Terms Used Within Identifying Changes in Background Water-Quality Conditions, Arkansas River and Fountain Creek, near Pueblo, Colorado

Dissolved solids as an indicator

Dissolved solids was chosen as the indicator parameter in this study for several reasons. First, dissolved solids in the river have been a concern to local water managers, water suppliers, and farmers in the lower Arkansas River Basin for many years. Second, dissolved solids also are a good indicator of the general water quality in the study area and can be accurately estimated from available specific-conductance data (Cain, 1987; Ortiz and others, 1998). Third, dissolved-solids concentrations are sensitive to changes in streamflow, water operations, and source contributions. Lastly, the use of dissolved-solids concentration data combined with streamflow allows for analysis of mass loading as a parameter of interest.

Specific conductance as an indicator of dissolved solids

There is a strong linear relation between instantaneous measurements of specific conductance and dissolved-solids concentration in the study area. In addition, real-time specific-conductance data are measured directly from field instruments at 15-minute intervals and uploaded to the database every 4 hours. The average value throughout a single day is expressed as the daily mean value. As such, the daily mean specific-conductance value can be used to estimate a daily mean dissolved-solids concentration using the regression equations in Table 2 of the USGS SIR 2004-5024.

What is a flow-adjusted concentration?

The concentrations of many constituents are affected by streamflow. Typically, adjustment of the concentration data to account for varying streamflow is needed before further analysis to evaluate differences in concentration is done. The effects of streamflow can pose a problem when evaluating changes in specific conductance and dissolved solids because of the inherent inverse relation between these variables and streamflow. The use of flow-adjusted concentrations by the residuals method provides a way to remove the source of much of the variance due to the influence of streamflow. Residuals analysis regresses the measured value on some function of streamflow and uses the residuals from the regression (the observed measurement minus the predicted value) as flow-adjusted values. The residual method works well if a functional form of the measured value to streamflow regression produces a reasonable fit. To estimate daily specific-conductance values from streamflow, instantaneous specific-conductance data and continuous daily streamflow data were input into a multi-parameter regression model. Regression equations for the model can be found in Table 3 of the USGS Scientific-Investigations Report 2004-5024. See the report for a more detailed explanation of flow-adjusted concentrations.

Explanation of tolerance limit plots

Tolerance limit is the daily threshold defined by the confidence in the test and the fixed proportion of the population covered by the test. As defined for the use here, a 95 percent confidence and a 97.5 percent population coverage were selected. An exceedance of the tolerance limit occurs when the current value exceeds the threshold. Seven exceedances within a 10-day period has been determined to constitute a change in water quality.

Current value represents the estimated concentration or load for the time frame displayed. The current values are estimated using daily specific conductance and/or daily streamflow data obtained from real-time monitors at each site location.

Mean value represents the daily average for each day of the calendar year calculated from as many as 13 to 18 years of background data. As an example, the mean value for 18 years of data collected on January 1 was 280.

Y-axis (vertical) format will show the estimated concentration in milligrams per liter or the estimated load in tons per day. Plotting intervals will vary depending on the range of data values.

X-axis (horizontal) format will show a series of date stamps that are dependent on the duration of the plot(s) selected.

What do the tolerance limit plots of concentration tell me?

Scenario A: Dissolved-solids concentrations (A1) exceed the tolerance limit for a period of time and flow-adjusted dissolved-solids concentrations (A2) exceed the tolerance limit. The results indicate the exceedances are likely attributable to changes in source loadings to the river and less likely due to changes in streamflow.

Scenario B: Dissolved-solids concentrations (B1) and flow-adjusted dissolved-solids concentrations (B2) did not exceed the tolerance limit. The results indicate that there is no change in water quality as indicated by dissolved solids with respect to historic conditions.

Scenario C: Dissolved-solids concentrations (C1) exceed the tolerance limit for a period of time; however, flow-adjusted dissolved-solids concentrations (C2) did not exceed the tolerance limit. The results indicate that the exceedances in concentrations are likely associated with changes in streamflow that may be attributable to changes in water operations. This scenario also indicates that the exceedances are not likely attributable to changes in source loadings.

           Examples

Glossary of Terms

Specific conductance is a measurement of the ability of a solution to conduct electrical current (measured at 25 degrees Celsius). The presence of charged ionic species in solution makes the solution conductive. The conductance of a solution increases as ion concentrations increase; therefore, the specific-conductance measurement provides an indication of ion or dissolved-solids concentration of the solution.

The background period at each site was defined by the number of consecutive calendar years that continuous daily specific-conductance data were available for a site. These data were available for 16 calendar years at the three primary study sites (1986-2001), 19 calendar years at Arkansas River at Portland (1983-2001), and 13 calendar years at Arkansas River at Pueblo (1989-2001). See USGS Scientific-Investigations Report 2004-5024 for a more detailed explanation.

Identification of significant changes require a statement of the risk that is acceptable to the scientist, manager, or decision maker. This risk, or significance level, is the probability of incorrectly rejecting the null hypothesis when in fact it is true. Whereas the risk does not depend on the data, the p-value provides information on the strength of the scientific evidence. The p-value is the probability of obtaining the computed test statistic, or one even less likely, when the null hypothesis is true. The evidence for rejection of the null hypothesis increases as the p-value becomes smaller. A p-value of 0.05 was used for this study.

The null hypothesis (Ho) is what is assumed to be true about the system under study prior to data collection, until indicated otherwise (Helsel and Hirsch, 1992).

Tolerance limits were selected for this application because they provide an estimation of the range that should contain a certain percentage of each individual measurement in the population. Within the context of a one-tailed test, an upper tolerance limit statistically defines what percentage of the population measurements will not exceed a defined upper limit. For the purposes of the study, and in consultation with the Southeastern Colorado Water Activities Enterprise, a 95-percent degree of confidence with a 97.5 percent population coverage was selected to determine the k-value for this report. See USGS Scientific-Investigations Report 2004-5024 for more detailed explanation.

A K-value is a numerical factor used to adjust the width of the tolerance interval as designated by the standard deviation of the data. The K-value defines the interval that includes at least a proportion of the population within the stated confidence. That is to say, you can define a K-value to assure that you have included x percent of the population (say 95) at a y percent confidence (say 95). As such, you can increase or decrease x and/or y as required for the application at hand. As x and y increase, the K-value becomes increasingly larger resulting in an increasingly larger width of the tolerance interval.

The use of flow-adjusted concentrations by the residuals method provides a way to remove the source of much of the variance due to the influence of streamflow; the concentrations of many constituents are affected by streamflow. Typically, concentration data need to be adjusted to account for varying streamflow before further analysis is done to evaluate changes in concentration. Residuals analysis regresses the measured value on some function of streamflow and uses the residuals from the regression (the observed measurement minus the predicted value) as flow-adjusted values. See USGS Scientific-Investigations Report 2004-5024 for more detailed explanation.

Daily loads were calculated as the product of estimated daily dissolved-solids concentration in milligrams per liter, the mean daily streamflow in cubic feet per second, and a conversion factor of 0.002697. The result of the calculations is expressed in units of tons per day.

Dissolved solids was chosen as the indicator parameter in this study for several reasons. First, dissolved solids in the river have been a concern to local water managers, water suppliers, and farmers in the lower Arkansas River Basin for many years. Second, dissolved solids also are a good indicator of the general water quality in the study area and can be accurately estimated from available specific-conductance data. Third, dissolved-solids concentrations (as estimated from specific conductance) are sensitive to changes in streamflow, water operations, and source contributions. Lastly, the use of dissolved-solids concentration data combined with streamflow allows for analysis of mass loading as a parameter of interest.

Dissolved solids consist of minerals, organic matter, and nutrients that have dissolved in water. The major components of dissolved solids of natural waters include bicarbonate, calcium, sulfate, hydrogen, silica, chlorine, magnesium, sodium, potassium, nitrogen, and phosphorus in the form of phosphate. The major dissolved solids in the Arkansas River near Pueblo are calcium (cation) and sulfate and bicarbonate (anions). Minor constituents that are normally present in trace concentrations in streams include iron, copper, zinc, boron, manganese, and molybdenum.

Daily values represent the average for a given day computed from as many as 96 measurements. These data are published annually in the USGS Water-Data Reports for Colorado (U.S. Geological Survey, 1984-2003).

Instantaneous data are single measurements in time. These data are published annually in the USGS Water-Data Reports for Colorado (U.S. Geological Survey, 1976-2003). Instantaneous specific-conductance data dating back to January 1975 were retrieved for all study sites on the Arkansas River (U.S. Geological Survey, 1976-2001). Similarly, instantaneous specific-conductance data dating back to January 1982 were retrieved for Fountain Creek at Pueblo (U.S. Geological Survey, 1983-2001).

Native water originates from within the Arkansas River Basin.

Transmountain water (non-native) originates from outside the basin and is conveyed by ditches or tunnels to the Arkansas River Basin.

Those entities with the oldest direct-flow water rights are said to have the senior water rights.

The Colorado Water Quality Control Commission (2001) refers to dissolved solids as salinity. Batie and Healy (1983) described the high-salinity problem in this area as the most pervasive problem associated with irrigated agriculture in the United States.

For each study site, daily mean dissolved-solids concentrations were sorted by date and plotted to evaluate the variability for any given day of the calendar year (fig. 4). The daily variability in concentration for any single date shows relatively little change regardless of the season for the sites located on the Arkansas River. The average daily variability in daily mean dissolved-solids concentrations at the Arkansas River at Portland site (1983-2001) and Arkansas River above Pueblo site (1986-2001) was less than 170 milligrams per liter while the daily variability at the Arkansas River at Pueblo site (1989-2001) and Arkansas River near Avondale site (1986-2001) was less than 320 milligrams per liter. Daily variability at the Fountain Creek at Pueblo site (1986-2001) was more pronounced (fig. 4) during the summer as dissolved-solids concentrations increased in response to reduced streamflow due to irrigation and/or ground-water pumping demands and dissolved-solids concentrations decreased in response to increased streamflow due to storm runoff. The average daily variability at the Fountain Creek at Pueblo site from April through September was about 550 milligrams per liter. The daily variability decreased to 308 milligrams per liter from October through March.