JAN FEB MAR APR MAY 



JUN JUL 



AUG 



SEP OCT 



NOV 



DEC 



Figure 2.— Mean dafly sea-suriace temperature (°C) at Ocean Weather Station V during 1954. Values indicated by heavy 

 dots and lines were derived by adding 3.3°C to original daily values. 



vessel data from 2° quadrangles surrounding the two 

 locations of OWS-V. The attempt failed, however, because 

 inadequate sampling frequency by month produced large 

 variability in the comparison of properties from the two 

 areas. However, in the computation of anomalies for the 

 cooling and heating portions of the annual cycle which will 

 be discussed in a later section, corrections were made for 

 the position change. The corrections used were the 

 differences of meteorological properties based on the 

 20-yr means (1949-68) for each month between the two 

 locations. 



Data Summarization 



AU acceptable three-hourly observations of the surface 

 meteorological properties were used to calculate a daily 

 mean value, excepting the total cloud amount. 



The daily mean of the total cloud amount was 

 determined from those observations taken during daylight 

 or twilight hours. Also, the predominant low cloud type 

 was determined for each day to be either 1) cumulus (code 

 numbers 1-4 in the U.S. Weather Bureau ship code) or 2) 

 stratus (code numbers 5-9). The low cloud type which was 

 observed most frequently during the day, or in case of an 

 equal number of both types, that observed closest to local 

 noon, designated the type for the day to be used in the 

 cloud correction. 



Monthly mean values listed in Appendix I are the 

 arithmetic means of the daily values. These values were 

 used to calculate the processes listed in Appendix II except 

 that the monthly wind stress components are mean values 

 of the daily stress components. Mean daily meteorological 



properties were used to obtain the results listed in 

 Appendix III. 



All properties and processes which were computed 

 using corrections or substitutions from merchant vessel 

 surface marine observations are annotated by an asterisk 

 (*) in Appendices I, II, and III. 



ACCURACY OF 

 HEAT EXCHANGE COMPUTATIONS 



Accuracy of the air-sea interaction values derived from 

 marine surface meteorological properties depends both on 

 the correctness of the empirical formulae and the quality of 

 the data used. Surface meteorological data from ocean 

 weather ships are generally of the best quality obtainable 

 at sea. This is true also for the three-hourly observations at 

 OWS-V except for the data inadequacies previously 

 discussed. We also mentioned the uncertainties connected 

 with the empirical formulae. 



Estimates of the radiation from sun and sky, Q(S), have 

 been uncertain because marine cloud observations are of a 

 subjective nature. Better measurements of cloudiness such 

 as the amount of opaque clouds or the percent of possible 

 sunshine are not reported, and information about the 

 thickness of the cloud layers is generally not observable 

 from ships. There have also been a variety of empirical 

 expressions to correct the clear sky radiation ranging from 

 linear to cubic functions of the cloudiness. When some of 

 these expressions, having been derived in mid-latitudes 

 and over land, are used over the tropical Pacific, erroneous 

 radiation estimates may result as reported by Quinn and 

 Burt (1968). 



