ERF1v2 watersheds were originally created by USGS, based on U.S. Environmental Protection Agency's River Reach File 1 (RF1). These watersheds are a consistent national layer of drainage areas corresponding to RF1 segments. The digital segmented network based on watershed boundaries, ERF1_2, includes enhancements to the U.S. Environmental Protection Agency's River Reach File 1 (RF1) (USEPA, 1996; DeWald and others, 1985) to support national and regional-scale surface water-quality modeling. Alexander and others (1999) developed ERF1, which assessed the hydrologic integrity of the digital reach traces and calculated the mean water time-of-travel in river reaches and reservoirs. ERF1_2 serves as the foundation for SPARROW (Spatially Referenced Regressions (of nutrient transport) on Watershed) modeling. This version of the network expands on ERF1 (Version 1.2; Alexander, et al., 1999) and includes the incremental and total drainage area derived from 1-kilometer (km) elevation data for North America. The ERF1_2 coverage extends the earlier drainage area founded on the 1-kilometer data for North America (Verdin, 1996; Verdin and Jenson, 1996). A 1-kilometer raster grid of ERF1_2 projected to Lambert Azimuthal Equal Area, NAD 27 Datum (Snyder, 1987), was merged with the HYDRO1K flow direction data set (Verdin and Jenson, 1996) to generate a DEM-based watershed grid, ERF1_2WS_LG. The watershed boundaries are maintained in a raster (grid cell) format as well as a vector (polygon) format for subsequent model analysis. Both the coverage, ERF1_2, and the grid, ERF1_2WS_LG, are available at: URL:http://water.usgs.gov/lookup/getspatial?erf1_2. ERF1_2 is joined with Course Proximity of Water to NLCD Agriculture tabular data base results. The following describes this tabular data: Nation-wide assessment of proximity of agricultural fields to water bodies summarized by watersheds. The National Agricultural Statistics Service, 2002 Census of Agriculture was used to estimate the amount of individual crops grown within the surface water margins, and within each watershed.

ERF1_2 was designed to be a digital database of river reaches capable of supporting regional and national water-quality and river-flow modeling by the water-resources community. ERF1, on which ERF1_2 is based, is used at the U.S. Geological Survey to support national-level water-quality modeling with the SPARROW approach (see Alexander and others, 2000; Smith and others, 1997). In the current and earlier analyses, the reach network is used to determine flow pathways between sources of point and nonpoint pollutants (e.g., fertilizer use, municipal wastewater discharges) and downstream water-quality monitoring locations in support of predictive water-quality models of stream nutrient transport. The following describes the purpose of the Course Proximity of Water to NLCD Agriculture data joined to the ERF1v2ws data set: The objective of the overall watershed characterization analysis was to describe the nation's agricultural proximity to water summarized for individual watersheds across the nation using a Geographic Information System (GIS) and publicly available datasets. The Enhanced River Reach File 2.0 watersheds (ERF1v2) were selected to summarize proximity information for the nation. For use with the GEOSTAC Database, this data set has been compiled to simplify pesticide risk assessment and to provide a common data set for all stakeholders.

This data set, ERF1_v2ws, is joined with The Course Proximity of Water to NLCD Agriculture tabular data set results. Please reference the The Course Proximity of Water to NLCD Agriculture database metadata for detailed information regarding the data. Any use of trade, product, or firm names is for descriptive purposes only and does not imply endorsement by the U.S. Government. Although this Federal Geographic Data Committee-compliant metadata file is intended to document the data set in nonproprietary form, as well as in ArcInfo format, this metadata file may include some ArcInfo-specific terminology.

The version of RF1 used to compile ERF1_2 was an early edition of a USGS RF1 translation and was updated by USEPA (USEPA, 1996). The capabilities of the enhanced version of RF1 (ERF1_2) and the current USEPA version have not been evaluated. The user is referred to the USEPA version [URL: http://www.epa.gov/owow/monitoring/georef/history.html] for discussions of streamflow accuracy and general background on the origin of RF1.

David L. Lorenz of the U.S. Geological Survey assembled initial versions of the documentation for this data set. Acknowledgements: The original USGS authors would like to thank Richard Smith, a co-developer of the SPARROW approach, Kristine Verdin, an Stephen Char, all of the U.S. Geological Survey, for providing technical assistance. The reviewers of this report, Dave Stewart, and Mike Wieczorek, are also acknowledged for their significant contributions.

The direct drainage area of each reach (RCHAREA) was computed by Smith and others (1997) using a Theissen method. It considers soley the distance to the nearest reach and does not consider any topographic features. It provides reasonably good estimates for higher-order streams, but is less accurate for lower orders. See process steps 1, 4, and 8, to assess the accuracy of MEANV and RCHTOT.

This data set is believed to provide complete and accurate connectivity among the streams represented. These arcs representing streams are intended to be presented in an FNODE to TNODE orientation. The arcs are pointing in the downstream direction.

Reservoir and lake boundaries were removed from the original RF1 data set. They are represented by a single trace to satisfy the connectivity requirements of the dataset. Centerlines were manually delineated within the RF1 shoreline boundaries (Alexander and others, 1999). The centerline reaches approximate the thalweg of each reservoir or lake. Shorelines of major estuaries and some coastal areas were added to the data set to serve as a stopping location for trace routines. Existing reaches extending within estuaries were deleted.

(1) Estimation of reach time-of-travel: The reach time-of-travel (RCHTOT) in units of days was computed as the quotient of reach length and reach velocity for 59,364 reaches with positive values of velocity such that RCHTOT = C1 * LENGTH / MEANV (1) where C1 is a conversion factor of 0.00003797 ft-day/meter-second, LENGTH is the length in meters of the Arc reach segment, and MEANV is the mean velocity in feet per second. Reaches with missing time-of- travel are assigned a value of zero.

(2) Addition of coastal shorelines and estuary boundaries: Major estuaries and coastal shorelines were added from the National Oceanic and Atmospheric Administration USEPA Reach File Version 1.0 (RF1) (NOAA, 1997) for the conterminous United States (CONUS) as enhanced for NOAA's Strategic Environmental Assessments Division. Shorelines and estuary boundaries were snapped to existing reaches and assigned ERF1 and E2RF1 values of gt 80000. For identification purposes, the shorelines and estuary boundaries maintained the USEPA items for identification purpose. In addition, TERMFLAG = 3 was assigned to these shoreline segments.

(3) TERMFLAG check: A check was made for all coastal shoreline reaches where E2RF1 gt 80000, and TERMFLAG = 3; TERMFLAG for disconnected reaches and reaches flowing outside continental US (TERMFLAG = 2); and TERMFLAG for reaches ending at the shoreline to terminate modeling applications (TERMFLAG = 1). [E2RF1 81893, TERMFLAG = 3; correction date: 20000601] [Additional TERMFLAG changes were made to terminal and transport reaches. Correction date: 20020522].

(4) Missing data: Reaches with missing or zero MEANQ or MEANV values (2739) were assigned a value of -9999.99. RESTOT and RCHTOT that resulted in a negative values were also assigned a value of -9999.99.

(5) Split reaches: ERF1_2 was edited to split reaches containing WQN and NAWQA sites. An automated graphical Arc AML was used for visual verification and splitting of the remaining reaches. A total of 768 reaches were split. Of the split reaches, the upstream ones were assigned a new identification number (E2RF1) and a STAID. One reach, E2RF1 29302, was assigned STAID 05020500, which was within the original 500m limit set by the AML, but remains unsplit. The radius limit was extended to 2000m to include data sampling sites that fit the criteria needed for modeling. Arc topology was updated following these changes. Manual verification of ERF1_2 arcs was completed by visual inspection for attributes E2RF1 and STAID agreement.

(6) E2RF1 check: Reaches containing multiple values for arcs were selected based on frequency of occurrence. A total of 40 reaches were selected. Arcs in ERF1_2 were edited to remove several dangle and duplicate arcs; flow direction was flipped for several arcs; several reaches were consolidated by removing pseudo nodes; STAID 01389500 was assigned to reach 61353, E2RF1 65607; TERMFLAG values were reassigned for a number of arcs, and TYPE = R was changed for one arc; several vertices were adjusted to enhance topology, and one arc was deleted to correct topology. The following new E2RF1 values were assigned to river reaches: 65608 to Waccamaw R; 65609 to *C; 65610 to Intracoastal Waterway; 65611 to Julington Cr; 65612 to Munuscong R; 65613 to Detroit R; 65614 to Rio Grande; 65615 to Sandy R; 65616 to Washougal R; 65617 to Kentack Slough. New E2RF1 values were assigned to the following shoreline reaches which were also renamed: 81893 to Delaware R; 81894 to Pacific Ocean; two segments 81895 and 81896 to L. Borgne; and 81897 to Pungo R. Arc topology was updated following these changes. [Correction date: 20010013] [E2RF1 44108, Green R, was assigned a new E2RF1 value, 65781; E2RF1 34672, PNAME "YELLOWSTONE", was changed to "MISSOURI R"; E2RF1 29130, PNAME "HEART R", was changed to "MISSOURI R"; the spelling of PNAME "KANKAKEE R" was corrected for reaches E2RF1: 20401, 20402, 20405, 40406, 20407, 20408, 20409, 20411, 20412, 20420, 20421, 20436. Correction date: 20020522].

(7) TNODE, FNODE flipped: Reach flow direction for ERF1_2 was flipped. These arcs are currently in a FNODE to TNODE orientation. The arcs are pointing in a downstream direction. Arc topology was updated following this change. [E2RF1 10128, *B, and E2RF1 6219, Stono R, arcs were flipped to update topology. Correction date: 20020522].

(8) Recalculate RCHTOT: An AML was used to recalculate the reach time of travel for flow for upstream and downstream portions of reaches in ERF1_2 that were split. The split reaches were identified by selecting the frequency of occurrence for ERF1. The stream reach time of travel [RCHTOT] values were recalculated. [Correction date: 20011102] [The MEANV value for E2RF1 29130 was recomputed to 2.61, and RCHTOT = 0.19611. Correction date: 20020522].

(9) Assign FRAC: Item FRAC was assigned to ERF1_2 using an Arc AML program. The AML selects the fractional diversion of the load for reaches that share to-nodes. This fraction was applied to upstream sources and instream measurements of loads. Some special-case calculations were added to the AML based on specific knowledge of the location: ERF1 10527 Santee River, FRAC = 0.99835; ERF1 10534 Cooper River, FRAC = 0.00165; ERF1 22100 Mississippi River outflow channel, FRAC = 0.21170; ERF1 22197 Mississippi River, FRAC = 0.78830.

(10) Assign NAWQA pesticide site attributes and create site attribute file (ERF1_2.SITES): An AML program was used to create a point coverage, PEST_SIT. The PEST_SIT coverage was defined as Albers Equal-Area Conic projection. Sixty-three sites were added to ERF1_2; eight reaches were split and STAID values were assigned to the upstream reaches. A manual check was conducted to verify the association between the site and reach. In addition, several STAIDs were altered to maintain consistency with earlier versions of ERF1_2. These STAIDs were modified in both the point coverage SAMP_SIT, and the river reach coverage ERF1_2. The STAID values that were changed are as follows: 385234087071800, changed to 385234087071801; 391732085414400, changed to 391732085414401; and 393306086585200 changed to 393306086585201. Arc topology was updated following these changes. The time of travel for both reaches (RCHTOT) and reservoirs (RESTOT) was recalculated to account for the newly split reaches. FRAC attributes were also assigned to the newly split upstream reaches. ERF1_2.SITES associates a split reach with a monitoring site type. Monitoring types were associated with 723 reaches.

(11) Update NOAA Estuarine Drainage Area values: ERF1_2 reaches missing EDACDA attribute values were populated with NOAA Estuarine Drainage Area data by creating a shapefile of ERF1_2 containing missing EDACDA values and selecting the closest approximated polygon value to the missing reach.

Process_Description:

(12) Generate watersheds for ERF1_2 reaches:

The creation of a watershed grid for ERF1_2 reaches involved several preparation steps. Critical for future SPARROW modeling applications was the exact alignment of the coordinates for the reach network line grid and the boundary mask grid with the flow direction grid, NA_FD_G2 (modified version of USGS HYDRO1K flow direction data set, clipped to the conterminous U.S.) (Brakebill, J.W., 1999). The line coverage, ERF1_2_A, was projected from Albers Equal-Area, NAD 83 datum to Lambert Azimuthal Equal-Area, NAD 27 datum (ERF1_2_L), and was used to create a line grid, ERF1_2_LG, with the attribute "VALUE" set to the unique reach identification number, E2RF1. The window and snap options were set to the extents of the HYDRO1K flow direction grid, NA_FD_G2, with the cell size set to 1000 (Verdin, K., USGS, EROS Data Center, pers. comm., 7/2000). The boundary mask was registered to the coordinates of the flow direction grid by resampling the polygon grid of the mask with the cell size, window, and snap options set to the extents of NA_FD_G2.

Using the flow direction data set (NA_FD_G2), and the mask of the U.S. (MASK_LG), a 1-kilometer cell-size raster grid of reach watersheds (ERF1_2WS_LG) was created (Environmental Systems Research Institute, 1992). The cell size, window, and snap options were set to the extents of NA_FD_G2, and the watershed boundary attribute "VALUE" was correlated with the unique reach identification number, E2RF1, that is associated with the coverage, ERF1_2. ERF1_2WS_LG has a Lambert Azimuthal Equal Area projection, NAD 27 datum.

Several "no value" areas or "holes" were formed during the creation of the watershed grid where low elevations occur, as along coastal areas, and in the western U.S., where most of the closed basins are found. To edit these areas, a vector polygon coverage of the watershed grid, ERF1_2WS_LG, was created and an AML was invoked that assigned a unique identification number to the ERF1_2WS-LG basins draining to the interior cells that were unlinked to ERF1_2 reaches. A value greater than 95000 was assigned to these basins. Those cells adjacent to coastal watersheds were assigned a value greater than 80000 by selecting the basin value that contained the longest shared boundary. A new grid was created containing the edited watershed cells and was manually checked for errors. The final processing step was to assign a value of -9999 to the NODATA areas of the watershed grid:

outgrid = con(isnull(existing ws grid),-9999,existing ws grid)

A total of 93 closed basins were assigned cell values where E2RF1 gt 95000.

ERF_1v2ws

Metadata: