660 MUNK [chap. 18 



accordance with classical theory. Actually the correction is completely negligible 

 except at the highest frequencies included in the figure. 



Fig. 3 is based on a composite of many measurements in the Camp Pendleton 

 area (see Munk, Snodgrass and Tucker, 1959, charts 2.12, 3.1, 3.2, 3.3, 4.1, 4.2, 

 4.3, 4.4, 5, 10.1, 10.3). The peak at 0.8 c/ks and the troughs at 0.5 and 2.3 c/ks 

 are found on all analyses. This peak will be discussed under "shelf resonance". 

 The broad humps between 10 and 20 c/ks are not nearly so well established. 

 They appear to depend on the depth of water, and accordingly on the distance 

 off-shore (see "surf beat", section 4). The sharp peak at 47 c/ks (period 21^) is 

 typical of a long-period swell from the south. New peaks appear once or twice 

 a week at about 40 c/ks, then increase in frequency by 5 to 10% per day, until 

 they disappear into the high background at about 125 c/ks. The wandering 

 peaks are associated with storm systems in the South Pacific and Indian Oceans 

 (Munk and Snodgrass, 1957). The dispersive shifts of these frequency peaks is 

 in marked contrast to the fixed spectral signatures of the long waves. 



Fig. 3 is to serve only as an idealized description of the background activity. 

 On any given day the spectral densities may be up or down by an order of 

 magnitude (although the relative spectrum is surprisingly invariable). The 

 spectral structure varies from place to place, particularly with regard to shelf 

 resonance. Finally it must be stated that the terms "surf beat" and "shelf 

 resonance" are by no means generally accepted ; these names imply causes that 

 may not be borne out by future work. 



4. Surf Beat 



A typical frequency is a cycle per minute. The existence of relatively promi- 

 nent waves in this frequency range was inferred by Munk (1949) and Tucker 

 (1950) from a visual examination of low-frequency wave records. The waves 

 were attributed to the fluctuating height of groups of sea and swell waves 

 breaking on the shore, hence the name "surf beat". Tucker and Munk both 

 found a clear relation between the amplitudes of the surf beat and of the 

 ordinary waves, approximately 1 to 10. A similar result was obtained by J. E. 

 Dinger for wave records from Barbados. In all these cases the wave recorders 

 were in quite shallow water. During the fall of 1959, Snodgrass, Miller and Munk 

 obtained several hundred successive spectra (not published) from a wave 

 recorder at 100 m depth on the exposed (western) shore of San Clemente 

 Island, 60 miles off-shore from the coast of southern California. In no instance 

 do we find a spectral peak of surf-beat frequency. Rather, the spectrum is flat 

 at about 10~2 cm^/c/ks for frequencies below the swell and above the shelf 

 resonance. This suggests that the surf-beat activity is characteristic of shallow 

 water. A re-examination of all .available spectra did, in fact, indicate a decrease 

 in the spectral activity between 10 and 40 c/ks with increasing depth of water, 

 as indicated in Fig. 3. 



Munk (1949) suggested that the surf beat may consist of long waves generated 

 in the surf zone by variable "radiation pressure" of the incoming breakers, and 



