This dataset contains polyline features representing historical (1930s) and contemporary marsh-forest boundaries across three sites in New Jersey, New York, and Massachusetts. Historical boundaries were digitized from georeferenced NOAA T-sheets (1932-1938), while current boundaries were derived from the National Wetlands Inventory (NWI) polygons (2004-2009). To quantify marsh-forest boundary movement, custom baselines were created for use in the Digital Shoreline Analysis System (DSAS v.6.0.168), from which transects were cast perpendicular to intersect both historical and current marsh boundaries. These intersections enable calculation of net boundary movement and annual rates of change. The dataset includes all spatial layers required for DSAS analysis, including formatted baselines, boundary polylines, and transect outputs. No personal or sensitive information is included, and all source data are publicly available from NOAA and the USFWS. 

1. Deriving the Historical Marsh-Forest Boundary

Historical georeferenced T-sheets from the early 1900s (1930–1938) were downloaded from NOAA’s National Geodetic Survey website. These T-sheets were originally mapped using the NAD 1927 datum and were later georeferenced and projected to the WGS 1984 coordinate system by NOAA during digitization. The NJ T-sheet was originally created in 1932, the NY T-sheet was originally created in 1933, and the MA T-sheet was originally created in 1938.

The T-sheets were imported into ArcGIS Pro (v.3.1.0), where the marsh-forest boundary was manually digitized. A polyline feature class was created to trace the boundary, following the visual distinction between the marsh (represented by hash marks) and the forest (depicted as asymmetrical tree symbols).

2. Deriving the Current Marsh-Forest Boundary

The current marsh-forest boundary was derived using wetland polygon layers downloaded from the National Wetlands Inventory (NWI). Polygons classified as wetlands by the NWI were used to create a polyline feature class that traced the outermost boundary of the wetland polygons. The NJ NWI wetlands polygon layer was originally created in 2009, the NY NWI wetlands polygon layer was originally created in 2004, and the MA NWI wetlands polygon layer was originally created in 2008.

3. Preparation for DSAS

To enable rate calculations and transect casting in the Digital Shoreline Analysis System (DSAS v.6.0.168), a baseline layer was created in ArcGIS Pro. This baseline served as a reference line from which transects could be cast. The baseline’s orientation, length, and position were customized for each study site to capture the extent of marsh-forest boundary changes.

The baseline was typically placed parallel to the marsh-forest boundary to accurately capture small-scale changes while also encompassing significant areas of forest retreat or marsh progression. To standardize the analysis, the baseline was positioned so that transects would first intersect the historical marsh-forest boundary, followed by the current marsh-forest boundary. This ensured that DSAS could interpret changes in the boundary as either:

• Positive: Current marsh-forest boundary closer to the baseline relative to the historical boundary.
• Negative: Historical marsh-forest boundary closer to the baseline relative to the current boundary.

The baseline layer was populated with the attributes required for use in DSAS, as specified in the readme file. Historical and current marsh-forest boundaries were combined into a single feature class, also formatted with the necessary attributes for DSAS and also detailed in the readme file.

Finally, both the baseline and the marsh-forest boundary feature classes were converted to GeoJSON format for use in DSAS.

4. DSAS Analysis: Casting Transects and Calculating Rates of Change

Both the baseline and marsh-forest boundary GeoJSON files were imported into DSAS (v.6.0.168), which automatically projected both files to EPSG: 32618 (WGS 84: UTM zone 18N) for analysis. Transects were cast at 20-meter intervals along the baseline, with transect lengths ranging from 700 to 1000 meters, depending on the site, to ensure full coverage of the marsh-forest boundary. Transect length did not affect rate calculations 

Rates of change were calculated using the Calculate Rates function in DSAS. The analysis used:

• Baseline Layer: The custom baseline shapefile.
• Shoreline Layer: The merged historical/current marsh-forest boundary shapefile.
• Transect Layer: The DSAS generated transect layer.

A shoreline intersection threshold of 2 was applied since only two marsh-forest boundaries (historical and current) were analyzed. Rates were visualized using the Shoreline Change Envelope (SCE) parameter and displayed with a color ramp to highlight spatial patterns of change. A 90% confidence interval was applied to the calculations.