sticking out of the bottom (proud objects). Most likely, these are errors induced 

 by the sensor platform traversing at a slightly tilted angle from horizontal. 



Finally, Figure 31 shows the distribution of apparent sizes of the anomalies. 

 The apparent size of an object is its equivalent radius, which is the radius of a 

 steel ball having the same dipole moment. Bell, DeProspo, and Prouty (1996) 

 found that ordnance caliber is almost equal to the measured dipole radius. 

 Figure 31 also shows that objects range from 5 to 50 cm, with the most common 

 clustering between 10 and 35 cm (4 to 14 in.). These sizes are consistent with the 

 caliber of ordnance recovered in the test raking operation by the Miss Kathy, but 

 the distribution is different. The distribution of the raked ordnance was dominated 

 by smaller pieces (i.e. , 8- to 13-cm range), and the raking operation only 

 recovered 24 objects, a sample size too small to use to evaluate the distribution of 

 size classes (Figure 32). 



A total of 240 anomalies were counted by AETC during their analyses. 

 Assuming that all the anomalies correspond to targets and that the detection swath 

 is 4 m to either side of the track, this amounts to an ordnance density of about 

 15.4 objects per hectare. The raking operation recovered ordnance at only one 

 tenth of this density, about 1.3 objects per hectare. The discrepancy may be due 

 to three factors. First, not all anomalies may be caused by actual ordnance but 

 rather by other sorts of metallic debris. This, however, is not likely to be 

 significant due to the average precision of fit which exceeded 0.98 of the 

 measured magnetic anomalies to simple dipole models. Most marine debris 

 would not be representative of simple dipole magnetic sources. Second, the 

 raking operation may have failed to recover many ordnance items on the sea- 

 floor. Preliminary tests in other locations have shown that many shells fall out of 

 the rakes before they can be retrieved onto the deck of the vessel. Also, a factor 

 due to the raking activity occurred in Borrow area 1 A, which is seaward of the 

 area evaluated by AETC. Most likely the difference is from the shallow depth 

 (10 cm) that the Miss Kathy was able to reach. Analysis of the depth of the 

 ordnance, Figure 29, shows that most of the ordnance is below 10 cm in the sand, 

 but buried shallower than 1.5 m. 



In summary, the AETC sensitivity analyses indicate that out to a range of 3 or 

 4 m from the survey track, a large piece of ordnance (e.g. , greater than a 4-in. 

 caliber shell) can be located within 10 - 20 cm accuracy (x, y, and z) relative to 

 the array using the survey data. Using a statistical sample of 100 magnetic 

 anomalies from the surveys, the distribution of apparent dipole orientations 

 indicates that the magnetic moments are largely induced and that the objects tend 

 to be lying flat, parallel to the bottom, rather than upright. Most objects appear to 

 be on the bottom or at fairly shallow depths. The computed target density was 

 about 16 items per hectare, over ten times greater than was computed from the 

 Miss Kathy raking operation. 



Magnetic Location Conclusions 



The cesium-vapor gradient magnetometer proved to be highly successful in 

 detecting and resolving the presence and location of ordnance-like objects in the 



Chapter 6 Magnetometer 27 



