EVALUATING THE ACCURACY OF A GRAVIMETRIC SURVEY 161 



in a decreases rather significantly (from 25 to 11%), the value D^^ increases 

 sharply (from 4 to 31%). Asa result, the value of the total mean square 

 error for gravity increment between neighbouring isoanomalies D^ varies 

 very Httle. It appears that with reduction in the network density by a half 

 the value of E^ increases only by 3% of the actual, and the value D^ becomes 

 even less than the actual. In other words, this reduction in the network 

 had practically no effect on the accuracy of the survey (of the determined 

 value Ejj^ and would somewhat increase the accuracy of the map of the 

 gravity isoanomalies. R. F. Volodarskii, the author of the report considered, 

 also came to the conclusion of the desirability for reducing the network 

 density for this type of survey. He showed this conclusively by reducing 

 the density network of a survey which had already been carried out and 

 by reconstructing the map for the gravity isoanomalies with the new smaller 

 density network. 



2. The gravimetric survey of the Atlymsk party No. 7 (55-32) 55-56 in 

 the region of Western Siberia in 1955-1956 (A. A. Serzhant). 



The aim of this work was to carry out geotectonic mapping of the area 

 and to find local gravitational anomalies within its limits. The author observes 

 that this was the first gravimetric survey to be carried out in this region, 

 there was, therefore, no information on possible dimensions and intensity 

 of local anomahes. 



The area of the survey was 7100 km^, and there were 1849 coordinate 

 points. Consecutive observation points were placed on profiles, the point - 

 intervals being 1 km. and the profile -intervals 4 km. The mean square 

 error in the values of gravity at the consecutive points was ±0.50 mgal. 

 The map with a cross -section of isoanomalies for every 2 mgal was drawn 

 to a scale of 1:200,000. On the map there were several large anomalies of 

 the force of gravity with value g-Q = 5 mgal and Iq = 10 km. Furthermore, 

 there are doubtful indications of the anomalies of smaller dimensions and 

 ampUtudes for which it can be assumed that g^^ = 2-2.5 mgal and /q ~ ^ ^^^ 

 (values of gQ and Iq in both cases are determined from a very small number 

 of anomalies). 



Since the point -intervals differ considerably from the profile -intervals, 

 it is desirable to calculate the accuracy of the survey individually for different 

 values of a. In the first line of Table 5 there are the actual values for the relative 

 errors for large anomalies in a direction perpendicular to the profiles, i.e. 

 for a equal to 4 km. In the second and third lines there are the errors for 

 a equal to 2 and 6 km. 



The data of Table 5 show that with decrease in the point -intervals from 4 

 to 2 km, the error in the observed field E^ decreases by about 1%. This is 



Applied geophysics 11 



