744 
application of the knowledge thus gained to the de- 
velopment of entirely new scientific forecasting tech- 
niques. Obviously, it is quite impossible at this time 
even to suggest how such a solution is to be achieved, 
but the following remarks are offered as one mete- 
orologist’s opinion as to the direction from which the 
problem should be approached. 
Approach to the Problem of Scientific Weather Fore- 
casting. It is evident from the preceding discussion of 
the essential limitations of the present forecasting tech- 
niques that the primary inadequacy which they share 
stems from the fact that they are all essentially extra- 
polation techniques which, by one method or another, 
statistical or synoptic, qualitative or quantitative, at- 
tempt to derive the future or prognostic weather pat- 
tern from the current and past patterns. None of them 
is based on adequate thermodynamic concepts either 
of the operation of the basic drive of the general circula- 
tion, or of the energy sources or transformations which 
are involved in sudden accelerations or changes of 
trend in the development of the weather pattern. They 
all fail at the point of anticipating such frequent ac- 
celerational developments before the process is well 
in progress, when it is too late for anything but a very 
short-range forecast. The basic problem of forecasting 
research is to derive some quantitative physical model 
of the general circulation which fits the statistical and 
synoptic facts of weather observation. 
There are two rather opposite points of view from 
which this primary forecast problem is most frequently 
approached. They represent two quite distinct philoso- 
phies of the basic character of the longer-period anom- 
alous fluctuations of the world weather patterns. One 
method is to make an intensive study, by means of a 
dense network of complete and frequent synoptic ob- 
servations within the area of the investigation, of the 
detailed structure and of the dynamic and thermo- 
dynamic activity of the atmosphere within a limited 
region from which complete data can be obtained. The 
purpose of such an investigation is primarily to deter- 
mine the mechanics of operation of the dynamic and 
thermodynamic processes by which limited cellular cir- 
culations of the atmosphere are locally generated, trans- 
ported, and dissipated. Intensive, relatively localized 
synoptic studies of this type tend to be favored because 
of the comparative ease of obtaining adequate observa- 
tional data for restricted areas. It is more or less tacitly 
assumed in such studies that the entire developmental 
process of the local circulation is sufficiently self-con- 
tained so that it can be essentially explained by condi- 
tions within the limited field of observation [5]. If this 
line of approach to the basic forecast problem is se- 
lected, the choice implies that scientific forecasting can 
be realized at best only as a short-range accomplish- 
ment, that the general circulation in its entirety exists 
only as the integration of a number of individual cells 
or centers of action which operate independently and 
more or less at random such that the entire system lacks 
any unifying principle or control which makes its char- 
acter or behavior over extended periods of time either 
distinctive or predictable. 
WEATHER FORECASTING 
The opposite line of approach to the basic forecast 
problem is to study the general circulation extensively, 
preferably over all of the globe from which even an 
approximate picture of the circulation pattern can be 
obtained. The time scale is likewise extended to deal 
with charts at 24-hr intervals or with mean charts 
for even more extended periods of time, to obtain a 
comprehensive picture of the large-scale fluctuations of 
the general circulation in its entirety over relatively 
long periods. The essential purpose of such an ex- 
tended investigation is primarily to determine the me- 
chanies of operation of the dynamic and thermodynamic 
_ processes by which the entire general circulation is 
maintained in its continuously and irregularly fluctu- 
ating pattern of intensity and form. Investigations on 
this scale tend to be avoided because of the relative 
inadequacy of observational data over large parts of 
the earth’s surface, and the large amount of work in- 
volved in obtaining and processing the data for analy- 
sis. It 1s more or less tacitly assumed in such studies 
that the general circulation in its larger features, at 
least on either hemisphere, operates essentially as an 
integrated unified system which is functionally imter- 
related in all its parts. 
If this second line of approach to the basic forecast 
problem is selected, the choice implies that scientific 
forecasting must be based primarily on a better phys- 
ical understanding of the operation of the general circu- 
lation as a whole, and that the development and move- 
ment even of the secondary cellular circulations, and 
of all the attendant weather phenomena, the primary 
concern of the short-range forecaster, depends prima- 
rily upon the dynamics or thermodynamics of the 
hemispheric flow pattern and cannot be explained or 
anticipated on the basis of only the regional conditions 
in the sector of the development. This point of view 
indicates that the greatest improvement in weather 
forecasting is to be expected in the extended or long- 
range rather than in the short-range categories. 
There are a number of indications, based on a great 
number of statistical, synoptic, and theoretical inves- 
tigations in meteorology and in related fields, that this 
second approach to the basic forecast problem is es- 
sentially the correct one. This fact does not at all imply 
that nothing is to be gained from synoptic or theoretical 
studies of the secondary or tertiary atmospheric circu- 
lations, but it does imply that if this type of study is 
restricted to a limited sector of the earth’s surface, fac- 
tors of primary importance to the weather development 
probably will be overlooked. The primary evidence for 
the essential correctness of the global rather than the 
regional line of approach to the forecast problem may be 
summarized briefly as follows: 
1. The failure of many years of intensive study of 
regional weather patterns to evolve either a physical 
model or a theory of cyclogenesis and anticyclogenesis 
(pressure changes) of any real practical value in fore- 
casting regional changes of the synoptic weather pat- 
terns. Gross inadequacy of world-wide observational 
synoptic data has prohibited any corresponding study 
of the general circulation im its entirety. 
