2. BACKGROUND 



Before selecting and justifying an approach which would be simple 

 but still give some hope of yielding useful results, it was necessary 

 to screen and evaluate a voluminous amount of literature on the subject. 

 Fortunately, a great part of this review had been done recently by 

 Laevastu (1952). 



Most theoretical approaches have been mainly concerned with 

 explaining the general, more permanent features of the horizontal 

 circulation patterns (see, e.g., Robinson 1953). The Russians have 

 recently applied correlation theory in an attempt to forecast detailed 

 current changes from a known field; however, our knowledge of the 

 initial state (and particularly its derivatives) is often rather sketchy. 

 Actually, the ocean responds quite rapidly to hourly and daily changes 

 in driving forces, and currents are known to be variable in space and 

 time (Knauss 1960). 



These considerations dictated use of a method which vv^ould account 

 for fairly rapid response and would stand up to daily verification. Many 

 attempts at current prediction have been disappointments because of 

 oversimplifications resulting from the assumptions made. It seems 

 logical to separate the total current into its elementary components to 

 see which should be neglected, which can be simplified, etc. This is 

 the attack which has been followed in this investigation. 



