Discharge Measurement

Two monitoring stations were established in October 2017: one at the upstream streamgage (05586647) and one at the downstream streamgage (05586745). Discharge measurements were made according to established USGS protocols (Turnipseed and Sauer, 2010) using an Acoustical Doppler Current Profiler tethered and pulled across the stream channel from the bridge or by wading and using a Sontek Flowtracker handheld Acoustic Doppler Velocimeter. Continuous record (15-minute interval) of the stream stage at each station was measured using a Design Analysis H-3613 radar unit and a Sutron Satlink 2 data-collection platform. Daily discharge values were calculated for the two monitoring stations according to established USGS methods and procedures (Turnipseed and Sauer, 2010). Discharge was measured periodically at each site, then correlated with stage readings to create a surrogate model used to report continuous discharge values. Wire-weight gages were also installed on the downstream side of the bridge at each location and functioned as the base gages. General site information of the monitoring stations is shown in table 1. Stage and discharge measurements for each site are available in the USGS National Water Information System database (U.S. Geological Survey, 2022a) using the station numbers provided in table 1.

Water-Quality Sample Collection and Analyses

Approximately 140 water samples were collected at each monitoring station between October 2017 and August 2021. Samples were collected on a weekly schedule and during select high-flow events. Sampling frequency was sometimes extended to a biweekly schedule when variance in concentrations was expected to be minimal, such as during periods of extended low flow or of sustained normal flow conditions, primarily during late fall and winter. Sample collection was typically at the center of the creek using a weighted-bottle sampler suspended from the bridge with a rope and lowered and raised by hand (U.S. Geological Survey, variously dated). The openings on the polypropylene sample-collection bottles were sized to collect water from the entire water column under various flow conditions and depths. Cross-section samples were collected during periods of high flow, higher stream velocities, and increased depth; these samples were collected using a DH-95 integrating sampler, a 4-wheel base and crane, and a polypropylene churn splitter (U.S. Geological Survey, variously dated).

Following collection, the water samples were stored in cooler on ice until processing and preservation before being shipped overnight in ice to the USGS National Water Quality Laboratory in Denver, Colorado, for analysis. Sample preparation and preservation included filtration of samples analyzed for the dissolved fraction (nitrate plus nitrite nitrogen and orthophosphorus) and the addition of sulfuric acid to samples analyzed for total phosphorus. Suspended-sediment samples were stored in the dark until shipment to the USGS Iowa Sediment Laboratory in Iowa City, Iowa.

Analytical methods for the sample analyses are given in table 2. References for descriptions of the analytical methods include Fishman (1993), Guy (1969), Patton and Kryskalla (2003, 2011).

Water-quality data collected for this study are available in the USGS National Water Information System database (U.S. Geological Survey, 2022a) using the station numbers provided in table 1.

Computation of Concentration, Flux, Loads, and Yields

Daily concentration and flux values of suspended sediment, phosphorus, and nitrate were computed using Weighted Regressions on Time, Discharge, and Season–Kalman (WRTDS–K; Zhang and Hirsch, 2019), which is part of the Exploration and Graphics for RivEr Trends (Hirsch and De Cicco, 2015) R-package incorporating continuous discharge and periodic water-quality constituent concentration data.

The default computation of daily concentration and flux estimates in Exploration and Graphics for RivEr Trends is completed using Weighted Regressions on Time, Discharge, and Season (WRTDS). This computation was calculated using a weighted regression model (eq. 1) for days on which a sample was collected, as shown in the following equation:

WRTDS calculates a bias correction factor (r) to estimate daily concentrations for days when a sample was not collected (Hirsch and De Cicco, 2015). A modification to the original WRTDS model is WRTDS–K, which was used in this report to compute daily concentration and flux values. WRTDS–K differs from WRTDS in two notable ways. First, sampled values are substituted for model-derived daily concentration values for days when samples were collected, which results in no difference between model values and sample values. Laboratory-derived concentration values were used in this report instead of estimated values because the laboratory values reflect the most accurate environmental value. Second, residuals (r) are calculated from the differences between predicted concentrations and observed concentrations. The residuals (r) take the place of random error (ε) in the model (eq. 2). The resulting equation is applied between sampling dates, with weighted regression being applied between sample dates as compared to the entire period (Zhang and Hirsch, 2019):

The flux values, in kilograms per day, are then computed by multiplying the daily estimated concentration by discharge values. The term “load” is synonymous to flux but the time component is often stated at the onset (that is, daily loads, annual loads, and so forth). Loads are shown as annual amounts in kilograms within this report. Yields are the loads divided by the drainage area of the basin and are represented in kilograms per square kilometer. All modeled derived estimates of concentration and flux for nutrient and sediment data calculated for this study for stations listed in table 1 are available in a USGS data release (Garcia, 2023); additional discharge and sample data are available from the USGS National Water Information System database (U.S. Geological Survey, 2022b, c).

Constituent yields are an informative measure by which to evaluate nutrient contributions from different portions of the study area. Yields are derived by dividing the loads for each station by the respective drainage areas to provide a comparable value from which to compare areas of different drainage areas. Yields can be reported in different ways, represent different time periods, and provide information characterizing the transport of constituents during storm events, annually, or over longer periods of time.

Concentrations

Approximately 140 water-quality samples were collected at each of the two monitoring stations during the study period. Summary statistics for the water-quality constituents (dissolved nitrate plus nitrite, total phosphorus, and suspended sediment) are shown in table 4 (hereafter, nitrate plus nitrite will be referred to simply as “nitrate”). The interquartile range (IQR) is the difference between the 75th percentile and the 25th percentile concentration values, represents the middle 50 percent of the concentration values, and indicates the variability of the concentrations. Nitrate and total phosphorus at the upstream station (05586647) had higher median concentrations (nitrate=3.04 milligrams per liter [mg/L], total phosphorus=0.25 mg/L) and IQR (nitrate=5.43 mg/L, total phosphorus=0.25 mg/L) concentrations than the median concentrations for samples collected at the downstream station (05586745) (nitrate=2.13 mg/L, total phosphorus=0.24 mg/L) and IQR (nitrate=3.51 mg/L, total phosphorus=0.21 mg/L). Conversely, the suspended-sediment median and IQR values for the downstream station (34.0 and 109.3 mg/L, respectively) were higher than for the upstream station (15.0 and 34.5 mg/L, respectively).

WRTDS–K was used to estimate concentrations for days on which a sample was not collected. Estimated daily concentrations of constituents for the two monitoring stations for the study period are shown in figure 3. Measured constituent concentrations by month for the study period are shown in figure 4.

Seasonal variability of nitrate concentrations can be seen in figures 3A and 4A. Nitrate concentrations were highest from January to July each year. The seasonal variability likely is due to a combination of environmental causes such as seasonal fluctuations in discharge and aquatic and terrestrial growth, and anthropogenic causes such as nutrient consumption, agricultural activities (including fertilizer application), and other processes within the watershed.

Variability in phosphorus and suspended-sediment concentrations did not follow the same strong seasonal variability as nitrate (figs. 3B,C and 4B,C). Phosphorus concentrations are more highly dependent on discharge (fig. 5). Variations in flow-concentration relations for phosphorus and discharge between the two stations was noticed. At the upstream station (05586647) periods of low flow and somewhat elevated phosphorus concentrations were present in late 2017, early 2018, late 2020, and early 2021, whereas concentrations were lower during this period at the downstream station (05586745) (fig. 3B). Higher phosphorus concentrations during periods of low flow at the upstream station (05586647) might reflect some local inflow or anthropogenic sources such as agricultural practices, wastewater or septic input. More focused monitoring that would target sub-watersheds or regulated inputs (such as septic inputs) upstream of station 05586647 would be beneficial to help narrow down the cause of the elevated concentrations.

Loads and Yields

A summary of the mean daily fluxes for the upstream (05586647) and downstream (05586745) monitoring stations is provided in table 5, and a summary of the annual loads and yields (by water year) for each of the three subbasins is provided in table 6. The study area subbasin constitutes approximately 57 percent of the total area of the downstream monitoring station. The loads and yields for the study area subbasin were computed by calculating the difference between drainage areas of the upstream and the downstream monitoring stations, 419.6 and 986.3 km², respectively (table 1). The area of the drainage basin between the two monitoring stations is approximately 567 km².

Nitrate loads and yields from the study area subbasin were lower than the loads and yields from the upstream subbasin for all years of the study. Furthermore, because the study area subbasin is marginally larger than the upstream subbasin, total phosphorus yields from within the study area were lower than within the upstream basin (table 6).

Unlike nitrate and phosphorus, suspended-sediment loads were greater from the study area subbasin than from the upstream subbasin values for all water years. The study area contributed between 50 and 79 percent of the suspended-sediment load measured at the downstream station (05586745) (table 6).

Time series plots of estimated daily yields for nitrate, phosphorus, and suspended sediment are shown in figure 6. Annual yields (the sum of the daily yields during the respective water-year period) depict differences in constituent transport affected primarily by yearly variation in precipitation and runoff as well as land-use practices that can change from year to year. Bar charts of estimated annual yields for nitrate, phosphorus, and suspended sediment are shown in figure 7, and estimated cumulative yields for the entire study period (2018–21), reported are shown in figure 8. The graphs for yield, unlike those for concentration, all show a similar pattern, with periods of greatest yield occurring during the periods of greatest flow (figs. 2 and 6). The seasonal pattern of nitrate concentration is muted when the flow is incorporated for yield computations, partly owing to the variability of the magnitude in yield values.

The yield data provide two important conclusions. First, the mean annual yield values indicate the downstream subbasin is exporting lower yields of nitrogen and phosphorus than the upstream subbasin because of the smaller contribution from the study area. Second, the study area is contributing a greater amount of suspended sediment than the upstream subbasin. The mean annual yield of nitrate at the upstream station (05586647) during water years 2018–21 was approximately 1.64 metric tons per square kilometer (t/km²). The mean annual yield of nitrate from the study area was approximately 0.60 t/km², a yield of about 36 percent of the mean annual yield at the upstream station. Similarly, the mean annual phosphorus yields for the upstream station and study area were approximately 0.21 and 0.13 t/km², respectively; the yield from the study area was about 62 percent of the upstream yield. Mean annual suspended-sediment yields showed an opposite pattern, with a mean annual yield at the upstream station of approximately 131 t/km² and a mean annual yield from the study area of approximately 166 t/km², which is about a 126-percent increase.

The cumulative yields during the study period help to indicate the effect the study area subbasin reach of the upper Macoupin Creek has on the overall watershed. The cumulative yields of nitrogen at the upstream station (05586647) during the water years 2018–21 was approximately 6.6 t/km², whereas the cumulative yield for the downstream station (05586745) was approximately 4.2 t/km² (fig. 8, table 6). The total yield of nitrate within the study area itself was approximately 2.4 t/km². Similarly, the total yield of total phosphorus for the upstream station was approximately 0.83 t/km², the total yield of total phosphorus for the downstream station was was approximately 0.67 t/km², and the total yield for the study area itself was approximately 0.53 t/km² during the study period. The total yield of suspended sediment for the upstream and downstream stations was approximately 524 and 605 t/km², respectively, and the total yield for the study area was approximately 663 t/km². Overall, the study area subbasin of the upper Macoupin Creek watershed reduces the yields from the watershed by 54 percent for nitrate and by 21 percent for total phosphorus while increasing the yield of suspended sediment by 10 percent. This increase in sediment in the study area likely is due to gully and streambank erosion determined in a previous study (Northwater Consulting and American Farmland Trust, 2019).

The decrease in yield of nitrate and phosphorus from the upstream (05586647) to downstream (05586745) monitoring sites is likely due to the smaller contribution (load) from the study area subbasin. Although it is outside the scope of this report to evaluate what environmental changes or land-use activities might contribute to these reductions, a concerted effort was made by collaborating partners to educate and encourage landowners within the study area subbasin to adopt nutrient and sediment BMPs during the study period. A further look at land-use practices and recent BMP implementation might explain some of the causes of these findings. Further study might also reveal trends within the watersheds that reflect other land-use changes.