SOUTH BAY HISTORICAL TIDAL MARSH GIS, version 2.0
This dataset integrates several sources of data describing the historical features of South Bay tidal marshes. The primary source is the maps of the United States Coast Survey (USCS; later US Coast and Geodetic Survey), a federal agency renowned for the accuracy and detail of its 19th-century maps of America's shoreline. In most parts of the country, these maps provide the best early pictures of coastal and estuarine habitats prior to substantial Euro-American modification.
The initial effort (Phase 1) resulted in a single polyline layer, and subsequent work improved the data by creating a polygon feature class and a single-channel polyline feature class (Phase 2). Each of these efforts is described briefly below.
In Phase 1, nine Coast Survey maps (T-sheets) were individually georectified and vectorized. The georectification included transforming historical latitude/longitude markings on the original map to the modern coordinate grid and, subsequently, a locally-derived adjustment to maximize correspondence between persistent physiographic features such as "ghost" channels in diked baylands, road intersections, and hilltops. Marsh features, in particular tidal channels and pannes, were vectorized using pattern recognition software and careful on-screen digitizing by GIS technicians. The T-sheets were then appended into a single polyline coverage.
Phase 1 also incorporated several additional sources to those described above -- including one additional Coast Survey map (T 2312) and 1939 aerial photography -- to fill gaps in either the spatial coverage or the level of detail in the T-sheets. These sources and the estimated associated certainty levels are recorded in the coverage attributes. Three distinct kinds of certainty are recorded: the certainty of our interpretation of the feature, its size/shape, and its location. The layer's attributes also record the original source, original surveyor, year of survey or flight, the GIS layer author, and the methods used.
In Phase 2, the initial polyline layer was then improved by creating a polygon feature class from the polyline layer. Wherever possible, errors in vectorizing were corrected and missing features were added. Additionally, large channel networks were separated and coded, landscape attribution was added, and the T-sheet layer was integrated with information existing in SFEI's EcoAtlas. For ease of use given the large amount of data, the single layer was tiled into pieces corresponding with the USGS 7.5 minute series quadrangles. Additionally, a single-channel polyline feature class was created for each quadrangle to show the channels that are not part of the polygon layer. These single-line tidal channels, while small, account for a substantional portion of overall channel . The polygon and polyline feature classes can be used together to represent tidal marsh features as polygons and single channels and polylines.
As with any map or GIS layer, confidence or certainty varies geographically due to differences in source data or methods. Certainty levels were recorded in the polyline feature class for each 7.5 minute quadrangle. This geodatabase provides direct information about the certainty level of different areas and features to allow the user to intelligently assess the applicability of the data for the chosen technical question.
The dataset provides a voluminous set of information about early geomorphic and ecological characteristics of South Bay tidal marshlands. Based upon our research on early surveyors and map accuracy, most of this dataset represents a consistent level of detail and accuracy. However, there are areas of less detail that should be recognized, particularly towards the upland edge of the extreme South Bay, and in several other specific areas. Also because several later sources had to be used to fill in these gaps, the extent of anthropogenic influence increases in these places, resulting in some modified channels and associated features such as landings and ditches. Technical users are strongly encouraged to read the process documentation to gain a stronger understanding of the strengths and limitations of the dataset.