sea-level values will differ from the line by less than two times the variability, 

 and practically all the values will differ by less than three times the variability. 



Table 4 gives the same statistical values for the common series 1940 through 

 1980. The date 1940 was chosen in order to give as long a series as possible that 

 would be coimion to all stations and, at the same time, include a sufficient number 

 of stations for a representative coverage of the coast. The formulas used to com- 

 pute the values in tables 3, 4, and 6 are given in appendix C. Table 3 should be 

 used for the best values at each station; table 4, for comparing stations. 



V. AREA MEANS 



The coastline of the United States was divided into five areas. Soutneast 

 Alaska and Hawaii were not included in this area study. This division was based on 

 a subjective balance between series coherence (Hicks and Shofnos, 1965b), equal 

 coastal lengths, and Gutenberg's (1941) concept of geologic regions. The areas, 

 together with their inclusive stations, are listed in table 5. Although Sandy Hook, 

 New Jersey, and Galveston, Texas, have continuous series dating long before 1940, 

 they were not used because they are undergoing anomalous localized subsidence as 

 determined independently by terrestrial leveling. This is not to say that the other 

 stations are not undergoing subsidence; only that Galveston differs greatly from the 

 gulf area pattern due to its subsidence. 



For each year, the yearly mean sea-level values at all stations in an area were 

 averaged for that year. An average series was thus obtained for ewery one of the 

 five areas. The resulting graphs, with straight lines connecting the averaged 

 yearly points, are shown in figures 5 through 9. The curve (dotted) is the result 

 of a seven-point triangular weighting array (1,2,3,4,3,2,1) applied to the averaged 

 yearly points on the graph. The response of the array (figure 3) attenuates, by 

 more than 90 percent, the amplitudes of all the meteorological and oceanographic 

 oscillations with periods less than 5 1/3 years (Hicks and Shofnos, 1965b). 



For each year, the average area values for all five areas were averaged for 

 that year. The resulting composite graph, with damping array applied, is shown in 

 figure 10. The area and national trends, standard errors, and variability are given 

 in table 6. 



14 



