Figure 15 shows the cumulative probability curves for the shipboard, 

 boom, and rawinsonde temperatures for the 2000-1059 GNfT subgroup for the 

 Rainier. The maximum D value and the calculated D^ value are indicated, 

 based on data obtained between June 19 and July 2. 



5.3 useless Mt. Mitchell 



The size of the Mt. Mitchell is the same as that of the Rainier and the 

 location of the ship observations relative to the boom as described in the 

 preceding section also apply to the Mt. Mitchell, Because of this similarity, 

 one would expect that the temperatures measured aboard the tv;o ships would 

 agree closely. However, the Mt. Mitchell shipboard temperatures show an un- 

 usual diurnal variation (fig. 16). For example, for the data when the wind 

 is blowing from the ship toward the boom (see table 5), which are divided 

 into daytime and nighttime subgroups beginning at sunrise and sunset respec- 

 tively, the average daytime shipboard temperature may be equal to the night- 

 time average, while the daytime boom temperatures average 0.7°C warmer, and 

 the adjusted rawinsonde temperatures average 0.5°C warmer. Also, of the 53 

 shipboard temperatures obtained when the wind was from the ship toward the 

 boom, 20 fall between 27.20 and 27.39 (fig. 17), while no more than 10 boom 

 temperatures fall into any single 0.2°C increment. If this peculiarity in 

 the shipboard temperature cannot be explained logically, the possibility of 

 using the boom values exclusively should be considered seriously. The fact 

 that the ship's temperature modifies the air temperature passing over it and 

 maintains that control is one explanation. Even in this case, admittedly, 

 the boom temperatures would better represent the environment. 



Figure 18 shows the lag coefficients used to derive the independent 

 s amp 1 e . 



Table 6, showing the results of the parametric and nonparemetric tests, 

 indicates that the K-S test rejects the null hypothesis for the shipboard- 

 boom comparison only for the 1100-2159 GMT subgroup. In all other comparisons, 

 the hypothesis is not rejected, and therefore it would probably be better to 

 use the boom data with the rawinsondes for all time periods. 



5.4 useless Discoverer 



On the Discoverer^ the shipboard observations were made aft of the 

 exhaust stacks near the rawinsonde inflation shelter about 220 ft from the 

 bow of the ship and about the same distance above the ocean surface as the 

 boom - conditions similar to those on the Oaeanographer . Figure 19 shows 

 the average differences between the shipboard and boom temperatures at rawin- 

 sonde release times. As is evident from this figure, the shipboard tempera- 

 ture is higher than the boom temperature for each hour in both groups and 

 considerably higher during the night, except for the 0600 GMT release time 

 when the wind was blowing across the ship toward the boom. These higher 

 shipboard temperatures were probably caused by the exhaust gases from the 

 ship that contaminated the shipboard observations; possibly these emissions 



