38 MEASURING THE RADIO REFRACTIVE INDEX 



2.4. Comparison Between the Direct and 

 Indirect Methods of Measurement 



Absolute accuracy may not be the only consideration in the comparison 

 between the direct and indirect method of measurement. Although 

 refract ometers may be capable of superior accuracy, the factors of require- 

 ments, economics, and availability of competent technical personnel may 

 outweigh this advantage. Refractometers are relatively expensive, some- 

 what complex, and require competent technical personnel to maintain, 

 calibrate, and operate them. 



In many cases where average values or long-term statistics are adequate, 

 the use of refractometers may not be indicated. The data on refractive 

 index structure derived from weather service data has long been used 

 successfully for the determination of average conditions. 



Where extreme accuracy is required, the use of refractometers is indi- 

 cated. Radar and radio navigation are examples where accurate esti- 

 mates of both surface values and gradient are necessary to determine the 

 refraction through the atmosphere. The necessity for true vertical 

 gradients would demand the use of a balloon-borne or dropsonde type of 

 refractometer. In many applications the indirect method may be suffi- 

 cient; however, determination of the fine structure of the refractive index 

 appears to be presently limited to some type of radio-frequency refrac- 

 tometer. 



2.5. Radiosonde Lag Constants 



2.5.1. Introduction 



The determination of N from radiosonde data is subject to all of the 

 errors inherent in the radiosonde observation. Recently Wagner [40] 

 has analyzed the errors in A^ arising from time lag of the sensing elements, 

 data transmission techniques, and significant level-selection criteria. Of 

 these sources of error, Wagner concludes that the time lag of the sensing 

 elements is the most serious source of error. Further, for the southern 

 California coastal inversions, Wagner concludes that only the time lag of 

 the humidity strip need be considered. A similar conclusion has been 

 reached by Clarke [41] for practical applications involving ship-borne 

 radar and over-water air-to-air communications. Although there is a 

 significant correction associated with the time lag of the U.S. radiosonde's 

 lithium chloride humidity-sensing element, there is also a time lag in the 

 temperature element, which, as will be shown, must also be taken into 

 consideration. The correction for the temperature element yields a two- 

 fold correction to A^ due to the actual error in temperature and the ancil- 

 lary correction in vapor pressure resulting from the more correct estimate 

 of the true saturation vapor pressure. This arises from the fact that 



