SECT. 1] SEISMIC REFRACTION AND REFLECTION MEASUREMENTS 17 



absence of sufficient thickness or velocity gradient, Rn simply merges with 

 Rs at the range which j)roduces a grazing ray at the 2-3 interface. 



6. Some Complications and Deviations from Simple Theory 



Although geologic structures which are encountered in exploration can often 

 be approximated by models which are relatively simple, there are usually 

 certain complications. For example, the interfaces are seldom found to be 

 plane boundaries. They are more apt to be undulating or broken by folding or 

 faulting. This has the effect of making the refraction lines and reflection 

 curves undulating or broken. Other complications arise from the fact that 

 some layers have lateral velocity variations, or have a velocity dependence on 

 direction of propagation. This latter effect is particularly common in laminated 

 rocks such as shales or schists. Many other problems associated with structural 

 or compositional inhomogeneities can easily be imagined. 



Fig. 1 1 . Ray path for a model in whicli the velocity in Layer 3 is less than that in Layers 

 2 and 4. 



Complications also arise from difficulties in identifying arrivals owing to 

 low signal strength or to coincidence of arrival times of two or more phases. 

 Unless there is an appreciable difference in the characteristic frequency of two 

 arrivals which reach a detector at almost the same time, it is usually difficult 

 to pick the beginning of the second one. Hence, the most reliable arrivals are 

 first ones, or those which come long after the motion from earlier ones has 

 appreciably died out (G3, for example, in Fig. 4). In certain structures there 

 may be layers for which the refracted waves are never the first arrivals. The 

 reliability of detecting these "masked" layers by refracted arrivals is particu- 

 larly dependent upon the closeness of shot spacing, since a single record cannot 

 be counted on for a precise pick. 



Thin layers, even if they are not masked, may be missed, or their velocity 

 and thickness inaccurately determined by profiles in which the shot spacing is 

 too wide. Field observations and model experiments have shown that, in a 

 layer which is thin compared with a wavelength of the seismic energy, reliable 

 refracted arrivals are obtained only over relatively short distances. Hence a 

 thin layer will produce reliable arrivals over only a short portion of a profile, 

 and close shot spacing is essential for detecting it. 



A low-velocity layer between two media of higher velocity is not detectable 

 by normal refraction techniques. Fortunately, occurrences of this type of 

 layering are believed to be uncommon. Fig. 11 is a ray diagram for a hypo- 

 thetical example. The time-distance graph for this model would give no 



2 — s. Ill 



