Page 142 - Green - Maritime Archaeology: A Technical Handbook. 2nd ed

Chapter 4: Conventional Survey 121
restrained further surveying. Large standard deviations were then assigned to those few remaining unreliable measurements. As the measurements were refined and estimates of the coordinates became more accurate, the resulting standard deviations for the coordinates and the residual values became uniformly small. In one particular case, a small but consistently larger than average standard deviation was noted for measurements to one particular point. On subsequent inspection of the point, it was discovered that the tag was attached to a plank that was slightly loose and as a result was causing unreliable measurements (for further details, see Atkinson et al., 1989).
C. THE DIRECT SURVEY METHOD SYSTEM
One of the first maritime archaeological applications of computer-based surveying techniques was developed in Norway during the Lossen wreck by the Norsk Sjøfartsmuseum in 1967 (Andersen, 1969). In this system three distance measurements were made from three reference points and a simple calculation of resection was made. This was done graphically and also using a Fortran program. Christensen (1969) noted that the system had the potential to save time while at the same time providing high accuracy. Because the site was in a rocky gully, it was possible to utilize elevated datum points making it possible to obtain three-dimensional coordinates.
A similar system, which he refers to as the direct survey method (DSM), was developed independently by Nick Rule during the Mary Rose excava- tion (Rule, 1989). The basis of the formulation is to solve the three equa- tions of the three spheres representing three distance measurements from the respective reference points. The equations have two solutions with one being the mirror image of the other. Additional distance measurements were added to the computation and the results averaged to give an esti- mate of error. This system was used in the survey work on the Mary Rose project. It is estimated that 450 diver days were taken up surveying 100 datum points including the underwater work of sighting, leveling, and sur- veying these data, together with the land-based work of recording, com- puting, and plotting the results. It is calculated that it took about 4.6 minutes to make one measurement. A total of 740 measurements were recorded from the site of which about 560 were useful measurements. Rule estimates that the DSM system takes about one quarter of the time taken to plot data using a ruler and compass (Rule, 1989). The author is extremely grateful to Nick Rule for the information provided as part of an extensive report on the DSM surveying technique to be published as part of the Mary Rose report (Rule, 1982). It appears, however, that a large proportion of the