410 



EXPLORATION GEOPHYSICS 



1100 12:00 



S'OO 



The proper density for the area is that density which yields a minimum 

 correlation with the topography. Ordinarily a rather definite selection can 

 be made of the density which gives reduced gravity values showing a 

 straight line on the graph of the profile. The method is quite advantageous 

 in arriving at the best density figure to use in the Bouguer correction. It 

 efifectively weighs the topography of a feature so tested, and the density 

 thus obtained can be used in the Bouguer reduction for an entire area. 



Accuracy of Data * 



The value of a gravity map depends largely on its accuracy. Various 

 types of tests have been made to determine this accuracy. Aside from 



mistakes such as misreading the 

 scale or error of computation, the 

 error arises from three principal 

 causes. The first is erratic fluctua- 

 tions in the instrument reading 

 from such causes as temperature 

 fluctuation, elastic-after-effect, ef- 

 fect of jars, and the like. The 

 second is inaccuracy of the calibra- 

 tion curve. The third is errors in 

 making the latitude or elevation 

 corrections due to inaccuracy of 

 surveying or use of an incorrect 

 elevation correction constant. 

 The magnitude of the first type of error may readily be determined by 

 noting the deviations of a succession of observations at the same station. 

 In making such a test it is evidently necessary to simulate closely actual 

 field conditions. The procedure, accordingly, is to occupy a station and 

 make an observation in the manner customarily employed in the normal 

 field work. The instrument is then transported in its truck a distance of 

 one mile and returned to the same station. Here a second standard ob- 

 servation is made. A series of eleven such observations plotted as a drift 

 curve are shown in Figure 245.$ These data were treated as follows for the 

 computation of the probable error of a single occupation of a station. The 

 best straight line is first passed through the observed values. This may be 

 readily done by a trial and error method or by the usual least square pro- 

 cedure. The deviations of the observed values from the drift line are then 

 computed, and finally the probable error is obtained by the standard pro- 

 cedure of multiplying the square root of the average square deviation by 

 the factor 0.67. In this test the mean deviation is found to be 0.06 milligal 

 and the probable error of a single observation 0.04 milligal. The largest 



*The method of determining accuracy has been supplied largely by L. M. and 

 F, W. Mott-Smith in a personal communication. 



% L. M. and F. W. Mott-Smith, "Advancements in the Use of the Gravimeter in Oil Explora- 

 tion," The Petroleum Engineer, July, 1939, pp. 85-97. 



2:00 3^00 4:C 



TIME 



Fig. 245. — Plot of repeated observations at 

 the base station. (L. M. and F. W. Mott-Smith, 

 The Petroleum Engineer.) 



