Data provided in this release support the findings in Choquette et al. (2019), utilizing methods for evaluating water-quality and daily-streamflow trends described also in Hirsch and DeCicco (2015 and 2018a) and Hirsch (2018). The trend results and model-input data focus on 10 locations in the Lake Erie watershed that have long-term (20 or more years) water-quality and streamflow monitoring records. The trend results include the years 1987 through 2016 or specified sub-periods during this time frame. The model-input data records spanned the time period 1974 through 2016 although record lengths varied by site, data type, and trend analysis. The water-quality records were provided by the National Center for Water Quality Research (NCWQR; Heidelberg University, Tiffin, Ohio) and the Indiana Department of Environmental Management (IDEM), and streamflow records were provided by the U.S. Geological Survey (USGS). The 10 water-quality trend sites were identified using abbreviated names of the nearby USGS streamgage that provided streamflow data for determining nutrient fluxes at these sites (see Site_map.pdf and Site-summary_table.csv). Trends in flow-normalized nutrient fluxes were determined using the method Weighted Regression on Time, Discharge, and Season (WRTDS) method (Hirsch and DeCicco, 2015, 2018a, and 2018b) and streamflow (discharge) trends were determined using the graphical-statistical method of Quantile-Kendall plots (Hirsch, 2018). The nutrient trend analyses focus on the parameters total phosphorus (TP, as P), soluble reactive phosphorus (SRP, as P), total nitrogen (TN, as N), nitrate plus nitrite (NO23, as N) filtered at NCWQR sites or unfiltered at IDEM sites, and total Kjeldahl nitrogen (TKN, as N). TN was calculated as TKN plus NO23. SRP was monitored at only 6 of the 10 trend sites. Additional information on field and laboratory methods appears in Choquette et al. (2019). The dataset is presented in two parts: 1. Nutrient and Streamflow Model-Input Data 2. Nutrient and Streamflow Trend Results References: Choquette, A.F., Hirsch, R.M., Murphy, J.C., Johnson, L.T., and Confesor, R.B. Jr., 2019, Tracking changes in nutrient delivery to western Lake Erie: approaches to compensate for variability and trends in streamflow: J. of Great Lakes Research, v. 45, no. 1, p. 21-39, https://doi.org/10.1016/j.jglr.2018.11.012. Hirsch, R.M., 2018, Daily streamflow trend analysis: U.S. Geological Survey Office of Water Information Blog, 38 p., at: https://owi.usgs.gov/blog/Quantile-Kendall/. Hirsch, R.M., and De Cicco, L.A., 2015 (revised). User Guide to Exploration and Graphics for RivEr Trends (EGRET) and dataRetrieval: R Packages for Hydrologic Data, Version 2.0, U.S. Geological Survey Techniques Methods, 4-A10. U.S. Geological Survey, Reston, VA., 93 p. (at: http://dx.doi.org/10.3133/tm4A10). Hirsch, R.M., and De Cicco, L.A., 2018a, Guide to EGRET 3.0 Enhancements: at https://cran.r-project.org/web/packages/EGRET/vignettes/Enhancements.html. Hirsch, R.M., and De Cicco, L.A., 2018b, EGRET release 3.0, and EGRETci release 2.0, at: https://cran.r-project.org/ .

This data set includes WRTDS nutrient flux trend results and the values of daily streamflow trend results displayed in the Quantile-Kendall plots. For 1995-2015 nutrient trends, the method of generalized flow normalization (FNG) was used which explicitly addresses non-stationary streamflow conditions. For 2005-2015 nutrient trends, the WRTDS trend analyses used the method of stationary flow normalization (FNS) because streamflow nonstationarity is difficult to assess over this shorter duration time frame. The 1995-2015 annual nutrient trends were determined for all five nutrient parameters (TP, SRP, TN, NO23, TKN), and monthly trends were evaluated only for SRP. The 2005-2015 annual nutrient trends were determined for the three parameters TP, SRP, and TN. For the water-quality parameter SRP, monitoring data and trend results were available for 6 of the 10 trend sites. Daily streamflow trends were evaluated for the time-period 1987-2016 at 9 of the study sites, applied as climatic years (years which start April 1 and end March 31) for a period of 29 climatic years (1988 – 2016). Details on the WRTDS method of generalized flow normalization appear in Hirsch and De Cicco (2018) and in Choquette et al. (2019). Details on the Quantile-Kendall plots and their construction appears in Hirsch (2018) and in Appendix A of Choquette et al. (2019). Details regarding interpretations of these trend results and the watershed characteristics upstream of these sites appear in Choquette et al. (2019). The results dataset is presented in 5 parts: 1. AnnualNutrientTrends_1995-2015.csv (WRTDS nutrient flux trend results) 2. MonthlySRPTrends_1995-2015.csv (SRP monthly flux trend results for 6 sites) 3. AnnualNutrientTrends_2005-2015.csv (WRTDS nutrient flux trend results) 4. DailyFlowTrends_1987-2016.zip (Annual streamflow trend results, by site, presented in the Quantile-Kendall plots) 5. Quantile-Kendall-plots.pdf (plots showing 1988-2016 streamflow trend results) References: Choquette, A.F., Hirsch, R.M., Murphy, J.C., Johnson, L.T., and Confesor, R.B. Jr., 2019, Tracking changes in nutrient delivery to western Lake Erie: approaches to compensate for variability and trends in streamflow: J. of Great Lakes Research, v. 45, no. 1, p. 21-39, https://doi.org/10.1016/j.jglr.2018.11.012. Hirsch, R.M., 2018, Daily streamflow trend analysis: U.S. Geological Survey Office of Water Information Blog, 38 p., at: https://owi.usgs.gov/blog/Quantile-Kendall/. Hirsch, R.M., and De Cicco, L.A., 2018, EGRET release 3.0, and EGRETci release 2.0, at: https://cran.r-project.org/

This data set includes WRTDS nutrient flux trend results and the values of daily streamflow trend results displayed in the Quantile-Kendall plots. For 1995-2015 nutrient trends, the method of generalized flow normalization (FNG) was used which explicitly addresses non-stationary streamflow conditions. For 2005-2015 nutrient trends, the WRTDS trend analyses used the method of stationary flow normalization (FNS) because streamflow nonstationarity is difficult to assess over this shorter duration time frame. The 1995-2015 annual nutrient trends were determined for all five nutrient parameters (TP, SRP, TN, NO23, TKN), and monthly trends were evaluated only for SRP. The 2005-2015 annual nutrient trends were determined for the three parameters TP, SRP, and TN. For the water-quality parameter SRP, monitoring data and trend results were available for 6 of the 10 trend sites. Daily streamflow trends were evaluated for the time-period 1987-2016 at 9 of the study sites, applied as climatic years (years which start April 1 and end March 31) for a period of 29 climatic years (1988 – 2016). Details on the WRTDS method of generalized flow normalization appear in Hirsch and De Cicco (2018) and in Choquette et al. (2019). Details on the Quantile-Kendall plots and their construction appears in Hirsch (2018) and in Appendix A of Choquette et al. (2019). Details regarding interpretations of these trend results and the watershed characteristics upstream of these sites appear in Choquette et al. (2019). The results dataset is presented in 5 parts: 1. AnnualNutrientTrends_1995-2015.csv (WRTDS nutrient flux trend results) 2. MonthlySRPTrends_1995-2015.csv (SRP monthly flux trend results for 6 sites) 3. AnnualNutrientTrends_2005-2015.csv (WRTDS nutrient flux trend results) 4. DailyFlowTrends_1987-2016.zip (Annual streamflow trend results, by site, presented in the Quantile-Kendall plots) 5. Quantile-Kendall-plots.pdf (plots showing 1988-2016 streamflow trend results) References: Choquette, A.F., Hirsch, R.M., Murphy, J.C., Johnson, L.T., and Confesor, R.B. Jr., 2019, Tracking changes in nutrient delivery to western Lake Erie: approaches to compensate for variability and trends in streamflow: J. of Great Lakes Research, v. 45, no. 1, p. 21-39, https://doi.org/10.1016/j.jglr.2018.11.012. Hirsch, R.M., 2018, Daily streamflow trend analysis: U.S. Geological Survey Office of Water Information Blog, 38 p., at: https://owi.usgs.gov/blog/Quantile-Kendall/. Hirsch, R.M., and De Cicco, L.A., 2018, EGRET release 3.0, and EGRETci release 2.0, at: https://cran.r-project.org/

Trends in nutrient fluxes and streamflow for selected tributaries in the Lake Erie watershed were calculated using monitoring data at 10 locations. Trends in flow-normalized nutrient fluxes were determined by applying a weighted regression approach called WRTDS (Weighted Regression on Time, Discharge, and Season). Site information and streamflow and water-quality records are contained in 3 zipped files named as follows: INFO (site information), Daily (daily streamflow records), and Sample (water-quality records). The INFO, Daily (flow), and Sample files contain the input data, by water-quality parameter and by site as .csv files, used to run trend analyses. These files were generated by the R (version 3.1.2) software package called EGRET - Exploration and Graphics for River Trends (version 2.5.1) (Hirsch and DeCicco, 2015), and can be used directly as input to run graphical procedures and WRTDS trend analyses using EGRET R software. The .csv files are identified according to water-quality parameter (TP, SRP, TN, NO23, and TKN) and site reference number (e.g. TPfiles.1.INFO.csv, SRPfiles.1.INFO.csv, TPfiles.2.INFO.csv, etc.). Water-quality parameter abbreviations and site reference numbers are defined in the file "Site-summary_table.csv" on the landing page, where there is also a site-location map ("Site_map.pdf"). Parameter information details, including abbreviation definitions, appear in the abstract on the Landing Page. SRP data records were available at only 6 of the 10 trend sites, which are identified in the file "site-summary_table.csv" (see landing page) as monitored by the organization NCWQR (National Center for Water Quality Research). The SRP sites are: RAIS, MAUW, SAND, HONE, ROCK, and CUYA. The model-input dataset is presented in 3 parts: 1. INFO.zip (site information) 2. Daily.zip (daily streamflow records) 3. Sample.zip (water-quality records) Reference: Hirsch, R.M., and De Cicco, L.A., 2015 (revised). User Guide to Exploration and Graphics for RivEr Trends (EGRET) and dataRetrieval: R Packages for Hydrologic Data, Version 2.0, U.S. Geological Survey Techniques Methods, 4-A10. U.S. Geological Survey, Reston, VA., 93 p. (at: http://dx.doi.org/10.3133/tm4A10).

Trends in nutrient fluxes and streamflow for selected tributaries in the Lake Erie watershed were calculated using monitoring data at 10 locations. Trends in flow-normalized nutrient fluxes were determined by applying a weighted regression approach called WRTDS (Weighted Regression on Time, Discharge, and Season). Site information and streamflow and water-quality records are contained in 3 zipped files named as follows: INFO (site information), Daily (daily streamflow records), and Sample (water-quality records). The INFO, Daily (flow), and Sample files contain the input data, by water-quality parameter and by site as .csv files, used to run trend analyses. These files were generated by the R (version 3.1.2) software package called EGRET - Exploration and Graphics for River Trends (version 2.5.1) (Hirsch and DeCicco, 2015), and can be used directly as input to run graphical procedures and WRTDS trend analyses using EGRET R software. The .csv files are identified according to water-quality parameter (TP, SRP, TN, NO23, and TKN) and site reference number (e.g. TPfiles.1.INFO.csv, SRPfiles.1.INFO.csv, TPfiles.2.INFO.csv, etc.). Water-quality parameter abbreviations and site reference numbers are defined in the file "Site-summary_table.csv" on the landing page, where there is also a site-location map ("Site_map.pdf"). Parameter information details, including abbreviation definitions, appear in the abstract on the Landing Page. SRP data records were available at only 6 of the 10 trend sites, which are identified in the file "site-summary_table.csv" (see landing page) as monitored by the organization NCWQR (National Center for Water Quality Research). The SRP sites are: RAIS, MAUW, SAND, HONE, ROCK, and CUYA. The model-input dataset is presented in 3 parts: 1. INFO.zip (site information) 2. Daily.zip (daily streamflow records) 3. Sample.zip (water-quality records) Reference: Hirsch, R.M., and De Cicco, L.A., 2015 (revised). User Guide to Exploration and Graphics for RivEr Trends (EGRET) and dataRetrieval: R Packages for Hydrologic Data, Version 2.0, U.S. Geological Survey Techniques Methods, 4-A10. U.S. Geological Survey, Reston, VA., 93 p. (at: http://dx.doi.org/10.3133/tm4A10).

Phosphorus, nitrogen, chloride, and suspended-sediment loads, in 24 U.S. tributaries of the Great Lakes, were estimated using U.S. Geological Survey (USGS) data from the Great Lakes Restoration Initiative (GLRI) monitoring program for the period Oct 2010 through Sept 2023 (USGS water years 2011-23). Specific water-quality constituents include total phosphorus, orthophosphate, particulate phosphorus, total nitrogen, nitrate plus nitrite, ammonium plus ammonia, chloride, and suspended sediment. Concentrations and loads, including actual and flow-normalized estimates, were estimated with Weighted Regression on Time, Discharge, and Season (WRTDS). Results are reported at daily and annual time steps in this data release and via an interactive web application (https://rconnect.usgs.gov/glritrends). This data release is an update of a previous version for water years 2011-20. Major changes include the addition of data from water years 2021-23 and the addition of chloride to the constituents.

Phosphorus, nitrogen, and suspended-sediment loads, in 24 U.S. tributaries of the Great Lakes, were calculated using U.S. Geological Survey (USGS) data from the Great Lakes Restoration Initiative (GLRI) monitoring program for the period Oct 2010 through Sept 2020 (USGS water years 2011–2020). Total phosphorus, orthophosphate, particulate phosphorus, total nitrogen, nitrate plus nitrite, ammonium plus ammonia, and suspended-sediment loads and changes in loads were determined by applying a weighted regression approach called WRTDS (Weighted Regression on Time, Discharge, and Season). The load results represent the total mass of nitrogen, phosphorus, and suspended sediment exported downstream at each tributary site. Daily loads were estimated using the WRTDS method with Kalman filtering. To determine the change in loads between the first and last water year in record, the annual load results were flow-normalized to standardize among years with varying flow dynamics. Three primary data tables are provided that describe phosphorus, nitrogen, and suspended-sediment conditions at 24 tributaries across the U.S. Great Lakes watershed: (1) annual results, (2) daily results, and (3) changes results. The annual table is organized by USGS water year (defined as Oct 1–Sept 30). The annual and daily tables also include time-weighted mean concentrations, mean discharge, and yield estimates, the latter being calculated by dividing loads by watershed areas. Descriptions for data tables are provided in the DataDictionary.csv file. Results are also displayed in an interactive web application (https://rconnect.usgs.gov/glritrends). This data release also contains a model archive. The “model_code.zip” file contains the R code and input data to run the models, generate the results, and display the results on a shiny application. The Readme.md file includes instructions of how to run the model and descriptions of all inputs, model arguments, and outputs. Each model (saved as an eList) contains the input data (USGS station information (i.e., site information), daily discharge, and water sample data), model parameters, estimated values, and bootstrapped results. All eLists are archived in the “model_archive” folder, which is organized by sample dataset (i.e., allsamples and midsamples) and are named with the parameter and station number (e.g., NH4_04024000.RData). The RData files used in the shiny application are archived in the “shinydata” folder.

Phosphorus, nitrogen, and suspended-sediment loads, in 24 U.S. tributaries of the Great Lakes, were calculated using U.S. Geological Survey (USGS) data from the Great Lakes Restoration Initiative (GLRI) monitoring program for the period Oct 2010 through Sept 2020 (USGS water years 2011–2020). Total phosphorus, orthophosphate, particulate phosphorus, total nitrogen, nitrate plus nitrite, ammonium plus ammonia, and suspended-sediment loads and changes in loads were determined by applying a weighted regression approach called WRTDS (Weighted Regression on Time, Discharge, and Season). The load results represent the total mass of nitrogen, phosphorus, and suspended sediment exported downstream at each tributary site. Daily loads were estimated using the WRTDS method with Kalman filtering. To determine the change in loads between the first and last water year in record, the annual load results were flow-normalized to standardize among years with varying flow dynamics. Three primary data tables are provided that describe phosphorus, nitrogen, and suspended-sediment conditions at 24 tributaries across the U.S. Great Lakes watershed: (1) annual results, (2) daily results, and (3) changes results. The annual table is organized by USGS water year (defined as Oct 1–Sept 30). The annual and daily tables also include time-weighted mean concentrations, mean discharge, and yield estimates, the latter being calculated by dividing loads by watershed areas. Descriptions for data tables are provided in the DataDictionary.csv file. Results are also displayed in an interactive web application (https://rconnect.usgs.gov/glritrends). This data release also contains a model archive. The “model_code.zip” file contains the R code and input data to run the models, generate the results, and display the results on a shiny application. The Readme.md file includes instructions of how to run the model and descriptions of all inputs, model arguments, and outputs. Each model (saved as an eList) contains the input data (USGS station information (i.e., site information), daily discharge, and water sample data), model parameters, estimated values, and bootstrapped results. All eLists are archived in the “model_archive” folder, which is organized by sample dataset (i.e., allsamples and midsamples) and are named with the parameter and station number (e.g., NH4_04024000.RData). The RData files used in the shiny application are archived in the “shinydata” folder.

The United States Geological Survey’s (USGS) SPAtially Referenced Regressions On Watershed attributes (SPARROW) model was developed to aid in the interpretation of monitoring data and simulate water-quality conditions in streams across large spatial scales. SPARROW is a hybrid empirical/process-based mass balance model that can be used to estimate the major sources and environmental factors that affect the long-term supply, transport, and fate of contaminants in streams. The spatially explicit model structure is defined by a river reach network coupled with contributing catchments. The model is calibrated by statistically relating watershed sources and transport-related properties to monitoring-based streamflow and water-quality load estimates. Streamflow and load estimates considered for use in regional SPARROW model applications (2012 base year) are described in Saad and others, 2019 (https://dx.doi.org/10.3133/sir20195069). Load estimation methods described in this report include the Beale Ratio Estimator and Fluxmaster models. This USGS data release contains all of the input and output files necessary to reproduce the load estimates considered for inclusion in the 2012 regional SPARROW models. Data preparation for input to the load estimation models is also fully described in the above-mentioned report.

The United States Geological Survey’s (USGS) SPAtially Referenced Regressions On Watershed attributes (SPARROW) model was developed to aid in the interpretation of monitoring data and simulate water-quality conditions in streams across large spatial scales. SPARROW is a hybrid empirical/process-based mass balance model that can be used to estimate the major sources and environmental factors that affect the long-term supply, transport, and fate of contaminants in streams. The spatially explicit model structure is defined by a river reach network coupled with contributing catchments. The model is calibrated by statistically relating watershed sources and transport-related properties to monitoring-based streamflow and water-quality load estimates. Streamflow and load estimates considered for use in regional SPARROW model applications (2012 base year) are described in Saad and others, 2019 (https://dx.doi.org/10.3133/sir20195069). Load estimation methods described in this report include the Beale Ratio Estimator and Fluxmaster models. This USGS data release contains all of the input and output files necessary to reproduce the load estimates considered for inclusion in the 2012 regional SPARROW models. Data preparation for input to the load estimation models is also fully described in the above-mentioned report.