Part Be Phase 59.0 62.0 19.0 eee ee ad 18.0 a MC “gee FRED: : salt ae etebye OM ek ged”, Sake ae dbictn tadtenince ht i med als. Observational Data, First Intensive (August 18 -~ September 1, 1957) NOTES: TABLE_ke : Aro cal CR fl iee al arege TABLES 4-18 HOURLY OBSERVATIONS AND DAILY SUMMARY. These Notes apply both to Table 4 and Table 19, which presents similar observational data Se 420 bee Pe 83.0 e. for the second Intensive Phas sandths of an inch, with the units and P represents station pressure and is given to thou tens omitted. 30.000. ere L76.0 exeRe 7 tower } ee “METER #2 iaRNe HetSag Reees Seas BeIiessSRETS «ale or oe LU peerSes In Tables 4 and 19, all values are preceded by 29, except 000, which represents The mercurial barometer (used daily to check the microbarograph) was calibrated January 30, 1958 and found to be 0.020 inch too low. in the Tables. This value should be added to those shown In addition, unreliability is introduced because the hourly values were read from the microbarograph and because of the lag in this instrument. Allowing for this factor, after 0.020 has been added te the values, the resulting values will all be correct within 0,020 10.0 (plus or minus) and half of the resulting values will be correct within 0.004. line extreme error of 0.020 represents the maximum 10-minute change that may be expected at Eniwetok, considering both the diurnal pressure curve and the changing synoptic situations. (More rapid change might accompany approach of a typhcoon or an intense tropical storm, but such eee SMteTe, BEEret eee ot Stas did not occur during these observational periods.) The ten minute period represents the maximum time-lag between the mercurial barometer and the microbarograph at times when the pressure is changing rapidly. (When it is changing very slowly the lag may be greater, but then the error amplitude is diminished very appreciably.) The value 0.004 is based on the assumption that rates of change of pressure over 5-10 minute periods are distributed normally about their mean. Finally, it should be noted that these error estimates allow for the fact that often in actual practice observers do not tap the microbarograph to permit the pen te adjust to the current pressure. TT and Ty were to be read to 0.1° F. according to standard instructions. It is evident, however, from the very high frequency of values ending in .0 or .5 that the observers usually read the temperature to the nearest graduated mark (.0 or 05). Allowing for this fact and for an extreme instrumental error of 0.3°, all values are correct within 0.5% 67 “This assumes there is no consistent bias, either instrumental or human, and that in borderline cases the observer can discriminate to 0.1*. 39