FISHERY BULLETIN: VOL. 70, NO. 3 



DATA AND PROCEDURES 



Monthly mean sea levels were obtained from 

 the National Ocean Survey' which has for many 

 years compiled these data for its tide stations. 

 The coincident period of record for Honolulu 

 and San Francisco is 65 years, 1905-69. The 

 coincident period for Hilo and Avila is 20 years, 

 1947-59 and 1961-67. It is broken and shortened 

 because of incomplete records at Avila. How- 

 ever, Avila was selected among other California 

 stations to pair with Hilo, so that the great circle 

 between them would be close to that joining 

 Honolulu and San Francisco and would cross es- 

 sentially the same currents. 



All of the tide gauges have moderately good 

 exposure. Hilo and Avila are small coastal har- 

 bors protected by artificial breakwaters. The 

 Honolulu gauge is in the outer of two harbor 

 areas just inside of the entrance channel. The 

 San Francisco gauge is at Fort Point just below 

 the Golden Gate Bridge on the south side of the 

 entrance to San Francisco Bay. Depths of 100 

 fm (183 m) are less than 10 km offshore from 

 Hilo and Honolulu tide gauges, less than 20 km 

 off Avila, and about 55 km off San Francisco. 



ELIMINATION OF TRENDS 



The sea level observations refer to the level 

 of the ocean surface relative to the adjacent land. 

 The annual mean sea levels at each of the four 

 stations increase irregularly with time as shown 

 in Figure 2. Such long-term trends as are due 

 to a combination of change in the total mass of 

 ocean water by melting (or accretion) of glaciers 

 and of local subsidence (or emergence) of the 

 land on which the gauge is located (Hicks and 

 Shofnos, 1965), need to be eliminated from the 

 data. 



The linear trends determined by least squares 

 regression, of 0.17 cm/year at Honolulu and 

 0.20 cm/year at San Francisco (Figure 2) are 

 essentially the same as reported by Hicks and 

 Shofnos (1965) for sea levels through 1962 at 



' Formerly the Coast and Geodetic Survey, Environ- 

 mental Science Services Administration. 



the same stations. The trend of 0.24 cm/year 

 at Avila is only slightly larger than the first two, 

 but the trend of 0.57 cm/year at Hilo is well over 

 twice as large. In a computer analysis of the 

 tide I'ecords at five stations in the Hawaiian 

 Islands, Moore (1971) found a pattern of near 

 zero trend at the older islands to the northwest 

 increasing consistently toward the southeastern 

 younger islands. Moore attributes the greater 

 trend at Hilo to subsidence caused by loading of 

 the crust by active volcanism. An interesting 

 study by Apple and MacDonald (1966) of arche- 

 ological features — native bait cups, net tanning 

 tubs, and playing boards — carved into a newly 

 submerged reef at Honaunau, Hawaii, further 

 indicates a century of subsidence like that indi- 

 cated by the recent tide gauge records. 



By an indirect method using the decay con- 

 stant of the autocorrelation coefficient Roden 

 (1966) examined the consistency of trends for 

 moving 30-year periods in the longer records 

 at Honolulu and San Francisco. He found that 

 the 30-year trends at Honolulu varied irregularly 

 from 0.12 cm/year to 0.25 cm/year. At San 

 Francisco the trends were low (about 0.05 cm/ 

 year) for the 1904-35 period, rising gradually 

 and leveling oflf at about 0.25 cm/year for 30- 

 year periods starting after 1915. Since not 

 enough is known about the possible meteorologi- 

 cal and oceanographic contributions to these 

 trends, only the long-term linear trends shown 

 in Figure 2 have been eliminated. 



ADJUSTMENT FOR 

 ATMOSPHERIC PRESSURE 



There is general agreement that for periods 

 of 1 month or more the ocean maintains an iso- 

 static equilibrium with changes in atmospheric 

 pressure. Assuming the average pressure over 

 the whole ocean to be constant, if the pressure 

 difference between two locations changes, the ele- 

 vation of the sea surface changes in a compensa- 

 tory manner so that there is no change in distri- 

 bution of pressure on the sea floor. There is no 

 balancing current associated with this portion 

 of the slope of the sea surface which is balanced 

 by the atmospheric pressure gradient. There- 



622 



