EXTENDED-RANGE WEATHER FORECASTING 
tion is an extraordinarily complex phenomenon. Those 
parts of the sun’s radiation which emanate from the 
sun’s corona and the uppermost chromosphere fluctuate 
with the sunspots, that is, on the average, with an 
eleven-year period. These radiations (corpuscular rays 
and electromagnetic waves) are, however, with one 
exception, absorbed in the ionosphere and hence, ac- 
cording to our present ideas, are of no influence upon 
the weather. Only the ultrashort electromagnetic waves 
with a wave length \ < 15 m penetrate to the earth’s 
surface. Since, according to Kiepenheuer [85], ultra- 
short waves act upon colloidal processes, it is possible 
that the single fluctuation of precipitation in the tropics 
is caused by a solar influence on the condensation and 
sublimation of the water vapor, whereas the tempera- 
ture fluctuation is secondarily produced by cloudiness 
and precipitation. In agreement with this assumption 
is the fact that the correlation of sunspots with tem- 
perature is of opposite sign and of less significance than 
is the correlation of sunspots with the amount of pre- 
cipitation or with the water level of large inland lakes. 
The radiation emanating from the photosphere, that 
is, the heat radiation, shows—as has been demonstrated 
several times [7, 9, 26]—no single oscillation in or out 
of phase with the sunspots. However, when the differ- 
ent mean annual trends in the time intervals 1919-23 
and 1924-30 are eliminated [9], the curve of the solar 
constant for the period from July 1918 to February 1937 
shows minimum values at about the time of sunspot 
extremes, and maximum values 114 to 214 yr before the 
sunspot extremes [19; 20, pp. 967-968]. These fluctua- 
tions agree quite well with the double oscillation of the 
meteorological elements as presented in the preceding 
section. However, it must be conceded that the double 
fluctuation in the most recently revised monthly means 
of the solar constant no longer appears so clearly [1]. 
An increase in the total energy radiated from the 
sun would mean an increase in the heat supply, par- 
ticularly for low latitudes during winter. The tempera- 
ture difference between the tropics and the polar region 
would thereby be increased, resulting in an increase of 
the general circulation of the atmosphere and a con- 
sequent increase of the pressure gradient between the 
subtropics and the subpolar low-pressure belt. The 
reality of the fluctuations of the solar constant is cor- 
roborated by the fact that during midwinter (January 
and February) in the period 1919-7, a strong positive 
correlation existed between the mean solar constant 
and the simultaneous mean values of the pressure 
differences between Valencia (Ireland),and Spitsbergen 
and between Rome (Italy) and Haparanda (Sweden) 
[17, pp. 97-105]. These correlation coefficients were 
+0.69 and +-0.65, respectively, both of which exceed 
the maximum chance value. 
The objection has been raised that the variations of 
the solar constant are too small to cause fluctuations 
of the general atmospheric circulation. However, ac- 
cording to the Stefan-Boltzmann law, a variation of an 
amplitude of 1 per cent of the solar constant (assuming 
incoming and outgoing radiation to be equal) causes a 
temperature change (temperature departure) of 0.7C 
821 
which refers to the entire earth’s surface. It is therefore 
quite possible that, owing to the inequalities of insola- 
tion (as a function of the geographical latitude) and 
the different heat capacities of land and ocean, the 
temperature change caused by a one per cent change 
of insolation may amount to considerably more than 
0.7C in some regions. 
Results of recent investigations [35, 37] suggest that 
the fluctuations of the atmospheric circulation within 
a sunspot cycle are caused not so much by the changes 
of the entire heat radiation as by the fluctuations of 
the sun’s radiation in the ultraviolet. This idea is sup- 
ported by the fact that the difference between the areas 
of the sun’s faculae / and those of sunspots L, both 
relative to their long-established mean values Fy and 
LI, undergoes the same double fluctuation within a 
solar cycle as does the atmospheric circulation [19]; 
this is shown in Fig. 4. Radiation emanating from the 
YEARS 
AFTER 
MAX. 
Fig. 4.—The mean course of the difference between the 
equivalent annual means of the areas of sun faculae (Ff) and 
sunspots (ZL), 100(F/Fo—L/Lo), for the sunspot cycle, from 1875 
to 1940. 
faculae is known to be particularly rich in ultraviolet. 
This relationship can be explained as follows: In the 
upper mixing layer above the stratospheric inversion 
layer, temperature differences are set up by differences 
in the absorption of ultraviolet radiation. These tem- 
perature differences are connected with pressure dif- 
ferences of the same sign in a manner similar to that 
found in the middle troposphere. If this assumption 
holds true, an increased ultraviolet radiation results 
in an increased pressure gradient from the equator 
toward the pole in the upper-air layers, because—dis- 
regarding the time around the summer solstice (see p. 
824)—the ultraviolet radiation as well as the total 
radiation affects primarily the lower latitudes. Because 
of the increase of the pressure gradient aloft, the plane- 
tary (zonal) circulation increases, and the poleward 
border of the subtropical high-pressure belt is displaced 
toward the pole. 
Solution of the Second Basic Problem. It follows from 
the discussion in the foregoing sections that weather- 
governing complexes, the existence of which has been 
proven (first empirical theorem), are not to be found 
in terrestrial processes. The terrestrial processes are 
either too rare (strong volcanic eruptions) or too weak 
(motions of the pole) to influence the course of the 
weather. There are, indeed, purely terrestrial regulatory 
