intersection, the GP0LYT2 program adjusted each of the shoreline points 

 enclosed within a projection cell based on true values of intersections versus 

 digitized and computed values for those same intersections. Values for each 

 shoreline point are thus correct in their position relative to known (true) 

 projection intersections and to known triangulation data (Figure 19). 



56. Following the digitizing process, each sheet was reviewed visually 

 with use of a raw data plot in which shoreline positions were shown at the 

 same scale as the original map. Plotted shorelines were superimposed on orig- 

 inal maps and checked for completeness and accuracy of tracking during 

 digitization. This review helped to minimize a potential source of human 

 error that could occur during the digitizing process. 



57. Other sources of potential error also were considered. The most 

 difficult of these to determine precisely was location accuracy of the MHWL on 

 source surveys and maps, on either (a) early surveys prior to approximately 

 1930 and (b) maps based on photogrammetric surveys. In discussing early sur- 

 veys, Shalowitz (1964) has stated: 



The accuracy of the surveyed line here consid- 

 ered is that resulting from the methods used in locat- 

 ing the line at the time of survey. It is difficult 

 to make any absolute estimates as to the accuracy of 

 the early topographic surveys of the Bureau. In gen- 

 eral, the officers who executed these surveys used 

 extreme care in their work. The accuracy was of 

 course limited by the amount of control that was 

 available in the area. 



With the methods used, and assuming the normal 

 control, it was possible to measure distances with an 

 accuracy of 1 meter (Annual Report, US Coast and Geo- 

 detic Survey 192 (1880)) while the position of the 

 planetable could be determined within 2 or 3 meters of 

 its true position. To this must be added the error 

 due to the identification of the actual mean high 

 water line on the ground, which may approximate 3 to 

 4 meters. It may, therefore, be assumed that the 

 accuracy of location of the high-water line on the 

 early surveys is within a maximum error of 10 meters 

 and may possibly be much more accurate than this. This 

 is the accuracy of the actual rodded points along the 

 shore and does not include errors resulting from 

 sketching between points. The latter may, in some 

 cases, amount to as much as 10 meters, particularly 

 where small indentations are not visible to the topog- 

 rapher at the planetable. 



44 



