WATER VAPOUR IN THE UPPER AIR 
By G. M. B. DOBSON and A. W. BREWER 
The Clarendon Laboratory, Oxford 
INTRODUCTION 
Water vapour in the upper air plays an important 
part in the dynamics of the atmosphere because (1) it 
is a source of atmospheric energy, (2) its presence af- 
fects the release of energy, (3) it is the origin of all hy- 
drometeors, and (4) it is the main constituent of the 
lower atmosphere to absorb infrared radiation. In ad- 
dition the water-vapour content of the air can be used 
in meteorological studies as a natural “‘tracer’’ element. 
The importance of water vapour as a source of atmos- 
pheric energy has been discussed by Normand [12], but 
generally when the temperature is below about —20C 
the water vapour is not thermodynamically significant. 
All hydrometeors originate from the condensation of 
water vapour in the upper air, but the production of 
large amounts of rain or snow requires the larger va- 
pour pressures, which occur only at temperatures above 
about —20C. Condensation at lower temperatures than — 
this may be of importance in inaugurating the colloidal 
instability of a cloud whereby the liquid cloud droplets 
can fall as snow, but the quantities of water contributed 
by the regions of low temperatures will not usually be 
significant. 
Measurements of the humidity of the free air have 
been made for many years using, for example, hair or 
goldbeater’s skin hygrometers, or wet- and dry-bulb 
psychrometers, but these are useful only when the 
amount of water vapour in the air is large. We do not 
propose to discuss these phenomena, which are already 
well known. 
For the remaining processes listed in the first para- 
graph, water vapour is important even if present in 
very small amounts, but since there has been no hy- 
grometer capable of measuring very small water-vapour 
pressures, this field of study has received little atten- 
tion. Since 1943 it has been possible, by a simple devel- 
opment of the well-known dew-point hygrometer, to 
obtain reliable measurements of the water content of 
the upper air at all levels which multiseated aircraft 
can reach, and a large number of measurements up to 
12 or 13 km have been made by the Meteorological 
Research Flight of the British Meteorological Office. 
No entirely reliable measurements of the distribution 
of water vapour at levels above about 13 km have yet 
been made, but the values obtained by Regener [14] 
appear to be very high.! The rare mother-of-pearl clouds 
which have been described by Stérmer [17] are usually 
formed at about 25 km and appear to be very tenuous 
water clouds, which would indicate that at times the 
air may be saturated at these levels. 
1. This was written before the results of Barrett, Herndon, 
and Carter [1] were available. Their results are discussed below. 
THE MEASUREMENT OF WATER VAPOUR IN 
THE UPPER ATMOSPHERE 
The difficulty of measuring the water-vapour content 
of the upper air lies in the extremely small amount of 
water vapour present, and sometimes even at moderate 
levels the air is too dry to permit measurement of its 
water content with ordinary hygrometers (see Fig. 9). 
One cubic metre of saturated air contains only 26 mg 
of water at a temperature of 220K, 1.6 mg at 200K, 
and 0.3 mg at 190K. Water-vapour densities of the 
order of 1 mg m~™ are found in the stratosphere over 
southern England. Since it is possible to deal with only 
a small fraction of a cubic metre of air, it will be real- 
ized how very small an amount of water or ice has to 
to be measured, and it is for this reason that most of 
the usual methods of measuring humidity fail when 
used in the upper air. 
Glueckauf [8] showed that a very thin film of poly- 
vinyl acetate containing a hygroscopic salt could be 
used as a very rapid hygrometer. The thickness of his 
films was about 0.1 »; the films changed their thickness 
rapidly with changing relative humidity (with respect 
to water). If illuminated by monochromatic light, the 
thickness of the film could be estimated from the reflec- 
tion coefficient. Unfortunately these films are very del- 
icate, and the method has not come into general use 
Glueckauf [9] has also obtained humidity measurements 
at high levels by correcting the hair hygrometer of a 
Dines meteorograph for its lag. 
Another possible hygrometer involves the use of a 
Wilson cloud chamber. Air with suitable nuclei is sup- 
plied to the chamber and expansions of increasing ratio 
are made until a fog is just formed. The first formation 
of a fog can be detected by the scattering from a strong 
beam of light passing through the cloud chamber. This 
method has not yet been fully developed, but tests 
made by Palmer [13] show that it should be effective 
down to a frost point of about 210K, below which the 
fog formed is too thin for convenient observation. 
A method of measuring the water content of the high 
atmosphere which has some unique advantages is to 
measure the absorption of the sun’s infrared energy by 
water vapour. Because of the very smail amount of 
water vapour in the upper atmosphere it is necessary 
to use the 6.3-» absorption band of water vapour since, 
except for the absorption band at 2.7 u, the bands of 
shorter wave lengths have absorptions which are too 
small. The 2.7-1 band is confused with a CO, absorp- 
tion band and cannot be used. 
The absorption gives information about the total 
amount of water vapour above the height of observa- 
tion, and the vapour pressure at various levels must be 
obtained by differences. If these differences can be com- 
311 
