740 
cant correlation between snow cover and subsequent 
anomalous conditions of the large-scale weather pat- 
terns. Hence these continental factors are at best only 
of minor secondary importance in extended or long- 
range forecasting. 
A great amount of statistical work has been per- 
formed by Helland-Hansen, Walker, C. E. P. Brooks, 
and many other investigators? in the effort to correlate 
anomalous conditions of sea-surface temperature and 
polar ice with contemporary, preceding, and subsequent 
anomalies of pressure (atmospheric circulation), air 
temperature, and rainfall. The upshot of most of this 
work can be summarized essentially as follows: 
a. Anomalous conditions of sea-surface temperature 
and polar ice are explained in general by present and 
past anomalies of atmospheric wind and temperature. 
b. Sea-surface temperature anomalies and cooling by 
melting ice are small in amplitude compared to air- 
temperature anomalies and show little tendency to 
persistence or displacement except as they are main- 
tained by anomalies of the atmospheric circulation. 
They cannot conceivably, from any quantitative con- 
sideration, be the primary cause of contemporary or 
subsequent air-temperature anomalies. 
c. Some significant correlation has been found be- 
tween spring ice conditions in the Greenland Sea and 
Barents Sea regions and subsequent summer, autumn, 
and winter weather in northern and central Europe. 
Since the sea-surface anomalies are utterly inadequate 
to account for the subsequent weather anomalies, it 
appears that both must be related to some primary 
factor of weather control which is best expressed in the 
highly anomalous state of the general circulation by 
which extreme ice conditions are initially produced. 
Certainly it is reasonable to conclude, granted the 
correctness of the above statements, that a further 
study of oceanographic influences is not a promising 
or direct line of attack on the basic problem of ex- 
tended or long-range weather forecasting. 
Solar control, either direct or indirect, of the normal 
seasonal features of the world weather patterns is axio- 
matic, but the question as to the extent to which the 
anomalous fluctuations of the general circulation are 
influenced or controlled by irregular solar activity is a 
highly controversial one. Literally scores of investiga- 
tions, statistical, synoptic, and theoretical, have been 
directed toward one or another aspect of this problem. 
Relationships of world weather patterns to sunspots or 
solar constant have been investigated by Abbot, Baur, 
Clayton, B. and G. Duell, Hanzlik, Helland-Hansen, 
K6ppen, Kullmer, Schell, Simpson, Tannehill, Walker, 
Willett, and many others. Direct solar effects on the 
higher atmosphere, notably on temperature, circulation, 
distribution of ozone, and ionization, have been ob- 
served or theorized about by equally many investiga- 
tors, including among others Craig, Dellinger, Dobson, 
G6étz, Haurwitz, Hulburt, Maris, Mimi, Shapley, Stet- 
son, and Wulfe. é 
In spite of the great amount of effort which has been 
2. See Mon. Wea. Rev. Wash., Supp. No. 39, pp. 27-57, 1940. 
WEATHER FORECASTING 
expended on this problem it remains unproved today 
whether solar activity plays a primary, secondary, or 
insignificant role in the irregular fluctuations of the 
atmospheric conditions of the troposphere (world 
weather patterns). There is no question whatsoever 
about the occurrence of numerous direct effects of sudden 
solar disturbances in the higher atmosphere, but any 
such effects in the lower troposphere, if they exist, are 
so indirect, masked, and complex that they have not 
been statistically confirmed. There is an imposing 
amount of statistical evidence for long-term fluctua- 
tions of the world weather patterns, particularly m the 
tropics, but also im the higher latitudes, that roughly 
parallel the single or double sunspot cycle. Since the 
parallelism is either not uniformly close or not consist- 
ent in phase over long periods of time, the evidence in- 
dicates that sunspots themselves are not a satisfactory 
index of the disturbing solar influences. It can be stated 
without qualification that up to the present no attempt 
to base extended or long-range weather forecasts di- 
rectly on any solar index, usually sunspots or solar 
constant, has attaimed any significant success. Prob- 
ably no primarily statistical attack on this problem can 
be expected to accomplish more at present. 
Assuming that there is an ultimate control or direc- 
tion of the major anomalous fluctuations of the general 
circulation by solar activity, then our failure to estab- 
lish the reality or nature of this control is readily ex- 
plained by our almost complete ignorance in the fol- 
lowing three areas: 
a. The physical nature and intensity of the disturb- 
ing solar influences which enter the outer atmosphere. 
b. The qualitative and quantitative effects which 
these influences produce in the higher atmosphere. 
c. The essential mechanics of the general circulation, 
including any possible mechanism of interaction be- 
tween the troposphere and the higher atmosphere. 
Probably the most representative opinion of meteor- 
ologists who have recently dealt with the broader as- 
pects of the anomalous solar-weather relationships [1, 2, 
17] can be expressed approximately as follows: Solar 
activity undoubtedly exerts some guiding influence, 
possibly even directing control, on the extended and 
long-range anomalous fluctuations of the general circu- 
lation. This influence or control is so indirect and so 
complex in the sum total of its manifestations that little 
progress is to be expected from any further primarily 
statistical attack on the problem. This problem calls 
for a major program of research in solar and atmospheric 
physics, a program which must be an integral part of 
any major attack on the basic problem of weather fore- 
casting. Such a program must be directed at all three 
of the basic areas of ignorance noted above. 
RECOMMENDATIONS FOR THE IMPROVEMENT 
OF WEATHER FORECASTING 
Possible Immediate Technical Improvements. The 
general problem of the improvement of weather fore- 
casting must be considered from two rather distinct 
angles. On the one hand there is the question of possible 
accomplishment in the immediate present or near future 
