A Digital Elevation Model (DEM) contains a series of elevations ordered from south to north with the order of the columns from west to east. The DEM is formatted as one ASCII header record (A- record), followed by a series of profile records (B- records) each of which include a short B-record header followed by a series of ASCII integer elevations (typically in units of 1 centimeter {0.01 meter]) per each profile. The last physical record of the DEM is an accuracy record (C-record).
The 7.5-minute DEM (30- by 30-m data spacing) is cast on the Universal Transverse Mercator (UTM) projection. It provides coverage in 7.5- by 7.5-minute blocks. Each product provides the same coverage as a standard USGS 7.5-minute quadrangle but the DEM contains over edge data. Coverage is available for many estuaries of the contiguous United States but is not complete.
7.5-minute DEMs have rows and columns which vary in length and are staggered. The UTM bounding coordinates form a quadrilateral (no two sides are parallel to each other), rather than a rectangle. The user will need to pad out the uneven rows and columns with blanks or flagged data values, if a rectangle is required for the user's application. Some software vendors have incorporated this function into their software for input of standard formatted USGS DEMs.
The data within the bathymetry file is floating point. When using the data within a GIS care must be taken to ensure that the data are being read as floating point and not integer data.
Acknowledgment of the National Oceanic and Atmospheric Administration- Nation Ocean Service would be appreciated in products derived from these data.
The datum for these bathymetric DEMs is not the same as that used by the US Geological survey (USGS) for land based DEMs which results in a discontinuity if the two datasets are merged together. Moreover, the shoreline for the USGS DEMs is indeterminate and not the same as that used for the Bathymetric DEMs.
The data within the bathymetry file is floating point. When using the data within a GIS care must be taken to ensure that the data are being read as floating point and not integer data.
Due to the variable orientation of the 7 1/2 minute quadrilateral in relation to the Universal Transverse Mercator (UTM) projection grid, profiles that pass within the bounds of the DEM quadrilateral may be void of elevation grid points and are not represented in the DEM. This condition occurs infrequently and is always the first or last profile of the dataset.
DEM's may contain void areas caused by elevations being above Mean High Water or on non-tidal land. Void elevations are assigned the value of -32,767. In addition, suspect elevation areas may exist in the DEM but are not specifically identified.
Only available data digitized before 1997 were used in this project. Additional sounding information may exist for areas which have holes in the bathymetric data set, but was not available at the time this project was completed. No additional updates or error corrections are planned for this data set.
It is estimated that the accuracy of the Bathymetric DEMs is 2% of depth or 1 meter for depths grater than 20 meters and 2 % of depth or 0.20 meters for depths shallower than 20 meters. THESE DEMs SHOULD NOT BE USED FOR NAVIGATION.
There are three types of DEM vertical errors: blunder, systematic, and random. These errors are reduced in magnitude by editing but cannot be completely eliminated. Blunders are errors of major proportions and are easily identified and removed during interactive editing. Systematic errors follow some fixed pattern and are introduced by data collection procedures and systems. Systematic error artifacts include vertical unsampled elevation shifts, relative spacing of the source soundings, misinterpretation of terrain surface caused by softness or poor reflectivity and by the resolution of the collected soundings (feet, feet & fractions, fathoms, fathoms & fractions, meters, tenths of meters etc.). Random errors result from unknown or accidental causes. The 1 degree (DSQ) DEMs are generated from 30 m grids on UTM projection. The RMSE difference between these surfaces is an estimate of the vertical accuracy of the DSQ DEMs.
Logsheets were kept at all stages of processing to track file names, dates, hydrographic survey coverage, and soundings or hydrographic surveys that were deleted as part of the quality control process. A short summary of the processing steps for each estuary is available on the individual data pages. In addition, a list of each of the surveys which were included is accessible through the individual data pages.
Original hydrographic data from the National Ocean Service and it predecessors was used exclusively as source data. Processing was performed on desktop computers using a variety of commercial and custom software systems. The two main software systems used were Digital Optimization of Grid Systems (DOGS) version 1.5x software developed by NOAA and MapInfo Professional version 4.5 published by MapInfo augmented by a MapInfo add-on named Vertical Mapper published by Northwood Geosciences. The processing sequence for each estuary started with the generation of a comprehensive source data set from the NOS archives, These source sounding were quality controlled to eliminate outliers and superseded surveys, optimized to reduce the number of data values, augmented with points representing the Mean High Water shoreline, and then gridded. The gridded data sets were converted from the internal proprietary grid format to Digital Elevation Model (DEM) format for public distribution.
Creating Point Sets of Hydrographic Survey Data: Sounding data obtained from Hydrographic Surveys were extracted using DOGS from the GEODAS CD distributed by the National Geophysical Data Center using each estuary's shoreline as a clipping boundary. Large estuaries were broken into several overlapping regions and subsets of points were extracted and processed. Most historic hydrographic surveys are included on the GEODAS CD. For those regions which were missing data (parts of southern Florida & Chesapeake Bays only), soundings were digitized from a hard copy Smooth sheets generated by the Hydrographic Surveys.
Editing and Quality Control of Bathymetry Point Data: The first step of the quality control was to review the data using the DOGS software. The data were first examined for surveys or parts of surveys that are redundant. For many areas there are more surveys than are actually useful. Entire surveys or large portions of surveys were deleted for the following reasons: - Another more recent survey fully covers the same area. - The survey has questionable values which can not be fixed by way of figuring out a mathematical update value for the entire survey, and there are too many bad points to pick out probable "good" values. - Sections of surveys were omitted if they were overlapped by more accurate surveys or by similar yet denser coverage. The second step was to display the data by depth values and to remove stray points that were obvious outliers from the surrounding data values.
Optimizing the Bathymetry Point Data using DOGS: The data was first triangulated in DOGS in order to optimize the set of points. The computer program DOGS can analyze a large set of bathymetric data and create from it a smaller set of optimized points which describes the bathymetry of a geographic region to within a user defined error. This smaller data set then can be used on its own as a representation of the area's bathymetry, or as input into other computer programs or Geographical Information Systems.
There are two calculation options for triangulation in DOGS: relative and absolute. Bathymetry point files were cut into sections for separate processing based on mean depth of 10 meters. A relative height error criteria of 0.01 was used for the triangulation of regions whose depths averaged above 10 meters. An absolute height error criteria of 0.1 meters was used for areas with average depths less than 10 meters. The two resultant files were saved as text files. After combining these files, the vertical error associated with the optimized data set was 1% of depth or 0.1 meters, whichever was greater. The optimized DOG file was imported into MapInfo's MapInfo Professional desktop mapping software.
Augmenting the data set with Shoreline Points: The final bathymetry was clipped to NOAA's 1:250,000 Coastal Assessment Framework (CAF) shoreline. A copy of the shoreline file was edited to create a point file from the vector vertices and optimized using DOGS to produce a more workable, smaller file, with little compromise to the shape of the shoreline. Mean High Water tide level was assigned to the shoreline points to give them a height. These Shoreline data points were added to the set of bathymetry points before doing the final triangulation in MapInfo.
Generation of a Gridded Bathymetry Dataset: Linear triangulation of the combined point files were done in MapInfo using the Vertical Mapper 2.0 partner product software. The resultant TIN file was then used to create a continuous grid file with 30 meter resolution on a UTM projection using a NAD27 horizontal datum. A second grid was created from the first by aggregating the 30 meter grid values to 90 meter resolution and exporting the 90m center points. A new TIN was created after reattaching the shoreline pts. The triangle side lengths and coincident point distances were small enough to disallow interpolation outside the 90m distance, so as to emulate a rectangular interpolation and still allow the shoreline to be represented without averaging out the values. The result of the new file is a geographic 3 arc second resolution grid. Both grids were cut to the estuary boundary shoreline. 7.5 minute and 1 degree sections were then created from these grids. The 7.5 minute grids have 30 meter resolution in a UTM projection using the NAD27 datum. The 1 degree grids have a 3 arc second resolution in a geographic projection (Latitude/Longitude) using the NAD27 datum. In their native form, both of these grids are in proprietary formats.
Creating DEM files: The MapInfo grid files were converted to a public domain USGS format DEM files using the NOAA DEM maker software. The formats available for downloading are 30 meter & 3 arc second DEMs in USGS DEM format. DEM's are viewed on interactive editing systems to identify and correct blunder and systematic errors. DEM's are verified for physical format and logical consistency.