TROPICAL METEOROLOGY 
By C. E. PALMER 
Institute of Geophysics, University of California 
INTRODUCTION: THREE METHODS 
OF ANALYSIS 
In 1947, when the University of California established 
its Institute of Geophysics at Los Angeles, one of the 
research projects initiated with the Institute was an 
investigation of the synoptic meteorology of the tropical 
Pacific Ocean. Although the literature contained only 
scanty references to the topic, it was known that many 
meteorological data had been collected during World 
War II and that new ideas had been fermenting in the 
area throughout the period of conflict. In order to 
obtain some insight into the general condition of synop- 
tic meteorology in the Pacific as a preliminary to the 
investigation the Director of the Institute at that time 
(Dr. J. Kaplan) issued a questionnaire to meteorologists 
who were known to have been associated with Pacific 
observation and research during the war, and answers 
from about half of those addressed were finally received 
in the Institute. It would serve no good purpose to 
enter here into a detailed analysis of the replies; how- 
ever, one striking feature revealed by the survey is 
relevant, since it turns out to be characteristic of the 
present state of tropical meteorology in general. The 
replies fell into three well-marked classes, distinguished 
by the method of approach to tropical weather prob- 
lems displayed, in some cases consciously, in others 
unconsciously, by the authors. 
The meteorologists whose replies fell into the first 
class tended to put the greatest emphasis on climato- 
logical concepts: even in their approaches to synoptic 
problems they were dominated by the results of statisti- 
cal analysis, by mean maps, by statistical notions like 
the trades, the monsoons, the doldrums, by the conviction 
that the day-to-day weather in the tropical Pacific 
differs very little from that revealed by monthly and 
annual means. They extended this approach to the 
forecasting problem. Some felt, for example, that the 
best guide to forecasting the tracks of individual ty- 
phoons and tropical storms might lie in the study of 
mean storm tracks. Their dynamic theories, when they 
had any, were expressed in terms of a hypothetical 
general circulation of the tropics, and this was the 
entity whose characteristics were to be explained by 
physical reasoning from first principles, not the vagaries 
of the daily weather map. For the purposes we have in 
mind here, we may term the method of approach 
which is based upon these conceptions, the climato- 
logical method. 
In the second group we must place those meteorolo- 
gists whose thinking was much under the influence of 
concepts derived originally from the study of high- 
latitude weather, the followers of the air-mass method. 
The more extreme replies of this class maintained more 
or less explicitly that the tropical atmosphere differed 
from that in higher latitudes only in having a higher 
temperature and an easterly instead of a westerly 
mean wind direction. The dynamics, and by this they 
usually meant the frontal dynamics, remained the same. 
In the tropics there were fronts, air masses with source 
regions and modifications as in higher latitudes, an 
equatorial front similar in principle to the polar front, 
and above all an occlusion process that resulted in the 
formation of typhoons and tropical storms. Small tem- 
perature differences became important, and the “‘slope”’ 
of systems a matter of earnest discussion. 
Finally we have the third group of replies, sent in by 
workers originating in or influenced by the Institute of 
Tropical Meteorology at San Juan, Puerto Rico, a 
school set up by the University of Chicago durmg 
World War II but now, so far as I know, no longer 
active. This group takes its origm from Gordon E. 
Dunn’s paper on easterly waves [25] and its members 
have at some time been closely associated with the 
University of Chicago, drawing their theoretical 
stimulus from C.-G. Rossby. It is natural to term the 
method followed by this group the perturbation method. 
The concepts of the climatological group, referring to a 
hypothetical general circulation of trades, monsoons, 
doldrums, ete., identifiable as such on the mean maps, 
were taken over by these meteorologists. However, the 
basic currents were now considered to be for the most 
part zonal but subject to perturbations, and the per- 
turbations to be accompanied by characteristic weather 
and pressure patterns, identifiable on daily weather 
maps. This group was much preoccupied with models of 
various kinds of perturbations. In dynamics, one gets 
the impression that they considered all dynamical prob- 
lems connected with the perturbations as already solved 
by Rossby’s work and that their great task was to 
explain the dynamics of the general and zonal circula- 
tion. 
The significance of the three broad classes of Pacific 
meteorologists revealed by this survey lies in the fact 
that, once one knows what to look for, the same 
divisions can be discerned in the work of all tropical 
meteorologists. It is not that they may be divided into 
those whose interests are primarily in climatology, or 
in synoptic meteorology, or in dynamic meteorology. 
On the contrary, the division rests upon distinct meth- 
ods of approach to all branches of tropical meteorology, 
statistical, dynamic, and synoptic alike. It is therefore 
convenient to regard writings on tropical meteorological 
topics as emanating from three distinct schools of 
thought, the climatological, the air-mass, and the per- 
turbation schools, even though it is only in the last 
class that the opinions have been common to a locally 
concentrated group of workers. And, of course, it must 
be remembered that there will be many borderline cases, 
859 
