The two images below illustrate the idea. In each, an identical portion of the aerial photograph corresponding to the 'midday' phase of the battle is shown. The Smoketown Road/Hagerstown Turnpike is toward the upper left; the sunken road stretches across the bottom center.
In the first image, the familiar base map available from the Antietam Park Bookstore (hereafter the APB map) is superimposed over the aerial photograph. In the second image, an outline of this document's map is again superimposed on the aerial photograph. Six points (labeled 'A' to 'F') have been added to compare areas between the two images.
Points 'A' and 'B' represent points that are easily visible on both the wartime and modern version (representing where Mumma's Lane meets the Sunken road and a property boundary on the Neikirk Farm). These points were used to place and scale the APB map on the photograph, and consequently fall right on target.
Point 'C' indicates the region where Mumma's Lane meets the Smoketown road. Note the APB version in the first image is bent considerably to the north and east of its proper location. In the second image, the outline of the Mumma Lane/Smoketown road intersection lines up perfectly with the aerial photograph. Likewise, a comparison between the two images of the other points shows that the map in the second image correlates very well to the aerial photograph, while the APB map in the first image is at best an approximation. Consequently, the map in the second image is more accurate than the wartime map of the first image.
The first attempt was to simply create a transparency of the historic map scaled to the same size as the US Geological Survey maps, also available in the Antietam Bookstore. When the wartime transparency was overlayed over the modern map, however, it became apparent that, although there was basic agreement between the two maps regarding major features like roads and fields, it was not possible to slide the transparently such that all features would 'fit' properly. For example, the Dunker Church, Upper and Lower Bridges are all features visible on both wartime and modern maps, and can be assumed to have not moved since 1862. Nonetheless, there was no way to place the wartime transparency over the modern map such that all three fell on top of each other. From this I concluded that the wartime version (drawn up in the 1890s) must represent only an approximation of the lay of the land, at least by modern standards1.
The Antietam National Park Web Site contains topographical data of the park in GRASS format. The information available includes data on the location of roads, buildings, elevation data, modern landuse data, and a high resolution aerial photograph. It should be pointed out that all the map data available here pertains to the modern state of the battlefield; historic features now removed (such as the North and West Woods) do not appear on these maps.
According to its User's Reference Manual, GRASS is "an integrated set of programs designed to provide digitizing, image processing, map production, and geographic information system capabilities to its users". It is a product of the US Army Corps of Engineers Construction Engineering Research Laboratories (USACERL). It is available on-line for free (but see the technical system notes) from the Internet. All information pertaining to obtaining, installing, and using GRASS can be found through the LINK in the bibliography.
Together, the GRASS software and the Antietam park data provided a powerful tool for constructing and analyzing a new wartime map of Antietam.
GRASS software can be divided into 3 categories:
In addition, there are two basic types of maps that GRASS can produce:
Of great importance is the size or resolution of the map. This describes the physical size each square in the matrix represents. By its nature, the data in each square must somehow represent an average of the terrain attribute it describes. The greater the resolution, the smaller each square is, and the overall better quality of the map. Typical values range from 1 meter to a side (very good) to about 10 meters to a side (poor).
Since Raster Maps are just arrays of numbers it is relatively straightforward to perform arithmetic and algorithmic operations on them.
Preparation of Maps for this Document
For this document, two basic maps were created:
For the creation of these maps, the following technique was used. A large scale (1:6300 or about 30"x42") version of the Antietam Battlefield Park's digital data aerial map was printed. Also, a composite of the Park's vector anti.roads, anti.landuse, anti.buildings, anti.hydrology, and anti.boundary with a grid overlay of 20 meter increments was printed at the same scale. Lastly, a series of transparencies of the APB map were made, again at the same scale. Using the modern maps as references for the absolute coordinates of features, over 1400 points where plotted by hand to represent the locations of the edges of various wartime fields, fencelines, woodlots, etc. In practice, about 70% to 80% of all wartime feature boundaries are actually still visible from the aerial photograph in the form of roads, tree lines, and ground cover. The remaining boundaries are extrapolated from the 'best fit' of the APB map over that particular region. A database containing the coordinate pairs of these 1400 points and a database of how the points are connected together to make up the wartime features were then run through some software to create ASCII data suitable for input into the GRASS map construction routines. The end results are the 1862.landuse and fences maps, which by the nature of their construction are in very close agreement with the actual lay of the land. They therefore represent a whole new map of the Antietam Battlefield, not not just a copy of the original flawed survey map done in the 1890s.
Virtually all the maps included in this document have the modern anti.roads map overlayed in white. Thus, the original goal of a map showing both wartime features alongside modern reference points has been met.
1 NB: an examination of the recently published McElfresh Antietam map reveals that it closely follows the APB map, and therefore suffers from the same defects.