SHORT-RANGE WEATHER FORECASTING 
By GORDON E. DUNN 
U.S. Weather Bureau, Chicago, Illinois 
INTRODUCTION 
Literature. Literary productivity in meteorology dur- 
ing the past two decades has reached the highest rate 
in meteorological history. Much of it has been descrip- 
tive and theoretical in nature, some of it with fore- 
casting implications, but remarkably little has been di- 
rectly concerned with or has contributed significantly 
to the improvement of forecasting. Indeed, the few 
modern textbooks supposedly directed primarily toward 
analysis and forecasting devote comparatively little 
space to the practical problems of forecasting. A tend- 
ency exists for the theoretical meteorologist, and even 
the analyst, to remain aloof from the vicissitudes and 
discouragements of the practical forecaster. Research 
workers apparently believe their obligations have been 
fulfilled when their results and suggestions have been 
passed along to the forecaster, while the latter, as a rule, 
has neither the time nor the facilities to test the sug- 
gested practical applications of current research. Thus, 
unfortunately, the gap between the theoretical meteorol- 
ogist and the forecaster remains unbridged. 
The Forecasting Problem. The problem of forecast- 
ing may be divided into three separate but closely re- 
lated phases: (1) analysis, (2) prognostication of pres- 
sure patterns, and (3) forecasting the weather. Air- 
mass analysis may be defined as the study of preceding 
and current meteorological conditions over a prescribed 
area for the purpose of deriving a satisfactory and 
logical explanation of the physical processes and 
weather actually observed. The analysis should imclude 
a determination of the structure, location, direction and 
rate of movement of fronts, the characteristics of the 
various air masses with particular reference to temper- 
ature and moisture, and an explanation of any precipi- 
tation areas. Analysis of the area required for forecast- 
ing at the district level has become too great a burden 
for forecasters and the desirability of central analysis 
centers 1s now generally accepted. 
In addition to the more normal functions, analysis 
centers (or extended-forecast units) should provide dis- 
trict forecast centers with analyses of broad or large- 
scale features of the general circulation as determined 
by the “‘centers of action” or “weather controls,” which 
often lie some distance off the normal forecasting chart. 
These analyses would include hemispheric wave lengths 
and indications of blocking, since these often influence 
weather developments far away within very short 
periods of time. . 
Complete utilization of analysis centers must await 
further development of facsimile smce the coding and 
decoding of analyses is time consuming and, more im- 
portant, the consequent smoothing results in a serious 
loss of character in the analysis. However, transmission 
of prognostic charts and many types of analyses by 
facsimile is already satisfactory in the limited areas 
where this facility is in use. 
The preparation of prognostic surface-pressure charts, 
with indicated frontal positions, and prognostic con- 
stant-pressure charts comprises the second phase of the 
forecast problem. These charts are derived by more or 
less mechanical means and are subjectively evaluated 
and modified at the analysis central and then trans- 
mitted to forecast centers. Duplication of the same proc- 
ess of analysis at individual forecast centers represents 
an excessive waste of time and a group of specialists, 
with an adequate staff for plotting the necessary addi- 
tional maps and diagrams, should provide forecasters 
with better analyses than are obtainable by any other 
method. Prognostic surface charts for periods up to 30 
hr now maintain a very high standard of accuracy and 
satisfactory progress is being made toward 54-hr sur- 
face prognostics. The preparation of the 700-mb prog- 
nostic chart has, apparently, not met, so far, with the 
success which might be expected from acquired ex- 
perience and available techniques. The preparation of 
both surface and upper-air prognostic charts has not 
improved to the point where the district forecaster is 
relieved of the duty of checking and recomputing pre- 
dicted frontal and pressure system positions for his own 
district on the basis of additional and later reports and 
his greater experience in his own area. 
The third, and most difficult and important, phase 
of the forecast problem is the prediction of the weather. 
Because of the local nature of many forecast problems 
and the extreme variability and complexity of weather 
over any large area such as the United States, some 
decentralization of forecasting is required. This article 
will deal primarily with the preparation of prognostic 
charts and the prediction of weather, with particular 
reference to the middle latitudes, and will be concerned 
only casually with analysis. However, it should be 
emphasized that there is no distinct demarcation be- 
tween these three phases of weather forecasting. 
Present State of Short-Range Weather Forecasting. 
Douglas [19] stated in 1931 that, because of the stag- 
gering complexity of the atmosphere, forecasting was 
largely a matter of experience and judgment and thus 
almost wholly on an empirical basis. Byers [11] in his 
chapter on forecasting techniques remarks: ‘The ability 
to forecast weather accurately comes as much from ex- 
perience as from study. The rules for prediction cannot 
be stated simply and it is extremely difficult to attain 
success in forecasting through formal study or instruc- 
tion.” In an earlier edition, he further said: ‘‘In general, 
it may be stated that... forecasting is the application 
of all the forecaster’s knowledge of meteorology, aug- 
mented by thermodynamic calculation to the problem 
at hand.”’ Willett [58] in a similar chapter and in like 
747 
