THE FORECAST PROBLEM 
within the framework of our present: basic knowledge of 
atmospheric circulations and with our present fore- 
casting techniques. On the other hand there is the ques- 
tion of the ultimate approach to true scientific weather 
forecasting by means of extended research in the basic 
problems of dynamic meteorology and the development 
of new objective forecasting techniques. 
Tt is emphasized in the preceding discussion that im- 
portant progress in the basic problem of weather fore- 
casting is not to be achieved by mere extension, elabora- 
tion, or refinement of any of the present great variety 
of forecasting techniques. On the other hand, it is quite 
certain that some improvement of accuracy and uni- 
formity of forecasting performance beyond that now 
attamed can be achieved on the basis of our present 
basic knowledge and forecasting techniques. The pos- 
sibility of the technical improvement of present fore- 
casting procedures may be considered at three points: 
1. Weather Observations. Our knowledge of present and 
past weather conditions, on which are based both the 
forecast of the future weather and the verification of 
past forecasts, depends almost entirely upon the analy- 
sis of synoptic weather observations. Obviously it is 
important that these observations be adequate for the 
uses to which they are to be applied. By and large it 
may be stated that observational techniques are far 
more advanced than forecasting techniques in mete- 
orology, hence primary emphasis in meteorological re- 
search for forecasting purposes should definitely be 
shifted from the former to the latter. However, meteor- 
ological observations are far from being all that they 
should be. In this connection it can be stated that at- 
tention should be devoted primarily to the reliability 
and distribution of weather observations, rather than 
to mereased accuracy, elaborateness, or density of ob- 
servations in favored local areas. The greatest need, for 
the improvement of general or extended forecasting as 
well as for forecasting research, is for simple reliable 
observational techniques which can be applied uni- 
formly to large areas of the earth’s surface and to the 
upper atmosphere, to fill in the extensive global gaps 
which still exist im available synoptic information. This 
calls for instruments which are as foolproof and free of 
calibration difficulties as possible, and the further de- 
velopment of automatically recording and transmitting 
weather-station equipment for use in relatively inac- 
cessible areas. Any possible simplification of raob and 
rawind techniques is highly desirable. 
More elaborate or exact weather-measuring devices 
are not justified for general use because the small-scale 
random variability of weather conditions renders exact 
measurement of no significance. For example, useful 
as it would be to observe the field of vertical motion in 
the atmosphere, exact measurements of this quantity, 
even if they could be made, would be of little value be- 
cause the local variations are very large relative to the 
prevailing motion. 
It is only for short-range forecasting of local condi- 
tions for special purposes that it is possible to justify 
a dense network of special observations, such as the 
radar technique for the measurement of condensation 
741 
forms. Insofar as it appears advisable, detailed study 
of local forecast problems and problems of atmospheric 
physics can be set up with full equipment in selected 
limited areas, but at the present time it is unlikely that 
such studies will contribute as much to the solution of 
the basic problem of general weather forecasting as 
will adequate observations to eliminate the great gaps 
in available synoptic information in many parts of the 
world. The primary inadequacy of our present fore- 
casting knowledge lies in the field of the dynamics of 
the general circulation rather than in the field of at- 
mospherie physics. 
The proper implementation of an adequate world- 
wide network of synoptic weather observations is com- 
pletely dependent upon the existence of a strong inter- 
national meteorological organization which can specify 
uniformity in instrumentation, observational tech- 
niques, hours of observation, and the coding and trans- 
mission of synoptic data, even to the distribution of 
international funds to support such a program. There 
is no possibility of realizing the necessary world-wide 
network of synoptic weather observations except as it 
is internationally financed according to the ability of 
the different nations to support it. 
2. Standardization of Forecasting Techniques. It is 
evident from the discussion of forecasting techniques 
(see p. 733) that there is practiced today a great variety 
of such techniques, many of which are overlapping, 
essentially redundant, or of very limited applicability, 
and for which little or nothing in the way of rigorous 
objective evaluation has been attempted. It was noted 
that the essential basis of almost all short-range or daily 
forecasting consists of the empirical extrapolation, in 
the form of two-dimensional prognostic charts, of the 
current two-dimensional synoptic weather patterns, in- 
fluenced by certain physical factors such as topography 
and the modification of air-mass properties or of frontal 
structure, and occasionally aided by the use of some 
climatological statistics. The three-dimensional distri- 
bution of the weather elements, both current and prog- 
nostic, is expressed by the selection of suitable two- 
dimensional sections for analysis and prognosis. With 
increasing range of weather forecasts to the extended 
and long-range categories, the forecasting techniques 
are based increasingly on the extrapolation of mean 
synoptic rather than instantaneous synoptic patterns 
and on the use of statistical aids. 
The greatest improvement of weather forecasting 
which can be expected within the framework of our 
present basic knowledge almost certainly is to be ef- 
fected by the standardization and simplification of the 
multiplicity of techniques which are employed for the 
prognostic extrapolation of synoptic weather patterns, 
particularly in the short-range and daily forecast cate- 
gories. This applies both to the techniques of extrapola- 
tion and to the preliminary presentation and analysis 
of the synoptic data. It is possible to present synoptic 
weather data in an infinite variety of forms, as regards 
the two-dimensional sections which are selected for 
analysis, the combinations of weather elements which 
