7635 HYDROGRAPHIC MANUAL PaGL 706 



7635. Analysis of Data 



For analyzing the various data, it is not possible to prepare a table of relative 

 weights to apply to the different situations that might arise. Experience is the best 

 guide and much depends on the conditions prevailing at the time of the survey. All 

 three-point sextant fLxes should be considered precise locations until proved otherwise. 

 The locus of one sextant angle should likewise be considered a fixed line of position 

 unless proved erroneous. Bearings to near objects will generally be accurate. The 

 acoustic returns from the nearest station will generally be found more reliable than 

 those from more distant stations, except that consistently good returns may be received 

 from certain stations while those from other stations may be just as consistently un- 

 reliable. A position determined by a good intersection of three distance arcs may 

 generally be accepted as a reliable position, but it must be remembered that an in- 

 accurate velocity or other factors may cause the arcs to intersect at a position which 

 is not the correct one. 



7636. Sources of Error in R.A.R. Positions 



Before the final positions are accepted, all conflicting data shoidd be resolved, if 

 possible. Such conflicts may be due to determinable errors, such as those arising in 

 the plotting or in the reduction of the records, or they may be due to certain inde- 

 terminable errors inherent in the method itself. The various sources of error may be 

 grouped as follows: 



a. Scaling the chronograph tape.- — A large error is sometimes made in identifying the signal olfset 

 on the tape. If this occurs at one position only of a series of positions, and for one distance arc, the 

 error is generally recognizable, but where made at all of the positions of the series it is harder to 

 detect. A large error in all of the time distances at one position is usually easily detected since the 

 resultant arcs plot with a large triangle of error and not in accord with the dead reckoning. An 

 error of 1 second in reading the tape has been found in this manner. (See also 6853.) 



Errors are also sometimes made in scaling the initials on the chronograph tapes. Such an error 

 is more difficult to detect for it may result in a comparatively small triangle of error that may appear 

 to be caused by the use of an erroneous velocity of sound. Errors are occasionally made in marking 

 the bomb return on the tape at the time it is received, especially if this occurs during excessive static 

 conditions or other radio interference, when it is difficult to distinguish the bomb returns from the 

 strays or false indications. In such cases a re-examination of the tape will sometimes disclose the 

 correct return, but changes should not be made in the distances in the Bomb Record unless the correct 

 bomb return can be identified on the tape with reasonal^le certainty. The above are a few of the 

 reasons why the chronograph tapes must be available to the smooth-sheet plotter (see 6854). 



b. Plotting and reduction of records. — Errors in drawing the distance arcs on the smooth sheet are 

 not common. An error is occasionally made by plotting the time increment as a decrement from the 

 next larger distance circle or vice versa, where the position falls approximately midway between them. 

 A distance arc may be plotted occasionally from the wrong station. Mathematical errors are some- 

 times made in converting actual elapsed times to those used in plotting. For this reason, where 

 certain distance arcs appear doubtful, the mathematical conversions should be rechecked in the 

 Bomb Record. (See 8311.) 



c. The use of an erroneous velocity of sound. — These cases are not uncommon and they are particu- 

 larly difficult to detect since an incorrect velocity will proportionally enlarge or reduce the distances 

 from all stations, and the distance arcs will still plot with a very small triangle of error. In analyzing 

 such cases it should be remembered that the correct position will be inside the triangle of error only 

 where the stations are arranged in a triangle around it. A change in velocity will probably alter the 

 distances to all stations proportionally. Therefore, in case of an error in velocity the correct position 

 will be distant from each plotted arc an amount proportional to the distance of each arc from its 

 respective station. 



