614 



GROEN AND GROVES 



[chap. 17 



the order of 7-h period or less would not. The gravest modes of many gulfs and 

 semi-enclosed seas have periods of the order of a few hours, and similar con- 

 siderations apply. 



A. The Day-to-Day Variation of Sea- Level 



Observation of surges at the low-frequency end is simplified by the fact that 

 their frequencies and those of the tide do not overlap. Consequently, to elimi- 

 nate the tide one can use an averaging technique (such as described by Groves, 

 1955) as well as subtraction of the predicted tide. The former has the advantage 

 that an identical scheme can be used at all places, the computations are simpler, 



SEA LEVEL (Cj,) 

 -Of San Froncisco 



Fig. 2. Simultaneous plots of sea-level along the California coast. Dates correspond to 

 0000, 120°W meridian time. The dashed portions of the curves indicate uncertain 

 data. "C51" refers to the 51-ordinate averaging scheme used (Groves, 1955). (After 

 Groves, 1957. By courtesy of the American Meteorological Society.) 



and one is not limited to ports for which predictions have been made. The non- 

 linear difficulties discussed in section 1-B are usually not serious. The main 

 limitation of the averaging technique is that waves whose periods are two days 

 or less are considerably attenuated. 



A series of such sea-level records was examined by Groves (1957), from which 

 Fig. 2 was taken. Simultaneous three-month records of sea-level at ports along 

 the California coast are shown. There is considerable coherence between 

 adjacent ports, and some even between the northernmost (San Francisco) and 



