extended to obtain the information and disseminate it rapidly to their own troops. But while the war acted as a stimulus to the obtaining of data about the regions in which aircraft of all kinds are used and through which projectiles pass, in another way it has been a great hindrance to an advance in meteoro- logical science. Not only were the international committees largely broken up but in the latter stages of the war the publica- tion of observations was prohibited; each nation treated its own work as more or less confidential, and although all restrictions are now removed it is not easy to obtain and assimilate the im- portant papers that were written during the war.
The Upper Atmosphere. During the years before the war observations on the temperature and humidity of the air strata were rapidly accumulating, more particularly from a network of stations spread over Europe, and since the W. and N. of Europe is subject during the winter to the passage of many deep cyclonic depressions, the conditions of temperature in cyclones and anticyclones up to a height of some 20 km. (125 m.) had become known. The brief tables which were all that were available to Cleveland Abbe in 1909 had been supplemented by much information, drawn up and arranged for the European results by Lt.-Col. E. Gold (M.O. No. 2100, Geophysical Memoirs, No. 5), by Dr. Wegener for the Continent (Die Temperatur- verhaltnisse in der frcien Atmosphdre, III. Band, Heft 2/3, Leipsic, 1909) and for Russia by Dr. Rykatchew (Meteorologische Zeitschrift, Jan. 1911). In 1916 a summary of the information available about the upper air was drawn up for the Meteoro- logical Office but not published. It quoted freely from Gold's paper but included the results of observations up to 1916. This summary together with certain theoretical matter was published in 1919 under the title " Characteristics of the Free Atmos- phere " (M.O. 22oc, Geophysical Memoirs, No. 13), and from it the following abstract summarizing our present knowledge of the strata from o to 20 km. is mostly taken.
Temperature. As the surface of the earth is left the temperature of the air decreases with increasing height, and when the great varia- tions of climate and of the conditions prevalent in different parts of the earth are considered it is remarkable how uniform is the fall of temperature, now commonly called the lapse rate. The height to which it extends is variable, but in all places in which observations have been made, the lapse rate iJp to 8 km. has been found close to , 6 C. per kilometre. This holds, not indeed exactly but approxi- mately, for summer and winter and for places as far apart as the equator and Ross Bay in lat. 78 S. Thus in Batavia the lapse rate up to'8 km. is 6-1 per km.; at Petrograd it is 5-8. In England in the winter it is 5-8, in the summer it is 6-0. These are means but the rule holds quite well even for the individual case, for if in one part of the 8 km. the lapse rate is small this is usually compensated for by its being large in the other part. The only important exception that has been found so far is in regions and at times where the temperature is extremely low, as in Siberia or Canada in the winter. In such instances the bottom layer is unduly cold and the lapse rate is negative over the first 2 km., so that the rule would make the upper air temperature too low. Also it must be remembered that the daily variation of temperature does not extend upward more than one or two km., so that the mean for the day rather than the precise temperature at the moment should represent the surface tempera- ture. This layer, in which temperature falls with increasing height, is called the troposphere.
At a certain height, which varies with the latitude, with the barometric conditions and with the season, the fall of temperature ceases, and the air up to the greatest heights that have been explored remains at a nearly uniform temperature in the vertical direction. This upper part in which there is no lapse rate is called the strato- sphere. The boundary between the two parts is found at about 16 km. near the equator and at 10 km. in northern Europe. Over England its mean height is 10-5 km., falling to rather below 10 km. in the winter, and rising to over 1 1 km. in the summer. In the centre of a deep cyclone the value may easily fall to 8 km. ; in an anticyclone it may exceed 12 km.
The temperature of the stratosphere is below 200 A. over the equator and in tropical regions; it is above 220 A. in northern Eurcjpe. In Canada it seems to be lower in the summer than in the winter. These anomalies are roughly expressed by the rule that the mean temperature of the air column taken with regard to height from o up to 19 km. is approximately the same in all parts of the earth. There is probably a physical reason for this, and it explains the unexpectedly low temperature above 14 km. over the equator and the curious reversion of temperature between summer and winter over Canada (Toronto) where the seasonal range is very large. The annual range of temperature in the troposphere does not
differ very greatly from the range at the surface; in island and coastal climates like England it is rather greater in the upper parts than at the surface; in continental climates the surface has the greatest range. In the stratosphere the range is much reduced and, as already stated, appears in Canada to be reversed although enough observations are not yet available to make this absolutely certain.
Whether or not there is any regular diurnal change of temperature above 2 km. height is uncertain ; all that can be said is that if there be any its amplitude is certainly less than I C.
The mean annual temperatures are given in the accompanying Table I. In Europe the probable error of any value is about 1 C.; for Canada and the equator owing to paucity of observations it is greater, especially above 15 km., where it may reach perhaps 3 C. Over Europe the mean temperature does not change from 14 to 20 km. and does not change much over Toronto. Over the equator the lowest temperature, which is about 193 A., is not reached under 16 or 17 kilometres.
TABLE I. Mean Temperature.
The values are in the Absolute scale with the first "2" omitted. 273-0 =oC. =32 F.
1
1
c a
1-1 0)
J3
d
-6
te
u
3
.0
Ul
a a e
.g
- S ^
|
S
jd bio
"
g
"S
&*(/5
(2
(0
3
B
&
M
o
3 O*
"8
&
$
a
w
!H
>
s w
8-
W
14
23-5
22-O
20-5
18-7
18-9
19-1
17-9
19-6
17-7
19-1
12-5
3-o
13
23-4
21-8
20 6
19-3
18-7 ig-3
17-6
19-6
16-4
19-2
14-0
I I-O
12
20-7
21-6
20-0
18-3
18-8
19-5
16-8
18-3
16-1
18-4
16-2
I9-0
II
2O-O
20-5
20-9
19-2
19-6:20-2
18-1
18-4
18-5
19-1
19-3
27-O
IO
21-3
21-2
23-2
21-9
22-2
24-3
22-3
21-8
22-7
22-2
23-2
35-o
9
8
24-4 24-8 29-8I3O-2
28-2
33-8
26-8 33-1
27-5 33-6
30-0
27-8
34-8
26-9 33-6
27-3
33-9
27-2
33-4
29-3
35-9
43-o 51-0
7
37-1 '38-0
40-2
40-8
40-7
44-3
42-1
41-2
41-2
40-7
43-5
58-0
6
43-3 45-o
47-0
47-9
47-8 51-4
49-3
48-8
49-4
47-8
50-9
65-0
5
49-8 52-0
53-8
54-8
54-8,58-1
56-i
55-6
56-2
54-6
57-7
72-0
4
55'7 58-4
60-4
61-0
61-7 64-3 62-4
61-9
62-9
61-1
64-1
79-0
3
61-3 64-0
66-6
66-9
67-7 69-8 68-4
67-6
69-2
67-0
69-6
85-0
2
66-7,7o-3
71-7
71-7
73-2 74-5 73-8 73-o
75- 1
72-4
74-8 90-0
I
71-075-3
77-0 ;6-S
78-0 78-5 78-2 77-6
80-7
76-8 78-3 : 95-Qj
Pressure and Density. The temperature of the air having been found by observation, the pressure and the density are easily found up to the height to which the observations extend. In the same way the mean pressures and mean densities can be determined from the mean temperatures without appreciable error provided the mean pressure at the surface is known.
In the lower strata the pressure at any particular height is natural- ly most dependent upon the surface pressure, but since the air is lighter, bulk for bulk, when it is warm the pressure decreases less rapidly than usual in a warm area, and the pressure at any given height depends more and more upon the temperature of the under- lying air as that height increases. Thus it comes to pass that in the hot regions of the earth, say in the belt included between the two tropics, the pressure at the height of 9 km. is very much less than it is at the same height over temperate latitudes, and the pressure gradi- ent which causes the prevailing westerly winds of the cirrus level is thus produced. At a height of 20 km. the surface pressure has ceased to have much effect, and it requires a rise of nearly 20 mb. l in the surface pressure to produce a rise of I mb. at 20 km., whereas a change of 1-5 C. in the temperature of the air column will produce that effect. It has been stated that the mean temperature of the air column up to 19 km. is much the same in all parts of the world, and it follows that the same level is one of nearly uniform pressure.
The pressures are given in Table II. at various stations for heights up to 20 kilometres. The values for Canada and the equator at heights above 15 km. are not very reliable owing to paucity of data.
The densities are given in Table III. The variations in the density became of great consequence during the war on account of their influence on the range of projectiles; they depend on the connexion which has been found to exist between temperature and pressure.
Statistical Methods. Statistical methods have been much in vogue of late years and it is necessary to indicate how the method of correlation has been used for forecasting and for elucidating meteorological problems. A large number of cor- relation coefficients have been determined between various meteorological events, and the values of many of them are given in the Computer's Handbook, M. O. 223, Section V, Tables, published by the Meteorological Office.
The advantage of a correlation coefficient in estimating the connexion if any between two events is that it expresses the connexion as a decimal which must lie between r and i, and
'The average pressure of the atmosphere at sea-level being reckoned as i bar, = 1,000 millibars (mb.). i mb. =0-0295306 mercury in. at 32 F. in lat. 45.
