Data Processing

Orthoimages of Assateague Island were acquired in digital format from the U.S. Department of Agriculture (USDA), U.S. Geological Survey (USGS), and Virginia Geographic Information Network (VGIN). They were used to extract the data included here.

Variability in the quality of the imagery from one date to another, from one image to another taken on the same date, and from one part of an image to another part of the same image exists due to many factors including variation in sun angle, atmospheric conditions, season, reflectivity of the earth surface, camera angle, camera lens, camera settings, film sensitivity, water turbidity, water depth, tide level, waves, and error in image rectification. No adjustments were made to account for tide level. Based on areas where no long-term, statistically significant changes were observed, the standard errors of linear regression are as follows.

• For the back-island shoreline: 23.28-m average, 0.16-m minimum, and 569.47-m maximum;
• For the sand line, 24.21-m average, 0.30-m minimum, and 119.86-m maximum;
• For the open-ocean shoreline, 9.18-m average, 0.47-m minimum, and 32.68-m maximum.

Table 1. Information about the orthoimagery used in the Assateague Island study, Maryland and Virginia. [MD, Maryland; VA, Virginia; DOQQ, digital orthophoto quarter quads; NAIP, National Agriculture Imagery Program; VBMP, Virginia Base Map Program; USGS, U.S. Geological Survey; USDA, U.S. Department of Agriculture; VGIN, Virginia Geographic Information Network]

Open-Ocean Shorelines: A line was hand-digitized at the approximate water line by using ArcMap 10.1 and a scale of approximately 1:2,000. The line placement was intended to represent the still water level, such as tide-level, at the time of acquisition and was fit through the average position of waves and swash apparent on the beach. No adjustment was made for varying tide levels.

Back-Island Shoreline Points: Shore-normal transect lines spaced 20 m apart were laid over the orthoimages by using ArcMap 10.1 and a scale of approximately 1:2,000. Points were hand-digitized at the intersections of the apparent back-island shoreline and the transect lines. Only one back-island shoreline/transect point was digitized per transect line.

Back-Island Shorelines: These lines are the back-island shoreline points connected together to form a line. The lines falling between the points or transect lines do not follow the natural shoreline. They have been included for use with software that requires data in line format.

Sand Polygons: The orthoimagery was subset, or clipped, to the Assateague Island area by using ERDAS Imagine 9.3. Ten to 40 image classes were generated by using the unsupervised classification method with the following settings: principal axis initializing, 1 standard deviation, 10 iterations maximum, 0.95 convergence threshold, skip factor of 1 for both x and y (no skipping), and zeroes not classified. The resulting classes were visually examined and coded as either sand or not sand. The sand-coded raster datasets were converted to polygons by using ArcGIS 10.1. Polygons outside the island area were deleted. Polygons with areas less than 100 square meters (m²) were deleted. The open-ocean shoreline edge of the polygons were trimmed and filled to the open-ocean shorelines digitized from the same image to simplify the irregular and misclassified areas caused by wet sand and waves.

Sand Lines: The sand area polygons previously described were converted to lines by using ArcGIS 10.1. The lines that, in general, composed a continuous line of the inland extent of the sand, were kept. A rule was used to determine which lines were kept around islands of vegetation seaward of the sand line. The vegetation island had to be 100 m² or larger, had to be the closest vegetation island, in the shore-normal direction, of that size to the continuous sand line, and the distance between the continuous sand line and the vegetation island had to be less than or equal to the shore-normal width of the vegetation island for 50 percent or more of its longshore length. The remaining lines were deleted.

Transects and Baselines: These have been used for various USGS studies and are included here for convenience (Hapke and others, 2010).  

Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. Government.