750 
front to accelerate from an almost stationary condition 
to 40-60 mph within 12-18 hr. 
Temperature-change charts have relatively little fore- 
cast value but are required for weather bulletins for 
press and radio, for preparation of shippers’ forecasts 
and other information for commercial and agricultural 
interests, and for the general public. 
Analogues. The system of locating a number of pre- 
vious weather charts which are most nearly analogous 
to the current chart, and then determining the prog- 
nostic chart by analogy, has been used successfully by 
a number of forecasters, particularly J. J. George. The 
past maps are known as “‘analogues.”’ Weather charts 
for several decades are catalogued on the basis of their 
dominant characteristics, for example, Colorado lows 
or Alberta highs. If, for example, on some particular 
December day those are the two principal character- 
istics of the current weather chart, the December, and 
also the January and November, file of past situa- 
tions is inspected and the five to ten charts with greatest 
similarity to the current chart are selected and studied. 
A choice of two or three is finally made, primarily on 
the basis of history and secondarily on the basis of de- 
tail. “Good” analogues should provide reasonably good 
prognostic charts since the numerous hydrodynamic 
and thermodynamic parameters that describe the cur- 
rent weather situation are conveniently integrated in 
the analogue. Where selection of the analogue is made 
mechanically (7.c., by machine) the system has been 
a complete failure, since historical sequence is much 
more important than actual location and intensities of 
pressure systems. Manual selection, which permits con- 
sideration of both surface and upper-air patterns, does 
provide better results but time and patience are re- 
quired. Because of the persistence of weather types, the 
best analogue may often be found from the charts 
within the three or four weeks previous to the current 
chart. The purpose of the analogue procedure is to 
provide the imexperienced forecaster with assistance, 
but, in practice, he does not have the experience re- 
quired for the selection of the best analogues and for 
the recognition of the significant differences which are 
always present between the analogue and the current 
chart. The use of analogues will never prove satisfac- 
tory until the upper-air patterns and historical se- 
quences are included in the parameters upon which 
their selection is based. 
Types. Attempts have been made to type weather 
maps and situations since the beginning of forecasting. 
Many of these attempts have met with limited or no 
success because of the infinite variations of weather 
situations and the superficiality of some methods of 
typing. 
Studies of storm types, frequencies, and normal storm 
tracks, such as those of Bowie and Weightman [7] in 
the United States and of Van Bebber [54] and Braak 
[8] in Europe, are widely used by forecasters. The San 
Francisco forecasters were moderately successful for 
many years in forecasting for the Pacific Coast by di- 
viding their weather into three main steering types: 
northwesterly, westerly, and southwesterly. This typing 
WEATHER FORECASTING 
was expanded and refined by the Meteorology Depart- 
ment at the California Institute of Technology under 
the leadership of I. P. Krick. 
The C. I. T. typing method is based on the location 
of the “center of action,” here the semipermanent 
eastern Pacific high. Krick, Elliott, and colleagues [12] 
have classified North American weather into five main 
types, with a considerable number of subtypes, de- 
pending upon the location of the Pacific high. It was 
found that a type tended to persist for six days (five 
to seven days) and then to repeat or change to another 
type. Composite charts, by seasons, were derived for 
each of the six days of each type. As indicated by 
Elliott [20], changes in index and in the various arrange- 
ments of the large upper-level waves (meridional flow 
patterns) and different degrees of expansion of the ring 
of strongest westerlies (zonal flow patterns) are regu- 
larly associated with the shift and change in intensity 
of the Pacific high. While typing of this kind is most 
helpful in extended forecasting, it has also been found 
of value i 24-72-hr forecasting as far east as central 
United States. It is understood that further investiga- 
tions are under way to develop composite upper-air 
maps analogous to the composite surface charts and to 
determine the limits of variation of upper-air situations 
with individual surface types. The principal limitations 
of the type technique are (1) the frequent and extensive 
variations of a single type, and (2) the difficulty in 
recognizing some types in complex situations. Typing 
on a somewhat more circumscribed scale, such as that 
of Saucier [49] in connection with Texas-West Gulf cy- 
clones, can provide the district forecaster with a valu- 
able tool. 
USE OF UPPER-AIR DATA 
Radiosonde Diagrams. For a number of reasons, in- 
cluding the requirements for the preparation of certain 
specialized and detailed forecasts for very short-term 
periods as well as thorough analysis of local weather 
conditions, it is desirable and necessary to make a de- 
tailed analysis of individual radiosonde observations 
within and adjacent to the forecast district. The anal- 
ysis enables the forecaster to locate cloud layers and 
their thicknesses as well as areas of freezing rain and 
snow (when present) and also to ascertain stability con- 
ditions in order to forecast the probability of cloud, 
cloud heights, and convectional shower activity. For 
convenience, the forecaster may wish to classify the 
lower troposphere according to the customary degrees 
of stability, namely, stable, conditionally unstable, and 
absolutely unstable. During certain periods and seasons, 
the forecaster will wish, as a matter of course, to de- 
termine the lifting condensation level, the convection 
condensation level, and the energy available. For a full 
discussion of stability and instability see any standard 
textbook, particularly. Petterssen [42, Chap. II]. 
Two methods exist for determining the stability of the 
atmosphere: the parcel and the slice methods. The 
parcel method is less exact in that as a parcel of air 
rises, other air is entrained from the outside and when 
the air is conditionally unstable (when the exact degree 
ae 
