2012-2014 contour-derived mean high water shorelines of the Massachusetts coast used in shoreline change analysis

Explanation of the methods used to delineate shoreline features that are a part of this update for the Massachusetts Office of Coastal Zone Management Shoreline Change Project:
This data release has two methods of shoreline extraction: the profile method and the contour method. The profile method is further broken down into two varieties: shorelines that are extracted along open-ocean coasts and those that are sheltered.
Profile open-ocean coast (bias): datum-based mean high water (MHW) shoreline. The elevation of MHW was obtained from Weber and others (2005). These data have an associated uncertainty table that provides the horizontal uncertainty associated with the shoreline, a proxy-datum bias value describing the unidirectional horizontal offset between the MHW shoreline and the historical proxy-based high water line (HWL) shorelines, and the uncertainty associated with the calculated proxy-datum bias value. These shorelines are polyline-M shapefiles.
Profile not open-ocean (no bias): datum-based mean high water (MHW) shoreline. Since Weber and others (2005) only covers open-ocean coast, all MHW elevations for these data come from NOAA's vdatum (version 3.8; https://vdatum.noaa.gov/). These data have an associated uncertainty table that provides the horizontal uncertainty associated with the shoreline. These shorelines are polyline-M shapefiles.
Contour method: this method is used along sections of the coast that were too crenulated for the for the profile method. The elevation of MHW was used from Weber and others, 2005 when available. In areas not covered by Weber and others (2005), NOAA's vdatum (version 3.8; https://vdatum.noaa.gov/) is used to determine MHW. Once this value is determined, the contour line of that value is extracted from the DEM surface in the area of interest. These shorelines are polyline shapefiles.
Not included in this data release is another method of shoreline delineation. A brief explanation of this method is provided to convey the importance of the information contained in the uncertainty tables, even if that information is not actively used in all cases in this data release.
Proxy-based historical shorelines: Vector shorelines digitized from georegistered T-sheets using standard editing tools in ArcMap provide a proxy-based high water line (HWL) feature that is not tidally-referenced. Individually these shorelines are stored as polyline shapefiles. In previous analyses (see Himmelstoss and others, 2011 listed in cross references) these data were published as a merged file with profile method data extracted from lidar. Therefore the published data are all polyline-M, but the historic HWL shorelines contain no linear referencing. Visually identified HWL-type proxy shorelines are virtually never coincident with datum-based MHW-type shorelines. In fact, HWL shorelines are almost universally estimated to be higher (landward) on the beach profile than MHW shorelines. Not accounting for this offset will cause shoreline change rates to be biased toward slower shoreline retreat, progradation rather than retreat, or faster progradation than in reality (for the typical case where datum-based MHW shorelines are more recent data than the proxy-based HWL shoreline dates).
The Digital Shoreline Analysis System software used to compute rates detects when proxy-based and datum-based shorelines are present and uses linear referencing to retrieve the information on bias and uncertainty stored in the DBF table associated with the profile method shorelines and correct for the proxy-datum bias offset.

In areas where the sandy coastline was complex and/or curvilinear, our preferred profile method of extracting a lidar shoreline was not an optimal choice as it is best suited for long, linear sections of coast. Instead, a contour method was used to extract the operational mean high water (MHW) shoreline.
The height of MHW was determined from vdatum provided by NOAA. (https://vdatum.noaa.gov/). We continued the practice set out by Weber and others (2005) of using one MHW value for a continuous section of coast (as opposed to using a continuously varying value). We chose this value such that it is always within 15 cm of the value returned by vdatum at any point along the coast. For example, MHW = 0.6 m was used for all of Buzzards Bay even though vdatum shows it varying slightly over the basin. This is the contour value used to delineate the shoreline and is recorded as the contour attribute in this shapefile.
Two data sets were used to extract MHW shorelines using the contour method. The dataset from which the majority of contour shorelines were extracted was the 2013-2014 U.S. Geological Survey Coastal and Marine Geology Program lidar data set. There were some gaps in these data along the coasts of Martha’s Vineyard and Tuckernuck Island so a second data set, the 2012 U.S. Army Corps of Engineers Post-Sandy Lidar data, was used to fill in the gaps. The lidar data were downloaded in LAS format (lidar data exchange file format) from NOAA’s Data Access Viewer: https://coast.noaa.gov/dataviewer/#/ in the UTM Zone 19 projection with the NAD83 horizontal datum and NAVD88 vertical datum.
This process step and all subsequent process steps were performed by Amy Farris.

A series of ArcGIS tools (ArcGIS version 10.4.1) were used to extract a MHW contour from the lidar data. First, the LAS files were loaded into ArcMap using the ArcToolbox tool LAS to Multipoint, and feature classes were created using the tool Multipart to Singlepart. Location and elevation data were then added to each point in the attribute table using Add XY Coordinates. There was often overlap between the search boxes when downloading the data from the NOAA website and this resulted in repeated points in the singlepart features. These duplicate points were deleted using the Delete Identical tool in ArcToolbox. The feature classes were then divided into two subsets using the Subset Features tool. This ArcToolbox function divided the data into a training set that randomly included 90% of the points and a test set that included 10% of the points. The 10% test set was used to estimate the uncertainty associated with the gridding process. This uncertainty was found to be significantly less than the other sources of uncertainty discussed in the horizontal accuracy report section of this metadata file. The 10% omitted from the surface construction were initially going to be used to estimate a surface error for the DEM, but later in the process of extracting these shoreline features, it was decided to implement the total positional uncertainty described in the horizontal accuracy report section of this metadata file instead. The files with 90% of the points were brought into feature datasets to create terrain surfaces. The tool Terrain to Raster was then used to build 1 m x 1 m digital elevation models (DEMs) using a Natural Neighbours interpolation method.

To extract the MHW shoreline from each DEM surface, the operational MHW elevation was entered into the ArcGIS tool Contour as the contour value, and the Smooth Line tool was used to smooth the resulting contour with a 10 m smoothing tolerance and the PAEK smoothing algorithm. The shoreline was then manually edited with the lidar data displayed with categorized elevation values to highlight the MHW values. Segments of the contour that did not fall near MHW were removed. Sometimes the lidar data did not extend down to MHW, resulting in gaps in the contour.

The feature class was exported out of the geodatabase to shapefile format by right-clicking on the feature class name in the table of contents of the ArcMap project and choosing Export > Shapefile.

The data were projected in Esri's ArcToolbox v10.5 > Data Management Tools > Projections and Transformations > Feature > Project. Parameters: input projection = UTM zone 19N (NAD83); output projection = geographic coordinates (WGS84); transformation = WGS_1984_To_NAD_1983_1.

Added keywords section with USGS persistent identifier as theme keyword (20200810). Tweaked a thesaurus name (20211117).