Le
RH is a calculated value based on TT and TT,.
purposes.)
(P is an insignificant factor for our
It follows that for the dry-bulb and wet~bulb temperatures experienced at Eniwe t«
all RH values are correct within 6%, and 9 out of 10 are correct within 42% {assuming normal
error distribution and allowing for 1% error in conversion).
N is probably too high, especially at night, in all instances in which it largely depen
on an observation of 10 Cs.
An exception would be when 10 Cs was also observed at one of th:
other stations (BRUCE, KEITH, ELMER or MACK).
It is noted that 10Cs was seldom reported at
these other Eniwetok locations and that at several widely scattered stations in the tropical
Pacific that take rawinsondes it has become customary to enter 10 Cs persistently on the
primary basis of presence of a moist layer high aloft and on a secondary basis of real or
imagined visual observations, including a slight diminution of starlight that can equally we
be attributed to the high moisture content of the lower air.
Cloud observations involving 10 Cs are not always reliable, as noted above.
Low cloud
heights are probably correct within 200 feet during daylight because of the high frequency o.
local air traffic.
At night they are probably correct within 400 feet.
Estimated middle cl
heights are probably correct within 2000 feet. All cloud-height values are given in hundred
of feet.
Thus the entry "18" represents 1800 feet.
Direction of cloud movement is to four |
points of the compass.
DDFF is given to 16 points of the compass, with speed in knots for one-minute intervals
Assuming no persistent bias, speeds are correct within 10% and directions are correct within
1 point (plus or minus).
T,T,
xx and TT,
nin were taken from the hourly values.
For this reason, on afternoons with f
clouds the true T,T, may have been as much as 1° higher than those shown;
while during the
nighttime and very early morning T,T, may have been as much as 1° lower than the values show.
whenever there were showers.
(Lowest temperatures on tropical atolis are apt to occur momen
tarily during showers, evidently because of overturning of the air combined with the effect
evaporation.)
This source of unreliability is additive to that for TT (above).
RR is accurate within 0.01 inch, assuming care was taken in the observations.
In any
event, the representativeness of the catch is a factor that lowers the reliability decidedly
more than do any inaccuracies in measurement.
(See Table 34 and the notes therefor.
These
make it clear that RR values in Table 4 are decidedly too low.)
TIMES OF RAINFALL are biased by one to a few minutes in that there was no recording gag
and the observer would seldom notice to the minute (especially at night) the exact time of
inception or termination of rain.