This dataset contains annual peak and seasonal maximum streamflow data from the USGS National Water Information System (NWIS) and the results of an analysis of spatial and temporal patterns for those data for the conterminous U.S. An interpretation of the analysis of these data will be published in a journal article. The streamflow data were compiled for the years 1966 to 2015 for 415 streamgages that are part of the USGS HCDN-2009 network. The HCDN-2009 network contains streamgages in watersheds with minimal anthropogenic change. The dataset contains the annual peak and the maximum daily streamflow for the months of October through December (OND), January through March (JFM), April through June (AMJ), and July through September (JAS). The maximum daily streamflow for OND, JFM, AMJ, and JAS was determined as the largest daily streamflow over those months. This dataset contains the following files for the annual peaks and maximum daily streamflow. The files labeled "raw" contain the water year and the raw annual peaks from NWIS and the selected maximum daily streamflow from NWIS. The files labeled "SPI" contain the water year and the standardized peaks and maximum daily streamflow by using the Standardized Precipitation Index method. The files labeled "cluster_assignments" contain (in columns) the gage name (two letter abbreviation of the state in which the gage is located and the USGS site number), the latitude and longitude (NAD 83) of the gage, an integer that is an identifier of the cluster that the gage was assigned, and the Pearson correlation coefficient between the standardized streamflow and the average standardized streamflow of all gages in its cluster. The annual peaks were clustered into 4 groups, the OND maximum flows were clustered into 7 groups, the JFM maximum flows were clustered into 9 groups, and the both the AMJ and JAS maximum flows were clustered into 5 groups. The files with the term "cluster_means" in the file name contain the water year and the mean of the standardized streamflow of all gages in the clusters of peaks and maximum daily streamflow for OND, JFM, AMJ, and JAS.

This data release contains results of a study investigating changes in streamflow seasonality associated with hydroclimatic variability in the north-central United States, including nine States (Illinois, Iowa, Michigan, Minnesota, Missouri, Montana, North Dakota, South Dakota, and Wisconsin). Peak-flow records from unregulated U.S. Geological Survey streamgages were used to evaluate changes in streamflow seasonality over 75-, 50-, and 30-year trend periods through water year 2020. The streamgages in each of the nine states used in the analysis and the results of the seasonal characteristics and statistical analyses are provided in tabular form (in csv file format) in file "Results.zip" under "Attached Files" below.

This data release contains results of a study investigating changes in streamflow seasonality associated with hydroclimatic variability in the north-central United States, including nine States (Illinois, Iowa, Michigan, Minnesota, Missouri, Montana, North Dakota, South Dakota, and Wisconsin). Peak-flow records from unregulated U.S. Geological Survey streamgages were used to evaluate changes in streamflow seasonality over 75-, 50-, and 30-year trend periods through water year 2020. The streamgages in each of the nine states used in the analysis and the results of the seasonal characteristics and statistical analyses are provided in tabular form (in csv file format) in file "Results.zip" under "Attached Files" below.

This dataset contains gage information for 75 Hydroclimatic Data Network-2009 basins in the conterminous United States and associated annual runoff volume, winter-spring runoff volume, and winter-spring runoff timing data 1920-2014, as well as trend results for WSCVD and WSV for periods 1920-2014, 1940-2014, and 1960-2014.

This metadata record describes the observed and estimated hydrologic metrics for the 1980 to 2019 period for U.S. Geological Survey streamgage locations across the Conterminous United States. The datasets are arranged in four files: (1) CONUS_Observed_Estimated_HMs_Annual_Monthly.csv, (2) CONUS_Bootstrap_Validations_for_Models.csv, (3) CONUS_Streamflow_Gages_for_Models.csv, and (4) Data_Dictionary_Flow_Metrics.csv. The CONUS_Observed_Estimated_HMs_Annual_Monthly.csv file contains the following six attributes: (1) the U.S. Geological Survey streamgage identification number, (2) calendar year, (3) observed hydrologic metric, (4) estimated hydrologic metric, (5) hydrologic metric abbreviation, and (6) aggregated level 2 ecoregion. The observed hydrologic metrics were calculated using collected streamflow daily values from U.S. Geological Survey streamflow gaging stations (U.S. Geological Survey National Water Information System, http://dx.doi.org/10.5066/F7P55KJN), and the estimated hydrologic metrics were estimated by cross-sectional time series random forest modeling methods by Miller, M.P., Carlisle, D.M., Wolock, D.M., and Wieczorek, M., 2018, A database of natural monthly streamflow estimates from 1950 to 2015 for the conterminous United States: Journal of the American Water Resources Association, 54(6), 1258-1269 [Also available at https://doi.org/10.1111/1752-1688.12685]. Forty-seven hydrologic metrics representing magnitude, frequency, duration, and timing were calculated. The hydrologic metric abbreviations, definitions, units, and citations are detailed in the Data_Dictionary_Flow_Metrics.csv file. The low- and high-flow magnitudes were calculated from the 10th and 90th percentile non-exceedence streamflows divided by the drainage area, respectively. The low- and high-flow frequencies were calculated as the number of pulses below the 10th and above the 90th percentile values, respectively. The low- and high-flow durations were calculated from the length of time (in days) that the streamflow was below the 10th percentile or above the 90th percentile, respectively. The low- and high-flow seasonality values were calculated based on frequency of occurrence in different seasons (for more details, please see Eng, K., Carlisle, D.M., Grantham, T.E., Wolock, D.M., and Eng, R.L., 2019, Severity and extent of alterations to natural streamflow regimes based on hydrologic metrics in the conterminous United States, 1980-2014: U.S. Geological Survey Scientific Investigations Report 2019-5001, 25 p. [Also available at https://doi.org/10.3133/sir20195001]. The CONUS_Streamflow_Gages_for_Models.csv file contains the U.S. Geological Survey list of streamflow gaging stations used in cross-sectional time series random forest models. The CONUS_Bootstrap_Validations_for_Models.csv file lists the U.S. Geological Survey streamflow gaging stations used in the bootstrapped validation data sets used to assess model performance. In addition, bootstrap validation also assesses model robustness by testing various calibration configurations. These bootstrap validation data sets may contain random amounts of observations that are outside the range of the observations used in the calibration, and/or observations that are not independent from one another. There are no missing values in any of the files. The three data files are in a comma separated value text format.

This metadata record describes the observed and estimated hydrologic metrics for the 1980 to 2019 period for U.S. Geological Survey streamgage locations across the Conterminous United States. The datasets are arranged in four files: (1) CONUS_Observed_Estimated_HMs_Annual_Monthly.csv, (2) CONUS_Bootstrap_Validations_for_Models.csv, (3) CONUS_Streamflow_Gages_for_Models.csv, and (4) Data_Dictionary_Flow_Metrics.csv. The CONUS_Observed_Estimated_HMs_Annual_Monthly.csv file contains the following six attributes: (1) the U.S. Geological Survey streamgage identification number, (2) calendar year, (3) observed hydrologic metric, (4) estimated hydrologic metric, (5) hydrologic metric abbreviation, and (6) aggregated level 2 ecoregion. The observed hydrologic metrics were calculated using collected streamflow daily values from U.S. Geological Survey streamflow gaging stations (U.S. Geological Survey National Water Information System, http://dx.doi.org/10.5066/F7P55KJN), and the estimated hydrologic metrics were estimated by cross-sectional time series random forest modeling methods by Miller, M.P., Carlisle, D.M., Wolock, D.M., and Wieczorek, M., 2018, A database of natural monthly streamflow estimates from 1950 to 2015 for the conterminous United States: Journal of the American Water Resources Association, 54(6), 1258-1269 [Also available at https://doi.org/10.1111/1752-1688.12685]. Forty-seven hydrologic metrics representing magnitude, frequency, duration, and timing were calculated. The hydrologic metric abbreviations, definitions, units, and citations are detailed in the Data_Dictionary_Flow_Metrics.csv file. The low- and high-flow magnitudes were calculated from the 10th and 90th percentile non-exceedence streamflows divided by the drainage area, respectively. The low- and high-flow frequencies were calculated as the number of pulses below the 10th and above the 90th percentile values, respectively. The low- and high-flow durations were calculated from the length of time (in days) that the streamflow was below the 10th percentile or above the 90th percentile, respectively. The low- and high-flow seasonality values were calculated based on frequency of occurrence in different seasons (for more details, please see Eng, K., Carlisle, D.M., Grantham, T.E., Wolock, D.M., and Eng, R.L., 2019, Severity and extent of alterations to natural streamflow regimes based on hydrologic metrics in the conterminous United States, 1980-2014: U.S. Geological Survey Scientific Investigations Report 2019-5001, 25 p. [Also available at https://doi.org/10.3133/sir20195001]. The CONUS_Streamflow_Gages_for_Models.csv file contains the U.S. Geological Survey list of streamflow gaging stations used in cross-sectional time series random forest models. The CONUS_Bootstrap_Validations_for_Models.csv file lists the U.S. Geological Survey streamflow gaging stations used in the bootstrapped validation data sets used to assess model performance. In addition, bootstrap validation also assesses model robustness by testing various calibration configurations. These bootstrap validation data sets may contain random amounts of observations that are outside the range of the observations used in the calibration, and/or observations that are not independent from one another. There are no missing values in any of the files. The three data files are in a comma separated value text format.

Time series of monthly minimum and maximum temperature, precipitation, and potential evapotranspiration were derived for 1,469 watersheds in the conterminous United States for which stream flow measurements were also available from the national streamflow database, termed the Hydro-Climatic Data Network (HCDN), developed by Slack et al. (1993a,b). Monthly climate estimates were derived for the years 1951-1990.The climate characteristic estimates of temperature and precipitation were estimated using the PRISM (Daly et al. 1994, 1997) climate analysis system as described in Vogel, et al. 1999.Estimates of monthly potential evaporation were obtained using a method introduced by Hargreaves and Samani (1982) which is based on monthly time series of average minimum and maximum temperature data along with extraterrestrial solar radiation. Extraterrestrial solar radiation was estimated for each basin by computing the solar radiation over 0.1 degree grids using the method introduced by Duffie and Beckman (1980) and then summing those estimates for each river basin. This process is described in Sankarasubramanian, et al. (2001). Revision Notes: This data set has been revised to update the number of watersheds included in the data set and to updated the units for the potential evapotranspiration variable. Please see the Data Set Revisions section of this document for detailed information.

Peak-flow frequency analysis is crucial in various water-resources management applications, including floodplain management and critical structure design. Federal guidelines for peak-flow frequency analyses, provided in Bulletin 17C, assume that the statistical properties of the hydrologic processes driving variability in peak flows do not change over time and so the frequency distribution of annual peak flows is stationary. Better understanding of long-term climatic persistence and further consideration of potential climate and land-use changes have caused the assumption of stationarity to be reexamined. This data release contains input data and results of a study investigating hydroclimatic trends in peak streamflow (peak flow) in the Central United States, including nine states (Iowa, Illinois, Michigan, Minnesota, Missouri, Montana, North Dakota, South Dakota, and Wisconsin). Peak flow records from unregulated U.S. Geological Survey (USGS) streamgages were used to evaluate changes over 30-, 50-, 75-, and 100-year trend periods, all ending in water year 2020. This data release contains station lists of the streamgages used in each of the nine states, the peak streamflow input data and peak streamflow analysis results, and the climate input data and climate analysis results. See "Station_Lists.zip" on the landing page for station lists (in text file format) for each state included in the study.

This dataset contains mean-annual metrics quantifying intermittent streamflow, climate, topography, land cover and geology for 540 USGS GAGES-2 watersheds experiencing non-perennial flow.

This data release contains trend results computed on the basis of modeled and observed daily streamflows at 502 reference gages across the conterminous U.S. from October 1, 1983 through September 30, 2016. Modeled daily streamflows were computed using the deterministic Precipitation Runoff Modeling System (PRMS), and five statistical techniques: Nearest-Neighbor Drainage Area Ratio (NNDAR), Map-Correlation Drainage Area Ratio (MCDAR), Ordinary Kriging of the logarithms of discharge per unit area (OKDAR), Nearest-Neighbor nonlinear spatial interpolation using flow duration curves (NNQPPQ), and Map-Correlation nonlinear spatial interpolation using flow duration curves (MCQPPQ). Observed daily streamflow data for the study gages were retrieved from the National Water Information System (NWIS). Study gages were selected from among Hydro-Climatic Data Network 2009 (HCDN-2009) gages in the GAGES-II dataset considered to be minimally affected by regulation, diversion, mining, or other anthropogenic activities. Results include trends in annual and monthly means, annual percentiles (1, 5, 10, 25, 50, 75, 90, 95, 99), annual 1-day high, 3-day high, and 7-day low, and annual snowmelt-related runoff timing for a subset of snowmelt dominated basins. Bias and volumetric efficiency statistics between observed and modeled streamflows also are provided.