points, transects, beach width: Barrier island geomorphology and shorebird habitat metrics at 50-m alongshore transects and 5-m cross-shore points: Smith Island, VA, 2014

Full methods are provided in the associated Methods OFR (Zeigler and others, 2019). The Jupyter notebook used for processing is distributed with this dataset (extractor_smi14.ipynb).
smi_trans.shp
NASC transects from Himmelstoss and others (2010) were modified for the purposes of this study. Transects were extended inland to cover the width of the barrier, and additional transects were added to fill alongshore gaps greater than 50 m.
Transects are extended by using the last two vertices of each transect to programmatically place the end of the line 3,000 m beyond the end of the original line segment. Python (version 3) with the modules ArcPy and Collections in an ArcGIS Pro 2.0 environment are used for programming (see function core/functions_warcpy.ExtendLine in bi-transect-extractor (Sturdivant, 2019)). The process transforms the data from WGS84 to NAD83 UTM Zone 18N using WGS_1984_(ITRF00)_To_NAD_1983_2011 (WKID: 108354, accuracy: 0.1 m). It uses the coordinates of the last two vertices of each feature and calculates a new second coordinate at a specified distance (3,000 m) along the prolongation of the line from the first coordinate.
Next, transects are manually added to fill alongshore gaps greater than 50 m. To do so we copy the extended NASC transects, replace their attribute values with fill values (-99999), move groups of them to fill gaps, delete any that are unchanged, and then merge the altered file back to the NASC transects. ID values ('sort_ID') are assigned that order the transects consecutively along the oceanside shoreline using the ArcGIS Spatial Sort tool in spatial increments. At Smith Island, these sort_ID values are not all consecutive along the shoreline.
We eliminate transect overlap in certain locations by manually clipping the transects to the first intersection point with an overlapping transect. While doing so, we prioritize the original NASC transect geometries. Overlapping transects retain the azimuths of the original lines but in some cases are shortened.

Full methods are provided in the associated Methods OFR (Zeigler and others, 2019) and technical steps can be found in the source code (Sturdivant, 2019). The Jupyter notebook used to create the point dataset (extractor_smi14.ipynb) is distributed in this data release. All steps described below were performed in the Jupyter notebook extractor_smi14.ipynb, which used Python 3 and ArcPy distributed with ArcGIS Pro 2.0. Manual steps indicated in the notebook were executed in a session of ArcGIS Pro 2.0.
The term "shoreline" is used to describe multiple components. "Shoreline polygon(s)" (smi14_shoreline.shp in larger work) outline the boundary between land and water for the entire study area. They are composed of the oceanside shoreline and the back-barrier shoreline. "Oceanside shoreline" refers to the MHW contour line adjacent to the open ocean and bounded by either tidal inlet or the study area extent. In some cases, we may use MHW shoreline interchangeably with oceanside shoreline. "Back-barrier shoreline" refers to the boundary between land and water on the inland side of the island; it usually follows the mean tidal level (MTL) contour line. We also refer to "shoreline points," which are precise MHW positions defined in the smi14_SLpts dataset in the larger work.
smi14_pts.csv, part 1
First, we calculate values that apply to entire cross-shore transects. Supplemented NASC transects are populated with the shoreline change rate values from the original NASC transects. Lines that are not present in the NASC transects are populated with fill values (-99999) for LRR, TRANSECTID, and TRANSORDER fields.
MHW position and foreshore slope along transect:
Each transect is assigned the foreshore slope (Bslope) from the nearest oceanside shoreline point within 25 m (see larger work). The MHW shoreline easting and northing locate the intersection of the transect with the oceanside shoreline. These values are populated for each transect as follows (using Python 3 distributed with ArcGIS Pro 2.0, especially the modules ArcPy, NumPy, and Pandas): (1) Create a line representing the oceanside shoreline by converting the shoreline polygons to lines, clipping them at the inlet lines, and selecting only those segments that intersect shoreline points (from the morphology feature positions, smi14_SLpts in larger work); (2) get 'SL_x' and 'SL_y' at the intersection point between the transect and the oceanside shoreline (created from smi14_shoreline.shp in larger work); (3) find the closest shoreline point to that intersection point; and (4) assign the slope value from the shoreline point to the transect ('Bslope'). In rare situations, this results in a transect with a slope value but without a shoreline position.
Dune positions along transects:
'DL_x', 'DL_y', and 'DL_zMHW' are the easting, northing, and height above MHW, respectively, of the nearest dune toe point (smi14_DLpts.shp in larger work) within 25 meters of the transect. Likewise, 'DH_x', 'DH_y', and 'DH_zMHW' are the easting, northing, and height above MHW, respectively, of the nearest dune crest point within 25 meters (smi14_DCpts.shp in larger work). We find the XYZ positions of the nearest dune crest and dune toe within 25 meters for each transect using bi-transect-extractor (Sturdivant, 2019) to execute the following, repeated for both dune crest and dune toe: (1) evaluate the distance from the transect to every dune crest/toe point and find the shortest of these distances (distanceTo geometry method in ArcPy); (2) if the distance is less than 25 m, return the elevation from the point and the XY position of the point ‘snapped’ to the transect (snapToLine geometry method); (3) if there are no points within 25 m of the transect, populate the transect with null/fill values (Pandas); and (4) convert the elevations to the MHW datum by subtracting the MHW offset value (0.34 m based on Weber and others, 2005).
'Arm_x', 'Arm_y', and 'Arm_zMHW' are the easting, northing, and height above MHW, respectively, where an artificial structure crosses the transect in the vicinity of the beach. These features are meant to supplement the dune toe dataset by providing an upper limit to the beach in areas where dune toe extraction was confounded by the presence of an artificial structure. Values are populated for each transect as follows: (1) prioritizing areas where dune toe features were not successfully extracted, use orthoimagery, supplemented with the DEM, to manually digitize line segments on the oceanside face of artificial impediments to sediment ("armoring"), such as sand-fencing, sandbags, seawalls, etc.; (2) get the positions of intersection between the digitized armoring lines and the transects (Intersect tool from the Overlay toolset); (3) extract the elevation value at each intersection point from the DEM (Extract Multi Values to Points tool from Spatial Analyst; see Source Information for DEM); and (4) convert the elevations to the MHW datum by subtracting the MHW offset value (0.34 m based on Weber and others, 2005).
Beach width and height:
Upper beach width ('uBW') and upper beach height ('uBH') are calculated based on the difference in position between two points: the position of MHW along the transect ('SL_x', 'SL_y') and the dune toe position or equivalent (usually 'DL_x', 'DL_y'). All calculations are performed by the function calc_BeachWidth_fill in bi-transect-extractor (Sturdivant, 2019). In some cases, the dune toe is not appropriate to designate the "top of beach" so beach width and height are calculated from either the position of the dune toe, the dune crest, or the base of an armoring structure. The dune crest is only considered a possibility if the dune crest elevation (DH_zMHW) is less than or equal to 2.5 m, which was selected for this site through expert judgement. They are calculated as follows, relying primarily on NumPy and the snapToLine geometry method in ArcPy, and using Pandas for data storage/organization:
(1) Find the position along the transect of an orthogonal line drawn to the dune point ('DL_x', 'DL_y' and 'DH_x', 'DH_y'). (2) Calculate distances from MHW to the position along the transect of the dune toe ('DistDL'), dune crest ('DistDH'), and armoring ('DistArm'). (3) Conditionally select the appropriate feature to represent "top of beach." Dune toe is prioritized. If it is not available and 'DH_zMHW' is less than or equal to maxDH, use dune crest. If neither of the dune feature positions satisfy the conditions and an armoring feature intersects with the transect, use the armoring position. If none of the three are possible, 'uBW' and 'uBH' are given a NoData value of -99999. (4) Copy the distance to oceanside shoreline and height above MHW ('Dist--', '---zMHW') to 'uBW' and 'uBH', respectively.
Distance to inlet:
Distance to nearest tidal inlet ('Dist2Inlet') is computed as alongshore distance of each sampling transect from the nearest tidal inlet. Inlets are manually delineated during the creation of the shoreline polygon file. This distance includes changes in the path of the oceanside shoreline rather than just a straight-line distance between each transect and the inlet and reflects sediment transport pathways. It is measured using the shoreline polygon(s) and the delineated tidal inlets (smi14_inletLines.shp in larger work) as follows: (1) split the shoreline polygon(s) at the tidal inlets by converting the polygon(s) to polylines including the inlet lines (Feature to Line in Data Management); (2) retain only the oceanside segments of the shoreline by deleting all segments that do not intersect any shoreline points (disjoint geometry method in ArcPy data access module); (3) if the shoreline is bounded on both sides by an inlet, measure the distance to both and assign the minimum distance of the two or if the shoreline meets only one inlet (meaning the study area ends before the island ends), use the distance to the only inlet (cut, disjoint, and length geometry methods and properties in ArcPy)
Island widths:
Barrier Island Width ('WidthLand') is calculated as the aggregate above-water width between the back-barrier and oceanside shorelines along the transect. 'WidthLand' only included regions of the barrier within the shoreline polygon(s) (smi14_shoreline.shp in larger work). We also measure the shore-to-shore extent of the island, which includes space occupied by waterways ('WidthFull') and the width of only the most seaward portion of land within the shoreline ('WidthPart'). These are calculated as follows:
(1) Clip the transect to the shoreline polygon(s) (Clip in the Analysis toolbox); (2) for 'WidthLand', get the length of the multipart line segment from "SHAPE@LENGTH" feature class attribute, which will include only the remaining portions of the transect; (3) for 'WidthPart', convert the clipped transect from multipart to singlepart and get the length of the first line segment, which should be the most seaward; (4) for 'WidthFull', calculate the distance between the first vertex and the last vertex of the clipped transect (Feature Class to NumPy Array with explode to points, Pandas groupby, NumPy hypot).
The shoreline polygon may be artificially clipped where the site boundary or the available data do not include a back-barrier shoreline, which prevents the accurate calculation of island width values (e.g. Coast Guard Beach in Massachusetts). In these cases, transects that intersect a false back-barrier shoreline were manually identified and assigned fill values (-99999) for WidthLand, WidthFull, WidthPart, Dist_MHWbay as applicable.

Full methods are provided in the associated Methods OFR (Zeigler and others, 2019) and technical steps can be found in the source code (Sturdivant, 2019). The Jupyter notebook used to create the point dataset (extractor_smi14.ipynb) is distributed in this data release. All steps described below were performed in the Jupyter notebook extractor_smi14.ipynb, which used Python 3 and ArcPy distributed with ArcGIS Pro 2.0. Manual steps indicated in the notebook were executed in a session of ArcGIS Pro 2.0.
smi14_pts.csv, part 2
Nourishment, Construction, Development:
We manually assigned coded values for the attributes Nourishment, Construction, and Development by comparing the transect positions to ancillary datasets in ArcGIS 10.5. These datasets included the inventory of habitat modification (Rice, 2015), available aerial imagery, and the development layer included in this data release. See the associated Methods OFR (Zeigler and others, 2019) for details and example figures.

Full methods are provided in the associated Methods OFR (Zeigler and others, 2019). All steps described below were performed in the Jupyter notebook extractor_smi14.ipynb, which used Python 3 and ArcPy distributed with ArcGIS Pro 2.0. Manual steps indicated in the notebook were executed in a session of ArcGIS Pro 2.0.
smi14_pts.csv, part 3
The point dataset represents 5 m sampling of the land along each shore-normal transect (smi_trans.shp; see Zeigler and others, 2019). The 5-m point file is created from the supplemented NASC transects as follows: (1) Manually shorten overlapping transects to the first point of intersection. Prioritize preserving original NASC transects and those that are oriented most orthogonally to the shore. (2) Clip the transects to the shoreline polygon(s) of the barrier island, retaining only those portions of the transects that represent land. (3) Create a point along each transect every 5 m starting from the oceanside shoreline.

Full methods are provided in the associated Methods OFR (Zeigler and others, 2019). All steps described below were performed in the Jupyter notebook extractor_smi14.ipynb, which used Python 3 and ArcPy distributed with ArcGIS Pro 2.0 (Sturdivant, 2019). Manual steps indicated in the notebook were executed in a session of ArcGIS Pro 2.0.
smi14_pts.csv, part 4
Point identifier:
We populate the 5-m points with a numerical identifier ('SplitSort') of the 5-m data sampling points at a particular study site, sorted by order along oceanside shoreline and by distance from the oceanside shoreline. 'SplitSort' values are populated by sorting the points by 'sort_ID' and 'Dist_Seg' (see below).
Distances:
'Dist_Seg' and 'Dist_MHWbay' measure the distance of the point from the shoreline adjacent to the open ocean and the shoreline adjacent to the back-barrier waterbody respectively. 'Dist_Seg' is calculated as the Euclidean distance between the point and the shoreline ('SL_x', 'SL_y'). 'Dist_MHWbay' is calculated by subtracting the 'Dist_Seg' value from the transect 'WidthPart'. A NoData value for WidthPart prevents the calculation of Dist_MHWbay. This is performed by the function prep_points in bi-transect-extractor (Sturdivant, 2019).
'DistSegDH', 'DistSegDL', and 'DistSegArm' measure the distance of each 5-m point from the dune crest, dune toe, and dune armoring position respectively along a particular transect. They are calculated as the Euclidean distance between the 5-m point and the given feature.
Elevation and slope:
'ptZ' and 'ptSlp' are the elevation (NAVD88) and slope at the 5-m cell corresponding to the point. We use the 5-m DEM (see Source Information) to generate a slope surface (Slope tool in 3D Analyst). The elevation and slope values are assigned to the points using the Extract Multi Values to Points tool in Spatial Analyst, which is called by bi-transect-extractor (Sturdivant, 2019). 'ptZmhw' is calculated from 'ptZ' by subtracting the MHW offset value (0.34 m based on Weber and others, 2005).
Transect-averaged elevation:
We calculate the per-transect mean and maximum barrier elevation (mean_Zmhw, max_Zmhw) from the 5-m elevations (ptZmhw). Mean barrier elevations are calculated for only those transects having less than 20 percent missing values within the 5-m points. Locations not satisfying this criterion are assigned a fill value. This is performed by the function aggregate_z in bi-transect-extractor (Sturdivant, 2019).
Habitat variables:
Variables for vegetation type (VegType), vegetation density (VegDens), substrate type (SubType), distance to foraging habitat (DisMOSH), and geomorphic setting (GeoSet) are populated from the corresponding raster layers provided with the larger work (files Smi14_VegType.tif, Smi14_VegDen.tif, Smi14_SubType.tif, Smi14_DisMOSH.tif, Smi14_GeoSet.tif). They are assigned to the points using the Extract Multi Values to Points tool in Spatial Analyst, which is called by bi-transect-extractor (Sturdivant, 2019).
The values are recoded as follows using bi-transect-extractor, which is documented in the file extractor_smi14.ipynb.
SubType: 7777:'{1111, 2222}', 1000:'{1111, 3333}’ VegType: 77:'{11, 22}', 88:'{22, 33}', 99:'{33, 44}' VegDens: 666: '{111, 222}', 777: '{222, 333}', 888: '{333, 444}', 999: '{222, 333, 444}'
To calculate the values, the data are passed between Pandas dataframe format and ArcGIS geodatabase feature class. Once all values have been calculated, the Pandas dataframe is saved in comma-separated values (CSV) format. Curly brackets ('{}') around two values are used in the modeling process.

Full methods are provided in the associated Methods OFR (Zeigler and others, 2019). For the detailed documentation of the processing, see the files extractor_smi14.ipynb and the source code bi-transect-extractor (Sturdivant, 2019).
smi14_ubw.tif
The width of the upper beach (from oceanside mean high water (MHW) shoreline to either dune crest, dune toe, or coastal armoring/development) is calculated along the supplemented NASC transects.
Calculate beach width along transect:
To calculate the beach width for the transects, upper beach width ('uBW') is calculated based on the difference in position between two points: the position of MHW along the transect ('SL_x', 'SL_y') and the dune toe position or equivalent (usually 'DL_x', 'DL_y'). It is calculated using bi-transect-extractor (Sturdivant, 2019) and documented in extractor_smi14.ipynb as follows:
(1) Find the position along the transect of an orthogonal line drawn to the dune point ('DL_x', 'DL_y' and 'DH_x', 'DH_y'). (2) Calculate distances from MHW to the position along the transect of the dune toe ('DistDL'), dune crest ('DistDH'), and armoring ('DistArm'). (3) Conditionally select the appropriate feature to represent "top of beach." Dune toe is prioritized. If it is not available and 'DH_zMHW' is less than or equal to maxDH, use dune crest. If neither of the dune feature positions satisfy the conditions and an armoring feature intersects with the transect, use the armoring position. If none of the three are possible, 'uBW' is given a NoData value of -99999. (4) Copy the selected distance to oceanside shoreline and height above MHW ('Dist--') to 'uBW'.
Assign values to raster
A transect ID raster is created from the transect file. Transects are modified from the supplemented transect file (smi_trans.shp) by manually shortening overlapping transects to the first point of intersection. Overlapping transects were ignored if the area of overlap did not intersect land. When a transect overlaps with a supplementary transect, the supplementary transect is shortened to the intersection point. The beach width values from each transect are assigned to the cells that represent that transect using the JoinField tool in the data management toolbox, which is also an automated process performed by bi-transect-extractor.
The raster file is exported to GeoTiff using the Export Raster tool. Raster cells outside the bounds of the data are assigned a NoData value of 255.

Added keywords section with USGS persistent identifier as theme keyword.