This data set has been merged into a county files for ease of processing and decision making.
In digitizing historical shorelines from hard copy maps, MGS had adopted an operational definition of the headward extent of a stream - the point at which the stream was represented on the map by a single line versus a double line delineating opposite stream banks. That point, dependent as it is on the scale of the map, does not necessarily represent the head of tide. A similar criterion was used in extracting the 1990 shoreline to the extent that the contractor could fit a double line on photography displayed at a scale of 1:24,000. Reaches upstream of that point were excluded. Again, the headward extent of a digitized stream does not necessarily correspond with head of tide.
Shorelines in the 1990 data set bounded several different wetlands classifications (Cowardin et al., 1979): 1. E 1 UB L (Estuarine -- Subtidal -- Unconsolidated bottom -- Subtidal) 2. E 2 US M (Estuarine Intertidal Unconsolidated shore Irregularly exposed) 3. E 2 US N (Estuarine Intertidal Unconsolidated shore Regularly flooded) 4. E 2 US P (Estuarine Intertidal Unconsolidated shore Irregularly flooded) 5. M 2 US M (Marine Intertidal Unconsolidated shore Irregularly exposed) 6. M 2 US N (Marine Intertidal Unconsolidated shore Regularly flooded) 7. R 1 UB V (Riverine -- Tidal -- Unconsolidated bottom -- Permanent tidal (tidally influenced, freshwater system) 8. R 2 UB 8 (Riverine -- Lower perennial -- Unconsolidated bottom -- Eusaline) In addition to designating UB L and UB V as water, M 2 or E 2 US (Unconsolidated shore) M (Irregularly flooded) or N (Regularly flooded) classifications were considered water, given that those areas were more likely to be under water than above. E 2 US P (Irregularly flooded) was considered upland, given that those areas were more likely to be above water than under.
The production of accurately georeferenced DOQQs depends on sufficient ground control. In open water, ground control is necessarily lacking, because land areas, if they exist at all, are small and/or inaccessible. Two quarter quads in the data set, Deal Island NW and Barren Island NE, contain so much open water that they are not properly tied to the adjoining tiles, Nanticoke SW and Honga NW, respectively. Consequently, corresponding line work (shoreline) had to be visually corrected to maintain consistency within the data set.
For shoreline vectors extracted from the existing wetlands coverage, EDI stripped wetlands vectors of their linear attributes (line classes). All shorelines and the DOQQ tile boundary were displayed over the DOQQ (raster) from which they were originally interpreted and reassigned attributes. Shoreline segments were classified using the following four categories: beach, structure, vegetated, and water's edge. All four categories are linear features, except for "beach," which can be both linear and polygonal. The DOQQ tile boundary was arbitrarily assigned one of the four categories so that it could be extracted with the shoreline vectors. Shoreline vectors were extracted from the original vector set by line class, using only the four shoreline categories. The extracted vectors were then displayed to detect shoreline breaks or other consistencies. Errors were corrected, and shoreline vectors were re-extracted. After extraction, vector sets were cleaned by deleting any extraneous lines (non-shoreline vectors) that had been mistakenly assigned one of the four categories before extraction. Beach polygons were attributed at this time, and the DOQQ tile boundary was assigned to a fifth category - unclassified. The final quality control check consisted of two steps. First, each tile (quarter quad) was displayed individually to check for unclassified shoreline vectors. Second, if the photography on which the DOQQs were based was flown in the same year, vectors from adjacent tiles were merged into a single, 7.5-minute quadrangle vector set. Otherwise, multi-year shorelines for the same quad were stored as separate objects within the TNTmips project (.rvc) file. Merged vector sets were displayed to check for class consistency (proper edge matching) between adjacent tiles. Any necessary corrections were made before saving the final vector sets.
Shorelines derived directly from DOQQs were interpreted according to the procedure outlined in the wetlands inventory metadata.
The DOQQs were already georeferenced to the Maryland State Plane Coordinate System (North American Datum of 1983, meters). Georeferencing was transferred automatically upon delineating either wetlands or shorelines.
The second step in creating the state wide file was to add the Maryland State Boundary to the linear shorelines. The linear shorelines were merged with the most recent Political State Boundary. Unfortunately, the final output of this merge also left a file with undershoots, dangles, and mismatches. Undershoots were just a lack of connection between the linear shoreline and the State Boundary. Therefore, across connector was added and attributed. Dangles on the other hand were lines digitized from the orthophoto that extended outside the Political State Boundary. These lines were closed and attributed as a cut off to digitizing effort. And, an attribute field was added to describe to which state the linear feature belongs(Maryland, Delaware, Washington, or Virgin). The State Wide file was once again validated for topology.
Each .rvc file contains one or two "objects" shoreline vectors for particular years. Naming and describing objects also follow conventions. The object name "_19xx" always indicates the shoreline year. The object description gives the exact date on which the aerial photography from which the DOQQs were derived was flown. For example, "1994 shoreline from DOQQs flown 4/8/1994". Line and polygon attribute tables, both named "TYPE," are associated with each object. They contain the attributes described in Section 5.1. A second table "TypeStyle" contains the style assignment for the TYPE Table. TypeStyle points to another project file, ShoreStyles.rvc, which contains common line and polygon styles for the entire data set.