OBSERVATIONAL STUDIES OF GENERAL CIRCULATION PATTERNS 
masses in the northeastern oceans is strongly suggested 
by the work of Elliott and Smith [21]. 
In studying the sources and transformations of atmos- 
pheric moisture Jacobs showed that the greatest evapo- 
ration in winter takes place over western oceans. Ac- 
cording to him this latent heat appears to be released 
to the atmosphere largely in the eastern portions of the 
oceans where condensation reaches its maximum. 
Another method of investigating the availability of 
moisture for atmospheric processes resulted from the 
technique of isentropic analysis. By this means Namias 
and Wexler [55] were able to construct monthly mean 
isentropic charts for summer over the United States 
and thereby indicate the normal upper-level sources of 
moisture. Unfortunately, due more to the discontinu- 
ance of the isentropic technique than to the lack of 
upper-air data, such charts for other areas and seasons 
have not been constructed. 
The transport of various meteorological quantities 
by methods other than isentropic analysis has received 
considerable attention in recent years. Thus, Priestley 
[41] has suggested a systematic study of the transport 
of heat, moisture, and momentum through the use of 
radiosonde and radiowind measurements at a large 
x 
° 
PRESSURE 
DIFFERENCE 95 
IN 
MILLIBARS 
2913 27 10247 21 7 21 4 18 2 163013 27 Il 25 8 22 5 19 3 I73I 14 2812 26 9 23 6 206 203 I7 | 15 
JUN JUL AUG SEP OCT NOV lien FEB MAR APR eM 
1937 
WEEK NOV DECIJAN FEB MAR-APR MAY 
BEGINNING] jo36 
559 
the normal map computed from a long series of data 
achieves a certain character. 
The variations of the circulation about its normal, 
considering weekly or monthly mean charts, is far 
ereater than that which one would obtain if the cir- 
culation types followed each other in purely random 
sequence. This fact is reflected in the amazingly large 
variability of means for a month, year, or decade and 
probably for still longer periods up to the ice ages, as 
stressed by Willett [58] and Tannehill [49]. 
The observed sequence of types in any particular 
month or season more often than not takes the form of 
an alternation between two or more mid-tropospheric 
flow patterns, one of which persistently recurs and 
dominates the mean over longer periods. It is presum- 
ably this persistent recurrence which led the Mul- 
tanovsky School of long-range forecasting [39] to draw 
conclusions with regard to “natural synoptic periods,” 
to the opposition of types between natural periods, and 
to the relative constancy of length of the natural period 
during one “synoptic season.”® 
Some idea of the weekly variations of the strength 
of the zonal circulation at sea level may be obtained 
from Fig. 10. The absolute variations at higher eleva- 
35°N TO 55°N ZONAL CIRCULATION INDEX 
1938 
Fie. 10.—Weekly averages of zonal westerlies index, November 1936—May 1938. 
network of individual stations. The importance of the 
transport and convergence of momentum and heat in 
producing local wind accelerations has been stressed by 
Namias and Clapp [38] and by Starr [48]. More study 
can profitably be made of the significance of thermal 
convergence in producing local transformations of en- 
ergy. 
NONSEASONAL (IRREGULAR) VARIATIONS 
IN THE GENERAL CIRCULATION 
Qualitative Synoptic Studies. The ‘normal”’ state of 
the general circulation, as shown by normal hemisphere 
maps, is one which is never strictly encountered in any 
given daily map nor, for that matter, in any weekly 
mean or monthly mean. Indeed, the normal map is 
composed of the averages of circulations which appear 
remarkably different from week to week and even be- 
tween corresponding months of different years. In any 
given region, however, there appears to be a preferred 
circulation type for a given month or season, so that 
tions appear to increase up to the level of the jet stream, 
but the percentage variation of normal strength is 
probably quite similar at all levels. If the average zonal 
wind speed for the hemisphere is considered as a whole 
and for weekly periods, this variation is of the order of 
50-150 per cent of the normal speed. Regional and 
shorter-period variations may, of course, exceed these 
values, and localized jet streams reaching three times 
their normal value have been recorded. There is some 
suggestion in the distribution of zonal indices of mid- 
troposphere that the frequency curve of index values 
is skewed in a direction such that high values (relative 
to normal) are less frequent than low. Perhaps this is 
suggestive of an atmospheric braking mechanism 
whereby some form of stability permits very low zonal 
wind speeds to persist but kills off unusually fast 
currents. 
5. Consult “General Aspects of Extended-Range Fore- 
casting” by J. Namias, pp. 802-813 in this Compendium. 
