English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

In support of nutrient reduction efforts, nitrate (as nitrate plus nitrite) and phosphorus loads and yields were computed for selected streams in Iowa based on continuously monitored sensor data for 2008–17 and 2014–17, respectively. Sample data were used to assess nitrate sensor bias and to create phosphorus-turbidity surrogate models. Where needed, nitrate loads were corrected for site-specific sensor bias, which was determined to be as high as 9.25 percent. Nitrate loads presented in this report using continuous (generally 15-minute interval) data were on average 4 percent less, but as much as 38 percent less, than annual loads computed from daily mean nitrate concentrations not corrected for sensor bias. Streamflow-based phosphorus models had poorer fit (adjusted coefficient of determination values less than 0.75) than turbidity-based models (adjusted coefficient of determination approximately 0.9). However, alternate models based on streamflow were used to obtain a more complete annual phosphorus load despite seasonal and fragmentary sensor data. Mean annual nitrate yields for 18 selected sites (96 siteyears) ranged from 1.68 to 164 pounds per square mile per day (lb/mi2/d), compared to 19.4 lb/mi2/d average statewide yield needed to achieve the nitrate-reduction goal. Mean annual phosphorus yields for selected sites on the Maquoketa River, South Raccoon River, and West Nishnabotna River range from 1.57 to 7.19 lb/mi2/d, compared to 1.06 lb/mi2/d average statewide yield needed to achieve the phosphorus-reduction goal. Introduction The U.S. Geological Survey (USGS) and the Iowa Department of Natural Resources (IADNR) cooperatively studied nitrate and phosphorus loads using surrogate relations in several streams. Eutrophication of local and downstream waters, specifically an abundance of the nutrients nitrogen and phosphorus, is a high-priority issue in the Mississippi River Basin (Mississippi River/Gulf of Mexico Watershed Nutrient Task Force, 2017) and local Iowa waters (Iowa Department of Agriculture and Land Stewardship and others, 2017a). Nutrient reduction goals in Iowa focus on nitrate loads as the dominant form of nitrogen, particularly for agricultural nonpoint sources from a landscape with pervasive subsurface drainage. Though laboratory and sensor measurements for nitrate include nitrite, nitrite is negligible in Iowa surface waters (Garrett, 2012), and the term “nitrate” is used for nitrate plus nitrite throughout this report. Using consistent methods applied across large scales, Iowa streams rank among the largest contributors to total Mississippi River Basin nutrient loads (Aulenbach and others, 2007; Robertson and others, 2009), but more accurate methods are needed locally to compute site-specific loads and track annual progress toward nutrient reduction goals within the State (Iowa Department of Agriculture and Land Stewardship and others, 2017b; Jones and others, 2018). The nutrient reduction strategy in Iowa (Iowa Department of Agriculture and Land Stewardship and others, 2017a), as in other Midwest States (Anderson and others, 2016; Illinois Environmental Protection Agency and others, 2015; Ohio Environmental Protection Agency and others, 2016), calls for large reductions in nutrient delivery to the Gulf of Mexico; the reduction goal for Iowa is 45 percent. Several traditional monitoring and load-computation approaches may be able to detect such large changes, but there also is a need to detect lesser interim changes. Load-calculation methods based on infrequent (weekly to monthly) samples may not be accurate enough to assess interim progress toward load reduction goals because loadcalculation estimation errors can be quite large. Aulenback and others (2007, updates through 2016) reported load estimates for the Iowa River at Wapello, Iowa (05465500), for 1978–2016 with mean annual 95-percent confidence intervals for predicted fluxes of 42 percent for phosphorus and 76 percent for nitrate (Aulenbach and others, 2007); Garrett (2012) reported confidence intervals of 41 percent for phosphorus flux and 31 percent for nitrate flux for 10 streams in Iowa during 2004–08. Many factors affect load-calculation errors. For example, Johnes (2007) highlights basin characteristics, such as low base-flow index and high population densities, that make 2 The Use of Continuous Water-Quality Time-Series Data to Compute Nutrient Loadings for Selected Iowa Streams, 2008–17 load estimates more sensitive to errors. Lee and others (2016) describe greater sensitivity to errors for constituents with strongly positive or log-curvilinear relation between concentration and streamflow. Although greater sensitivity to errors was common for some constituents (total phosphorus and suspended sediment), the relation was ultimately site-specific. Several researchers concluded no single model performed well across multiple sites or years (Schilling and others, 2017a; Preston and others, 1989; Lee and others, 2016), and that any of these models can be grossly inaccurate or biased when model assumptions are not met (Stenback and others, 2011; Preston and others, 1989; Hirsch, 2014). In contrast, monitoring with continuous data improves accuracy of load calculation (Cassidy and Jordan, 2011; Duan and others, 2014; Jones and others, 2012; Jones and others, 2018; Pellerin and others, 2014; Reynolds and others, 2016; Rozemeijer and others, 2010; Terrio and others, 2015). Jones and others (2012) demonstrated about a 10-fold improvement in the accuracy of assessing phosphorus regulatory compliance, as percent of time exceeding a concentration-based criterion, based on daily versus monthly data. Reynolds and others (2016) showed precision improved 9–12 percent for multiple metrics—mean concentration, exceedance of concentrationbased criterion, peak concentrations, and total loads—with nitrate sampling frequency increased from monthly to daily. However, Cassidy and Jordan (2011) determined none of the typical sampling regimes (weekly, daily, event-triggered, or random) allowed load calculations accurate enough to assess interannual changes in phosphorus loads for very small (3–5 square kilometers) basins. Though current phosphorus sensor technologies (Warwick and others, 2013) are not as field robust or accurate as ultraviolet nitrate sensors, many of the benefits of continuous data can be obtained with surrogate-derived concentration data (Baldwin and others, 2012; Jones and others, 2012; Rasmussen and others, 2009; Stubblefield and others, 2007). Surrogates, by definition, use indirect data intended to provide information about something that is difficult to measure directly, but often with a direct or uncomplicated association between the surrogate and the parameter of interest. However, the relation between a surrogate and parameter of interest should be more than statistical correlation and have some physical basis. For phosphorus, turbidity makes sense physically because, in many Iowa streams, total phosphorus loads are dominated by particulate-bound phosphorus (Garrett, 2012). Turbidity is a measure of the scatter of light from particles in the water, particularly fine-grained sediment. Several studies have determined turbidity to be a good predictor for total phosphorus, that additional parameters associated with dissolved constituents or biological processes can improve turbidity-based models, and that site specificity is important (Christensen and others, 2006; Rasmussen and others, 2005; Schaepe and others, 2014; Schilling and others, 2017b). Purpose and Scope The purpose of this report is to describe procedures for computing a time series of concentrations and loads of nitrate and total phosphorus based on data from water-quality sensors in selected Iowa streams collected during 2008–17. Techniques and guidelines are discussed to develop and evaluate an unbiased nitrate load time series based on continuous sensor measurements verified with sampled data, and total phosphorus loads primarily based on empirical statistical regression between sensor measurements and sample results, with alternate models and filling procedures for periods with gaps in sensor data. Continuous sensor data, sample results, regression models, and modeled continuous data are presented for selected sites. Continuous nitrate, turbidity, and water temperature data are summarized through time and through a range of seasonal and streamflow event conditions. Sample concentration data for nitrate, nitrogen, orthophosphate, phosphorus, and suspended solids are presented. Statistical regression models relating sensor data and sample results are described, with detail on model selection and diagnostics. Time-series nutrient (nitrate and phosphorus) concentrations and loads presented in this report and computed moving forward using the methods presented can be used to monitor changes in stream nutrients. Sites Water-quality samples and sensor data summarized in this report were collected in selected Iowa stream sites collocated with existing streamflow-gaging stations (fig. 1, table 1). The drainage area of nitrate and turbidity monitoring sites across the State vary from 3 to 12,500 square miles (mi2). Of these, three sites with turbidity monitoring were selected for development of phosphorus surrogate relations—Maquoketa River near Green Island, Iowa; South Raccoon River at Redfield, Iowa; and West Nishnabotna River at Randolph, Iowa (USGS stations 05418720, 05484000, and 06808500, respectively). Methods for Data Collection and Computation In-stream sensors were used to record nitrate, turbidity, and temperature at 15-minute intervals during 2008–17. At selected sites with turbidity sensors, discrete water samples were collected during 2014–17 for calibration of phosphorus models. Quality-control samples were collected at all sites to assess sensor performance and consistency in data collection.

The use of continuous water-quality time-series data to compute nutrient loadings for selected Iowa streams, 2008–17