INDIRECT MEASUREMENT 25 



Experimenters at the National Physical Laboratories [4] have shown that 

 when an Assnian hygrometer is used with thermometers accurate to 

 within 0.1 °C, the optimum expected accuracy in the determination of the 

 water vapor pressure is 0.2 mbar. Errors three times as great can be 

 expected at extreme conditions. Hence, it would appear that measure- 

 ment of water vapor limits the accuracy of the determination of the 

 refractive index to approximately ±1.0 A'^ units. 



Temperature, pressure, and humidity are standard measurements of 

 the world's weather services. Methods of measuring these parameters 

 are fairly standard. Usually the temperature is read from mercurial or 

 alcohol thermometers, the pressure from mercurial barometers, and the 

 huniidity from a conversion of wet- and dry-bulb thermometers. The 

 degree of accuracy of these measurements is usually a function of the care 

 exercised by the observer. Thermometers protected by radiation shields 

 are usually accurate to within ±0.1 °C, barometers to within ±1.0 mbar. 

 Reading errors can easily be in excess of the instrument error, especially 

 in the determination of humidity where the wet-bulb depression is subject 

 to many sources of error. Contamination of the wick or water, insuffi- 

 cient wetting, and inadequate aspiration are con^mon sources of error. 

 The wet-bulb determination can be used below freezing, if proper pre- 

 cautions are observed. 



Automatic-recording systems have been devised for measurement of 

 temperature, pressure, and humidity. The simplest are the hygrother- 

 mograph and the microbarograph, in which the sensors are connected by 

 mechanical linkages to pens or chart recorders. The pressure is recorded 

 by means of an aneroid capsule, the temperature by means of a bimetal 

 strip or a curved Bourdon tube, and the humidity by means of a hair 

 hygrometer. The accuracy of such devices limits the determination of 

 N to within 2 or 3 A^ units. 



Sensors producing electrical outputs are used to measure meteorological 

 parameters in a variety of automatic-recording systems, such as strip- 

 chart recorders, punched-paper tape, and magnetic-tape recorders. The 

 more common temperature sensors include resistance thermometers, 

 thermocouples, and thermistors. The platinum resistance thermometer, 

 an international standard, is capable of measuring temperature in still 

 air to within ±0.05 °C if used in a well-compensated bridge circuit. 

 However, the platinum resistance thermometer is velocity-sensitive in a 

 moving air stream. Thermocouples avoid this problem and yield short- 

 term accuracies of approximately ±0.1 °C with time constants measured 

 in milliseconds. 



The total atmospheric pressure can be measured by a variety of elec- 

 trical sensors. The simplest device is the pressure potentiometer where 

 an aneroid capsule is mechanically linked to a potentiometer. One or 

 two millibars can be considered the limit of accuracy in a differential 



