Accuracy of shorelines derived from paper maps: In general, the accuracy of a particular map is related to its date. Recent maps are more accurate than older maps. Error can also be introduced by physical changes such as shrinking and swelling in material on which the original data appear.
Accuracy in the interpretation of shorelines derived from aerial photographs: older aerial photographs are often of poor quality, especially along the shore. The beach and swash zone may both appear bright white as a result of their high albedo. The shoreline must be projected through these areas, resulting in the probable introduction of some error.
Error introduced through optical transfer methods: minor errors may be introduced during optical transfer. The error may increase where there is a lack of recognizable features that can be used for guiding the projected image to exactly the right place on the map.
Error introduced through digital rectification: Error may be introduced during the digital rectification process when there is lack of ground control that can tie an aerial photograph to the base map, or when a photograph is particularly distorted.
Many of the shorelines were derived from USGS Digital Orthophoto Quads (DOQs), and all of the digitally rectified aerial photographs are based on them. The DOQ horizontal positional accuracy and the assurance of that accuracy depend, in part, on the accuracy of the data inputs to the rectification process. These inputs consist of the digital elevation model (DEM), aerotriangulation control and methods, the photo source camera calibration, scanner calibration, and aerial photographs that meet National Aerial Photography Program (NAPP) standards. The vertical accuracy of all DEMs are equivalent to or better than a Level 1 or 2 DEM, with a root mean square error (RMSE) of no greater than 7.0 meters. Field control is acquired by third order class 1 or better survey methods sufficiently spaced to meet National Map Accuracy Standards (NMAS) for 1:12,000-scale products. Aerial cameras have current certification from the USGS, National Mapping Division, Optical Science Laboratory. Test calibration scans are performed on all source photography scanners. Adjacent DOQs, when displayed together in a common planimetric coordinate system, may exhibit slight positional discrepancies across common DOQ boundaries. Linear features, such as streets, may not be continuous. These edge mismatches, however, still conform to positional horizontal accuracy within the NMAS. The Horizontal_Positional_Accuracy Value (see below) is an root mean square error (RMSE) of image positions to ground coordinates as determined by the orthorectification software used. (Source: Metadata for USGS DOQs, 1995) RMS error for georectified images are generally less than 7 meters.
Process for shorelines derived from Digital Orthophoto Quarter Quads (DOQQ): The shorelines were interpreted as the contact between wet and dry sand, and were digitized at a scale of 1:3000 using GIS software. DOQQs are differentially corrected photographs, each one covering one-quarter of the area covered by a USGS topo sheet. The DOQQs that make up the area being studied were flown in 1995 and 1996 at a scale of 1:40,000. They are color infrared and 1 meter resolution.
Process for shorelines derived from LIDAR: 1. Transfer data sets to NT workstation: a) raw ALTM 1225 (serial number: 99D117) flight data recorded/downloaded using Exabyte 8mm tape drive, b) airborne GPS data collected at 1 Hz using Novatel receiver, and c) ground-based GPS data collected at 1 Hz usingTrimble 4000SSI receivers.
2. Generate LIDAR point file from above three data sets using Optech'sREALM 2.27 software (Optech Inc., 2000). This is a 9-column ASCII dataset with the following data format: time, first pulse easting, northing, and HAE, last pulse easting, northing, and HAE, first pulse intensity, last pulse intensity.
3. View decimated LIDAR point file using Surfer or TerraModel software to check data coverage (i.e., sufficient overlap of flight lines and point spacing).
4. Collect and process GPS ground data within target area(s) for ground control checks and for elevation biascorrection estimation.
5. Generate second aircraft trajectories using Optech's REALM 3.02 software for 2 days, May 24 and 25 (Optech Inc., 2001).
6. Use POSProc (Applanix Corp., 1999) to compute an optimally accurate navigation solution for the data files.
7. Substitute POSProc aircraft trajectory and navigation solution into REALM 2.27 and regenerate LIDAR point file.
8. Extract calibration area data set from LIDAR point file for quality control/instrument calibration checks.If necessary, iteratively adjust calibration parameters (pitch, roll, and scale) and reprocess sample data set. Then regenerate entire LIDAR point file (9-column ASCII dataset).
9. Transfer LIDAR point file from NT workstation to UNIX. 9-column ASCII dataset has the following information: column data 1 GPS time 2 first return x value 3 first return y value 4 first return z value 5 second return x value 6 second return y value 7 second return z value 8 first return intensity value 9 second return intensity value
10. Parse the 9-column LIDAR point file into smaller 3.75-minute quarter-quad components.
11. Grid the quarter-quad point files with software written in-house. We are able to grid simultaneously the four following attribute data: a. first return z b. first return intensity c. second return z d. second return intensity ...and output the data into one of two formats: a. an ArcInfo ASCII raster file. Using this format, each one of the four attributes listed above must be output to a separate file for import into ArcView (must have Spatial Analyst extension) or ArcInfo. This format consists of a matrix of attribute values preceded by six lines of header information including: 1. number of columns 2. number of rows 3. x coordinate of the lower-left cell 4. y coordinate of the lower-left cell 5. cell size 6. null value b. a raw 4-byte binary raster file. Using this format, we can generate multi-band, band interleave files containing one, two, three, or all four of the attribute data referred to above. Additionally, we output a header file in ERMapper's ".ers" format for each of the binary files so that the data can be viewed in ERMapper or ArcView, with the appropriate ECW plug-in. These header files contain the same information as the ArcInfo-format header files (except the coordinate values are of the upper-left cell) plus datum and projection information. The gridding software written in-house uses a weighted inverse distance algorithm to interpolate cell values. It also provides the option of converting from HAE toNAVD88 or ITRF97 by subtracting the appropriate geoid model from the internal HAE array before writing to output.
12. Extract +0.6m MSL contour. The grids generated above were opened in ERMapper and a 0.6m contour line was calculated and displayed. This contour polyline is very complex in places, so we used it as a guide to digitize a new 0.6m contour line, which is the shoreline we've provided.
Process for West and East Bay shoreline extraction from Lidar DEM: The shoreline was extracted by manually digitizing a line that represents approximately the mean higher high water elevation. Through comparisons with tide gauges in West and Christmas Bays the 0.3 m level relative to NAVD88 was determined to be the mean higher high water level. This elevation generally coincides with the water line in images of the DEM. Where low marsh is present the line is along the seaward edge, which is consistent with shorelines interpreted on historical aerial photography. The DEM was contoured with the 0.3 m line and overlain on an image of the DEM. Because of the very subtle topography, variations in vegetation height, and errors in the lidar data, the 0.3 m contour line is highly broken up into short segments in places. To acquire a continuous shoreline, therefore, the line was digitized using the 0.3 m contour line and the DEM image for guidance. However, the 0.3 m contour line was extracted where it was continuous. This method provided a continuous shoreline, where justified, while preserving the greatest amount of detail and accuracy.
Process for shorelines derived from optical transfer methods: The shoreline was interpreted as the boundary between wet and dry sand. Using a zoom transfer scope, Saltzman projector, or reducing pantograph, aerial photographs are either enlarged or reduced to the precise scale of a topographic map, and are mechanically projected onto the map using light and mirrors. the shoreline is drawn directly onto the map using a pencil. The shorelines may then be digitized for use in GIS software.
Process for shorelines derived from maps: Land/water interfaces are digitized directly from paper maps using a digitizing tablet and GIS software.
Process for shorelines derived from GPS: Shorelines derived from GPS were surveyed using a dual-antenna real-time kinematic differential global positioning system (DGPS) mounted on a four-wheel-drive all-terrain vehicle (fig. 6). Kinematic refers to the continuous movement of the GPS antenna as the vehicle is driven along the beach, and real-time differential means that the corrected position of the GPS antenna is received at the time of the survey. These GPS terms are explained in more detail in Morton and others (1993). During kinematic beach surveys, horizontal positions were collected at a 1-s sampling interval, which translates to an average alongshore spacing of approximately 15 ft at high speed and 10 ft at low speed. Static positions were recorded for 5 min at the beginning and end of each beach segment. Beach segments were limited in length by natural features, such as large drainage channels, or physical barriers, such as cables or revetments across the beach, that prevented continuous lateral movement. Within a beach segment, way points were recorded to mark the positions of prominent (reference) features (drainage channels or houses on the beach) or the locations where the surveyed shoreline feature changed from one type to another. Most of the way points were photographed, and field notes were recorded for future reference. The entire 1996 DGPS survey of the shoreline took 4 days. The segment between Sabine Pass and Bolivar Roads was conducted on February 7 and 9, whereas the segments between Bolivar Roads and the Brazos River were completed on May 14 and 15. The raw DGPS data were converted to State Plane, South-Central Zone, NAD 27 datum, survey feet. Several files of positions were collected in a nondifferential mode when the differential receiver was unable to provide corrected positions. These files were corrected in postprocessing using differential corrections from the Texas Department of Transportation Houston Area Regional Network station.
Sources
BLACK AND WHITE PHOTOMOSAICS: Tobin Research Inc.
AERIAL PHOTOGRAPHS: U. S. Dept. Agriculture, U. S. Army Corps Engineers, Texas Highway Department, Natl. Oceanic and Atmospheric Administration, Texas Forest Service, Texas General Land Office, Environmental Protection Agency, International Boundary and Water Commission, Jack Ammann
MAPS: Natl. Oceanic and Atmospheric Administration, U. S. Army Corps Engineers, U. S. Geological Survey
Meaning of "source" values: 1. Shoreline derived from aerial photographs that have had their shorelines transferred to a topo sheet using optical transfer devices. 2. Shoreline derived from digitally rectified aerial photographs. 3. Shoreline derived from a USGS Digital Orthophoto Quad (DOQ). 4. Shoreline derived from LIDAR data. 5. Shoreline derived from ground GPS survey. 6. Shoreline derived from a paper map.
Dates: The exact date was not known for all of the lines. Some have only a year, others have a month and year but no day.