Table 1.— Monthly mean sea-suriace temperature (T) and standard deviation (o) of the means at OWS-V 

 (A) and in a 2° quadrangle centered at lat 31°N, long. 164°E (B) for the year 1954. 



Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec. 



by Kraus (1972) and falls within the range of determina- 

 tions made during the last 10 yr. 



On a provisional basis we have also calculated the 

 turbulent transfer processes that reflect changes in 

 stability by using Deardorff s correction for the transfer 

 coefficients at neutral stability (Appendix III). We feel that 

 the use of corrections for the ocean-wave spectrum in the 

 routine calculation of turbulent transfer processes from 

 marine surface data would be premature. 



Q(C), transfer of sensible heat. — Estimates of the 

 turbulent flux of sensible heat between the sea and the 

 atmosphere suffer from the same deficiencies as the water 

 vapor flux. The sensible heat flux is proportional to the 

 sea-air temperature difference and the wind speed. This 

 process is of relatively small magnitude in comparison with 

 the other air-sea interaction processes. 



Wind Stress 



Again, the turbulent flux of momentum across the sea 

 surface is subject to uncertainties discussed above. In 

 addition, because the magnitude of the stress is proportion- 

 al to the square of the wind speed the climatological mean 

 approach used in the calculation of the water vapor flux and 

 the sensible heat flux should not be used in the calculation 

 of the momentum flux. In this paper, the resultant stress 

 components are the mean values of the stress components 

 computed from individual wind observations using 



T, =p CdW, W (8) 



T, =p Cd w,W (9) 



w, and Wy are the components of the wind in the zonal and 

 meridional directions and W is the magnitude of the wind 

 speed. For the density of air we used p = 0.00123 g cm"^ 



PROCESSING OF DATA 



Before the summarization of meteorological properties 

 for the computation of air-sea interaction processes, 

 several deficiencies in the three-hourly observations at 

 OWS-V had to be corrected. The most troublesome 

 deficiencies are gaps in the data record and errors in the 

 sea-surface temperature. A shift in location of OWS-V in 

 1955 introduces another deficiency in that comparisons of 

 air-sea interaction processes after 1955 with those before 

 1955 are difficult. Procedures to overcome these deficien- 

 cies are described below. 



Data Gaps 



Large data gaps in the time series were not common. 

 However, there was a 13-day gap from 2 to 14 May 1970 



when no observations were taken. All properties except 

 sea-surface temperature were measured from 16 March 

 1952 to 31 March 1953. For this period, monthly mean 

 sea-surface temperatures from merchant vessel observa- 

 tions (National (Climatic Center, Tape Data Family 11) in a 

 2° quadrangle centered at lat. 31°N, long. 164°E were 

 substituted and used in the heat exchange computations. 

 Agreement between OWS-V and merchant vessel monthly 

 mean surface temperatures is good (Table 1). 



The sea-surface temperature data were also missing 

 from 1 May 1963 to 21 June 1963, and daily bucket 

 temperatures collected aboard the Ocean Station Vessel for 

 the National Marine Fisheries Service (Yong 1971) were 

 substituted. 



Wet bulb temperatures for the entire month of 

 December 1955 were missing. This data gap was filled by 

 computing the saturation vapor pressure of the air from 

 merchant vessel dew-point temperatures interpolated to 

 the position of OWS-V. The saturation vapor pressure was 

 computed by the ideal gas law formula for moist air 

 (Longley 1970). 



Erroneous Data 



Erroneous sea-surface temperature values were detect- 

 ed during the initial pass through the data by computing a 

 16-point running mean. Those values which differed by 

 more than 5°C from the running mean were rejected. A 

 second quality control check was performed on the daily 

 mean sea-surface temperatures for each year by using 

 harmonic analysis as a curve-fitting technique (Seckel and 

 Yong 1970). Fourier analysis was carried out to the 13th 

 harmonic with a fundamental period of 365 days. Daily 

 values that deviated more than 1°C from the expected 

 value were rejected. Three separate 21-day periods in 

 1954—13 June to 8 July, 17 October to 5 November, and 28 

 November to 17 December— were found to contain 

 sea-surface temperatures which were consistently about 

 3°C lower than surrounding data points. It was found that 

 the same vessel was on station during these periods and we 

 assumed that an erroneously calibrated thermometer was 

 used on this ship with an error of 3.3°C. The erroneous 

 temperatures were corrected by adding 3.3°C (Fig. 2). 



Position Change 



A change in location of OWS-V from lat. 31°N, long. 

 164°E to lat. 34°N, long. 164°E occurred in March 1955. 

 Although the locations are separated by only three degrees 

 of latitude, spatial differences of meteorological properties 

 are of the same mag^nitude as the interyear differences that 

 are of interest to us. An attempt was made to correct the 

 pre-1955 data to the new latitude by comparing merchant 



