40 30 * o 20 *

465 Ill 10 O DC 7 CO 5 O 8 * 4 o O * 111 3 2 o O O >- -1- o o < Q 1 - -L * -L- O O 27 m33 0.7 zH m 0.5 m 33 0.4 I 1955-79 Si 0.3 s 0.2 M 1980-94

30 d 0.1 < o January February March April May June July August September October November December m z 33 (/> Figure 22. Daily mean streamflow at station 07130500 (Arkansas River below John Martin Reservoir), 1955-79 and 1980-94. Table 10. Step-trend results on the daily mean streamf low at station 07130500 (Arkansas River below John Martin Reservoir) between 1955-79 and 1980-94

[ft3/s, cubic feet per second; N, number of values; p value is the significance level of the test; <, less than; NS, trend not statistically significant; I, statisti­ cally significant increasing trend; D, statistically significant decreasing trend]

1955-79 1980-94 Month Median Median p value Significance1 streamflow N streamflow N (ft3/s) (ft3/s) January 3 775 2 465 <0.01 D February 3 700 3 420 <.01 D March 4 775 3 465 <.01 D April 82 750 460 450 .61 NS May 159 775 470 465 <.01 I June 415 750 566 450 <.01 I July 490 775 878 465 <.01 I August 385 775 514 465 <.01 I September 90 750 348 450 <.01 I October 78 775 232 465 <.01 I November 15 750 3 450 <.01 D December 3 775 2 465 <.01 D *A statistically significant trend was defined as having a p value less than or equal to 0.05.

This change is very similar to the change in specific conductance in the reservoir inflow, as indicated by conductance that occurred at station 07099400 after the the specific-conductance trends at station 07124000; construction of Pueblo Reservoir (fig. 10). The annual (2) the establishment of long-term storage and a maximum specific conductance at station 07130500 permanent pool in John Martin Reservoir; and decreased from a range of about 4,500 to 5,000 |O.S/cm (3) an increase in the mixing of water with different in 1955-79 to a range of about 2,500 to 3,800 ^iS/crn specific-conductance values in the reservoir. The in 1980-94 (fig. 23). The annual minimum specific establishment of the 10,000-acre-ft permanent pool conductance increased from a range of about 500 to and the implementation of the 1980 operating plan 1,000 j^S/cm in 1955-79 to a range of about 1,000 eliminated the complete drawdown of reservoir to 1,500 j^S/cm in 1980-94 (fig. 23). Specific storage, which frequently occurred prior to 1980. conductance generally decreased after 1979 in all The increase in storage resulted in an increase in months, except June (fig. 24); however, the only the mixing of low- and high-specific-conductance statistically significant decreases were in September- water in the reservoir and narrowed the range of April (table 11). Seasonally, the median specific conductance during the growing season decreased specific conductance in the reservoir outflow. The significantly (p=0.02) from 2,180 |iS/cm in net result of these conditions was increased specific 1955-79 to 2,050 (^S/cm in 1980-94. Similarly, conductance during times when specific conductance the median specific conductance in the winter- historically was lowest and decreased specific conduc­ storage season decreased significantly (p<0.01) tance during times when specific conductance histori­ from 3,650 |iS/cm in 1955-79 to 2,640 j^S/cm in cally was highest. The 1980 operating plan and the 1980-94. 10,000-acre-ft permanent pool might have changed The changes in specific conductance the timing and magnitude of the annual minimum that occurred at station 07130500 since 1979 specific conductance at station 07130500. The annual probably were caused by: (1) Decreased specific minimum specific conductance, prior to 1980,

38 Relations of Streamflow and Specific-Conductance Trends to Reservoir Operations in the Lower Arkansas River, Southeastern Colorado generally occurred in June or July and was associated lowest specific-conductance water. Dannie McMillan with snowmelt runoff from the upper basin. Prior (city of Lamar, oral commun., 1996) reported that the to 1980, reservoir storage generally was depleted dissolved-solids concentration in the municipal water by the time of snowmelt runoff, and the low-specific- supply for Lamar has increased since 1980, although conductance water did not mix with the elevated there are no data to quantify the change. The change specific-conductance water in the reservoir. After in the timing of the minimum specific conductance at 1980 and with a permanent pool in the reservoir, station 07130500 may have affected the quality of the low-specific-conductance runoff mixed with the ground water that the city of Lamar uses as its munic­ contents of the reservoir, thereby causing the annual ipal water supply. minimum specific conductance to occur later in the The changes in specific conductance that summer. occurred at station 07130500 did not change The city of Lamar typically has diverted the salinity-hazard classification of the water about 2,000 acre-ft/yr from the river to provide for irrigated agriculture. Water at station 07130500 additional recharge to the alluvial aquifer where was classified as Class C3 (high salinity hazard) and its municipal well field is located. Water generally Class C4 (very high salinity hazard) before and after has been diverted during June in order to obtain the 1980.

6,000

5,500 DC LJJ Implementation of John Martin Reservoir 1980 operating plan LJJ 5,000 h- z LJJ O 4,500 t . DC - LJJ Q_ I* . gf ** */ ** ^ *'«... 8 f CO ,n 4,000 r *.* *.:* * * v*«

[±! LJJ 3,500 CO O Oco DC LU 52 LJJ 3,000 « * . . :>g * * " * * **» »*:*.». " -. .:/ Z LJJ *» ^ *».* *** * / ' t LU Q 2,500 . . . % « ; , f . \ t "* «:%** * S.»** : " 2,000 *$* ! O Q . .. v \ * * O 1,500 : : %- - i : . . O >:* ' t i * * ?i---.-:.:....* O »«». « *»*» ,t*»»« '.-* t 1,000 . : t * O LU . I Q_ W 500

1955 1960 1965 1970 1975 1980 1985 1990 1995 Figure 23. Specific conductance at station 07130500 (Arkansas River below John Martin Reservoir), 1955-94.

RELATIONS OF STREAMFLOW AND SPECIFIC-CONDUCTANCE TRENDS TO RESERVOIR OPERATIONS 39 IN THE ARKANSAS RIVER 30 3D 5,500

71 82 5,000 II 36 8 J2 54 95 CD Q) 33 43 39 * 42 DC 4,500 82 70 O o LU ° Q) LU 53 Q. Q. 0 101 4,000 LU O * 9 DC o LU o Q_ 10 Q. 3,500 C 10 §£3,000yj LU I 12 T 13 I JC LU Q. O LU 12 13 (0 i DC T O J o 3D ^ g 2,500 "J in I O CM I I 10 * O 2,000 =) Q * * O O I 1,500 O I

O I I LU Q_ CO 1,000 1955-79

500 1980-94

January February March April May June July August September October November December

Figure 24. Monthly specific conductance at station 07130500 (Arkansas River below John Martin Reservoir), 1955-79 and 1980-94. Table 11. Step-trend results on specific conductance at station 07130500 (Arkansas River below John Martin Reservoir) between 1955-79 and 1980-94

[|0,S/cm, microsiemens per centimeter at 25 degrees Celsius; N, number of values; p value is the significance level of the test; <, less than; NS, trend not statistically significant; D, statistically significant decreasing trend]

1955-79 1980-94 Median Median Month specific specific p value Significance1 N N conductance conductance (uS/cm) (|iS/cm) January 3,910 43 2,780 10 <0.01 D February 3,890 36 2,840 10 <.01 D March 3,550 39 2,720 12 <01 D April 3,300 71 2,620 14 <.01 D May 2,720 82 2,400 13 .30 NS June 1,690 82 2,050 12 .10 NS July 1,650 101 1,550 13 .94 NS August 1,860 95 1,560 10 .21 NS September 2,220 70 1,530 11 .04 D October 2,700 53 1,880 11 .02 D November 3,400 54 2,340 10 <.01 D December 3,800 42 2,590 11 <.01 D 1A statistically significant trend was defined as having a p value less than or equal to 0.05.

At Lamar The large increase in the median daily Streamflow in July (52-411 ft3/s) (table 12) is a function of Station 07133000 (Arkansas River at Lamar) the John Martin Reservoir 1980 operating plan is located about 20 mi downstream from John Martin and the release of stored water for downstream Reservoir. Water quality at station 07133000 is delivery to irrigators in Kansas. Prior to the 1980 important from an agricultural perspective because operating plan, Colorado and Kansas irrigators several large irrigation canals divert water from the downstream from John Martin Reservoir generally Arkansas River in the reach between station 07133000 received their shares of water very early in the and the Colorado-Kansas State line. Streamflow irrigation season. After the 1980 operating plan and specific-conductance data were available at was implemented, Kansas irrigators could delay station 07133000 for 1964-94. the delivery of their water until later in the irriga­ The median annual Streamflow at station tion season. 07133000 increased significantly (p<0.01) from about Specific conductance at station 07133000 39,100 acre-ft/yrin 1964-79 to about 67,100 acre-ft/yr (fig. 26), like specific conductance at station in 1980-94. The daily mean Streamflow at station 07130500 (fig. 23), changed substantially after 1979. 07133000 generally increased in all months after The median specific conductance at station 07133000 1979, except April (fig. 25). Although 11 of the decreased about 12 percent from 4,000 |LiS/cm in 12 increases were significant at a 95-percent confi­ 1964-79 to 3,510 jiS/cm in 1980-94. The annual dence level, the magnitude of the increases generally maximum specific conductance decreased from a was small (table 12). The increased Streamflow was range of about 5,500 to 8,000 jiS/cm in 1964-79 to largely attributable to increased inflow to John Martin a range of about 4,500 to 5,000 jiS/cm in 1980-94 Reservoir, as discussed in the "At Las Animas" (fig. 26). The annual minimum specific conductance and "Below John Martin Reservoir" sections. increased from a range of about 500 to 1,500 |LiS/cm

RELATIONS OF STREAMFLOW AND SPECIFIC-CONDUCTANCE TRENDS TO RESERVOIR OPERATIONS 41 IN THE ARKANSAS RIVER 10,000 7,000 480

5,000 496 4,000 465 496 480 450 450 3,000

2,000 496 465 465 1,000 450 480 465 o 700 o 496 111 500 CO 400 en 496 in 300 o_ 496 200 LU u_LU o 100 CO 70 Z) 465 420 o 50 40 30 O 20

111 tr

LU S >-

< Q 1 0.7 0.5 0.4 1964-79 0.3 1980-94 0.2 D

0.1 January February March April May June July August September October November December

Figure 25. Daily mean streamflow at station 07133000 (Arkansas River at Lamar), 1964-79 and 1980-94.

in 1964-79 to a range of about 700 to 2,000 jiS/cm and trends in specific conductance at station 07133000 in 1980-94 (fig. 26). Specific conductance generally were very similar to the variations and trends at decreased or remained relatively constant after 1979 station 07130500 and were attributable to decreased in all months, except for June (fig. 27). The decreases specific conductance in the inflow to John Martin in specific conductance were statistically significant Reservoir, the establishment of long-term storage for February through April and September through and a permanent pool in John Martin Reservoir, and December (table 13). Seasonally, the median specific the mixing of low- and high-specific-conductance conductance in the growing season decreased signifi­ water in the reservoir. The changes in specific conduc­ cantly (p=0.01) from 3,400 jiS/cm in 1964-79 to tance that occurred at station 07133000 did not change 2,995 jiS/cm in 1980-94. Similarly, the median the salinity-hazard classification of the water for irri­ specific conductance in the winter-storage season gated agriculture. Water at station 07133000 was clas­ decreased significantly (p<0.01) from 4,900 jiS/cm in sified as Class C3 (high salinity hazard) and Class C4 1964-79 to 4,375 jiS/cm in 1980-94. The variations (very high salinity hazard).

42 Relations of Streamflow and Specific-Conductance Trends to Reservoir Operations in the Lower Arkansas River, Southeastern Colorado Table 12. Step-trend results on the daily mean streamflow at station 07133000 (Arkansas River at Lamar) between 1964-79 and 1980-94

[ft3/s, cubic feet per second; N, number of observations; p value is the significance level of the test; <, less than; NS, trend not statistically significant; I, statistically significant increasing trend]

1964-79 1980-94 Month Median Median p value Significance1 streamflow N streamflow N (ft3/s) (f^/s) January 10 496 23 465 <0.01 I February 10 448 22 420 <.01 I March 6 496 20 465 <.01 I April 33 480 33 450 .81 NS May 22 496 28 465 .04 I June 56 480 78 450 <01 I July 52 496 411 465 <0l I August 32 496 62 465 <01 I September 10 480 21 450 <.01 I October 7 496 15 465 <.01 I November 6 480 26 450 <01 I December 12 496 27 465 <01 I !A statistically significant trend was defined as having a p value less than or equal to 0.05.

8,500 1 i ' i i i ' i i i i i i i i

DC 8,000 LJJ ^^, Implementation of John Martin LJJ ^^"^ Reservoir 1980 operating plan

^ 7,000 - - LJJ O - DC LJJ CL 6,000 . 4

§CO . ;

* * .

roco*»en §oo o . DEGREESATCELS125 ' . ONDUCTANCE,MICROSIEMIN *.< * :* < - . .. * . :* * ' .- - >. : v o . . * * ^ ' * f \ " , ' Y -

* * - . t * : « » * * * * - . . : :. . : \ t v - ' . .- . 0 . . . v* * " . j : O " LL ' . * \ \ LJJ 1 '00° - »*» CL CO

n 1 1 1 1 1 1 1 1 I 1 I I I 1 I I I I I ! I 1964 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 Figure 26. Specific conductance at station 07133000 (Arkansas River at Lamar), 1964-94.

RELATIONS OF STREAMFLOW AND SPECIFIC-CONDUCTANCE TRENDS TO RESERVOIR OPERATIONS 43 IN THE ARKANSAS RIVER i i i i i i i i i i e HZ] I o

T3 C CO

CO ^ E = CO

I o o o CO CO s I CO

o 5 * EH o - 'o 4 CD CD 00 CL O) O) DH o £ g c ca O

8 8 CM N. co in -^ co I SniS130 S33H03Q 9ZIV H3131AII1N30 0) H3d SN31AI3ISOHOIIAI Nl '30NV10nQNOO Oldl03dS il

44 Relations of Streamflow and Specific-Conductance Trends to Reservoir Operations in the Lower Arkansas River, Southeastern Colorado Table 13. Step-trend results on specific conductance at station 07133000 (Arkansas River at Lamar) between 1964-79 and 1980-94

[(iS/cm, microsiemens per centimeter at 25 degrees Celsius; N, number of values; p value is the significance level of the test; <, less than; NS, trend not statistically significant; D, statistically significant decreasing trend]

1964-79 1980-94 Median Median Month specific specific p value Significance1 N N conductance conductance (nS/cm) (\iS/cm) January 4,640 12 4,350 9 0.28 NS February 5,000 17 4,500 10 .05 D March 4,980 16 4,480 10 .01 D April 3,900 25 3,350 13 .04 D May 3,800 26 3,780 11 .62 NS June 2,200 27 2,500 17 .35 NS July 1,900 23 1,890 13 .69 NS August 3,000 21 2,850 11 .32 NS September 4,000 18 3,320 13 .01 D October 4,540 14 3,520 10 <.01 D November 4,540 18 3,880 10 .01 D December 4,500 11 4,260 10 <.01 D 1 A statistically significant trend was defined as having a p value less than or equal to 0.05.

SUMMARY River water onto fallow fields to maintain soil mois­ ture. Alternatively, this water could have been stored In the lower Arkansas River, the specific during the winter and then released to the river for conductance increases downstream from a median the downstream irrigators to use during times when of about 500 ]iS/cm near Pueblo to about 3,900 jiS/cm streamflow was insufficient to meet irrigation needs. at Lamar. The increase is largely attributed to the However, under Colorado water law, storage of water consumptive use of surface water and ground water that is diverted with direct-flow water rights is not for agricultural irrigation and the concomitant increase permitted. Therefore, the WWSP was created to allow in the dissolved-solids concentration. The operations several irrigation canal companies downstream from of two main-stem reservoirs on the lower Arkansas Pueblo Reservoir to store their direct-flow water in River (John Martin Reservoir, constructed near Pueblo Reservoir, John Martin Reservoir, and in Las Animas in 1948, and Pueblo Reservoir, several private off-channel reservoirs during winter constructed near Pueblo in 1975) have the potential and to use this water during the crop-growing season. to alter the specific conductance in the Arkansas River Under the WWSP, winter water storage is allowed by streamflow management. A change in specific from November 15 to March 15. Generally, WWSP conductance could affect the intended use of the water water is released from storage at times when stream- as an agricultural or domestic water supply. flow is not large enough to meet irrigation demands. The most notable aspect of the operation of This situation usually occurs in early spring or Pueblo Reservoir, in terms of its effect on the historic late summer and autumn. Winter water was stored streamflow regime of the lower Arkansas River, is the every year from 1975 to 1994, except during the Winter Water Storage Program (WWSP). In the winter 1977-78 winter-storage season. During 1975-94, (November-March), prior to the implementation of the median annual volume of water that was stored the WWSP in 1975, irrigators in the lower Arkansas in Pueblo Reservoir as part of the WWSP was about River Valley generally diverted appropriated Arkansas 42,200 acre-ft.

SUMMARY 45 Storage of irrigation water in John Martin record varied between stations; therefore, trend Reservoir is by agreement between the States of test results for different stations were not directly Colorado and Kansas, under the terms of the Arkansas compared. Data from the station in the upper basin River Compact. The Arkansas River Compact is an and from the three stations located between Pueblo agreement between Colorado and Kansas, signed in Reservoir and John Martin Reservoir were analyzed 1948, which ensures both States will receive their for trends that may have occurred after 1974, which percentage share of Arkansas River flows. Provisions corresponds to the completion of Pueblo Reservoir in were made in the Compact for the rate of release of 1975. Data from the two stations located downstream stored water, without reference to the volume of stored from John Martin Reservoir were analyzed for trends water assigned to each State. To ensure that each State that may have occurred after the implementation of a received its share of stored water, release demands had new reservoir operating plan in 1980. to be made concurrently. Historically, following the At the station (07096000) in the upper basin, winter-storage period (November-March), reservoir streamflow increased significantly and specific storage usually was drawn down to empty or almost conductance decreased significantly after 1974 during empty very early in the irrigation season, often by the the low-flow months, January, February, and March. middle of April. Because of the unsatisfactory nature These trends apparently were caused by the increased of this operation, a resolution was adopted by the importation of low-specific-conductance water from Arkansas River Compact Administration in 1980. This the Colorado River Basin into the Arkansas River. resolution is commonly referred to as the 1980 oper­ The median volume of water imported from the ating plan. Under the new plan, any water not immedi­ Colorado River Basin into the upper Arkansas River ately called for and released to downstream irrigators increased from about 62,900 acre-ft/yr in 1964 74 to is stored in separate storage accounts for the States of 103,000 acre-ft/yr in 1975-94. This transmountain Colorado and Kansas. Either State can call for the water generally is held in storage in upper basin reser­ release of its stored water independent of the other. voirs as long as possible in order to minimize evapora­ The 1980 operating plan has contributed to increased tive losses. However, during winter, stored water may long-term storage of water in John Martin Reservoir. be released to the river and possibly stored farther Prior to 1980, reservoir storage generally was depleted downstream in Pueblo Reservoir in order to create by the end of April. Since 1980, reservoir storage has upper basin storage space for the importation of trans­ increased substantially. mountain water during the coming snowmelt-runoff Streamflow and specific-conductance data that season. The imported water released to the river in the were collected at six main-stem Arkansas River sites winter tended to dilute the dissolved-solids concentra­ were evaluated with a step-trend analysis to determine tion of the more mineralized base flow and decreased if the operation of Pueblo Reservoir or John Martin specific conductance at station 07096000. Overall, Reservoir affected streamflow or specific conductance the median specific conductance at station 07096000 in the Arkansas River. The nonparametric Mann- decreased about 19 percent from 307 |iS/cm in Whitney-Wilcoxon rank-sum test was used for trend 1964-74 to about 250 ^iS/cm in 1975-94. analysis. The analysis of streamflow trends was done At the three stations located between Pueblo because streamflow and specific conductance gener­ Reservoir and John Martin Reservoir, streamflow ally are correlated; therefore, changes or trends in and specific conductance primarily were affected by specific conductance often can be explained in terms Pueblo Reservoir operations, but changes in the quan­ of the associated change or trend in streamflow. Data tity and quality of inflow from the upper basin and collected at five streamflow-gaging stations on the from Fountain Creek may have had some effect. At lower Arkansas River and at one station on the upper stations 07099400 and 07109500, which are located Arkansas River were analyzed for trends. The station 0.4 and 24 mi downstream from Pueblo Reservoir, in the upper basin was included in the analysis to streamflow generally increased during most months of differentiate between trends in the lower basin that the growing season and decreased during November were caused by differences in the quantity or quality through February after 1974. The streamflow trends at of inflow from the upper basin or were caused by these two stations largely were attributed to the opera­ reservoir operations in the lower basin. The period of tion of the WWSP.

46 Relations of Streamflow and Specific-Conductance Trends to Reservoir Operations in the Lower Arkansas River, Southeastern Colorado Specific conductance at station 07099400 basin flowed past station 07124000. After 1978, changed markedly after the construction of Pueblo that percentage increased to about 26 percent. Reservoir. Specific conductance at station 07099400 Although specific conductance tended to decrease decreased during most months between September at station 07124000 after 1974, the water continued and April and increased during the high-flow months, to have a high to very high salinity hazard classifica­ May through August. This trend was caused by the tion for irrigated agriculture. mixing of seasonally low-specific-conductance water At the two stations located downstream from with seasonally high-specific-conductance water in John Martin Reservoir, specific conductance was Pueblo Reservoir, thus narrowing the annual range affected by changes in John Martin Reservoir opera­ in specific conductance in the reservoir outflow. tions, increased reservoir inflow, and a decrease in the The median specific conductance had a significant specific conductance of the reservoir inflow. All of decrease of about 21 percent from 625 [iS/cm these factors resulted in increased streamflow and in 1966-74 to 496 (iS/cm in 1975-94. These decreased specific conductance downstream from changes in specific conductance seem to have the reservoir. At stations 07130500 and 07133000, improved the suitability of the Arkansas River located 0.2 and 20 mi downstream from John Martin as a domestic water supply in the 8.5-mi reach Reservoir, specific conductance decreased signifi­ between Pueblo Reservoir and the diversion cantly during most months from September through point for the St. Charles Mesa Water District. April and had no significant changes during May Few trends, except for increased specific through August. The median specific conductance conductance in June, August, and December, were at station 07130500 decreased significantly from detected in specific conductance at station 07109500, 2,700 jiS/cm in 1955-79 to 2,260 jiS/cm in 1980-94. which is located 24 mi downstream from Pueblo Similarly, the median specific conductance at Reservoir. It seems that the combined effects of station 07133000 decreased significantly from water storage and mixing in Pueblo Reservoir 4,000 jiS/cm in 1964-79 to 3,510 jiS/cm in 1980-94. and the increased inflow of relatively high-specific- These trends were very similar to the trends that were conductance water from Fountain Creek accounted for observed immediately downstream from Pueblo the few observed trends. The small amount of specific- Reservoir and largely were attributable to increased conductance data may have contributed to the detec­ storage and increased mixing of seasonally low- tion of few significant trends. and seasonally high-specific-conductance water At station 07124000, located 120 mi down­ in John Martin Reservoir. These factors tended to stream from Pueblo Reservoir, streamflow increased increase the minimum specific conductance and during all months after 1974. Although specific decrease the maximum specific conductance in the conductance tended to decrease during all months, reservoir outflow. Overall, the changes in specific most monthly trends were not statistically significant conductance that occurred at stations 07130500 and at a 95-percent confidence level. However, the median 07133000 did not change the salinity-hazard classifi­ specific conductance decreased significantly from cation of the water for irrigated agriculture; water at 3,000 jiS/cm in 1961-74 to 2,500 jiS/cm in 1975-94. both stations retained its high to very high salinity- The increase in streamflow at station 07124000 hazard classification. probably was caused by a combination of factors, including the WWSP and associated changes in seasonal irrigation practices and the increased impor­ REFERENCES CITED tation of Colorado River Basin water for irrigation. Prior to the construction of Pueblo Reservoir and Abbott, P.O., 1985, Description of water-systems the beginning of the WWSP, most winter stream- operations in the Arkansas River Basin, Colorado: flow was diverted and consumed upstream from U.S. Geological Survey Water-Resources Investiga­ station 07124000. With the adoption of the WWSP tions Report 85-4092, 67 p. and the increased use of transmountain water for irri­ Arkansas River Compact Administration, 1980, Thirty- gation, streamflow at station 07124000 increased second annual report: Lamar, Colorado, 50 p. significantly. Prior to 1978, about 14 percent of the Bradley, J.V., 1968, Distribution-free statistical tests: annual streamflow at the upstream end of the lower Englewood Cliffs, New Jersey, Prentice-Hall, 388 p.

REFERENCES CITED 47 Cain, Doug, 1985, Quality of the Arkansas River and Searcy, J.K., and Hardman, C.H., 1960, Double-mass irrigation-return flows in the lower Arkansas River curves; manual of hydrology, part 1, general surface- Valley, Colorado: U.S. Geological Survey Water- water techniques: U.S. Geological Survey Water- Resources Investigations Report 84-4273, 85 p. Supply Paper 1541-B, 65 p. 1987, Relations of specific conductance to stream- Taylor, O.J., and Luckey, R.R., 1974, Water-management flow and selected water-quality characteristics of studies of a stream-aquifer system, Arkansas River the Arkansas River Basin, Colorado: U.S. Geolog­ Valley, Colorado: Ground Water, v. 12, no. 1, ical Survey Water-Resources Investigations p. 22-38. Report 87-4041,93 p. U.S. Department of the Interior, 1994, Salinity update: Edelmann, Patrick, and Cain, Doug, 1985, Sources of water Denver, Bureau of Reclamation, Colorado River and nitrogen to the Widefield aquifer, southwestern Salinity Program Coordinator, 17 p. El Paso County, Colorado: U.S. Geological Survey 1996, Annual operating plans Fryingpan- Water-Resources Investigations Report 85^162, Arkansas Project: 16 p. 81 p. U.S. Environmental Protection Agency, 1986, National Helsel, D.R., and Hirsch, R.M., 1988, Discussion of secondary drinking-water regulations, part 143, "Applicability of the t-test for detecting trends section 143.3: Code of Federal Regulations, in water-quality variables," by R.H. Montgomery and J.C. Loftis: Water Resources Bulletin, v. 24, Title 40, parts 100-149, p. 587-590. p. 201-204. U.S. Geological Survey, 1959, Quality of surface waters 1992, Statistical methods in water resources: of the United States, 1955: U.S. Geological Survey Amsterdam, Elsevier Science Publishers, 529 p. Water-Supply Paper 1402, parts 7 and 8, 539 p. Hurr, R.T., and Moore, I.E., 1972, Hydrogeologic charac­ 1960, Quality of surface waters of the United teristics of the valley-fill aquifer in the Arkansas River States, 1956: U.S. Geological Survey Water-Supply Valley, Bent County, Colorado: U.S. Geological Paper 1452, parts 7 and 8, 469 p. Survey Hydrologic Investigations Atlas HA^61, 1961, Quality of surface waters of the United scale 1:62,500, 2 sheets. States, 1957: U.S. Geological Survey Water-Supply Lewis, M.E., and Edelmann, Patrick, 1994, Physical, chem­ Paper 1522, parts 7 and 8, 499 p. ical, and biological characteristics of Pueblo Reservoir, 1962-65, Surface water records of Colorado Colorado, 1985-89: U.S. Geological Survey Water- (published annually). Resources Investigations Report 94^097, 71 p. 1963a, Chemical analyses of surface waters in Miles, D.L., 1977, Salinity in the Arkansas Valley of Colorado, October 1959 to September 1962: 56 p. Colorado: U.S. Environmental Protection Agency 1963b, Quality of surface waters of the United and Colorado State University, Interagency Agree­ States, 1958: U.S. Geological Survey Water-Supply ment, EPA-IAG-D4-0544, 80 p. Paper 1573, parts 7 and 8, 588 p. Nelson, G.A., Hurr, R.T., and Moore, I.E., 1989a, Hydro- 1964, Quality of surface waters in Colorado, geologic characteristics of the valley-fill aquifer in October 1962 to September 1963: 83 p. the Arkansas River Valley, Prowers County, Colorado: U.S. Geological Survey Open-File Report 89-254, 1965a, Water quality records in Colorado, 1964: scale 1:62,500, 3 sheets. 84 p. 1989b, Hydrogeologic characteristics of the valley- 1965b, Quality of surface waters of the United fill aquifer in the Arkansas River Valley, Crowley States, 1959: U.S. Geological Survey Water-Supply and Otero County, Colorado: U.S. Geological Survey Paper 1644, parts 7 and 8, 507 p. Open-File Report 89-255, scale 1:62,500, 3 sheets. 1966-75a, Water-resources data for Colorado 1989c, Hydrogeologic characteristics of the pt. 1, Surface water records (published annually). valley-fill aquifer in the Arkansas River Valley, 1966-75b, Water-resources data for Colorado Pueblo County, Colorado: U.S. Geological Survey pt. 2, Water quality records (published annually). Open-File Report 89-256, scale 1:62,500, 3 sheets. 1976-95, Water resources data, Colorado, water Richards, L.A., ed., 1954, Diagnosis and improvement of years 1975-94 v. 1: U.S. Geological Survey Water- saline and alkali soils: U.S. Department of Agriculture Data Reports CO-75-1 to CO-94-1 (published Handbook 60, 160 p. annually).

48 Relations of Streamf low and Specific-Conductance Trends to Reservoir Operations in the Lower Arkansas River, Southeastern Colorado

* U.S. GOVERNMENT PRINTING OFFICE: 1998 673-083 / 20034 Region No. 8