Research Methodology

III. Research Methodology

This section on research methodology outlines the strategies used to accomplish the goals of this project.

3.1 Documentary Research

Research on the Chemung Battlefield began with the identification of primary historic records and secondary sources related to the battle and the Sullivan-Clinton Campaign. Appendix II contains a list of archival sources reviewed for this study. Researchers examined these documents and maps to identify landmarks, positions, and topographic features related to the Battle of Chemung. The review also keyed on information that would help researchers understand the combatants' motives and morale, both of which are closely related to how the battle unfolded and its aftermath. Research teams recorded any references or remarks relating to landscape features on a common form. This form recorded the name of the feature, the location of the feature, the relevance to the battle, the bibliographic source of the statement, the perspective of the source, any additional comments, and if the reference had a map or illustration of the feature. Researchers entered the records into an Access database for use in comparison and GIS mapping (See Appendix III).

A feature's name was often taken from the source's reference to the feature, such as "hill," "ridge," or "ambuscade." We standardized these names for use in the defining features database and the GIS. If descriptions of features had different names, such as hill versus mountain, but the locations and overall descriptions were similar, we concluded that the sources were referring to the same feature and they were given the most widely used name.

Researchers also recorded the contextual information of the source or witness, including bibliographic referencing. This contextual information included the source perspective, rank, and military unit. Perspective was defined as the source's "side" - British, Continental, or Native American. Rank and military unit represents a further breakdown of the perspective information. The contextual information allowed for comparisons of various sources. Analysts could compare different views of the same feature, or visualize how the same participant saw different features on the battlefield. The contextual information also could be used to track individuals across the battlefield for the duration of the battle, providing there was enough information from the witness. The identification of military unit helped to place the individual in the battlefield based on general troop movements. The placement of individuals allowed researchers to determine the reliability of the source's perspective of certain features and movements. For example, a Continental troop in Lt. Co. Adam Hubley's brigade was given more credence in describing the movement of Continental troops at the ambuscade/ ambush point than those of General Clinton's troops who were not with General Sullivan's troops at the Battle of Chemung. Those troops who were describing a feature or movement for which they were not directly associated were probably relaying secondary information, and although it may be correct, a direct participant's observations of an action were given precedence. Although we prioritized some sources given the context, it does not mean others were ignored. Relevance to the battle was an attempt to categorize the function of the feature or the role it had in the battle. It was a basic description of the feature provided by the witness. It was also an initial step towards performing a KOCOA analysis. This relevance can be further categorized using KOCOA analysis.

The recording of each observation of the different features also allowed for a comparison of views. The back and forth comparison of various sources' observations on features permitted a commonality between sources and general view of the landscape to develop. This comparative interpretation was illustrated on a series of GIS maps to further develop a comparison to known features. The alignment of historically identified features and the present landscape with added information, such as historic images, maps, and soil maps helped to show the changes in the battlefield's landscape and determine which present day features could be related to the battlefield's features.

3.2 Initial GIS Mapping

A GIS map for the project was constructed using the software program ArcGIS 9.2, and USGS 7.5 minute quadrangles [text deleted] as initial basemaps. Aerial photographs of the project area taken in 2006 were also used as a layer to display the landscape's current setting. Historic aerial photographs from 1939 were also utilized to help identify changes in the battlefield's landscape during the 20th century. Digital elevation models (DEMs) for the quadrangles were also used and interpolated in ArcMap 9.2 to conduct slope and viewshed analysis. Basic locations for defining features were mapped into a GIS based on reference to the modern landscape as identified on the USGS quadrangles. Previous excavations and field-testing conducted by the Public Archaeology Facility over the past 30 years were added to the GIS maps (Figures 34, 35, and 36). These data included boundaries of testing, locations of any identified sites, cultural affiliation of sites, soil types encountered during testing, and bibliographic reference of project reports. Testing for these projects included shovel test pits (STPs) and test unit excavation. Intensive remote sensing or metal detecting surveys were not used in this testing. The data related to these past projects can inform on the presence or absence of disturbance in areas of the battlefield. The presence or absence of intact soils associated with defining features of the battle can initially determine if a defining feature has good integrity.

Integrated with this GIS basemap was an initial mapping of the defining features identified in the documentary research. The initial locations and boundaries of the features were based on the descriptions of distances and sizes related to other defining features and other topographic features. Boundaries were also initially compared to the topographic and aerial photographs to relate the general locations of the defining features to the present landscape. The initial mapping allowed a basis for field research and further analysis and interpretation of the documentary accounts. The accounts provided an initial insight that field mapping and further GIS analysis would later help to either confirm or contradict.

3.3 Field Methods

After the defining features identified from the historic records were initially located in the GIS, researchers prioritized these features based on relevance to the battle and potential for preservation. Researchers assessed this potential using the nature of the feature and any information related to disturbance or integrity attained from previous testing. Field crews mapped primary defining features as well as features with less priority as they walked the battlefield. Since this project did not include subsurface testing, features were identified based on measurements provided in historic sources, relation to larger features (hills), and presence above the surface or as part of the terrain.

The survey team logged each Battlefield feature using a Trimble GeoWm 2008 series global positioning system (GPS) running Windows Mobile and ArcPad 7. The data dictionary used was the one developed by the National Park Service for the Revolutionary War/War of 1812 documentation project. Not all items in the data dictionary were relevant to this study. The team used this dictionary in order to be consistent with the previously collected NPS data. The GPS system parameters were set to those used in previous studies: PDOP Mask: 6.0; SNR Mask: 6.0; Elevation Mask: 15: Minimum satellites: 4 (NPS 2001).

Photographs were taken of the battlefield, including the locations of defining features and general shots of the overall battlefield landscape. Whenever possible, the survey team took multiple shots in a 360-degree pan from a central point. These shots were used to create an overall perspective of the landscape surrounding a defining feature.

3.4 GIS Analysis

The results of the field mapping helped to refine the locations of the defining features and identify their contexts. The data recorded by the GPS and the photographs were integrated into the GIS base map. The data aided in the comparison of the different accounts and data sets reviewed in this study. The comparative analysis of historic sources, present location and condition of landscape features, and known changes to the battlefield's landscape allowed for a refinement of the locations and boundaries for the battle's defining features. Specialized GIS analysis was conducted to further refine these boundaries and locations.

Digital Elevation Models (DEM) were interpolated for slope in the GIS in conjunction with historic maps to refine the location and shapes of defining features. The overlay of the historic maps over current aerial and recent maps of the battlefield showed possible changes in the landscape. Historic maps of the area from 1779, 1788, 1792, 1802, 1804, 1834, 1857, 1869, 1902, and 1948 (Figures 37-46) and a series of 1939 aerial photographs (Figures 47-53) were georeferenced into the GIS as individual layers. The review of the historic maps was important for determining changes in the landscape, such as courses of waterways and land use. The historic aerial photos from 1939 helped to confirm more recent alterations of the landscape and the presence or absence of wooded areas as related to agricultural fields. The determination of slope using DEMs helped to determine possible paths used by Continental forces in their approach to the Village of New Chemung. The defiles noted by Continental forces were identified using slope analysis.

The general mapping of the defining features was further refined using the military terrain analysis extension for ArcGIS 9. The extension allows for the interpolation of viewsheds based on DEMs based on various conditions, such as observer height, target height, azimuth, distance of observation, and elevation angle of observation. These attributes for the interpolation of viewsheds allows for experimentation in the positioning of observers and changes in the landscape, such as accounting for trees or obstacles. The military analysis allows for the creation of range rings marking the range of fire for weaponry. This analysis was integrated into the viewshed analysis to create a type of predictive model for the location of fire for the battle. By integrating the range of fire with a viewshed analysis, factors related to terrain and attributes of the shooter can be changed and accounted for in the analysis. Within the attributes of the viewshed, observer height can be changed for those troops standing or hidden at a lower height behind an obstacle, such as the breastworks. The distance range was established as the range of fire by limiting the distance of observation with the range of fire for a weapon.

For this study, muskets had an assumed range of 100 yards, and rifles had an assumed range of 300 yards. Muskets are described as being accurate within 50 to 80 yards of a target and rifles within 200 to 500 yards (Babits 1998: 12-15). The use of the 300 yards for the rifle was used as an average and the 100 yards for the musket was used to determine the range in which the majority of fired musket balls would be clustered across the battlefield. Due to increased accuracy and firing range of rifles, they were more likely to be aimed and fired for accuracy, whereas muskets were fired in an attempt to place fire in the battle with little regard to aim or accuracy (Babits 1998: 12-15). Those musket balls at the maximum extent of the range of fire would most likely not be accurate on their target. Positions of the observer were akin to the positions of shooters. These positions were dispersed across the presumed positions of troops across the battlefield based on historic descriptions and maps of the battlefield and the correlation of defining features in the GIS. The placement of musket fire was used for both British and Continental positions. Continental forces were documented as having rifles (Cook 1887: 72). The British and their native allies were not documented as having rifles and rifles were not standard issue for the British military during the Revolutionary War. However, the Native American warriors may have acquired rifles through trading, raiding, or other means. Rifle positions were placed along the Chemung Ambuscade to allow for possibility that Native Warriors had rifles. The horizontal angle or azimuth of fire was shortened to fit the most likely trajectory of the shooter. Multiple observer points could also be input in a single viewshed to allow for an overall range of fire for an area of the battlefield or for a line of troops (Figure 19).