might be able to use the facility for its own 

 work, although the form later taken by the 

 STOR program made it inappropriate to use 

 the station. 



Mr. Richardson was stationed in the Islands 

 from January 20 to June 20, 1958. During this 

 period he installed tide gauges at Wreck Bay, 

 San Cristobal Island, and Darwin Bay, Genovesa 

 (Tower) Island. The Wreck Bay station included 

 a laboratory for salinity titrations. It was 

 impossible to maintain the remote Darwin 

 Bay station, and it was finally dismantled. The 

 Wreck Bay station functioned well, with the 

 help of Ecuadorian Navy personnel trained by 

 Mr. Richardson, from its establishment in Jan- 

 uary 1958 through April 1959. It was renovated 

 in October 1959 and is now being operated 

 by the Inter-American Geodetic Survey for 

 a further 2 years. Surface temperature, density, 

 and sea level measurements become available 

 to Scripps Institution from time to time. 



STATISTICAL ANALYSIS OF OCEAN- 

 ATMOSPHERE RELATIONSfflPS 



As mentioned above, some of this work 

 was concerned with areas adjacent to rather 

 than in the area of the tropical tuna fishery, 

 because of the ready availability of suitable 

 records. It is thought that the experience 

 gained in analyzing these records will be 

 valuable when similar series of data are 

 readily available for places in the eastern 

 tropical Pacific; some such series already 

 exist, but not in the most suitable form for 

 analysis. 



Statistical Prediction of Surface 

 Temperature in the Northeastern Pacific 



(G. I. Hoden and G. W. Groves) 



The 1921-38 series of monthly means of 

 surface temperature, wind velocity, and in- 

 coming radiation, available for 5-degree rec- 

 tangles off the coast of Washington and be- 

 tween California and Hawaii, were used. 

 Autocorrelations, crosscorrelations, energy 

 spectra, and coherence and phase relationships 

 were used to show how surface temperature 

 could best be predicted from wind velocity 

 and radiation. 



It was concluded: that useful forecasts could 

 be made for only 1 or 2 months ahead; that 

 they would be slightly innproved by taking 

 wind as well as temperature data into account, 

 although radiation data were of very little 

 use; that little would be gained by using data 

 of more than one past month in addition to the 

 present month, for making such predictions; 

 and that optimum prediction by such statistical 

 methods alone could account for only about 40 

 percent of the variance in the next month's 

 average temperature. 



A paper on this investigation, entitled "On 

 the statistical prediction of ocean tempera- 

 tures," was published in Journal of Geo- 

 physical Research, vol. 65, no. 1, p. 249-263, 



1960. 



The work was supported in part {about 30 

 percent) by the California Marine Research 

 Committee. 



Nonseasonal Variations in Sea Level Along 

 the West Coast of North America fC. /. Roden) 



Monthly sea level anomalies at 20 tide gauge 

 stations, including 5 along the west coast of 

 Mexico, were analyzed for per sistence and re- 

 lation to monthly anomalies in atmospheric 

 pressure and sea surface temperature. The 

 purpose of this investigation was to find out 

 how well sea level responds to the inverted 

 barometer effect, i.e., whether an increase 

 in pressure is accompanied by a corresponding 

 drop in sea level, and also whether there is 

 any relation between the integrated thermo- 

 dynamic processes represented by sea level 

 and the sea surface temperature. The station 

 records varied in length from 6 to 37 years. 



The results of this investigation can be 

 summarized as follows: 



1. There is an immediate, inverse, and 

 strong response of sea level anomalies to 

 those in atmospheric pressure; the response 

 is much more pronounced in high than in low 

 latitudes. 



2. There is a direct and moderate response 

 between anonnalies of sea level and sea sur- 

 face temperature; the response is somewhat 

 better in low than in high latitudes. 



3. There is some agreement between anom- 

 alies in sea level and wind. 



4. Sea level anomalies are much more 

 persistent in low than in high latitudes. 



5. The prediction of sea level is less 

 successful than that of sea surface tempera- 

 ture. 



A paper on this investigation, entitled "On 

 the nonseasonal variations in sea level along 

 the west coast of North America," was pub- 

 lished in Journal of Geophysical Research, 

 vol. 65, no. 9, p. 2809-2826, 1960. 



The work was supported in part (about 65 

 percent) by the California Marine Research 

 Committee. 



13 



