must therefore be regarded with caution, except 

 at lat. 2° N., long. 157° W. where daily observa- 

 tions from Christmas Island have been used. 



North of lat. 15° N. sampling is more ade- 

 quate, particularly from Hawaii northward, in 

 the latitudes that contain the shipping lanes 

 from North America to the Far East and to 

 Hawaii. Reliable results in table B are those 

 from lat. 17° N., long. 167° W., an area which 

 contains the daily Johnston Island observations, 

 and from lat. 32° N., long. 142° W., which, in 

 addition to a large number of observations from 

 merchant vessels, contains those from the U.S. 

 Coast Guard Weather Station November. 



Of the interpolated meteorological properties 

 in table B, the cloudiness and the wind speed 

 are most subject to error due to insufficient 

 observations. The cloud cover may materially 

 change during timespans of less than an hour 

 and from day to day. A large number of obser- 

 vations is therefore necessary to provide a 

 measure of the monthly cloudiness. Only in the 

 latitudes north of Hawaii are the observations, 

 on average, more frequent than once per day. 



In the trade wind zone, similar wind speeds 

 and directions have a greater persistence than 

 the amount of cloud cover but may show rela- 

 tively large variations in timespans of a few 

 days. South of lat. 15° N. the interpolated winds 

 and cloudiness of table B are heavily biased by 

 the values obtained from a few monthly obser- 

 vations which, although they may be correctly 

 measured, do not necessarily reflect the month- 

 ly mean value. 



Despite the sparse data south of lat. 15° N., 

 interpolated results of the smoothed charts are 

 presented in table B. As a consequence large 

 evaporation rates of more than 700 cal. cmr^ 

 day"' may appear as, for example, in the east- 

 ern portion of the region at lat. 2° N. during 

 September and October 1963. These evapora- 

 tion rates are due to relatively few observations 

 with wind speeds of more than 10 m. sec."' (see 

 table A), The choice in the preparation of table 

 B was either to leave the area in question blank 

 ortopresent values based on the limited obser- 

 vations. The latter course was chosen since it 

 draws attention to the high winds which may 

 occur in this area and cause high evaporation 

 rates. 



Inadequacies in the quality of marine meteor- 

 ological observations are well known. Visual 

 inspection of the data used in this report again 

 served to draw attention to the primary causes 

 which reduce the quality of data. Because the 



quality can be so easily improved, the causes 

 for the inadequacies and their effect on the heat 

 exchange results are reviewed. Errors in the 

 marine surface data are not so much due to in- 

 correct reading of instruments and recording of 

 observations as to improper placement and cali- 

 bration of the instruments, improper techniques 

 of measurement, and errors in data reduction 

 at sea. 



To illustrate--high values for the wet and dry 

 bulb air temperatures indicate that the instru- 

 ments were placed in a warm location aboard 

 ship or that the psychrometric measurements 

 were not made in air freshly flowing off the sea. 

 It was previously stated that negative values in 

 the minimum sea-air vapor pressure differ- 

 ences of table A are probably due to erroneous 

 observations. Improper techniques in the psy- 

 chrometric measurement are believed to be the 

 primary cause. If the dry and wet bulb temper- 

 atures are not noted when the minimum wet bulb 

 temperature is reached, an overestimate of the 

 vapor pressure of the air results. Consequent- 

 ly, the sea-air vapor pressure difference will 

 be too low and may be negative. 



In an analysis of sea surface temperatures 

 taken aboard merchant ships, Saur (1963) found 

 that deviations from mean values were large 

 and that the "injection temperatures," which 

 are often used in marine weather observations, 

 are measured in enginerooms and tend to be too 

 high. The large temperature deviations, in 

 terms of use in oceanography or meteorology, 

 may be due to improper placement and calibra- 

 tion of engineroom thermometers. Saur sug- 

 gested a correction of -0.7° C. for average sea- 

 water temperatures obtained from the marine 

 surface meteorological observations. 



Errors in air and water temperatures affect 

 primarily the sea-air vapor pressure differ- 

 ence and so the heat of evaporation. Overesti- 

 mating the sea-water temperature by 0.7° C. 

 would mean an overestimate of the saturation 

 vapor pressure of water by 0.8 mb. at 15° C. 

 and 1.6 mb.at 28° C. Errors of similar magni- 

 tude occur in the vapor pressure of the air ow- 

 ing to wrong psychrometric measurements. The 

 sea-air vapor pressure differences of table A 

 are generally below 10 mb. and often below Smb. 

 The possible error in the heat of evaporation 

 may therefore be relatively large. 



Visual inspection of data also indicated that 

 erroneous data reduction is the probable cause 

 for wrong wind speeds and directions. For ex- 

 ample, in areas with a large number of obser- 



