Lauren E. Koenig Janet R. Barclay Alison P. Appling 20250915 tabular digital data Online U.S. Geological Survey https://doi.org/10.5066/P14GXPOW This data release contains input data used to test multi-scale modeling approaches for predicting water temperature in streams and rivers. The target application aims to predict stream temperature at relatively fine spatial resolution across a river basin by combining a machine learning (ML) model (optionally, with additional input data) configured at a relatively coarse spatial resolution with an ML model and additional input data at the fine resolution. The fine- and coarse- spatial resolutions were represented by two common river hydrography datasets. In each of 12 focal river basins, the coarse spatial resolution was represented by the Geospatial Fabric for Hydrologic Modeling (GFv1.1; Bock et al. 2020) and the fine spatial resolution was represented by the NHDPlusv2.1 flowline network. The model inputs include water temperature observations as well as information about the spatial relationships among river segments; stream and catchment characteristics; and daily meteorological drivers. Certain input datasets were prepared only for the fine spatial resolution (i.e., NHDPlusv2.1; flowline distance matrix, water temperature observations) to enable development of fine-resolution models for each basin, whereas other datasets were prepared at both spatial resolutions to provide additional data to inform the multi-scale modeling experiments (river attributes, meteorological data). The 12 focal river basins (436 - 958 km^2) are distributed across the conterminous United States and include Battle Creek, CA; Black Earth Creek, WI; Brandywine Creek, PA and DE; East River, CO; the Lower Delaware River, PA and NJ; the Lower West Branch Delaware River, PA and NY; Manhan River, MA; Neversink River, NY; Rancocas Creek, NJ; the South Fork McKenzie River, OR; Trinity River, TX; and the Upper South River, GA. This data release includes nine files that contain the model input data at one or both of the spatial resolutions described above for each river basin: 1. nhdv2_distance_matrix.npz: File includes a matrix indicating the upstream distances among flow-connected segments in the NHDPlusv2.1 river network. 2. nhdv2_nhgf_crosswalk.csv: File includes a crosswalk table that maps NHDPlusv2.1 flowlines to GFv1.1 flowline segments. 3. nhdv2_temp_observations.parquet: File includes water temperature observations summarized to daily values and aggregated to the NHDPlusv2.1 flowlines. 4. nhdv2_static_attributes.parquet: File includes attribute features that represent characteristics of the river segment, its catchment, or the upstream watershed area, summarized to the NHDPlusv2.1 flowlines. 5. nhgf_static_attributes.parquet: File includes attribute features that represent characteristics of the river segment, its catchments, or the upstream watershed area, summarized to the GFv1.1 segments. 6. nhdv2_inputs_io.zip: Zipped parquet file contains model input data including river segment characteristics and daily meteorological data, summarized to the NHDPlusv2.1 flowlines. 7. nhgf_inputs_io.zip: Zipped parquet file contains model input data including river segment characteristics and daily meteorological data, summarized to the GFv1.1 segments. 8. spatial_data.gpkg: File contains the spatial data representing the basin boundaries, the GFv1.1 segments and catchments, and the NHDPlusv2.1 flowlines and catchments. 9. source_code.zip: Compressed file contains R code used to generate the model input datasets. These data were created for water quality modeling methods research and development. 19790101 20211231 observed Complete None planned Conterminous United States -122.2934 -72.4015 44.1619 32.6125 None machine learning deep learning water resources water temperature modeling downscaling rivers USGS Metadata Identifier USGS:67d34d8cd34e1acf3979d0a9 None Manhan River Neversink River Upper South River Black Earth Creek Trinity River East River South Fork McKenzie River Battle Creek Brandywine Creek Lower Delaware River Lower West Branch Delaware River Rancocas Creek None. Please see 'Distribution Info' for details. None. Users are advised to read the dataset's metadata thoroughly to understand appropriate use and data limitations. Lauren E Koenig U.S. Geological Survey Data Scientist mailing

425 Jordan Road

Troy NY 12180 518-285-5695 lkoenigsnyder@usgs.gov This data compilation was supported by the U.S. Department of Energy, Office of Science, Office of Biological and Environmental Research, Environmental System Science Data Management Program, as part of the ExaSheds project, under Award Number 89243021SSC000068. Work was also supported by the U.S. Geological Survey, Water Availability and Use Science Program. These data were generated using a Linux computer system and the open-source R and Python programming languages. The R targets library (v1.3.0; Landau 2021) was used as a workflow manager system to facilitate efficient dataset preparation. A complete description of R libraries and versions used is provided within the "docker" subdirectory of the source_code.zip folder. Andrew R. Bock Marilyn Santiago Michael E. Wieczorek Sydney S. Foks Melissa A. Lombard 2020 ver. 3.0, November 2021 vector digital data https://www.sciencebase.gov U.S. Geological Survey GIS Features of the Geospatial Fabric for the National Hydrologic Model, version 1.1, GFv1.1.gdb.zip https://doi.org/10.5066/p971jagf Janet R. Barclay John R. Mullaney 2021 publication Reston, Virginia US Geological Survey https://doi.org/10.3133/sir20215116 William Landau 20210115 v1.3.0 publication Journal of Open Source Software vol. 6, issue 57 https://doi.org/10.21105/joss.02959 In this data release, two datasets contain information about 1) the spatial relationships among river segments within the NHDPlusv2.1 hydrography, and 2) the spatial relationships between segments represented in the NHDPlusv2.1 and GFv1.1 datasets, respectively. A matrix containing the upstream river distance between NHDPlusv2.1 flowlines encodes the spatial relationships among flowlines in the network. The river distances (i.e., the flow-connected, upstream distance between pairs of flowline outlets) were calculated using the flowline lengths reported in the NHDPlusv2.1 dataset. The calculated river distances were manually checked and confirmed for a few selected flowlines from one focal river basin (the Neversink River); otherwise, no formal attribute accuracy tests were conducted. The river distances reported in the nhdv2_distance_matrix.npz file rely on network attributes from the NHDPlusv2.1 flowline dataset to construct the river-network topology. Any errors in the NHDPlusv2.1 network attributes (i.e., "fromnode", "tonode", "lengthkm") will therefore affect the accuracy of the calculated river distances. Next, a crosswalk table represents the spatial relationships between segments represented in the two river hydrography datasets. For each river basin, results from the national NHDPlusv2.1 - GFv1.1 crosswalk table (Koenig et al. 2025) were visually inspected using maps of the crosswalked river segments to ensure suitability for the target application. Several steps were taken to ensure the quality of the observational site matches to the NHDPlusv2.1 hydrography. For the national water temperature dataset (Oliver et al. 2024), monitoring sites were omitted if they were located more than 200 meters from the assigned NHDPlusv2.1 flowline. This cutoff distance of 200 meters was selected to balance the tradeoff between using a too-wide radius that may improperly attribute site locations to a nearby flowline, and a too-narrow radius that could exclude sites with locations that do not directly intersect the NHDPlusv2.1 hydrography. For other water temperature datasets in this compilation that included fewer monitoring sites (CDEC; Neversink NGWOS, Terry et al. 2022; Carroll et al. 2023), the site-NHDPlusv2.1 flowline matches were visually inspected as a quality control measure. Additional quality control methods applied to the individual water temperature datasets are described in Process Step 3. When available and provided by the data contributor, daily time series data were used in lieu of aggregating sub-daily water temperature data as part of this data compilation. Any errors or inconsistencies in the provider-reported daily summary values will therefore affect the accuracy of the daily minimum, mean, and/or maximum water temperatures. No formal attribute accuracy tests were applied to the river attribute and meteorological driver datasets. Datasets were checked for logical consistency. Note that one river attribute variable representing river channel confinement only contains data for seven of the 12 focal river basins. Across the 12 focal basins, the river network was represented by 131 GFv1.1 segments and 3,469 NHDPlusv2.1 flowlines. Of these, six GFv1.1 segments and 91 NHDPlusv2.1 flowlines were missing associated catchments in the source data (GFv1.1 or NHDPlusv2.1, respectively). To ensure completeness, river attribute data and meteorological driver data were imputed for any flowlines lacking catchments by assuming equivalent conditions between the flowline in question and the nearest neighboring flowline with data. Andrew R. Bock Marilyn Santiago Michael E. Wieczorek Sydney S. Foks Melissa A. Lombard 2020 ver. 3.0, November 2021 vector digital data https://www.sciencebase.gov U.S. Geological Survey GIS Features of the Geospatial Fabric for the National Hydrologic Model, version 1.1, GFv1.1.gdb.zip https://doi.org/10.5066/p971jagf Digital and/or Hardcopy 20211129 publication date GFv1.1 The GFv1.1 river flowline vector data were used to represent the coarse-resolution river hydrography. U.S. Environmental Protection Agency (USEPA) U.S. Geological Survey (USGS) 2015 vector digital data https://www.epa.gov/waterdata/nhdplus-national-hydrography-dataset-plus 100000 Digital and/or Hardcopy 20150101 publication date NHDPlusv2.1 NHDFlowline feature class vector data were used to represent the fine-resolution river hydrography. Michael E. Wieczorek Shannon E. Jackson Gregory E. Schwarz 2018 v4.0, August 2023 tabular digital data https://www.sciencebase.gov U.S. Geological Survey https://doi.org/10.5066/f7765d7v Digital and/or Hardcopy 20230802 publication date Wieczorek et al. 2018 River and catchment characteristics used as feature attributes in machine learning modeling applications. Lauren E. Koenig Alison P. Appling Margaux J. Sleckman 2025 tabular digital data https://www.sciencebase.gov U.S. Geological Survey https://doi.org/10.5066/p14dcshv Digital and/or Hardcopy 20250314 publication date Koenig et al. 2025 The national NHDPlusv2.1 - GFv1.1 crosswalk table was subsetted for use in Process Step 1 of the data compilation workflow. United States Geological Survey 2016 tabular digital data http://waterdata.usgs.gov/nwis/ U.S. Geological Survey https://doi.org/10.5066/f7p55kjn Digital and/or Hardcopy 20240514 20240723 publication date USGS NWIS Field measurements of river channel width at USGS streamflow gages. Samantha K. Oliver Alison P. Appling David D. Watkins Rasha A. Atshan Jordan S. Read 2024 tabular digital data https://www.sciencebase.gov U.S. Geological Survey https://doi.org/10.5066/p9emwz35 Digital and/or Hardcopy 20241015 publication date Oliver et al. 2024 Water temperature observations for all 12 focal river basins. Neil C. Terry Martin Briggs Dana M. Kushner Hallie E. Dickerson Andrew Baldwin Brett R. Trottier Adam B. Haynes Charlie Besteder Robin L. Glas Daniel H. Doctor Christopher L. Gazoorian William E. Odom Joshua R. Benton 2022 tabular digital data https://www.sciencebase.gov U.S. Geological Survey https://doi.org/10.5066/p9r3tyoz Digital and/or Hardcopy 20230428 publication date Neversink NGWOS Daily mean water temperatures for monitoring sites in the Upper Neversink River basin (from the WTmatrix.csv file). Rosemary Carroll Alexander Newman Curtis Beutler Kenneth Williams Dylan O'Ryan 2023 tabular digital data Environmental System Science Data Infrastructure for a Virtual Ecosystem (ESS-DIVE) U.S. DOE Lawrence Berkeley National Laboratory Watershed Function Science Focus Area https://doi.org/10.15485/1779721 Digital and/or Hardcopy 2023 publication date Carroll et al. 2023 Daily mean water temperatures for monitoring sites in the East River basin. California Department of Water Resources 20241104 tabular digital data California Data Exchange Center California Department of Water Resources https://cdec.water.ca.gov Digital and/or Hardcopy 19991229 20230930 observed CDEC Water temperature data collected and published by the CA Department of Water Resources for two monitoring locations: NORTH FORK BATTLE CREEK NEAR MANTON (BNF) and SOUTH FORK BATTLE CREEK NEAR MANTON (BAS). John T. Abatzoglou 20111221 raster digital data International Journal of Climatology vol. 33, issue 1 online Wiley https://doi.org/10.1002/joc.3413 Digital and/or Hardcopy 20111221 publication date gridMET Source of gridded surface meteorological data for each date between 1979-01-01 and 2021-12-31. Ryan McManamay Chris DeRolph 2019 tabular digital data online Figshare https://doi.org/10.6084/m9.figshare.c.4233740.v1 Digital and/or Hardcopy 20190207 publication date McManamay and DeRolph 2019 Information used to estimate river channel confinement for individual NHDPlusv2.1 flowlines. David L. Blodgett 2023 ver. 2.0, February 2023 vector digital data https://www.sciencebase.gov U.S. Geological Survey https://doi.org/10.5066/p92u7zut Digital and/or Hardcopy 20230201 publication date WBD_20201006 Source data for vector polygons representing HUC12 boundaries, based on a snapshot of the Watershed Boundary Dataset from October 2020. HUC12 polygons were retrieved from a web service using the get_huc function from the nhdplusTools R library. Step 1: Format river-network data. The focal river basins were defined using lists of hydrologic unit codes that represent the HUC12 units from a snapshot version of the Watershed Boundary Dataset (WBD_20201006). The HUC12 spatial data were retrieved using the get_huc function from the nhdplusTools R library (version 0.7.1) and then combined to define the spatial boundaries of each river basin. This workflow assembles data at two spatial resolutions represented by different river hydrography datasets. The NHDPlusv2.1 dataset represents a relatively fine-resolution version of the river network relative to the Geospatial Fabric (GFv1.1) dataset, which captures the mainstem streams and rivers of the United States and so has fewer segments. To prepare data at both resolutions, the relevant spatial datasets were downloaded and formatted. The national-extent geodatabase file containing the GFv1.1 spatial data was downloaded from ScienceBase (GFv1.1.gdb.zip; Bock et al. 2020) using the sbtools R library (version 1.2.0). The file was unzipped, and both the river flowline (nsegment_v1_1) and hydrologic response unit (nhru_v1_1) vector classes were isolated for further processing. The national-extent flowline and hydrologic response units (HRU) data were each subsetted to the focal river basins for use in the remainder of the data compilation workflow. A river-network crosswalk table that links GFv1.1 segments to NHDPlusv2.1 flowlines was used to define the list of unique NHDPlusv2.1 identifiers to retrieve for each basin. The crosswalk table was developed by subsetting a national crosswalk dataset (Koenig et al. 2025) that indicates the NHDPlusv2.1 flowlines that correspond to a given GFv1.1 segment (i.e., “on-network” flowlines). Within this workflow, additional information was added for each basin to also indicate the NHDPlusv2.1 flowlines that drain to a given GFv1.1 segment. Those NHDPlusv2.1 flowlines that are within the basin boundaries but do not intersect a GFv1.1 segment were considered “off-network” flowlines. For each river basin, the unique NHDPlusv2.1 common identifiers (i.e., comid) from the expanded crosswalk table were used to retrieve the flowline and catchment spatial data using the get_nhdplus function from the nhdplusTools R library. A river distance matrix was created to represent the upstream and downstream distances among flow-connected NHDPlusv2.1 flowlines. First, the network topology was constructed using the get_tocomid function in the nhdplusTools R library. The distance matrix was calculated by formatting the network vertices and edges to facilitate representation of the NHDPlusv2.1 flowline network as a graph using the igraph R library (version 1.4.1). The flowline length attribute (“lengthkm”) in NHDPlusv2.1 was used to calculate the upstream and downstream river distances (in meters) along the directed graph network. Although multiple configurations of the river distance matrix were created to represent the upstream, downstream, and bidirectional river distances, respectively, only the upstream river distances were saved and included in this data release. Additional river-network data were created in preparation for subsequent steps of the data compilation workflow, including river width and GFv1.1-NHDPlusv2.1 spatial weights. First, median river width was estimated for each NHDPlusv2.1 flowline by fitting an empirical relationship between channel width measured at USGS streamflow gages and the total length of upstream flowlines (i.e., the “arbolate_sum”; NHDPlusv2.1), following Barclay and Mullaney (2021). Each river basin was mapped to its corresponding HUC4 spatial unit for which all USGS streamflow gages were identified and matched to the nearest NHDPlusv2.1 flowline. Only gages with site type “stream” (USGS NWIS) and those located on dendritic flowlines (i.e., “streamorde” equals “streamcalc”; NHDPlusv2.1) were retained in the analysis because divergent river channels may have widths that deviate from basin-average geomorphic scaling patterns. For each gage, channel width measurements from the earliest available date to 2023-11-01 were downloaded using the dataRetrieval R library (version 2.7.13) and used to calculate the median channel width for that site. For each HUC4 unit, linear regression was used to find the best-fit relationship between the log-transformed site median width and the log-transformed arbolate sum values. The number of gages used to compute this relationship ranged from 63 in the East River basin (HUC4: 1402) to 496 for the Neversink River basin (HUC4: 0204), and the coefficient of determination (r^2) that describes the goodness-of-fit to the empirical data ranged from 0.47 for Battle Creek (HUC4: 1802) to 0.86 for the South Fork McKenzie River basin (HUC4: 1709). The best-fit relationship was then used to predict median width across all NHDPlusv2.1 flowlines within each river basin corresponding to a given HUC4 spatial unit. Finally, a dataset of spatial weights was created to describe the extent of overlap between GFv1.1 catchments and NHDPlusv2.1 catchments. GFv1.1 HRUs do not necessarily represent complete catchments that drain to a given GFv1.1 segment; in some cases, they may represent the land surface that drains to the left-bank or right-bank of a river segment, respectively. Therefore, any GFv1.1 HRUs that drain to the same GFv1.1 segment - inferred using the attribute "hru_segment_v1_1" from the GFv1.1 dataset - were combined to represent the complete catchment. The intersecting area between the newly created GFv1.1 catchments and the NHDPlusv2.1 catchments (in square kilometers) was computed using the calculate_area_intersection_weights function from the ncdfgeom R library (version 1.2.0). The data processing steps described above for Step 1 generally correspond to the 1_network.R file within the source_code.zip file. GFv1.1 NHDPlusv2.1 Koenig et al. 2025 USGS NWIS WBD_20201006 20241115 Step 2: Prepare water temperature observation data. Water temperature observations were compiled from four data sources: a national, multi-agency water temperature dataset (Oliver et al. 2024) and three sources of local data that were used to augment the national dataset in the Neversink River, East River, and Battle Creek basins, respectively. The national water temperature data and site locations were downloaded from Oliver et al. (2024). The national dataset includes an attribute that links each site to the best-match GFv1.1 segment but does not match the sites to the NHDPlusv2.1 flowline network. Therefore, each monitoring location was matched to a NHDPlusv2.1 flowline within this data compilation workflow. The sites were matched to NHDPlusv2.1 by minimizing the river distance between the site location and the downstream vertex of nearby NHDPlusv2.1 flowlines. This method was selected over alternatives (e.g., matching sites to the nearest flowline) because it replicates the methodology used by Oliver et al. (2024) and is considered more representative for machine learning modeling applications that predict water temperature at segment outlets. To increase the likelihood that sites were appropriately matched to the NHDPlusv2.1 hydrography, sites farther than 200 meters from the linked NHDPlusv2.1 flowline were dropped from the dataset. Daily water temperature records with undesired QC flags that may indicate poor data quality were also dropped from the dataset. Additional water temperature data collected by the USGS in the Upper Neversink River were used to augment the national dataset in the Neversink River basin. Site coordinates for the NGWOS Neversink monitoring locations were formatted as a spatial vector dataset and matched to NHDPlusv2.1 flowlines using the same methodology described above for the national dataset. The NGWOS Neversink dataset contains several monitoring sites located on small, headwater streams that are not represented in either the GFv1.1 or NHDPlusv2.1 hydrographic datasets. To increase the likelihood that monitoring sites are actually located along the best-match NHDPlusv2.1 flowlines, sites farther than 200 meters from the linked NHDPlusv2.1 flowline were omitted from the dataset. For the NGWOS Neversink data, all site locations and matches to NHDPlusv2.1 were visually inspected for quality control. Sub-daily water temperature data were summarized to daily mean values by the data provider (WTmatrix.csv file). Within this data compilation workflow, any daily mean temperature values less than -0.5 degrees Celsius or greater than 35 degrees Celsius were dropped from the dataset. This range of plausible temperature values was selected to maximize the likelihood that the reported water temperature is representative of in-stream conditions (e.g., the water temperature sensor is submerged and not exposed to air), and to maintain consistency with the accepted temperature range used by Oliver et al. (2024). The filtered data were further formatted for consistency with the other water temperature datasets. In the East River basin, additional, local water temperature datasets collected between 2019 and 2022 were leveraged from Carroll et al. (2023). Similar to other temperature datasets, each monitoring location from Carroll et al. (2023) was matched to the nearest NHDPlusv2.1 outlet, or the vertex at the downstream end of the flowline. All nine monitoring sites were visually inspected for quality control, which resulted in modifying the provider-reported site coordinates (and resulting best-match NHDPlusv2.1 flowline) for one site, Gunsight Bridge (GSB). The daily mean water temperature values summarized by the data provider were extracted from the individual Mean_Daily files, formatted for consistency, and inspected for quality control. Based on quality control checks, water temperature data were dropped from one site and date (GSB; 2021-12-03) where it appeared the summarized value had included an erroneous sub-daily water temperature measurement (-511 degrees Fahrenheit) that skewed the reported daily mean water temperature value. 16 site-dates were also dropped because the daily mean temperature was outside of the accepted range of -0.5 - 35 degrees Celsius. Otherwise, all data were retained but some records were flagged for additional quality control checks that could be considered in downstream modeling applications. Finally, water temperature data were collected from two sites in the Battle Creek basin monitored by the California Department of Water Resources (sites BAS and BNF). The site data and hourly water temperature data were queried from the California Data Exchange Center (CDEC) starting with the earliest available date (BNF: 1999-12-29; BAS: 2000-07-19) through 2023-09-30. The monitoring locations were matched to NHDPlusv2.1 flowlines by minimizing the river distance between each site and the flowline outlet. Hourly water temperature data were inspected for quality control prior to summarizing the data to daily mean, minimum, and maximum values. Any hourly water temperature records with values less than –100 degrees Fahrenheit or greater than 100 degrees Fahrenheit were flagged as implausible. In addition, rolling 24-hour means were calculated for each site and hourly records were flagged as unexpectedly high for that date (or unexpectedly low) if water temperature was greater than 75 degrees Fahrenheit (or lower than 32 degrees Fahrenheit) and differed from the mean water temperature over the previous 24 hours by more than 15 degrees Fahrenheit. Additional site date-ranges were manually flagged as suspect, for example, because of incomplete and noisy intervals or because of extended periods of water temperatures equal to zero degrees Fahrenheit during the summer and fall months (BAS: 2007-11-19 – 2007-11-20; BNF: 2022-07-13 – 2022-10-14, 2023-04-10 – 2023-04-13, 2014-12-15 – 2015-05-08, 2022-12-27 – 2023-04-11). All flagged hourly records were dropped from the dataset. The remaining temperature records were converted to degrees Celsius and for days with at least 10 hourly observations, daily minimum, mean, and maximum water temperatures were computed using the sub-daily data. The daily summary values were filtered to omit values outside of the accepted range of -0.5 - 35 degrees Celsius. Daily water temperatures from each of the datasets described above were combined and summarized to NHDPlusv2.1 flowlines so that the model-ready dataset consists of daily minimum, mean, and maximum values for each date and NHDPlusv2.1 flowline comid. Note that these daily minimum and maximum estimates incorporate both spatial variability (among sites) and temporal variability (within sites) and so cannot be interpreted strictly as a diel range in water temperature for that flowline. All water temperature observation data prior to 1979-01-01 were dropped from the dataset to match the earliest possible starting temporal domain of target modeling applications. Although the target modeling application has an anticipated end date of 2021-12-31, temperature data after this date were retained to maintain the possibility of extending the model temporal domain as new input datasets permit. The data processing steps described above for Step 2 generally correspond to the 2_observations.R file within the source_code.zip file. Oliver et al. 2024 Neversink NGWOS Carroll et al. 2023 CDEC GFv1.1 NHDPlusv2.1 20250603 Step 3: Compile and process model inputs and river attributes. This step generated 1) model input data that were needed to predict water temperatures in downstream modeling applications as well as 2) river- and catchment-attribute features that capture variability in annual climate, geology, hydrology, and land cover among river segments. Model input and attribute data were generated for two spatial resolutions represented by the NHDPlusv2.1 and GFv1.1 datasets, respectively. The model inputs included river characteristics - such as river width, slope, and elevation that were assumed to be static or unchanging throughout the entire temporal domain - in addition to select meteorological variables that were hypothesized to influence river water temperatures and varied from day to day. For the NHDPlusv2.1 flowlines, the static river characteristics data were generated by appending river width estimated in Step 1 to value-added attributes available in NHDPlusv2.1, including minimum and maximum elevation (converted to meters) and slope. The NHDPlusv2.1 river characteristics data were then reformatted to generate values representative of the GFv1.1 segments. For each GFv1.1 segment, slope was calculated as the length-weighted mean slope of all intersecting NHDPlusv2.1 flowlines that comprise that GFv1.1 segment (i.e., “on-network" flowlines inferred using the crosswalk table generated in Step 1). The GFv1.1 segment minimum elevation was taken as the minimum elevation across all on-network NHDPlusv2.1 flowlines, and segment width was assumed equal to the maximum river width across all on-network NHDPlusv2.1 flowlines. At both spatial resolutions, meteorological conditions were represented using data from gridMET, a dataset of daily, high-spatial resolution (approximately 4 km) surface meteorological data for the conterminous United States (Abatzoglou 2011). gridMET variables representing daily precipitation, minimum and maximum air temperature, downward surface shortwave radiation, evapotranspiration, and minimum and maximum relative humidity were retrieved for each day between 1979-01-01 and 2022-12-31 using the getGridMET function from the climateR R library (version 0.3.5). The downloaded data were aggregated to the NHDPlusv2.1 catchments and GFv1.1 catchments (see Step 1), respectively, by calculating an area-weighted mean of the grid cell values. The aggregated meteorological data were formatted, including converting temperature units from degrees Kelvin to degrees Celsius and converting precipitation accumulation to units of meters. Daily mean air temperature and daily mean relative humidity were calculated by taking the average of the daily minimum and maximum values, respectively. For both the NHDPlusv2.1 flowlines and the GFv1.1 segments, the dynamic meteorological data were combined with the river characteristics values described above to create a dataset of daily model inputs for each NHDPlusv2.1 flowline or GFv1.1 segment and each date over the temporal domain of the target modeling application (1979-01-01 to 2021-12-31). In the model inputs data, the river characteristics values for each segment were held constant throughout the temporal sequence. Meteorological data or slope estimates were not available for all segments using the workflow described above because, for example, some NHDPlusv2.1 flowlines lack associated catchments. As a requirement of target modeling applications, the model inputs dataset cannot have missing values, and so any missing data were imputed using values from the nearest flow-connected segment with data. A flag variable was appended to the dataset to indicate imputed values. In addition to the model input variables, a separate set of attribute variables representing river, catchment, and upstream-watershed characteristics were compiled for each NHDPlusv2.1 flowline and each GFv1.1 segment within the focal river basins. Of 164 attribute features, 162 were compiled from Wieczorek et al. (2018) and included characteristics related to river geomorphology, catchment geology, hydrology, soil characteristics, and riparian and catchment land cover. A list of selected attributes and their descriptions is included in the source_code.zip file (3_attributes/in/nhdv2_attribute_variables.csv). For each NHDPlusv2.1 flowline, the selected attribute variables were downloaded using the get_catchment_characteristics function in the nhdplusTools R library and formatted for consistency (e.g., by replacing –9999 or –9998 values used to represent missing data). An additional variable representing the area-normalized upstream reservoir storage (“TOT_NORM_STORAGE2013_AREANORM”) was computed by dividing the normalized reservoir storage associated with dams built during or before 2013 (“TOT_NORM_STORAGE2013”) by the cumulative catchment area throughout the upstream watershed (“TOT_BASIN_AREA”). An additional attribute feature represented estimates of river channel confinement, which was calculated as the ratio between floodplain width and river channel width. The requisite width data were garnered from McManamay and DeRolph (2019), which includes valley confinement data for NHDPlusv2 flowlines spanning the conterminous United States as part of a broader stream classification system. Using the valley confinement dataset, river width (in meters) was calculated from the flowline area and flowline length, and floodplain width (in meters) was calculated by dividing the valley bottom area by the valley bottom length for all flowlines where the valley bottom length was greater than zero. The resulting widths for some NHDPlusv2 flowlines were deemed implausibly small and so all widths were set to a minimum value of one meter prior to calculating channel confinement. Missing confinement values were imputed using the nearest neighboring segment with a non-missing confinement value. A flag variable was appended to the feature attributes dataset to indicate imputed values. Note that channel confinement was estimated for any NHDPlusv2.1 COMIDs represented in the "East" region in McManamay and DeRolph (2019), which included seven of the 12 focal river basins in this data release. No confinement estimates were made for basins within the other three regions in McManamay and DeRolph (2019) due to inconsistent formatting in the valley confinement datasets sourced for this data processing step which resulted in lower confidence in their interpretation. To generate comparable attribute data for the GFv1.1 segments, the NHDPlusv2.1 values were rescaled by using the crosswalk table and spatial intersection weights generated in Step 1 to calculate the sum, area-weighted mean, or length-weighted mean value based on the NHDPlusv2.1 flowlines that drain to a given GFv1.1 segment (sum, area-weighted mean) or overlap a given GFv1.1 segment (length-weighted mean). To represent characteristics that describe the full watershed upstream of a given segment, the total upstream routed accumulated values from Wieczorek et al. (2018) were subset based on the NHDPlusv2.1 flowline located at the downstream end of each GFv1.1 segment. The data processing steps described above for Step 3 generally correspond to the 3_attributes.R file within the source_code.zip file. gridMET Wieczorek et al. 2018 McManamay and DeRolph 2019 NHDPlusv2.1 GFv1.1 20250205 Vector 0.0197552932 0.0250859083 Decimal seconds North_American_Datum_1983 GRS 1980 6378137.0 298.257222101 nhdv2_distance_matrix.npz File includes a matrix indicating the upstream distances among flow-connected flowlines in the NHDPlusv2.1 river network. The file contains 3,469 rows and 3,469 columns, one for each NHDPlusv2.1 flowline that is connected to at least one other flowline in any of the 12 focal river basins. U.S. Geological Survey NULL The first column contains the unique, common identifier (comid) of the starting NHDPlusv2.1 flowline for the distance calculation. The flowline identifier is equal to the comid attribute from the NHDPlusv2.1 dataset. The integer values for this column represent unique identifiers and do not have units. NHDPlusv2.1 1259243 948020795 NA NULL After the first column, all other columns are named according to the unique, common identifier (comid) of the ending NHDPlusv2.1 flowline in the pairwise distance calculation. The flowline identifier is equal to the comid attribute from the NHDPlusv2.1 dataset. For ending NHDPlusv2.1 flowlines that are connected to and are located upstream of the starting NHDPlusv2.1 flowline, the column values provide the river distance - in the upstream direction - from the outlet of the starting flowline to the outlet of the ending flowline. U.S. Geological Survey Inf Inf indicates either: 1) the two flowlines are not connected by streamflow within the NHDPlusv2.1 network, or 2) the ending NHDPlusv2.1 flowline is located downstream of the starting NHDPlusv2.1 flowline in the pairwise distance calculation. U.S. Geological Survey 0 88277 meters nhdv2_nhgf_crosswalk.csv Comma-separated file (CSV) containing a crosswalk table that maps NHDPlusv2.1 flowlines to GFv1.1 flowline segments. U.S. Geological Survey basin River basin name U.S. Geological Survey BattleCreek Data pertains to the Battle Creek basin, CA. U.S. Geological Survey BlackEarth Data pertains to the Black Earth Creek basin, WI. U.S. Geological Survey Brandywine Data pertains to the Brandywine Creek basin in PA and DE. U.S. Geological Survey EastRiver Data pertains to the East River basin, CO. U.S. Geological Survey LowerDelaware Data pertains to the Lower Delaware River basin in PA and NJ. U.S. Geological Survey LowerWestBranchDelaware Data pertains to the Lower West Branch Delaware River basin in PA and NY. U.S. Geological Survey ManhanConnecticut Data pertains to the Manhan River basin, MA. U.S. Geological Survey Neversink Data pertains to the Upper Neversink River basin, NY. U.S. Geological Survey Rancocas Data pertains to the Rancocas River basin, NJ. U.S. Geological Survey SouthForkMcKenzie Data pertains to the South Fork McKenzie River basin, OR. U.S. Geological Survey TrinityHeadwaters Data pertains to the Trinity Headwater basin, TX. U.S. Geological Survey UpperSouth Data pertains to the Upper South River basin, GA. U.S. Geological Survey seg_id_nat The GFv1.1 segment identifier. The segment identifier is equal to the nsegment_v1_1 attribute from the GFv1.1 dataset. U.S. Geological Survey 713 53746 PRMS_segid The alternative GFv1.1 segment identifier. The alternative segment identifier is equal to the seg_id_nhm attribute from the GFv1.1 dataset. U.S. Geological Survey 1491 57013 COMID Unique flowline common identifier (comid) taken from the NHDPlusv2.1 dataset. NHDPlusv2.1 1259243 948020795 comid_on_nhm Logical variable indicating whether the NHDPlusv2.1 flowline intersects the GFv1.1 network and is considered "on-network." U.S. Geological Survey False The NHDPlusv2.1 flowline drains to the GFv1.1 segment but does not intersect it, i.e., the NHDPlusv2.1 flowline is an "off-network" flowline. U.S. Geological Survey True The NHDPlusv2.1 flowline intersects the NHDPlusv2.1 flowline, i.e., the NHDPlusv2.1 flowline is an "on-network" flowline. U.S. Geological Survey network_ref Character value that describes whether the NHDPlusv2.1 flowline intersects the GFv1.1 segment or just drains to it. U.S. Geological Survey NHGF off-network The NHDPlusv2.1 flowline drains to the GFv1.1 segment but does not intersect it, i.e., the NHDPlusv2.1 flowline is an "off-network" flowline. Attribute comid_on_nhm is FALSE. U.S. Geological Survey NHGF on-network The NHDPlusv2.1 flowline intersects the GFv1.1 segment, i.e., the NHDPlusv2.1 flowline is an "on-network" flowline. Attribute comid_on_nhm is TRUE. U.S. Geological Survey comid_order_down_to_up Integer value indicating the relative order of each on-network NHDPlusv2.1 flowline along its corresponding GFv1.1 segment, starting from the downstream end of the GFv1.1 segment and moving upstream. U.S. Geological Survey NA The NHDPlusv2.1 flowline does not intersect the GFv1.1 segment. Attribute comid_on_nhm is FALSE and attribute network_ref equals "NHGF off-network". U.S. Geological Survey 1 26 nhdv2_temp_observations.parquet Parquet file containing water temperature observations summarized to daily values and aggregated to the NHDPlusv2.1 flowlines. U.S. Geological Survey basin River basin name. U.S. Geological Survey Descriptions for each unique value are included in the entity description for nhdv2_nhgf_crosswalk.csv. COMID Unique flowline common identifier (comid) taken from the NHDPlusv2.1 dataset. NHDPlusv2.1 1259243 948020794 date Observation date in YYYY-MM-DD format. U.S. Geological Survey 1979-01-01 2023-09-29 date site_id Concatenated site identifiers that are associated with a given NHDPlusv2.1 flowline and date. The individual site identifiers are from the source dataset. U.S. Geological Survey Character vector that contains the concatenated identifiers for all sites associated with a given NHDPlusv2.1 flowline. Individual site identifiers are separated by a semicolon delimiter. data_source_cd Concatenated data sources contributing observations for a given NHDPlusv2.1 flowline and date. Individual data sources are separated by a semicolon delimiter. U.S. Geological Survey Oliver2024-wqp Data source is the Water Quality Portal, compiled by Oliver et al. (2024). U.S. Geological Survey Oliver2024-nwis_dv Data source is the USGS NWIS daily values service, compiled by Oliver et al. (2024). U.S. Geological Survey DOE-LBNL_mean_daily_files Data source is Carroll et al. (2023). U.S. Geological Survey Oliver2024-norwest Data source is the NorWeST dataset compiled by Oliver et al. (2024). U.S. Geological Survey Oliver2024-nwis_uv Data source is the USGS NWIS instantaneous values service, compiled by Oliver et al. (2024). U.S. Geological Survey Oliver2024-ecosheds Data source is the SHEDS temperature database, compiled by Oliver et al. (2024). U.S. Geological Survey ngwos-neversink Data source is the Neversink NGWOS dataset (Terry et al. 2022). U.S. Geological Survey CA-Water-CDEC Data source is the CA CDEC. U.S. Geological Survey mean_time Mean of reported observation times for water temperature observations compiled by Oliver et al. (2024). U.S. Geological Survey Mean of reported observation times for daily water temperature summaries reported in Oliver et al. (2024). Times are reported as HH:MM:SS in local time. Rows with NA indicate that the mean observation time could not be computed for this comid-date. mean_temp_c The average of daily mean water temperature values across all sites matched to a given NHDPlusv2.1 flowline. The site-level daily mean water temperature values represent the average water temperature of all observations taken on the corresponding site-date, and in most cases, were provided by the data source. U.S. Geological Survey NA Daily mean water temperature for this comid-date could not be computed from the available data. U.S. Geological Survey -0.5 34.3 degrees Celsius min_temp_c The minimum of daily minimum water temperature values across all sites matched to a given NHDPlusv2.1 flowline. U.S. Geological Survey NA Daily minimum water temperature for this comid-date could not be computed from the available data. U.S. Geological Survey -0.5 34.3 degrees Celsius max_temp_c The maximum of daily maximum water temperature values across all sites matched to a given NHDPlusv2.1 flowline. U.S. Geological Survey NA Daily maximum water temperature for this comid-date could not be computed from the available data. U.S. Geological Survey -0.5 34.3 degrees Celsius sd_mean_temp_c The standard deviation of daily mean water temperature values across all sites matched to a given NHDPlusv2.1 flowline. U.S. Geological Survey NA The standard deviation of daily mean water temperature could not be computed for this comid-date. U.S. Geological Survey 0 11.1 degrees Celsius flag Data quality indicator. The appropriate use of flagged observations is dependent on the downstream application. Users should therefore use caution to interpret these indicator codes and weigh their suitability for the target application. U.S. Geological Survey NA No data quality flag was given in the source dataset or added in this data compilation workflow. U.S. Geological Survey DRE Daily range exceedance. The range of temperature values from that site-day exceeds a range of 13.2. This threshold was used because 99.9% of daily ranges from NWIS (after 1999-12-31) are below this value. Oliver et al. 2024 ES[comment] Comment from EcoSHEDS that often indicates something about the quality of the data. A flag of "ES" without a comment in brackets indicates a generic flag from EcoSHEDS that did not have an associated comment. Oliver et al. 2024 P Provisional data subject to revision Oliver et al. 2024 UE Units were estimated based on reported water temperature ranges at that site-year. Oliver et al. 2024 daily estimates may be biased by incorrect unit conversion applied by data provider. The daily values from the data source could not be reproduced from the published sub-daily data, possibly due to a different unit-conversions algorithm being applied. See the "process_doe_eastriver_data.R" script within the source_code.zip file for more information. U.S. Geological Survey network_ref Character value that describes whether the NHDPlusv2.1 flowline intersects a GFv1.1 segment (i.e., "on-network") or drains to it (i.e., "off-network"). U.S. Geological Survey NHGF on-network The NHDPlusv2.1 flowline intersects a GFv1.1 segment, i.e., the NHDPlusv2.1 flowline is an "on-network" flowline. U.S. Geological Survey NHGF off-network The NHDPlusv2.1 flowline drains to a GFv1.1 segment but does not intersect it, i.e., the NHDPlusv2.1 flowline is an "off-network" flowline. U.S. Geological Survey nhdv2_static_attributes.parquet Parquet file includes attribute features that represent characteristics of the river segment, its catchment, or the upstream watershed area, summarized to the NHDPlusv2.1 flowlines. This file includes 167 columns, 165 of which represent river attribute data. 162 columns represent select flowline attributes defined in Wieczorek et al. (2018) and are not re-defined here. U.S. Geological Survey basin River basin name. U.S. Geological Survey Descriptions for each unique value are included in the entity description for nhdv2_nhgf_crosswalk.csv. COMID Unique flowline common identifier (comid) taken from the NHDPlusv2.1 dataset. NHDPlusv2.1 1259243 948020795 TOT_NORM_STORAGE2013_AREANORM The area-normalized upstream reservoir storage associated with dams built during or before 2013, computed in this workflow by dividing the normalized reservoir storage (“TOT_NORM_STORAGE2013”) by the cumulative catchment area throughout the upstream watershed (“TOT_BASIN_AREA”). U.S. Geological Survey NaN The area-normalized upstream reservoir storage could not be calculated because the normalized reservoir storage and cumulative upstream area both equal zero. U.S. Geological Survey 0.0 810.810810810811 acre-feet per square kilometer confinement_calc_mcmanamay The estimated channel confinement derived using river channel and floodplain data from McManamay and DeRolph (2019). U.S. Geological Survey NA Channel confinement estimates are not available for this NHDPlusv2.1 COMID. U.S. Geological Survey 0.194507976101719 6691.989913097075259 m/m flag_value_cd Indicator that the estimated channel confinement was imputed using data from a neighboring flowline. U.S. Geological Survey NA The value in the confinement_calc_mcmanamay attribute field was not imputed using data from a neighboring flowline. U.S. Geological Survey Entries include the identifier of the donor flowline as well as the distance from the target flowline to the donor flowline. nhgf_static_attributes.parquet Parquet file includes attribute features that represent characteristics of the river segment, its catchment, or the upstream watershed area, summarized to the GFv1.1 segments. This file includes 167 columns, 165 of which represent river attribute data. 162 columns represent select flowline attributes defined in Wieczorek et al. (2018) and are not re-defined here. U.S. Geological Survey basin River basin name. U.S. Geological Survey Descriptions for each unique value are included in the entity description for nhdv2_nhgf_crosswalk.csv. seg_id_nat The GFv1.1 segment identifier. The segment identifier is equal to the nsegment_v1_1 attribute from the GFv1.1 dataset. U.S. Geological Survey 713 53746 TOT_NORM_STORAGE2013_AREANORM The area-normalized upstream reservoir storage associated with dams built during or before 2013, rescaled from the NHDPlusv2.1 resolution. U.S. Geological Survey 0.0 426.420110246687 acre-feet per square kilometer confinement_calc_mcmanamay The estimated channel confinement derived using river channel and floodplain data from McManamay and DeRolph (2019), rescaled from the NHDPlusv2.1 resolution. U.S. Geological Survey NA Channel confinement estimates are not available for this GFv1.1 segment. U.S. Geological Survey 0.194507976101719 406.798290506926264 m/m flag_value_cd Indicator that at least one catchment-scale ("CAT_") attribute was imputed using data from a neighboring GFv1.1 segment. U.S. Geological Survey NA No attribute values were imputed for a given GFv1.1 segment. U.S. Geological Survey Entries include the identifier of the donor segment as well as the distance from the target segment to the donor segment. nhdv2_inputs_io.zip Compressed parquet file containing model input data including river segment characteristics and daily meteorological data, summarized to the NHDPlusv2.1 flowlines. U.S. Geological Survey basin River basin name. U.S. Geological Survey Descriptions for each unique value are included in the entity description for nhdv2_nhgf_crosswalk.csv. COMID Unique flowline common identifier (comid) taken from the NHDPlusv2.1 dataset. NHDPlusv2.1 1259243 948020795 seg_id_nat The GFv1.1 segment identifier. The segment identifier is equal to the nsegment_v1_1 attribute from the GFv1.1 dataset. U.S. Geological Survey 713 53746 date Observation date in YYYY-MM-DD format. U.S. Geological Survey 1979-01-01 2021-12-31 date est_width_m Median NHDPlusv2.1 flowline width estimated by an empirical relationship fit using USGS streamflow gages from the HUC4 unit that the basin falls within. See Process Step 1 for more information about width estimation. U.S. Geological Survey 0.212427491890999 224.722833056374611 meters min_elev_m_nhdv2 Minimum NHDPlusv2.1 flowline elevation, taken from the minimum elevation attribute from NHDPlusv2.1 (minelevsmo) and converted to meters. U.S. Geological Survey 0 3694.61 meters slope_nhdv2 NHDPlusv2.1 flowline slope, taken from the slope attribute from NHDPlusv2.1. U.S. Geological Survey 0.00001000 0.59590909 unitless daily_max_rh_gridmet Daily maximum relative humidity for the NHDPlusv2.1 flowline catchment and date, calculated as the area-weighted mean of all grid cell values within the catchment. Daily maximum relative humidity data were sourced from gridMET. U.S. Geological Survey 10.0999994277954 100.0000076293945 % daily_min_rh_gridmet Daily minimum relative humidity for the NHDPlusv2.1 flowline catchment and date, calculated as the area-weighted mean of all grid cell values within the catchment. Daily minimum relative humidity data were sourced from gridMET. U.S. Geological Survey 0.353469610214233 100.000007629394531 % daily_mean_rh_gridmet Daily mean relative humidity for the NHDPlusv2.1 flowline catchment and date, calculated as the average of daily_min_rh_gridmet and daily_max_rh_gridmet. See Process Step 3 for more information about gridMET data. U.S. Geological Survey 6.99999964237213 100.00000762939453 % daily_srad_gridmet Daily shortwave radiation for the NHDPlusv2.1 flowline catchment and date, calculated as the area-weighted mean of all grid cell values within the catchment. Daily downward surface shortwave radiation (srad) data were sourced from gridMET. U.S. Geological Survey 0.0 387.100006103516 W m^-2 pet_gridmet Daily potential evapotranspiration for the NHDPlusv2.1 flowline catchment and date, calculated as the area-weighted mean of all grid cell values within the catchment. Daily reference evapotranspiration data were sourced from gridMET. U.S. Geological Survey 0.0 15.0044651031494 mm daily_min_airtempC_gridmet Daily minimum air temperature for the NHDPlusv2.1 flowline catchment and date, calculated as the area-weighted mean of all grid cell values within the catchment. Daily minimum air temperature data were sourced from gridMET. U.S. Geological Survey -37.7500061035156 30.1317077636719 degrees Celsius daily_max_airtempC_gridmet Daily maximum air temperature for the NHDPlusv2.1 flowline catchment and date, calculated as the area-weighted mean of all grid cell values within the catchment. Daily maximum air temperature data were sourced from gridMET. U.S. Geological Survey -26.8499969482422 45.0500122070313 degrees Celsius daily_mean_airtempC_gridmet Daily mean air temperature for the NHDPlusv2.1 flowline catchment and date, calculated as the average of daily_min_airtempC_gridmet and daily_max_airtempC_gridmet. See Process Step 3 for more information about gridMET data. U.S. Geological Survey -30.3413528442383 36.2451263427735 degrees Celsius daily_precip_m_gridmet Daily precipitation accumulation for the NHDPlusv2.1 flowline catchment and date, calculated as the area-weighted mean of all grid cell values within the catchment. Daily precipitation data were sourced from gridMET. U.S. Geological Survey 0.0 0.328267517089844 meters flag_values Indicates whether missing slope and/or meteorological data were imputed from a neighboring NHDPlusv2.1 flowline. See Process Step 3 for more information. U.S. Geological Survey NA No slope and/or meteorological values were imputed for a given NHDPlusv2.1 flowline. U.S. Geological Survey For all values other than NA, this column indicates which NHDPlusv2.1 flowlines were used to impute meteorological values for a given flowline, as well as the river distance from the target flowline to the donor flowline. nhgf_inputs_io.zip Compressed parquet file containing model input data including river segment characteristics and daily meteorological data, summarized to the GFv1.1 segments. U.S. Geological Survey basin River basin name. U.S. Geological Survey Descriptions for each unique value are included in the entity description for nhdv2_nhgf_crosswalk.csv. seg_id_nat The GFv1.1 segment identifier. The segment identifier is equal to the nsegment_v1_1 attribute from the GFv1.1 dataset. U.S. Geological Survey 713 53746 date Observation date in YYYY-MM-DD format. U.S. Geological Survey 1979-01-01 2021-12-31 date est_width_m Estimated GFv1.1 segment width. Segment width was assumed equal to the maximum river width across all on-network NHDPlusv2.1 flowlines for a given GFv1.1 segment. See Process Steps 1 and 3 for more information about width estimation. U.S. Geological Survey 2.26963208978408 224.72283305637461 meters min_elev_m_nhgf Estimated GFv1.1 segment minimum elevation. Segment minimum elevation was assumed equal to the minimum elevation across all on-network NHDPlusv2.1 flowlines for a given GFv1.1 segment. See Process Step 3 for more information about elevation estimation. U.S. Geological Survey 0.0 2720.83 meters slope_nhgf Estimated GFv1.1 segment slope. Segment slope was calculated as the length-weighted mean slope of all on-network NHDPlusv2.1 flowlines for a given GFv1.1 segment. See Process Step 3 for more information. U.S. Geological Survey 0.000010 0.067766 unitless daily_max_rh_gridmet Daily maximum relative humidity for the GFv1.1 catchment and date, calculated as the area-weighted mean of all grid cell values within the catchment. Daily maximum relative humidity data were sourced from gridMET. See Process Step 1 for more information on derivation of GFv1.1 catchments. U.S. Geological Survey 14.7049255371094 100.0000076293945 % daily_min_rh_gridmet Daily minimum relative humidity for the GFv1.1 catchment and date, calculated as the area-weighted mean of all grid cell values within the catchment. Daily minimum relative humidity data were sourced from gridMET. U.S. Geological Survey 1.0 100.000007629395 % daily_mean_rh_gridmet Daily mean relative humidity for the GFv1.1 catchment and date, calculated as the average of daily_min_rh_gridmet and daily_max_rh_gridmet. See Process Step 3 for more information about gridMET data. U.S. Geological Survey 7.85246276855469 100.00000762939453 % daily_srad_gridmet Daily shortwave radiation for the GFv1.1 catchment and date, calculated as the area-weighted mean of all grid cell values within the catchment. Daily downward surface shortwave radiation (srad) data were sourced from gridMET. U.S. Geological Survey 0.0 385.64990234375 W m^-2 pet_gridmet Daily potential evapotranspiration for the GFv1.1 catchment and date, calculated as the area-weighted mean of all grid cell values within the catchment. Daily reference evapotranspiration data were sourced from gridMET. U.S. Geological Survey 0.0 14.9461135864258 mm daily_min_airtempC_gridmet Daily minimum air temperature for the GFv1.1 catchment and date, calculated as the area-weighted mean of all grid cell values within the catchment. Daily minimum air temperature data were sourced from gridMET. U.S. Geological Survey -37.3987640380859 30.1113220214844 degrees Celsius daily_max_airtempC_gridmet Daily maximum air temperature for the GFv1.1 catchment and date, calculated as the area-weighted mean of all grid cell values within the catchment. Daily maximum air temperature data were sourced from gridMET. U.S. Geological Survey -25.3113311767578 44.7109619140625 degrees Celsius daily_mean_airtempC_gridmet Daily mean air temperature for the GFv1.1 catchment and date, calculated as the average of daily_min_airtempC_gridmet and daily_max_airtempC_gridmet. See Process Step 3 for more information about gridMET data. U.S. Geological Survey -30.2788757324219 36.2309051513672 degrees Celsius daily_precip_m_gridmet Daily precipitation accumulation for the GFv1.1 catchment and date, calculated as the area-weighted mean of all grid cell values within the catchment. Daily precipitation data were sourced from gridMET. U.S. Geological Survey 0.0 0.278172668457031 meters flag_values Indicates whether missing meteorological data were imputed from a neighboring GFv1.1 segment. See Process Step 3 for more information. U.S. Geological Survey NA No slope and/or meteorological values were imputed for a given GFv1.1 segment. U.S. Geological Survey For all values other than NA, this column indicates which GFv1.1 segments were used to impute meteorological values for a given GFv1.1 segment, as well as the river distance from the target segment to the donor segment. basins Polygon feature layer within spatial_data.gpkg, a geopackage database file that contains several spatial datasets for each of the 12 focal river basins. U.S. Geological Survey basin River basin name. U.S. Geological Survey Descriptions for each unique value are included in the entity description for nhdv2_nhgf_crosswalk.csv. GFv1.1_segments Line feature layer within spatial_data.gpkg, a geopackage database file that contains several spatial datasets for each of the 12 focal river basins. U.S. Geological Survey basin River basin name. U.S. Geological Survey Descriptions for each unique value are included in the entity description for nhdv2_nhgf_crosswalk.csv. seg_id_nat The GFv1.1 segment identifier. The segment identifier is equal to the nsegment_v1_1 attribute from the GFv1.1 dataset. The integer values for this column represent unique identifiers and do not have units. U.S. Geological Survey 713 53746 GFv1.1_hrus Multi polygon feature layer within spatial_data.gpkg, a geopackage database file that contains several spatial datasets for each of the 12 focal river basins. U.S. Geological Survey basin River basin name. U.S. Geological Survey Descriptions for each unique value are included in the entity description for nhdv2_nhgf_crosswalk.csv. nhru_v1_1 A unique, sequential identifier for each HRU feature in the GFv1.1 dataset. GFv1.1 2134 109030 seg_id_nat The identifier of the GFv1.1 segment to which runoff is routed to from each HRU. The segment identifier is equal to the hru_segment_v1_1 attribute from the nhru_v1_1 feature layer in the GFv1.1 dataset. U.S. Geological Survey 713 53746 GFv1.1_catchments Multi polygon feature layer within spatial_data.gpkg, a geopackage database file that contains several spatial datasets for each of the 12 focal river basins. The GFv1.1 catchments represent a modified version of the GFv1.1 HRUs to represent the land area that drains to each GFv1.1 segment. See Process Step 3 in the Data Quality metadata section for more information. U.S. Geological Survey basin River basin name. U.S. Geological Survey Descriptions for each unique value are included in the entity description for nhdv2_nhgf_crosswalk.csv. seg_id_nat The identifier of the GFv1.1 segment to which runoff is routed to from the catchment. U.S. Geological Survey 713 53746 NHDPlusv2.1_flowlines Line feature layer within spatial_data.gpkg, a geopackage database file that contains several spatial datasets for each of the 12 focal river basins. U.S. Geological Survey basin River basin name. U.S. Geological Survey Descriptions for each unique value are included in the entity description for nhdv2_nhgf_crosswalk.csv. COMID Unique flowline common identifier (comid) taken from the NHDPlusv2.1 dataset. NHDPlusv2.1 1259243 948020795 seg_id_nat The GFv1.1 segment identifier. The segment identifier is equal to the nsegment_v1_1 attribute from the GFv1.1 dataset. U.S. Geological Survey 713 53746 network_ref Character value that describes whether the NHDPlusv2.1 flowline intersects the GFv1.1 segment or just drains to it. U.S. Geological Survey NHGF off-network The NHDPlusv2.1 flowline drains to the GFv1.1 segment but does not intersect it, i.e., the NHDPlusv2.1 flowline is an "off-network" flowline. U.S. Geological Survey NHGF on-network The NHDPlusv2.1 flowline intersects the GFv1.1 segment, i.e., the NHDPlusv2.1 flowline is an "on-network" flowline. U.S. Geological Survey NHDPlusv2.1_catchments Multi polygon feature layer within spatial_data.gpkg, a geopackage database file that contains several spatial datasets for each of the 12 focal river basins. U.S. Geological Survey basin River basin name. U.S. Geological Survey Descriptions for each unique value are included in the entity description for nhdv2_nhgf_crosswalk.csv. comid Unique catchment common identifier (comid) taken from the NHDPlusv2.1 dataset. NHDPlusv2.1 1259243 948020795 source_code.zip Compressed file contains R code used to generate the model input datasets. U.S. Geological Survey The entity and attribute section here describes the nine files linked on the project landing page. Koenig, L.E., Barclay, J.R., Appling, A.P., 2025, Distance matrices, river-network crosswalks, water temperature observations, and river attribute data for twelve river basins in the United States: U.S. Geological Survey data release, https://doi.org/10.5066/P14GXPOW. U.S. Geological Survey - ScienceBase mailing address

Denver Federal Center

Building 810

Mail Stop 302

Denver CO 80225 1-888-275-8747 sciencebase@usgs.gov Unless otherwise stated, all data, metadata and related materials are considered to satisfy the quality standards relative to the purpose for which the data were collected. Although these data and associated metadata have been reviewed for accuracy and completeness and approved for release by the U.S. Geological Survey (USGS), no warranty expressed or implied is made regarding the display or utility of the data for other purposes, nor on all computer systems, nor shall the act of distribution constitute any such warranty. Digital Data https://doi.org/10.5066/P14GXPOW None 20250915 U.S. Geological Survey Lauren E. Koenig Data Scientist mailing

425 Jordan Road

Troy NY 12180 5182855695 lkoenigsnyder@usgs.gov FGDC Biological Data Profile of the Content Standard for Digital Geospatial Metadata FGDC-STD-001.1-1999