OBSERVATIONAL STUDIES OF GENERAL CIRCULATION PATTERNS 
gionally low-index circulation patterns at a variable 
speed which synoptic experience suggests may be of 
the order of 60° of longitude per week. At the surface 
in the area of blocking, pronounced anticyclogenesis 
(or ecyclolysis) is observed in higher latitudes, while 
wave cyclones are forced to skirt the perimeter of the 
developing high-latitude anticyclone, either northward 
west of the developing upper-level ridge at high lati- 
tudes, or southeastward into the cold developing vortex 
of southern latitudes. A good description of the evolu- 
tion of the day-to-day westerly waves in the mid-tropo- 
sphere westerlies [2] during blocking is that they appear 
to become progressively shorter and increase in ampli- 
tude as the scene of the blocking action is approached. 
In this respect they may be likened to waves advancing 
on to a beach, or to the pleats in an accordion. 
Westward progression of characteristic circulation 
features, while by no means rare, does not appear to 
be as common as eastward progression. Thus a sudden 
increase in amplitude of a wave system in North 
America may be shortly followed by a similar increase 
over the Atlantic and Europe. The speed at which this 
readjustment occurs often exceeds that of the zonal 
speed of the air particles—a fact which of late has re- 
ceived considerable attention from theoreticians [43, 
61] following its discovery in synoptic practice [36]. 
The transition in temperate latitudes from strong 
to weak zonal circulation and back again has been 
referred to as the index cycle. On the basis of a careful 
synoptic and statistical study of seven years of North- 
ern Hemisphere data, Willett [45] has given a descrip- 
tion of the index cycle which can hardly be improved 
upon. His statements follow: 
. . four principal states of the index cycle are recognized, 
each of which can be briefly characterized essentially as 
follows: 
(1) Initial high index (strong sea-level zonal westerlies), 
characterized by (a) sea-level westerlies strong and north of 
their normal position, long wavelength pattern aloft; (6) 
pressure systems oriented east-west, with strong cyclonic 
activity only im higher latitudes; (c) maximum latitudinal 
temperature gradient in the higher middle latitudes, little 
alr mass exchange; and (d) the circumpolar vortex and jet 
stream expanding and increasing in strength, but still north 
of the normal seasonal latitude. 
(2) Initial lowering of sea-level high-index pattern, charac- 
terized by (a) diminishing sea-level westerlies moving to lower 
latitudes, shortening wave-length pattern aloft; (6) appear- 
ance of cold continental polar anticyclones in high latitudes, 
strong and frequent cyclonic activity in middle latitudes; 
(c) maximum latitudinal temperature gradient becoming 
soncentrated in the lower middle latitudes, strong air mass 
exchange in the lower troposphere in middle latitudes; and 
(d) maximum strength of the cireumpolar vortex and jet 
stream reached near or south of the normal seasonal latitude. 
(3) Lowest sea-level index pattern, characterized by (a) 
complete breakup of the sea-level zonal westerlies in the low 
latitudes into closed cellular centers, with corresponding 
breakdown of the wave pattern aloft; (6) maximum dynamic 
anticyclogenesis of polar anticyclones and deep occlusion of 
stationary cyclones in middle latitudes, and north-south 
orientation of pressure cells and frontal systems; (c) maxi- 
mum east-west rather than north-south air mass and tempera- 
561 
ture contrasts; and (d) development of strong troughs and 
ridges in the circumpolar vortex and jet stream, with cutting 
off of warm highs in the higher latitudes and cold cyclones in 
the lower latitudes. 
(4) Initial increase of sea-level index pattern, character- 
ized by (a) a gradual increase of the sea-level zonal westerlies 
with an open wave pattern aloft in the higher latitudes; (6) 
a gradual dissipation of the low-latitude cyclones, and a 
merging of the higher-latitude anticyclones into the sub- 
tropical high-pressure belt; (c) a gradual cooling in the polar 
regions and heating of the cold air masses at low latitudes to 
re-establish a normal poleward temperature gradient in the 
higher latitudes; and (d) dissipation of the high-level cyclonic 
and anticyclonic cells, with a gradual re-establishment of the 
circumpolar vortex jet stream in the higher latitudes. 
While the index cycle described above appears to 
be recognized in its essential character during most 
winters, its precise form of operation, time of onset, 
and even its length and intensity are subject to ap- 
preciable variations. Extensive studies directed along 
these specific lines appear to be long overdue. A begin- 
ning on this problem [85] was made with the help of 
hemisphere-wide upper-air coverage obtained during 
the six-year period 1944-49. These data suggested the 
following conclusions: 
1. There is a strong preference for the winter’s pri- 
mary index cycle to occur in February and March, the 
minimum index occurring around the beginning of 
March, and the beginning and end points being re- 
moved by about three weeks from the minimum. 
2. The total momentum of the mid-troposphere west- 
erlies around the hemisphere tends to reach a certain 
value characteristic of the season, and it is only the 
distribution of this momentum with latitude that varies. 
Hence the low index of temperate latitudes really con- 
sists of a displacement of the “high-index” strong 
westerlies (farther southward). 
3. Hach primary index cycle is usually associated at 
its onset with a strong wave of Atlantic blocking, with 
a tendency to form warm anticyclones in high lati- 
tudes and cold cyclones in low latitudes. However, the 
blocking appears to be only a necessary and not a 
sufficient condition for a primary index cycle. 
4. The intensity of long index cycles appears to be 
largely determined by the reservoir of cold air in polar 
regions preceding their onset. 
These four statements appear to be sufficiently 
backed by empirical evidence to be considered in any 
theoretical treatment of the problem of the evolution 
of the radically different states of the general circula- 
tion. A qualitative treatment along these lines has been 
given [35]. The reader is referred to this paper since the 
presentation of theories lies outside the scope of this 
article. Other earlier attempts to explain the index 
cycle have been given by Rossby and Willett [45]. 
The index cycle, consuming about six weeks, is natu- 
rally not the only important period of evolution of 
radically different circulation patterns. Indeed for many 
years the problem of monthly mean, seasonal, and even 
secular large-scale circulation anomalies has been of 
paramount concern to meteorologists. While such evo- 
lutions undoubtedly lie in the province of ‘Observed 
