Introduction and Background

Milford Lake, the largest reservoir by surface area in Kansas, has had confirmed harmful algal blooms (HABs) every summer from 2011 through 2017 and from 2020 through 2023 (fig. 1; Kansas Department of Health and Environment [KDHE], 2023). No HABs were detected or reported by the KDHE during 2018–19 despite greater than mean annual total streamflow. During development of the 2012 303(d) list, the KDHE determined that excessive nutrients in Milford Lake are impairing aquatic life and domestic water supply and contributing to HABs that result in the eutrophication and impairment of contact recreation within Milford Lake (Kansas Alliance for Wetlands and Streams, 2019). The KDHE developed a nutrient total maximum daily load (TMDL) for Milford Lake eutrophication and dissolved oxygen impairments in 2014 (KDHE, 2014). The Republican River is the primary inflow to Milford Lake and drains areas of Kansas, Nebraska, and Colorado. The Lower Republican River Basin includes only the Kansas part of the Republican River Basin (basin will be hereafter referenced as watershed) in Clay, Cloud, Dickinson, Geary, Jewell, Mitchell, Republic, Riley, and Washington Counties (fig. 2).

The Watershed Restoration and Protection Strategy (WRAPS) for the Lower Republican watershed was developed to outline a plan of restoration and protection goals and actions for the surface waters of the watershed. The original WRAPS plan was developed in 2011, and TMDLs were approved in 2014 (KDHE, 2014; Kansas Alliance for Wetlands and Streams, 2019). WRAPS defines watershed restoration for surface waters that do not meet water-quality standards; watershed restoration for areas of the watershed that need improvement in habitat, land management, or other attributes; and watershed protection for surface waters that currently meet water-quality standards but need protection from future degradation (Kansas Alliance for Wetlands and Streams, 2019).

In 2015, the Kansas Water Office Regional Advisory Committees were established to develop broad goals to guide activities to improve water quality and quantity. In 2018, the Kansas Water Office submitted an application to the Kansas Forest Service Regional Conservation Partnership Program (RCPP) for the Milford Lake watershed, which was approved for funding by the Natural Resources Conservation Service (Kansas Alliance for Wetlands and Streams, 2019). The focus of the RCPP is to reduce phosphorus and align with goals set by WRAPS and the Regional Advisory Committees. The focus includes additional funding for the implementation of best management practices (BMPs) in the Milford Lake watershed (Kansas Alliance for Wetlands and Streams, 2019). The Milford Lake watershed RCPP primarily focuses on the Lower Republican River watershed (Kansas Water Office, 2023). Because the majority watershed land use is cropland (Dewitz, 2023), the Environmental Quality Incentives Program, through the Natural Resources Conservation Service, provides funding for eligible Milford Lake watershed RCPP BMPs with a cropland and livestock focus. Some of the eligible core practices of the Environmental Quality Incentives Program include implementing comprehensive nutrient management plans, fencing, cover crops, access controls, watering facilities, heavy use areas, and residue and tillage management (reduced till or no till) practices (Unruh, 2020).

The Milford Lake TMDL-estimated mean annual loading to the reservoir is 4,875,835 pounds per year of total nitrogen (TN) and 1,216,912 pounds per year of total phosphorus (TP) based on data collected during 2008–12 (table 1; KDHE, 2014). The Milford Lake TMDL estimated a slower than designed sedimentation rate of 1,007 acre-feet per year (acre-ft/yr; equivalent of 2,014,000 tons per year; KDHE, 2014). To improve the trophic condition and water quality of Milford Lake and meet desired endpoints, the TMDL estimates a necessary watershed reduction in TP and TN of 88 and 86 percent, respectively (KDHE, 2014). The Milford WRAPS plan addresses only the Kansas part of load reductions needed using a mass balance evaluation to remove the part of the load originating in Nebraska, resulting in a TP reduction goal of 617,204 pounds (lb) and a TN reduction goal of 2,341,263 lb annually from Kansas sources (table 1; Kansas Alliance for Wetlands and Streams, 2019). The plan also outlines a 20-year sediment load reduction goal of 630,612 tons (Kansas Alliance for Wetlands and Streams, 2019). The Milford Lake watershed RCPP project goal is to reduce watershed phosphorus loading with riparian area BMP prioritization based on the Milford Lake WRAPS watershed plan (table 1; Kansas Alliance for Wetlands and Streams, 2019; Unruh, 2020).

Purpose and Scope

The purpose of this report is to quantify and characterize water-quality constituent concentrations and loads using models previously published in Leiker (2022) for the Republican River near Clay Center, Kansas (U.S. Geological Survey [USGS] station 06856600; hereinafter “Clay Center site”), located about 15 miles upstream from Milford Lake, during August 1, 2018, through July 31, 2023. Water-quality constituents of concern include suspended sediment (SS), TN, and TP and were selected for their relation to water supply and water-quality issues in Milford Lake, including nutrient and sediment transport, taste-and-odor events, potentially toxic cyano-HAB events, and subsequent downstream transport of contaminants. Water-quality monitoring is essential to measure nutrient load changes, maintain modeled relations, deliver data to the public in real time, and monitor potential changes that may be related to conservation efforts such as those implemented by the Milford Lake watershed RCPP in the Republican River. The water-quality information presented in this report is important to the Kansas Water Office because it quantifies and characterizes nutrient and sediment transport into Milford Lake. In addition, the methods and results used in this study contribute to understanding the effects of agricultural land-use practices on surface-water resources and could be applied to other sites regionally, nationally, and globally.

Methods

Continuous and discrete water-quality data were collected at the Clay Center site from August 2018 through July 2023. The annual periods are designated by the calendar year in which they end; for example, August 1, 2018, through July 31, 2019, is called 2019 for the purpose of this report. Streamflow, water temperature, specific conductance, pH, dissolved oxygen, and turbidity were measured continuously in stream. Discrete water-quality samples were collected over a range of streamflows (appendix 1, fig. 1.1). Previously published linear regression models (Leiker, 2022) were used to compute water-quality constituent concentrations and loads using continuous water-quality data.

Continuous and Discretely Monitored Water-Quality Variables

Continuously measured streamflow ranged from 35 to 23,000 ft³/s with a mean of 968 ft³/s during the study period (table 3). Continuously measured water temperature ranged from −0.2 to 37.7 degrees Celsius (°C) with a mean of 15.6 °C during the study period (table 3). Continuously measured specific conductance ranged from 145 to 1,500 microsiemens per centimeter at 25 °C (µS/cm) and had a mean value of 881 µS/cm (table 3). Kansas aquatic-life-support criteria require that pH values in streams not measure less than 6.5 or more than 8.5 standard units (KDHE, 2022a). Continuously measured pH ranged from 7.2 to 9.4 and had a median of 8.4 during the analysis period (table 3). Measured pH was never lower than 6.5 but exceeded 8.5 about 21 percent of the time during the study period, primarily during low-flow conditions (fig. 3). Continuously measured dissolved-oxygen concentrations ranged from 3.3 to 18.8 mg/L and had a mean of 10 mg/L (table 3). Kansas aquatic-life-support criteria require that dissolved-oxygen concentrations are not less than 5.0 mg/L (KDHE, 2022a). Dissolved-oxygen concentrations were less than the minimum aquatic-life-support criterion less than 1 percent of the time and generally were detected during periods of increased water temperature (greater than 20 °C; fig. 3). Dissolved oxygen is an important factor for the survival of aquatic organisms (USGS, 2020). Continuously measured turbidity ranged from 1.3 to 1,080 formazin nephelometric units and had a mean of 74.0 formazin nephelometric units (table 3).

Discrete water-quality samples were analyzed for TN, TP, and SS. TN concentrations (n=63) ranged from 0.51 to 10.6 mg/L with a mean of 2.77 mg/L (table 2). TP concentrations (n=63) ranged from 0.15 to 2.43 mg/L with a mean of 0.67 mg/L (table 2). SS concentrations (n=61) ranged from 5.0 to 2,350 mg/L with a mean of 405 mg/L (table 2). Although TP and TN did have new maximum and minimum values, mean values of TN, TP, and SS did not vary substantially from the datasets that were used to develop the linear regression models (Leiker, 2022).

Concentrations and loads for TN, TP, and SS were primarily computed using previously developed linear regression models (Leiker, 2022). Since publication of the models, 20 additional discrete samples have been collected at the Clay Center site during April 2021 through June 2023. These samples were not included in the original model calibration dataset for model development but were compared to model estimated concentrations. Of the 20 samples collected after model development, 6 TN, 2 TP, and 3 SS samples plotted outside the 90-percent prediction interval (appendix 1, fig. 1.2). All samples that plotted outside of the prediction intervals were new minimum or maximum values or near the upper or lower range for the sample dataset. In general, for the model validation results, model predicted values tended to overestimate concentrations compared to observed values in the lower range of conditions, and model predicted values near the upper range tended to underpredict observed concentrations. The potential effect on this analysis could result in underestimated loads. Future analyses would benefit from updated linear models to account for a greater range of observed conditions.

Computed Concentrations, Loads, and Comparison to Watershed Goals

Annual total streamflows ranged from 76,400 to 1,650,000 acre-ft/yr (table 4). Compared to the historical mean annual total streamflow at the Clay Center site, annual total streamflows in 2019–20 were 1.7–2.5 times greater than historical mean annual total streamflow and 2021–23 streamflows were less than the historical mean annual total streamflow by 25–88 percent (fig. 4). TN, TP, and SS concentrations and loads were monitored for their relation to water supply an water-quality issues in Milford Lake, including nutrient and sediment transport, taste-and-odor events, potentially toxic cyano-HAB events, and subsequent downstream transport of contaminants.

All constituents of interest (TN, TP, and SS) had the smallest annual mean concentrations and loads in 2023 and the largest in 2019. Annual mean TN concentrations ranged from 1.40 to 2.62 mg/L with a study mean of 2.04 mg/L. Annual mean TP concentrations ranged from 0.34 to 0.61 mg/L with a study mean of 0.47 mg/L. Annual mean SS concentrations ranged from 69 to 365 mg/L with a study mean of 210 mg/L. Total annual loads for TN ranged from 336,000 to 17,000,000 lb with a study total load of 30,800,000 lb. Total annual loads for TP ranged from 80,900 to 4,600,000 lb with a study total load of 7,860,000 lb. Total annual loads for SS ranged from 12,300 to 1,810,000 tons with a study total load of 2,810,000 tons (table 4).

Annual loads of TN, TP, and SS for the study period are compared to TMDL-defined mean annual loads from 2008 to 2012, Kansas nonpoint source reduction goals defined by WRAPS, and TMDL-defined watershed reduction goals in figures 5–7 (KDHE, 2014; Kansas Alliance for Wetlands and Streams, 2019). TN and TP annual loads were greater than the TMDL-defined mean annual load during 2019–20 and less than the TMDL-defined mean annual load during 2021–23 (figs. 5–6; KDHE, 2014). The study mean annual loads (August 2018 through July 2023) for TN and TP were greater than the TMDL-defined mean annual load and exceeded the watershed reduction goals and the Kansas nonpoint source reduction goals (KDHE, 2014; Kansas Alliance for Wetlands and Streams, 2019). Annual loads of TN and TP were below the Kansas nonpoint source reduction goal in 2022 and below the watershed reduction goal in 2023.

Annual mean streamflow during the study period ranged from 2,270 ft³/s in 2019 to 105 ft³/s in 2023. Compared to flow-duration streamflow statistics for the Clay Center site, those annual mean streamflows correspond to the approximately 7 and 90 percent exceedance rates, respectively (table 4; appendix 1, fig. 1.1; Granato and others, 2017). The data collected during the study period represent a wide range of conditions at the Clay Center site with the first year (2019) representative of high-flow conditions with frequent runoff and the last year (2023) representative of low-flow conditions with less frequent runoff. About 51, 58, and 68 percent of the total loads were delivered during the top 5 percent of loading days (91 days) for TN, TP, and SS, respectively. Nutrient reduction goals were only met in the final 2 years of the study period when annual mean flow conditions were lower than normal, indicating that goals may not be met during average or high-flow conditions (figs. 5–6). The wide range of flows captured during this study would allow for future comparison to varying flow conditions. Continued data collection could allow further investigation regarding the attainability of the nutrient reduction goals under different flow regimes.

SS loads were less than the TMDL-computed mean annual sedimentation rate for the entirety of the study period and much less than the designed annual reservoir sedimentation rate (August 2018 through July 2023; fig. 7; KDHE, 2014). In all years except 2019, the annual mean SS load, in addition to the study mean, was at or below the 20-year sediment load reduction target (Kansas Alliance for Wetlands and Streams, 2019). Given the substantial range of annual flow conditions combined with the annual SS loads frequently not exceeding the 20-year target during the study period, it is likely the sediment load reduction target is attainable under similar streamflow conditions.

Despite periods of increased runoff and nutrient loading into Milford Lake (figs. 4–6), no HABs were reported or detected during 2018–19 (fig. 1). Data collection and analysis efforts to compare inflow and nutrient loading at the Clay Center site to subsequent HAB formation and severity in Milford Lake is beyond the scope of this report. However, increased nutrient loading and availability of nutrients may support formation and proliferation of HABs (Anderson and others, 2002; Watson and others, 2015). The lack of reported or detected HABs during the period of increased runoff could be attributed to short water residence times that precluded sustained algal growth. HABs may also not have been observed due to increased runoff events causing system alterations (shifts in food webs, habitat changes, light limitations because of increased turbidity, and so on) that affect the extent to which HABs can accumulate; and (or) the timing, distribution, and bioavailability of the external nutrient load (Heisler and others, 2008; Roelke and Pierce, 2011; Watson and others, 2015).

Although annual loads at the Clay Center site generally decreased during the study period (figs. 5–7), corresponding reductions in annual streamflows (fig. 4) indicate that these reductions may primarily be related to less frequent runoff from the upstream basin. To better quantify the degree to which trends in constituent loads may be related to changes in upstream landscape practices, a flow-normalized procedure such as that used in the Weighted Regressions on Time, Discharge, and Season package is often recommended (Hirsch and others, 2010; Hirsch and De Cicco, 2015). Quantification of potential nonpoint source trends in water-quality generally require the collection of streamflow and water-quality data over decades (Hirsch and others, 2010). Continued water-quality monitoring and tracking of BMPs will be helpful to understand the success of RCPP efforts to reduce nutrient transport in the Milford Lake watershed.

Summary

Milford Lake, the largest reservoir by surface area in Kansas, has had confirmed harmful algal blooms every summer from 2011 to 2017 and 2020 to 2023. Milford Lake has been listed as impaired and designated hypereutrophic under section 303(d) of the 1972 Clean Water Act. Kansas Department of Health and Environment developed a nutrient total maximum daily load (TMDL) for Milford Lake eutrophication and dissolved oxygen impairments in 2014. The Watershed Restoration and Protection Strategy for the Lower Republican watershed was developed to outline a plan of restoration and protection goals and actions for the surface waters of the watershed. In 2018, the Regional Conservation Partnership Program for the Milford Lake watershed was funded by the Natural Resources Conservation Service, focusing on advancing implementation of best management practices in the Lower Republican River watershed.

Continuous and discrete water-quality data were collected at the Republican River at the Clay Center site from August 2018 through July 2023 from a range of streamflow conditions. Streamflow and in situ water-quality data were used to compute total nitrogen (TN), total phosphorus (TP), and suspended sediment (SS) concentrations and loads using previously developed linear regression models. TN, TP, and SS concentrations and loads were evaluated relative to water supply and water-quality issues in Milford Lake, including nutrient and sediment transport, taste-and-odor events, potentially toxic cyano-harmful algal bloom events, and subsequent downstream transport of contaminants.

The data collected during the study period represent a wide range of conditions at the Clay Center site; the first year (2019) was representative of high-flow conditions with frequent runoff and the last year (2023) was representative of low-flow conditions with less frequent runoff. TN and TP annual loads were greater than the TMDL-defined mean annual load during 2019–20 and less than the TMDL-defined mean annual load during 2021–23. The study mean annual loads (August 2018 through July 2023) for TN and TP were greater than the TMDL-defined mean annual load and exceeded the watershed reduction goals and the Kansas nonpoint source reduction goals. TN and TP annual loads fell below the Kansas nonpoint source reduction goal in 2022 and below the watershed reduction goal in 2023. Nutrient reduction goals were only met in the final 2 years of the study period when annual mean flow conditions were lower than normal, indicating the goals may be difficult to attain during average or high-flow conditions. SS loads were less than the TMDL-computed mean annual sedimentation rate for the entirety of the study period and much less than the designed annual reservoir sedimentation rate. With the exception of 2019, the mean annual SS load was less than the 20-year sediment load reduction target.

Although annual loads at the Clay Center site generally decreased over time, corresponding reductions in annual streamflow indicate that these reductions may primarily be related to less frequent runoff from the upstream basin. Quantification of potential nonpoint source trends in water quality generally require the collection of streamflow and water-quality data over decades. Continued water-quality monitoring and tracking of best management practices would be helpful to understand the success of Regional Conservation Partnership Program efforts to reduce nutrient transport in the Milford Lake watershed.