




































 History of Meteorology 7 (2015) 1 

 
 

 

 

 

 

 

 

 

 

 

Weather and Climate as Shape-Shifting Nouns: 

Gordian Knots of Understanding and Prevision 
 

James Rodger Fleming 

jfleming@colby.edu 

Colby College 

 

 

“[Technology] is cutting the Gordian Knot instead of untying it”  

— Carl-Gustaf Rossby
1
 

 

 “Why should nearly perfect forecasts be unattainable?” — Chaos” 

— Edward N. Lorenz
2
 

 

“Climate” deserves to be a keyword in the vocabulary of culture and society. It is 

arguably one of the most linguistically complicated words — a historically shape-shifting 

noun — whose meaning has changed and is changing, perhaps faster than the climate 

itself. Its nature, history, and vicissitudes are key concepts organizing our ideas of the 

aerial environment and our relationship to it.
3
 Although climate can be depicted and 

modeled, it cannot be directly visualized or forecast. The relationships of climate to 

meteorology and meteorology to climate studies are also dynamic. This article reviews 

the changing nature of ideas about climate over an extended time period, with special 

focus on developments in dynamic meteorology and dynamic climatology in the period 

1900-1960, and with an update for the twenty-first century based on the implications of 

chaos theory. It adds a temporal dimension to the science and philosophy of weather 

prediction and climate change and the ways we think about the interrelationships of 

weather and climate. 

It may be a truism for historians that anything that can be named had different 

meanings in different eras, but some philosophers, with notable exceptions, have 

overlooked this.
4
 When doing epistemology, it is important to think about the history of 

                                                             
1
 C.-G. Rossby, “Current Problems in Meteorology,” The Atmosphere and the Sea in Motion: Scientific 

Contributions to the Rossby Memorial Volume, Bert Bolin, ed. (Stockholm: Rockefeller Institute Press and 

Oxford University Press, 1959), 9-50, quoted on 36. 
2
 Undated lecture transparency in the Papers of Edward N. Lorenz, U.S Library of Congress, Manuscript 

Division. 
3
 J.R. Fleming, “Climate, Change, History,” Environment and History 20 (2014): 577-586. 

4
 Nietzsche and Whitehead are notable exceptions; see “epistemology” in the Internet Encyclopedia of 

Philosophy, http://www.iep.utm.edu/epistemo/ 

mailto:jfleming@colby.edu


 History of Meteorology 7 (2015) 2 

 
 

 

the objects under discussion and how the meaning of terms has changed. As Mike Hulme 

has written recently, “Like all powerful ideas, climate change can be deceptively simple 

to define and yet subject to a multiplicity of cultural meanings and technical 

interpretations.” It is important to clarify the historical context and to think about what 

we are thinking about when we use terms such as climate, climate change, meteorology, 

dynamics, and models.
5
 

 

Apprehension and Authority 

 

In Historical Perspectives on Climate Change (1998), I examined climate and climatic 

changes from the Enlightenment to the late-twentieth century, asking the following 

questions, among others: “How do people (scientists included) gain awareness and 

understanding of phenomena that cover the entire globe, and that are constantly changing 

on time scales ranging from geological eras and centuries to decades, years, and 

seasons?”  “How was this accomplished by individuals immersed in and surrounded by 

the phenomena?” and, “How were privileged positions created and defined?”
6
 Since 

atmospheric scientists lack the ability to observe the climate system in its entirety (as an 

astronomer might view a star or planet) or to experiment on it directly (as a chemist 

might view a reaction), the pathways to the many and varied scientific understandings of 

climate are not at all straightforward; vexed too are the interrelations between elite and 

popular perceptions of the environment. 

There are three major definitions of the term apprehension that can be applied to 

climate: (1) awareness and understanding, (2) anticipation and dread, (3) intervention and 

control. Climate is an elusive entity. It is more than the average condition of the 

atmosphere compiled from weather statistics; it is something much more fundamental 

than that: it is lived experience providing frameworks for the material possibilities of life, 

inseparable from the temporality and specificity of the social world. Collectively, climate 

shapes life in specific places in fundamental and dramatic ways; it is woven into the 

fabric of the human past and future; both the experience and knowledge of climate is 

shaped by and enframed in our lifestyles, our seasonality, our clothing and housing, our 

aspirations and our tragedies. Climate involves themes of dominant empires, colonial 

pride and pioneer mentalities; popular, religious and racial sentiments; imposed and 

imported ideologies; and resilience to extreme conditions of flood and drought as 

embodied in agricultural and technological practices. Climate is both intimate and 

universal, involving local experiences of global change.
7
 Its influences are as proximate, 

                                                             
5
 Mike Hulme, “Climate Change (concept of),” International Encyclopedia of Geography  editor-in-chief 

Douglas Richardson (Hoboken, NJ: Wiley-Blackwell/Association of American Geographers, 2015). J.R. 

Fleming, “Historical Perspectives on Shape-shifting Nouns: Climate, dynamics, models.” Invited talk for 

Models and Knowledge in Climate Sciences: Historical, Scientific and Philosophical Perspectives, Rotman 

Institute of Philosophy Annual Conference, London, Ontario, Canada, Oct. 2014, 

http://www.rotman.uwo.ca/new-video-posted-james-fleming-dynamic-climatology-historical-perspectives-

on-shape-shifting-nouns 
6
 J.R. Fleming, Historical Perspectives on Climate Change (New York: Oxford University Press, 1998, 

2005). 
7
 Intimate Universality: Local and Global Themes in the History of Weather and Climate, James Rodger 

Fleming, Vladimir Jankovic, and Deborah R. Coen, eds. (Sagamore Beach, MA: Science History 

http://www.rotman.uwo.ca/new-video-posted-james-fleming-dynamic-climatology-historical-perspectives-on-shape-shifting-nouns
http://www.rotman.uwo.ca/new-video-posted-james-fleming-dynamic-climatology-historical-perspectives-on-shape-shifting-nouns


 History of Meteorology 7 (2015) 3 

 
 

 

and personal, as a breath of fresh air, a drink of local water, or an attachment to place. Its 

vicissitudes are increasingly associated with collectively held apprehensions (fears really) 

of the destructive reach of global civilization on the planet.
8
 

How then were privileged positions on such an elusive entity established? One 

approach, popular in the eighteenth century, was through appeals to authority—

references to culture, to historical literature, to first impressions of explorers, or to the 

memory of the elderly [See Fenby, this volume].  This was the rhetorical strategy of 

Enlightenment and early-American writers who wanted to support a particular theory of 

cultural development or decline. In Réflexions critiques sur la poësie et sur la peinture 

(1719), Abbé Du Bos examined the causes of the rise and fall of the so-called illustrious 

ages in the arts and sciences, attributing them primarily to changes in les causes 

physiques (the nature of the air, land, soil and especially climate). Many Enlightenment 

era authors followed this lead, including the Baron Montesquieu and Thomas Jefferson, 

who attributed climatic change in Europe and America to human settlement, including 

forest clearing, marsh drainage and cultivation.
9
 

Also in vogue was the practice of collecting massive amounts of meteorological 

data over large areas and extended time periods in the hope of deducing climatic patterns 

and changes [See Munger, this volume].  Individual observers in particular locales 

dutifully tended to their journals, and networks of cooperative observers gradually 

extended the meteorological frontiers.  Beginning in earnest in the nineteenth century, 

scientists tabulated, charted, mapped, and analyzed observations to provide climatic 

inscriptions. This process profoundly changed climate discourse and established the 

foundations of the science of climatology. Climates were seen as stable, to be described 

by geographers and studied by statisticians. The noted Austrian climatologist Julius Hann 

wrote in 1897 that climate is “the sum total of the meteorological phenomena that 

characterize the average condition of the atmosphere at any one place on the earth’s 

surface.”
10

 This became the dominant working definition, still invoked today. 

A third approach to privileged knowledge was to establish from first principles 

what the climate ought to be and how it ought to change. Approaches drawn from 

physical, mathematical, geological, and astronomical evidence and principles tended to 

be most satisfying to those scientists working within a particular disciplinary perspective.  

Most scientists had one favorite causal mechanism and only grudgingly admitted other 

possible secondary causes of climate change [See articles by Ellingsen and Lehmann, this 

volume].  In the mid-twentieth century Tor Bergeron, an acolyte of the Bergen school of 

meteorology, defined a “dynamic climatology,” which supplemented statistical means 

with detailed explanations of air masses, fronts and other thermodynamic factors 

influencing the climate. According to Bergeron, climatology had been essentially the 

systematic compiling of statistics on the individual meteorological elements, without an 

organized attempt to get at the underlying dynamic or thermodynamic phenomena in their 

entirety. Since descriptive climatology offered no unifying picture of the prime 

                                                                                                                                                                                      
Publications/USA, 2006); Osiris 26, Klima.  James Rodger Fleming and Vladimir Jankovic, eds. (Chicago: 

University of Chicago Press, 2011). 
8
 Naomi Oreskes and Erik M. Conway, The Collapse of Western Civilization: A View from the Future (New 

York: Columbia University Press, 2014). 
9
 Fleming, Historical Perspectives. 

10
 Julius Hann, Handbook of Climatology, trans. R. deC. Ward (New York: 1903), 1. 



 History of Meteorology 7 (2015) 4 

 
 

 

thermodynamic forces controlling the climate, Bergeron applied concepts of air masses 

and fronts to outline a comprehensive dynamic climatology explaining how the several 

complete thermodynamic units, rather than the unrelated distribution of the individual 

meteorological elements, control the climate of a region. An older, static climatology may 

say how much it rained in a given year; dynamic climatology would say when, where, 

and why it rained.
11

 

Another approach to privileged climate knowledge has been through technology.  

In part with the invention and standardization of meteorological instruments and the 

networking of meteorological observers, but more recently in the explosive growth of 

transformative technologies of communication and aerospace in the twentieth century, 

weather and climate observations have been rendered both global and three dimensional. 

A number of transformative technologies — including radar, nuclear tracers, digital 

computers, sounding rockets, and weather satellites — fostered the emergence of the 

interdisciplinary atmospheric sciences after World War II. Scientists working with 

computers in meteorology at the time demeaned the older, descriptive and statistical 

forms of geography as hardly constituting a physical science. Today many scientists 

working at the interface of remote sensing and modeling are hoping, through advances in 

technology, to provide new privileged positions.  For most scientists the goal is better 

understanding of climate; for some it is also prediction and control. 

Lest we overlook the applied dimensions, in the 1960s, climatologist Helmut 

Landsberg described what he called “techno-climatology” as comprising the influences of 

climate on the infrastructure of modern life, both in the quotidian sense, but also in the 

applicability of climatic information to engineering and city planning and the effects of 

climate on commerce and industry [See articles by Janković and Vollset, this volume].  

.
12

 By including industry, Landsberg raised the possibility of combining the history of 

natural and artificial environments, through case studies of climate, climate modification, 

and climate control as it affects storage, warehousing and the efficiency, health, and 

safety of workers. 

Climate is bigger than any single journal, book or even disciplinary approach. Its 

history draws in part on the histories of particular sciences such as astronomy, chemistry, 

computer science, geography, geology, meteorology, paleontology, and physics; and in 

part from much broader historical currents. Historians with particular disciplinary 

interests have examined all of these fields before, at least to some extent. With global 

climatic change as the new focus, however, a new interdisciplinary picture begins to 

emerge that includes both elite and popular apprehensions. Coming full circle, modelers 

increasingly are appealing to the authority and consensus being generated by the 

Intergovernmental Panel on Climate Change (IPCC). In reality, however, climatologists 

sense that they have lost control of a subject they never fully mastered: the grand global 

narrative of scientific rationality wedded to human behavior, technological solutions, and 

policy innovations. New voices, untrained in the nuances of the atmospheric sciences, 

from the press, the public, the state, the environmental movement, have flooded the 

literature, adding polarizing voices, while venerable but vulnerable practices of peer 

review and journal publication have taken back seats to the new electronically facilitated 

                                                             
11

 Tor Bergeron, "Richtlinien einer dynamischen Klimatologie,” Meteorologische Zeitschrift 47 (1930): 

246-262. 
12

 Helmut Landsberg, Physical Climatology, 2
nd

 ed. (Du Bois, Penna.: Gray, 1964), 389. 



 History of Meteorology 7 (2015) 5 

 
 

 

‘peer-to-peer’ review and ubiquitous blogs, tweets and quacks. Just as the world turns to 

crucial questions of what is to be done, it appears certain that the climatologists have lost 

voice, ownership, authority, status – and in many cases their innocence, patience and 

even their tempers, while others, ever more strident, speak for them, ever more shrilly. 

Skeptics (a venerable position in philosophy and science) hold that we do not know as 

much as we think we do about climate, especially regarding the sensitivity of the climate 

to carbon dioxide, while so-called “deniers” (a form of name-calling), maintain that we 

should do nothing about climate change. Climate interventionists, on the other hand, 

propose to control weather systems and ‘fix’ the climate with heavy-handed technologies: 

seed the clouds to create or dissipate rain on demand, launch a fleet of space mirrors, 

pump sulfates into the stratosphere to cool the planet, capture the world’s carbon 

emissions and sequester them safely and economically for thousands of years.  As the 

climate changes ever so slightly (perhaps more next decade according to the IPCC?), 

economies stumble and mitigation plans move ahead haltingly, if at all. The current 

consensus on global climate change calls into question the trajectory of modern 

civilization: Will it empower or perhaps overthrow the status quo?  Or are climate and 

climate change shape-shifting nouns? Such ideas might sound like science fiction, but in 

fact they are part of a very old story rooted in human apprehensions and aspirations to 

control nature.
13

 

 

Models 

 

Models too are shape-shifting nouns.  Consider the ancient explanation of the solidness of 

a material proposed by the Greek philosopher Democritus: “The solidness of a material 

corresponds to the shape of the atoms involved. Water atoms are smooth and slippery; 

salt atoms are sharp and pointed; air atoms are light and whirling.” According to the 

epistemology of shape-shifting nouns, “We, in actuality, grasp nothing for certain.”
14

 

A digital computer is, in reality, just a really big and fast calculating machine, not 

a “model” itself, but a useful tool. In the late 1940s C.-G. Rossby, Jule Charney and their 

associates articulated the simplified differential equations of upper atmospheric flow and 

expressed them as finite difference equations suitable for machine calculation. This 

involved replacing the time and space derivatives with differences of finite magnitude, 

for example, one-hour time increments and 300-kilometer horizontal grid spacing. The 

meteorological problem then consisted of programing the computer to compute the 

change of the state of the atmosphere, hour after hour, in the system so defined. With a 

computer fast enough to complete these calculations in less than one hour, the expected 

changes can be added to the originally observed values to generate, via an iterative or 

“bootstrapping method,” a 12- or 24-hour or longer prognosis. This was in fact the 

method used in routine numerical weather prediction services, available, at least for the 

upper atmosphere, since 1954 in the United States and, to a limited extent, in Sweden. 

The next logical and tractable step at the time consisted in integrating the atmospheric 

equations for a dynamic, moist, heated atmosphere forever — the so-called “infinite 

                                                             
13

 J.R. Fleming, Fixing the Sky: The checkered history of weather and climate control (New York: 

Columbia University Press, 2010).  

14 Jerry Coffey, “Democritus Atom,” Universe Today, http://www.universetoday.com/60058/democritus-

atom 

http://www.universetoday.com/60058/democritus-atom
http://www.universetoday.com/60058/democritus-atom


 History of Meteorology 7 (2015) 6 

 
 

 

forecast” — proposed by John von Neumann, Jule Charney, and Harry Wexler — and 

first attempted by Norman Phillips. The result was not a forecast or a prediction of 

particular weather conditions, but a limiting case converging on the statistical features of 

the general circulation, independently of whatever initial conditions may have existed. 

This is not a forecast in the usual sense, but a model test. Modern climate models, with 

secularly changing boundary conditions, are distant (very distant) cousins of this process. 

Some may wonder how weather and climate are interrelated rather than distinct. 

Both, for example, are at the center of the debate over greenhouse warming and hurricane 

intensity. A few may claim that rainmaking, for example, has nothing to do with climate 

engineering, but any intervention in the Earth’s radiation or heat budget (such as 

managing solar radiation) would affect the general circulation and thus the location of 

upper-level patterns, including the jet stream and storm tracks. Thus the weather itself 

would be changed by such manipulation. Conversely, intervening in severe storms by 

changing their intensity or their tracks, or modifying weather on a scale as large as a 

region, a continent, or the Pacific basin would obviously affect cloudiness, temperature, 

and precipitation patterns, with major consequences for monsoonal flows and ultimately 

the general circulation. If repeated systematically, such interventions would influence the 

overall heat budget and the climate.  In such a perturbation analysis, weather and climate 

are distinctly interrelated. 

 

The Gordian Knot 

 

Atmospheric researchers have long attempted to untie the Gordian Knot of meteorology 

— that intractable and intertwined tangle of observational imprecision, theoretical 

uncertainties, and non-linear influences — that, if unravelled, would provide perfect 

prevision of the weather for ten days, of seasonal conditions for the next year, and of 

climatic conditions for a decade, a century, a millennium, or longer. In the first six 

decades of the twentieth century, from the dawn of applied fluid dynamics to the 

emergence of the interdisciplinary atmospheric sciences, three interconnected generations 

of scientists aimed to extend and improve atmospheric measurements and to predict the 

future state of the atmosphere. In the first decade of the twentieth century, researchers 

pinned their hopes on wireless telegraphy and the dawning of the aerial age. Several 

decades later, radio, aviation, rockets, digital computing, and Earth-orbiting satellites had 

opened up entirely new research horizons. Each generation of atmospheric researchers 

aspired to a more global meteorology, zealously incorporating capabilities provided by 

new technologies into their science as they worked to link theory with practice. Each 

generation experienced, in their own ways, the heady feeling that they were the direct 

beneficiaries of new technological breakthroughs and that they stood on the brink of a 

major revolution in the science and practice of meteorology. Their goal was to produce 

accurate information about the state of the entire atmosphere, complete mathematical 

portrayals of its varied and changing states, and useful and timely forecasts of its near- 

and long-term future. Their soaring aspirations faced a multiplicity of crushing practical 

limitations exacerbated by war, bureaucracy, economic downturns, prejudice, and 

technological limitations.
15

  

                                                             
15

 J.R. Fleming, Inventing Atmospheric Science (Cambridge, Mass.: MIT Press, 2016). 



 History of Meteorology 7 (2015) 7 

 
 

 

In 1895 Cleveland Abbe, éminence grise of American meteorologists and editor 

and translator of several volumes of geophysical papers, pointed out that meteorology 

needed a deductive treatise on the laws governing the atmosphere as complete and 

rigorous as the Celestial Mechanics of Laplace: 

 
Meteorologists can never be satisfied until they have a deeper insight into the mechanics 

of the atmosphere. Something more is needed than the most perfect organization for 

observing, reporting and publishing the latest news from the atmosphere. It is not enough 

to know what the conditions have been and are, but we must know what they will be, and 

why so. We must have a deductive treatise on the laws governing the atmosphere as 

complete and rigorous as the “Celestial Mechanics”º of LaPlace, and this will necessarily 

be a treatise on the application to the atmosphere of the general laws of force, or what is 

technically known as the dynamics and thermo-dynamics of gases and vapors.
16

 

 

Note that Abbe did not differentiate between weather and climate, but emphasized the 

laws governing the entire atmosphere. 

Less than a decade later, Vilhelm Bjerknes (1862-1951) initiated a neo-Laplacian 

program to measure current atmospheric conditions and to calculate the future state of the 

weather using the equations of hydrodynamics and thermodynamics. He said this in 1902 

in the Meteorologische Zeitschrift: “Each task of theoretical mechanics is, when it is 

placed in direct form, a prognostic, just like the most well-known task of practical 

meteorology. The goal is to predict the dynamic and physical condition of the atmosphere 

at a later time, if at an earlier given time, this condition is well known.”
17

 According to 

Bjerknes, the central problem of the science of meteorology is weather prediction by 

rational dynamical-physical methods. He wanted to place meteorology on solid 

observational and theoretical foundations. In 1904, he published “Weather forecasting as 

a problem in mechanics and physics.”
18

 Here he stated the problem of weather 

forecasting as an initial-value problem in mathematics involving the ideal gas law, the 

first law of thermodynamics, the conservation of mass, and the dynamical equations of an 

ideal compressible gas. He wrote that the necessary and sufficient conditions for a 

rational solution of the problem of meteorological prediction include a sufficiently 

accurate knowledge of the state of the atmosphere at a certain time and a sufficiently 

accurate knowledge of the laws according to which one state of the atmosphere develops 

from another. These statements bring together massive programs in observation, theory, 

and forecasting and are really philosophical statements of faith in rational mechanics, 

raising the fundamental question of how far we can see into the future. Starting from a 

detailed, if not perfect, set of atmospheric measurements, it should be possible to take a 

finite, if not perfect, step forward using the time-dependent equations of atmospheric 

motion. This was the first step, for Bjerknes, in unraveling the Gordian Knot of 

meteorology. Bjerknes’s use of the word sufficiently in these statements tempered his 

determinism. The term undoubtedly derives from his personal experience in trying to 

measure, with any precision, the initial state of the atmosphere over an extended area and 

                                                             
16

 Cleveland Abbe, “The Needs of Meteorology,” Science n.s. 1, no. 7 (1895), 181-182.    
17

 Vilhelm Bjerknes, “Cirkulation relativ zu der Erde,” Meteorologische Zeitschrift 19 (1902): 97-108, on 

108. 
18

 Vilhelm Bjerknes. “Das Problem der Wettervorhersage, betrachtet vom Standpunkte der Mechanik und 

der Physik,” Meteorologische Zeitschrift 21 (1904): 1-7. 



 History of Meteorology 7 (2015) 8 

 
 

 

from his exposure to the mathematical lectures of Henri Poincaré in Paris. At the time, 

Poincaré (1854-1912) was writing his three-volume Les méthodes nouvelles de la 

mécanique celeste in which he employed analogies with fluid flow to characterize all 

motions of mechanical systems. In his essay Science et méthod, Poincaré raised the issue 

of a computation’s sensitivity to initial conditions and undermined the notion of a 

perfectly precise observation of nature, noting that, “small differences in the initial 

conditions produce very great ones in the final phenomena.”  This is now recognized as 

one of the canons of chaos theory.
19

 

Bjerknes insisted on the necessity of taking measurements in the centimetre-gram-

second (c.g.s.) system and worked to rationalize the analysis of observations. He accepted 

a call to Leipzig in 1913 to establish a geophysical institute.  This gave him access to a 

dense network of upper-air observations. The tragedy and hardships of World War I 

made conditions insufferable in Germany, and in 1917 Bjerknes moved to Bergen, 

Norway with his family where his group established a network of stations sufficiently 

dense to identify cold fronts and warm fronts as components of an ideal wave cyclone 

model.  Because exact solutions to his system of atmospheric equations were impossible, 

Bjerknes promoted the use of graphical methods including streamline analysis. He 

founded the Bergen school of meteorology in 1919, where graphical methods prevailed. 

Working in concert with his many associates, and with regular funding from the Carnegie 

Institution, the Bergen school delineated the properties of air masses and described the 

life history of cyclones developing along the polar front. Bjerknes once remarked, 

"During 50 years meteorologists all over the world had looked at weather maps without 

discovering their most important features. I only gave the right kind of maps to the right 

young men, and they soon discovered the wrinkles in the face of Weather."
20

 If we 

include the dynamic climatology of Tor Bergeron, we might also add, “wrinkles in the 

face of climate.” Collectively, Bjerknes and his associates championed something larger 

than themselves, something useful and entirely new, a set of graphical methods to analyze 

and predict short-term changes in the weather. They worked diligently to spread their 

theoretical insights internationally, to collect and interpret more and better observations at 

higher and higher altitudes, and to codify graphical techniques of air mass analysis and 

weather forecasting for the use of national weather services. 

 

Further Unraveling the Knot 

 

Weather and climate were intimately connected in the work of Carl-Gustaf Rossby 

(1898-1957), who brought the Bergen methods to the United States, developed new 

methods of collecting atmospheric measurements aloft, and connected geophysical fluid 

dynamics with the practical needs of aviation and forecasting. Rossby established the first 

commercial airlines weather forecasting system and built the first graduate program in 

meteorology at M.I.T. in 1928 along with a connection to dynamic oceanographers at 

                                                             
19

 Henri Poincaré, Science et méthod (Paris: E. Flammarion, 1906). 
20

 Robert Marc Friedman, Appropriating the Weather: Vilhelm Bjerknes and the construction of a modern 

meteorology (Ithaca: Cornell University Press, 1989); Tor Bergeron, Tor, Olaf Devik, and Carl Ludvig 

Godske, “Vilhelm Bjerknes, March 14, 1862 - April 9, 1951,” In Memory of Vilhelm Bjerknes on the 100
th

 

Anniversary of His Birth, A. Eliassen and E. Hoiland, eds., Geofysiske publikasjoner 24 (1962): 7-25, quote 

from 18. 



 History of Meteorology 7 (2015) 9 

 
 

 

Woods Hole in 1930. He supervised the training of literally thousands of aviation cadets 

during World War II, developed a general theory of long waves and jet streams in the 

upper atmosphere, and established the University of Chicago school of geophysics and 

the Institute of Tropical Meteorology at the University of Puerto Rico. After the war, he 

placed the American Meteorological Society on a professional keel by founding the 

Journal of Meteorology and, working closely with Jule Charney, stimulated work on 

numerical weather prediction at the Institute for Advanced Study. In the final decade of 

his life, Rossby returned to his native Sweden, establishing the International 

Meteorological Institute in Stockholm where he worked on global environmental issues. 

In his signature contribution, he identified upper-air planetary (or Rossby) waves as 

potential keys to long-range prediction, treating them as idealized cases suitable for 

computation by digital computers.  

One of Rossby’s first projects in America in the 1920s was the construction of a 

rotating tank experiment in the basement of the U.S. Weather Bureau, a model of 

hemispheric circulation scaled to emulate the large-scale dynamics of the atmosphere. In 

this he was inspired by Helmholtz's hydrodynamical theory of vortex motions, Felix 

Exner’s dishpan experiments, and the “principle of similar movements.”
21

 Rossby’s tank 

was two meters in diameter, filled with colored fluids of different densities, and rotated 

three to four times per minute. Like the atmosphere, the rotating tank had large horizontal 

dimension compared to its vertical depth. To a first approximation, it was two-

dimensional. This was the key to his theoretical approach. Although the tank suffered 

mechanical failure, and his initial attempt to emulate the atmosphere and write non-

dimensional equations of its motion was inconclusive, this line of research eventually 

proved fruitful in the late 1940s when Dave Fultz (1921-2002) at the University of 

Chicago employed a two-layer wave tank to produce experimental results that informed 

numerical analysis of the large-scale circulations that occur in the Earth’s atmosphere and 

oceans. 

For Rossby, the research frontier was theoretical, and it circulated at the 500-

millibar level (an altitude of about 5.6 kilometers). The upper-level winds and planetary 

waves looked and behaved very much like the colored fluids in the wave tank 

experiments. They had the distinct advantage that, unlike surface weather conditions, the 

flow at high altitudes was, to a first approximation, frictionless and devoid of the 

complicating factors of water vapor, clouds, and diurnal heating.  That is to say, the 

equations of motion were rather simple to write and were well- behaved over longer 

periods.  This provided hope for longer-range forecasts, and, when digital computers first 

became available, for tractable calculations that could be accomplished faster than the 

weather was changing; that is to say, for operational numerical weather prediction. 

In his later years, Rossby actively pursued numerical weather prediction in 

Sweden in an era in which there was no Swedish word for digital computer. There he 

supported the first operational numerical forecast in the world generated by the BESK 

computer. He was fully engaged at the cutting edge of research, maintaining a large 

correspondence network of colleagues and welcoming visitors to the International 

                                                             
21

 Felix Exner, Dynamische Meteorologie, 2nd ed. (Vienna: Julius Springer, 1925), 341; C.-G. Rossby, “On 

the Solution of Problems of Atmospheric Motion by Means of Model Experiments,” Monthly Weather 

Review 54 (1926): 237–240. 



 History of Meteorology 7 (2015) 10 

 
 

 

Meteorological Institute in Stockholm.
22

 Rossby’s work took a global, environmental turn 

in the mid-1950s as he turned his attention to climate. He fostered new conversations 

among geoscientists of all stripes: oceanographers, geographers, and geologists. Moving 

onto issues of geophysics and global pollution, he founded the international 

environmental journal Tellus.  He was interested in climatic change and variability on all 

time scales, including the grand cycles of ice ages and interglacials. He spoke 

increasingly of the atmosphere as a milieu, directly influencing all of human experience 

and warned of the increasing stress pollution was placing on it. He stood in awe of the 

atmosphere’s dynamical and chemical complexity and called for an attitude of respect, 

“for the planet on which we live."  He eagerly anticipated the coming breakthrough, “a 

grand era in meteorology,” when artificial satellites can view the atmosphere from above. 

"Right now, we are like crabs on the ocean floor, he said. “What we need is a view from a 

satellite. Only from a satellite can we see the planetary waves.”
23

  His best student, Harry 

Wexler, was, just then, in the process of developing such capacity. 

 

Using Technology to Cut the Gordian Knot 

 

Harry Wexler (1911-1962) introduced a number of transformative technologies into 

meteorological practice, including radar, nuclear tracers, digital computers, sounding 

rockets, and weather satellites, that helped cut into, if not through, the Gordian Knot of 

prevision. This was the legendary “Alexandrian solution,” alluded to by Rossby in 1956, 

when he wrote, “Technology is cutting the Gordian Knot instead of untying it.”
24

 Still, 

theorists held out hope for more and more accurate measurements and a more perfect 

model, perhaps driven by large and fast computers. On this foundation, Wexler and his 

many colleagues, notably Jule Charney, prepared the foundations for the emergence of 

the interdisciplinary atmospheric sciences.  

As an employee of the U.S. Weather Bureau, Wexler worked to bring air mass 

analysis to the United States, and, as an army air force officer, worked with Rossby to 

train a generation of weather cadets during World War II. After the war, Wexler became 

head of research in the weather bureau where he nurtured every new technological 

development and every new international program of relevance to the atmosphere.  In this 

capacity, he worked to incorporate radar, sounding rockets, nuclear tracers, digital 

computers, and satellites into meteorological practice. He demonstrated scientific 

leadership as a member of (and often chair of) many national and international 

meteorological advisory committees, including the U.S. military’s Research and 

Development Board, the Advisory Committee on Reactor Safeguards for the U.S. Atomic 

Energy Committee, and the National Research Council’s Space Science Board. Wexler’s 

vision was global, as evidenced by his leadership during the International Geophysical 

Year of 1957-58.  He conducted experiments and coordinated measurements on the 

Antarctic ice sheet and established the weather bureau's Mauna Loa Observatory, which 

measured the concentrations of carbon dioxide and other trace gases. In 1962, the year of 

his untimely death, Wexler was on assignment from the Kennedy Administration, 

                                                             
22

 Bert Bolin, “Carl-Gustaf Rossby: The Stockholm Period, 1947–1957,” Tellus A–B 51 (1999): 4–12. 
23

 “Man's Milieu,” Time 68 (Dec. 17, 1956), 68-79. 
24

 C.-G. Rossby, “Current Problems in Meteorology.” 



 History of Meteorology 7 (2015) 11 

 
 

 

negotiating the World Weather Watch, a lasting international agreement on the peaceful 

exchange of weather information.
25

 

Wexler’s position allowed him access to the full range of atmospheric 

technologies. Radioactive clouds generated by outdoor nuclear testing introduced new 

tools that cut across the vast spectrum of atmospheric processes and, as Wexler put it, 

provided meteorologists with a “matchless tracing agent” of horizontal and vertical flow 

and diffusion applicable at all scales and all levels, from the surface layer to the 

stratosphere.
26

  At higher levels, the radioactive clouds sketched out the patterns of 

Rossby waves.  Wexler also used the new technology of weather radar to determine the 

structure of a hurricane that passed over Florida in 1945.  He contrasted traditional 

observational and synoptic mapping techniques with time-lapse images and tracking 

provided by high-resolution, long-range military radar.  The images revealed hitherto 

unseen details of squall lines and hurricane bands and stimulated novel explanations of 

their dynamics. Wexler pointed out that the radar screen was now providing a new 

“eyepiece” thorough which to view severe weather and was opening up a new research 

field involving radar in meteorology.
27

 Wexler also served as weather bureau liaison for 

the Institute for Advanced Study computer project in Princeton, New Jersey, a project led 

by John von Neumann to investigate the theory of dynamic meteorology in order to make 

it accessible to high speed, electronic, digital, automatic computing. Subsequently, 

Wexler was in charge of institutionalizing and operationalizing numerical weather 

prediction and general circulation modelling within the U.S. Weather Bureau. Sounding 

rockets came under Wexler’s purview as scientific probes and observational platforms to 

investigate the upper atmosphere and photograph clouds from above. Wexler served as 

chair of several influential committees on this subject, and his extensive records from 

meetings show the evolution of issues, including the official naming of the new 

atmospheric layers being discovered.
28

  The meteorological satellite era, long anticipated 

by Rossby, officially began on April 1, 1960, when TIROS 1, the first Television and 

Infra-Red Observing Satellite, reached orbit. TIROS 1 carried two shuttered television 

cameras (photosensitive vidicon tubes) that recorded images of clouds on tape for later 

transmission to the ground. Wexler served as chief scientist for the program. 

Technology aside, the question remains: What can be accomplished and what 

cannot, especially in forecasting and prediction?  Techniques pioneered by the Bergen 

school made it possible to forecast partial rotations of a single vortex, initially for two to 

three days and later, with some skill, for three to seven days. However, it is not possible, 

due to turbulence and non-linear interactions, to forecast multiple rotations of multiple 

vortices, that is, to generate a long-range weather forecast for periods of weeks, months, 

                                                             
25

 J.R. Fleming, “Polar and Global Meteorology in the Career of Harry Wexler, 1933–1962,” Globalizing 

Polar Science: Reconsidering the social and intellectual implications of the International Polar and 

Geophysical Years, R.D. Launius, J.R. Fleming, and D.H. DeVorkin, eds. (New York: Palgrave, 2010), 

225-241. 
26

 Harry Wexler, “Introductory Talk,” Conference on Atomic Energy and Meteorology, U.S. Weather 

Bureau, December 19-20, 1949, Harry Wexler Papers, box 16, U.S Library of Congress, Manuscript 

Division. 
27

 Harry Wexler, “Structure of Hurricanes as Determined by Radar,” Annals of the New York Academy of 

Sciences 48 (1947): 821-844. 
28

 “Notes on Upper Atmospheric Nomenclature,” Harry Wexler Papers, box 32, U.S Library of Congress, 

Manuscript Division. 



 History of Meteorology 7 (2015) 12 

 
 

 

or seasons.  Prediction of planetary, or Rossby wave upper-air patterns, have been 

extended to five to seven days. Still, scientists may know where the ridges and troughs 

are, but cannot anticipate all their changes and vicissitudes with perfect certainty. In 

climate-related affairs, an “infinite forecast” can provide the overall patterns and the 

statistics of the general circulation, but at the expense of all detail. Yet it is not possible to 

predict the future state of the perturbed climate system, whether from natural or human 

causes, let alone the behavior of weather systems or even modified clouds. We can 

intervene in natural systems, but we cannot control them, nor can we forecast them with 

any precision. The best scientists can do is to gauge sensitivities to changes in various 

factors that make up the models. 

An influential report noted in 1962, “The task of the atmospheric scientists is to 

make quantitative measurements of the properties that describe the atmosphere's 

successive states, to understand the physical processes by which the successive states are 

determined, to predict future states, and, if feasible, to influence future states in a 

beneficial manner.”
29

 Atmospheric scientists have addressed the first two tasks and have 

generated robust progress; the final two tasks, prediction and influence (or control) 

remain beyond their reach. Finally, they are helpless (not a word often used in science) to 

predict the timing of regime changes, or changes of state of the flow.  

 

New Knots 

 

The neo-Laplacian program hit a solid brick wall in 1960 when, using a small computer 

and a simple non-linear model, Edward Lorenz (1917-2008) introduced chaos theory into 

meteorology and by extension into climate modeling.
30

 He brought the novel 

understanding that chaos theory provided a new topology, an extreme sensitivity to initial 

conditions in a dynamical system of deterministic non-periodic flow.
31

 In chaos theory, 

future states of the weather and climate then become identifiable with the attractor of the 

dynamical system — but the dynamical system may have more than one attractor! No 

matter how sophisticated the technology becomes, chaos theory holds that perfectly 

accurate measurements and perfectly accurate forecasts will never be possible.
32

  

Atmospheric scientists since face a number of difficult conceptual choices as they 

come to grip with chaos theory and the challenges it poses for both weather and climate 

modeling. The technologies that propelled the profound transformation of atmospheric 

research during and just after World War II, are, in their improved versions, still in 

widespread use.  Each made significant cuts into the Gordian Knot of meteorology, 

perhaps none greater than improvement of forecasting skill due to the ultimate linking of 

satellite monitoring and numerical weather prediction. Ensemble forecasters too attempt 

to circumvent chaotic limits by using spaghetti plots of multiple model forecasts to 

reduce uncertainty. Yet the neo-Laplacian knot persists and a new chaotic knot has been 

identified, clearly recognized in the field of numerical weather prediction, but still only 

                                                             
29

 Sverre Petterssen, “The Atmospheric Sciences, 1961-1971,” Weatherwise (Oct. 1962): 185-187, 213. 
30

 Edward N. Lorenz, “The Statistical Prediction of Solutions of Dynamic Equations,” Proceedings of the 

International Symposium on Numerical Weather Prediction in Tokyo, November 7-13, 1960, S. Syono, ed. 

Tokyo: Meteorological Society of Japan, 1962. 
31

 Edward N. Lorenz, “Deterministic Nonperiodic Flow,” J. Atmos. Sci. 20 (1963): 130–141. 
32

 Edward N. Lorenz, The Essence of Chaos (Seattle: University of Washington Press, 1993). 



 History of Meteorology 7 (2015) 13 

 
 

 

dimly apprehended with regards to climate models. In 1994 J. B. Elsner and J.C. Honoré 

noted that almost two decades elapsed following the seminal work of Lorenz before the 

larger scientific community engaged with the quantitative implications of chaos theory in 

any serious way.
33

 

In one of his many lectures, Lorenz addressed the issue of “Solving the weather 

forecasting problem.”  He shared three possibilities: 

1. Learning to make essentially perfect weather forecasts. 

2. Learning to make the best attainable weather forecasts, even if they are far from 

perfect. 

3. Learning how good the best attainable forecasts are even if we don’t learn how to 

make them. 

Then, with a twinkle in his eye, Lorenz asked, “Why should nearly perfect forecasts be 

unattainable?” — “Chaos.”
34

 Option 1. is where we were; option 2. is where we are; and 

option 3. represents the future.  In effect Lorenz had identified a new Gordian Knot of 

meteorology to accompany the older one that had been somewhat unraveled and frayed, 

but not untied. The atmospheric sciences are still coming to terms with this challenge.  

In 1927 Sir Arthur Eddington said this about the new and revolutionary quantum 

theory of matter: “Our eyes once opened, we may pass on to a yet newer outlook of the 

world, but we cannot go back to the old outlook.” This implicit acknowledgment of 

shape-shifting concepts in science applies equally well to new and emerging 

understandings of weather and climate, as informed by chaos theory. 

                                                             
33

 J. B. Elsner and J. C. Honoré, “Ignoring Chaos,” Bulletin of the American Meteorological Society 75 

(1994): 1846-1847. 
34

 Undated lecture transparency in the Papers of Edward N. Lorenz, U.S Library of Congress, Manuscript 

Division. 


