Although sea-level series are usually not strictly linear, a least-squares 

 straight line of regression has many advantages (figure 2). For practical applica- 

 tion, it provides a single-value rate at each station for the apparent secular 

 trend. A straight line also facilitates a convenient measure of yearly variability 

 in the "standard error of estimate" computation. Higher order polynomials may pro- 

 vide a better fit. It is stressed that there is no evidence or reason why sea level 

 should obey a straight line or any other polynomial . Thus, prediction by extrapo- 

 lation of these lines is questionable. 



Multiple regression of sea level on the meteorological and oceanographic 

 parameters, coupled with a spectral analysis for the periodicies, probably offers 

 the best avenue for understanding the processes involved and eventual prediction. 

 For this prediction method to be successful, however, the parameters themselves must 

 be predictable a formidable task at even one station. 



Damping yearly variability using a predesigned filter with specific response 

 characteristics (figure 3) reveals both tendencies and the apparent secular trend 

 very well (figure 4). It would seem that damping, by largely eliminating the mete- 

 orological and short-term oceanographic effects, would then enable separation of the 

 eustatic (glacial-eustatic and thermal volumetric) component from the vertical land 

 movement (including, for this discussion, tectonic, isostatic, elastic, compaction, 

 subsurface withdrawal, shelf loading, vertical component of plate movement, et 

 cetera) component. Unfortunately, sea-level series obtained from tide station 

 measurements are relative . That is, the series shows changes in sea level relative 

 to the adjacent land at that location. A series showing a rise in apparent secular 

 sea level (figure 4), for example, could be interpreted as: (1) that the ocean is 

 rising relative to the center of mass of the Earth while the land is rising at a 

 lesser rate relative to the center of mass, (2) that the ocean is rising while the 

 land is stationary, (3) that the ocean is rising while the land is sinking, (4) that 

 the ocean is stationary while the land is sinking, or (5) that the ocean is falling 

 while the land is sinking at a greater rate. 



II. APPLICATIONS --AN ELEMENTARY INTRODUCTION 



Although monthly means have been graphed and tabulated, the analysis portion of 

 this report has been limited to yearly mean sea level with emphasis on variations 



