A velocity of propagation is assumed to be 186, 218 statute miles/ 

 second at 29.92 inches Hg, which is probably correct to one part in 18, 600. 

 While the effects of changes in barometric pressure are calculable, they are 

 negligible compared to other errors in Shoran equipment; a change of 1 inch in 

 pressure will change the velocity of propagation about 2 miles/second, which is 

 about 1 part in 100,000. 



Random Errors 



The random errors cannot be eliminated. They are tabulated below 

 (in statute miles): 



(1) Setting and reading mileage verniers: ±0.002 



(2) Nonlinearity of calibrated phase shift network: ± 0.005 



(3) Zero adjustment in ground station: ± 0.002 



(4) Nonalignment of pulses due to insufficient gain 

 at long distances: ± 0.050 



It is not probable that all the errors will be accumulative at the same 

 time. Tlie first and third listed are of variable sign and value and affect dis- 

 tance measurements within the line-of-sight distances from the ground stations. 

 All four affect the results when greater than line-of-sight distances are meas- 

 ured. Errors in excess of 0.050 statute mile have been noted in lines as long 

 as 100 miles. At the longer ranges the pulses are reduced in size and shape, 

 so that the error will be dominated by nonalignment of pulses at long distances 

 and the measured range will be too long. 



The effects of the random errors is to give an uncertainty in position 

 corresponding to the magnitude of the errors and the angle at which the distance 

 arcs intersect. This uncertainty is illustrated in Figure VI-3. 



Long experience indicates the areas of uncertainty to be negligible 

 when the angles of intersection are between 30 and 150°. These limits are 

 often exceeded in surveys which extend along both sides of the base line and 

 its extension and at long range from shore. 



159 



artbitr ai.1littlc.Jnr. 



