Meteorologist, oceanographer, Python solvem probler, experimenter, Raspberry Pi tinkerer, geek dad, manager, space enthusiast, ham radio lover, science columnist, ADHDer, synthetic intellect curious, linux admin cosplayer.

GFDL Story Research: From ENIAC to the Nobel Prize

Research compiled: 2026-04-08

This document consolidates deep research on Joseph Smagorinsky, Syukuro Manabe, Richard Wetherald, and the GFDL story for a blog post tracing the path from the 1950 ENIAC forecast to the 2021 Nobel Prize in Physics.

1. Smagorinsky’s Immigrant Background

Joseph Smagorinsky was born 29 January 1924 in New York City, the youngest of five sons. His parents, Nathan Smagorinsky and Dina Azaroff, were Belarusian Jewish immigrants who had fled pogroms in Gomel (then in the Russian Empire, now Belarus). Nathan arrived in 1913 via the coast of Finland, passed through Ellis Island, and settled on the Lower East Side of Manhattan. He worked first as a house painter, then established a paint store that grew into a hardware business. Joseph’s siblings were Jacob (died in infancy), Samuel (b. 1903), David (b. 1907), and Hillel/Harry (b. 1919).

As a boy growing up in Depression-era New York, Smagorinsky was fascinated by the weather. During the 1930s he would visit the top of the New York Daily News building to look at the weather instruments displayed there. He took and passed the entrance examination for Stuyvesant High School, one of New York City’s three specialized science high schools, where he received an enriched education in mathematics and physics and further developed his interest in meteorology.

He enrolled at New York University (NYU) to study meteorology but was interrupted mid-sophomore year when he entered the U.S. Army Air Corps during WWII. He joined an elite group of cadet recruits selected for their talents in mathematics and physics, which led to six months of math and physics study at Brown University, followed by training in dynamical meteorology at MIT – where his instructor was Ed Lorenz, who would later pioneer chaos theory. During the war, Smagorinsky flew in the nose of bombers as a weather observer, making forecasts based on visible factors such as estimated wave size, air temperature, and wind velocity at the plane’s altitude. He provided flight forecasts for B-29 squadrons in Nebraska, then concluded his service as a weather reconnaissance officer in the North Atlantic.

After the war he returned to NYU, completing his B.S. (1947), M.S. (1948), and Ph.D. (1953, advisor: Bernhard Haurwitz).

Sources:

2. The ENIAC Forecast (1950) – Smagorinsky’s Role

In April 1950, Smagorinsky was part of the team led by Jule Charney that ran the first successful computer weather forecasts on the ENIAC at Aberdeen Proving Ground, Maryland. The team worked round the clock. The participants who made the first one-day nonlinear prediction were: Charney, Ragnar Fjortoft, John Freeman, George Platzman, and Joseph Smagorinsky.

Smagorinsky was among the jubilant meteorologists photographed in front of the ENIAC at the completion of the forecast. His wife Margaret Smagorinsky (nee Knoepfel), the first female statistician hired by the Weather Bureau, was one of the computer operators for the 1950 experiment, though her name was absent from the journal article detailing the results.

Sources:

3. The GFDL Founding Story

Creation (1955)

At John von Neumann’s instigation, the U.S. Weather Bureau created the General Circulation Research Section in 1955 and appointed Smagorinsky to direct it – at age 31 (he joined the Weather Bureau at age 29 in 1953; the Section was created two years later). Smagorinsky felt that his charge was to continue with the final step of the von Neumann/Charney computer modeling program: a three-dimensional, global, primitive-equation general circulation model of the atmosphere.

Initial Location

The Section was initially located in Suitland, Maryland, near the Weather Bureau’s Joint Numerical Weather Prediction Unit (JNWPU). In its early years, Smagorinsky collaborated with von Neumann, Charney, and Norman Phillips to develop a 2-level, zonal hemispheric model using a subset of the primitive equations.

Name Changes

  • 1955: General Circulation Research Section (Suitland, MD)
  • 1959: Renamed to General Circulation Research Laboratory (moved to Washington, D.C.)
  • 1963: Renamed to Geophysical Fluid Dynamics Laboratory (GFDL)

Move to Princeton (1968)

GFDL formed a collaboration with Princeton University in 1967 and moved to New Jersey in 1968, taking up residence in a low-slung building on Princeton’s Forrestal Campus. The move was motivated by the desire to take advantage of Princeton University’s leadership in early computing and the proximity of the Institute for Advanced Study (where von Neumann and Charney had worked). In 1968, Princeton created the precursor to today’s Program in Atmospheric and Oceanic Sciences (AOS), a graduate and postdoctoral program dedicated to climate science. GFDL became part of NOAA (National Oceanic and Atmospheric Administration) when that agency was created in 1970.

GFDL is not part of Princeton University – it is a federal laboratory under NOAA. But the collaboration has been deep and continuous: GFDL scientists hold visiting appointments at Princeton, and Princeton graduate students work at GFDL.

Sources:

4. Smagorinsky’s Personality and Leadership Style

Smagorinsky was described as exhibiting an “authoritarian style of rule tempered by protection of the scientists from disrupting outside influence” while celebrating the elitism and esprit de corps that characterized GFDL. He shielded researchers from government bureaucracy so they could focus on fundamental science.

Jerry Mahlman, who succeeded Smagorinsky as GFDL director, described Smagorinsky’s “almost relentless pursuit of excellence” and noted that he had no real interest in the “university scientific culture” that counted publications – instead he insisted on solving significant scientific problems.

Key aspects of his leadership:

  • International recruitment despite xenophobia: Shortly after WWII, with the nation still leery of Japan, Smagorinsky invited Suki Manabe, Yoshio Kurihara, and Kikuro Miyakoda to GFDL, valuing scientific merit over nationalism.
  • Computer procurement: He repeatedly secured the world’s fastest computers for GFDL through what colleagues described as unexplained influence in competitive resource allocation. GFDL ran on the IBM Stretch (world’s fastest 1961–1964), then the CDC 6600 (world’s fastest 1964–1969).
  • Talent scouting: He recruited Kirk Bryan (oceanography, 1961), who with Manabe produced the first coupled ocean-atmosphere model. He also brought in Isidoro Orlanski, Jerry Mahlman, and Isaac Held.
  • Long tenure: He directed GFDL for 28 years (1955–1983), building it into the world’s premier climate modeling laboratory.

Smagorinsky was recognized by NOAA as one of the ten most significant figures in the agency’s history.

Sources:

5. Smagorinsky’s 1963 Paper

The Paper

Smagorinsky, J. (1963). “General Circulation Experiments with the Primitive Equations: I. The Basic Experiment.” Monthly Weather Review, 91(3), 99–164.

What It Simulated

A nine-level, hemispheric, primitive-equation general circulation model. An integration for 60 days was made from initial conditions where random temperature disturbances were superimposed on a zonally symmetric, baroclinically unstable regime. The experiment displayed the scale-selective character of baroclinic instability, yielding zonal wave numbers 5 to 6, and predicted an index (energy) cycle with a period of 11–12 days for the first 40 days. The resulting mean zonal velocity profile was in good qualitative agreement with observation.

This paper fundamentally changed atmospheric modeling from quasi-geostrophic to primitive-equation frameworks.

The Computer

The model was run on the IBM 7030 Stretch (also known as IBM Stretch), which was the world’s fastest computer from 1961 to 1964. The GFDL modeling series known as MARKFORT began with this nine-level model on the Stretch.

The Smagorinsky-Lilly Subgrid Turbulence Model

In the same paper, Smagorinsky introduced a first-order subgrid-scale closure for turbulence: the eddy viscosity is expressed as the characteristic grid scale multiplied by a velocity scale. This was further developed by Douglas Lilly and James Deardorff and is now known as the Smagorinsky model (or Smagorinsky-Lilly model). It remains the most widely used approach in large eddy simulation (LES) and is applied across fluid dynamics far beyond meteorology – in engineering, combustion, oceanography, and astrophysics.

Sources:

6. Manabe’s Recruitment to GFDL

How Smagorinsky Found Manabe

Smagorinsky read a paper Manabe had written and invited him to the United States. Manabe had just completed his D.Sc. in meteorology at the University of Tokyo (1958), studying under Shigekata Shono. He arrived in Washington, D.C., from Japan in 1958 and formally joined the General Circulation Research Section in 1959.

Why Manabe Accepted

Manabe was drawn by several factors:

  • Computing power: Japan in the late 1950s was still recovering from war. There were no jobs, and computing resources were negligible. In America, Manabe had free use of a supercomputer.
  • Salary: His American salary was reportedly 25 times what he had earned in Japan.
  • Research freedom: Smagorinsky had assembled a young international team with an ambitious plan to build a General Circulation Model. Manabe recalled: “Smagorinsky’s and my role were complementary. He had the ambitious plan, and my job was to make it work. The computer was so feeble at the time… if we put everything into the model at once, the computer couldn’t handle it.”

Manabe’s Background

Born 21 September 1931 in Shinritsu Village, Ehime Prefecture, Japan. His grandfather and father were physicians who operated the village’s only clinic. He was expected to follow the family tradition but chose a different path, recalling: “Whenever there’s an emergency, the blood rushes to my head” and confessing he had a horrible memory and physical clumsiness that made him unsuitable for medicine. As a child he preferred “to gaze at the sky and get lost in thoughts.”

Sources:

7. Manabe on Japan vs. America

Manabe has been remarkably frank about why he left Japan and never went back permanently.

Key quotes:

  • On Japanese communication: “In Japan, if you ask a question you get ‘yes’ or ‘no.’ However, when the Japanese say ‘yes’ it doesn’t necessarily mean ‘yes.’ It could mean ‘no.’”

  • On why science requires directness: “Japanese people don’t want to say no. But in science this is not a very good thing – you have to say very clearly you disagree with each other.”

  • On why he didn’t return: “I don’t want to go back to Japan, because I am not able to live harmoniously.” He found himself caught up in bureaucracy and spent too much time coordinating research agencies. There were staff shortages and a cultural custom discouraging people from speaking up.

  • On American freedom: “In America, I can do whatever I want. I don’t care how other people feel.”

Note: The exact quote the user recalled – “In Japan, people are polite. They never disagree with anything. But they never do what you ask” – is a paraphrase. The verified quotes above convey the same essential meaning. Manabe did briefly return to Japan (1997–2001) to direct the Global Warming Research Division of the Frontier Research System for Global Change, but came back to Princeton in 2002.

Sources:

8. The 1967 Manabe-Wetherald Paper

Full Citation

Manabe, S. & Wetherald, R. T. (1967). “Thermal Equilibrium of the Atmosphere with a Given Distribution of Relative Humidity.” Journal of the Atmospheric Sciences, 24(3), 241–259.

What It Did

Manabe and Wetherald built a one-dimensional radiative-convective model that divided the atmosphere into multiple layers and calculated how heat moved between them. The model included accurate spectroscopy of CO2, ozone, and water vapor, atmospheric convection, and – critically – water vapor feedback. Previous studies (e.g. Arrhenius 1896, Callendar 1938, Plass 1956) had estimated CO2 warming but with crude radiation calculations and no feedback loops.

The key innovation was the treatment of relative humidity: instead of holding specific humidity constant (which would underestimate warming), they held relative humidity fixed at observed values. This meant that as the atmosphere warmed, it held more water vapor – a positive feedback that amplified the CO2-driven warming.

What It Predicted

  • Doubling CO2 from 300 to 600 ppm with fixed specific humidity: 1.3 degrees C warming.
  • Doubling CO2 with fixed relative humidity (including water vapor feedback): 2.36 degrees C warming (Table 5 of the paper).
  • Stratospheric cooling accompanying tropospheric warming – a distinctive fingerprint of greenhouse-gas-driven warming (as opposed to solar-driven warming, which would warm both).

How Accurate Was It?

Remarkably so. The slope of the CO2-vs-temperature regression line from 50 years of actual observations gives 2.57 degrees C per doubling – only slightly higher than their 2.36 degrees C prediction. The error in their estimate is less than 10%. Their result is well within the range reported by the IPCC Fifth Assessment Report.

The Computer

The calculations were performed on machines at GFDL during the mid-1960s. At this time GFDL was using the IBM Stretch and transitioning to the CDC 6600 (which became operational in 1964). The 1D radiative-convective model was computationally much less demanding than the full 3D GCM.

Reception

The paper was not widely noticed at first – it appeared in a specialist journal and climate change was not yet a public concern. But it was later voted by climatologists as the most influential climate research paper of all time (Carbon Brief survey).

Sources:

9. Richard Wetherald

Biography

  • Born: 28 March 1936, New Jersey, USA
  • Parents: Joseph Wetherald and Ruth M. Wetherald
  • Died: 9 October 2011, Trenton, New Jersey (age 75)

Wetherald worked at GFDL under the Climate Dynamics Group headed by Manabe. He was Manabe’s most important collaborator – their partnership spanning nearly four decades and producing the two most consequential papers in climate science (1967 and 1975). His work at GFDL, coupled with theoretical research, helped lay the foundation for a transition of greenhouse theory from “science fiction to science.”

He was named Meteorologist Professional of the Year for 2007 by Strathmore’s Who’s Who.

The Recognition Gap

Wetherald never received the major awards and recognition that Manabe did. He was not included in the 2021 Nobel Prize – the Nobel Committee cited only Manabe. This has been noted as one of many examples where long-standing collaborators in climate science go unrecognized. When Manabe won the Nobel, Wetherald had already been dead for a decade.

Sources:

10. The 1975 Manabe-Wetherald CO2-Doubling Study

Full Citation

Manabe, S. & Wetherald, R. T. (1975). “The Effects of Doubling the CO2 Concentration on the Climate of a General Circulation Model.” Journal of the Atmospheric Sciences, 32(1), 3–15.

What Was New

This was the first time a three-dimensional GCM was used to simulate the response of temperature and the hydrologic cycle to doubled CO2. The 1967 paper had used a 1D column model; the 1975 paper used a full 3D general circulation model.

Key Findings

  • Global mean surface warming: approximately 3 degrees C for a doubling from 300 to 600 ppm.
  • Arctic amplification: Much greater warming at the poles than at the equator. The increase of surface temperature in higher latitudes was magnified due to the recession of the snow boundary and the thermal stability of the lower troposphere, which limits convective heating. This was the first prediction of Arctic amplification in the literature.
  • Intensified hydrologic cycle: The model predicted that global warming would speed up evaporation and precipitation, producing more intense rainfalls.
  • Stratospheric cooling: Confirmed the 1D result from 1967.

Subsequent Validation

The 1975 result (ECS ~ 3 degrees C) became the anchor for the 1979 Charney Report, which evaluated models by Manabe and James Hansen and concluded that climate sensitivity was likely 3 degrees C +/- 1.5 degrees C. This estimate has remained largely unchallenged for over four decades.

Sources:

11. The 2021 Nobel Prize in Physics

The Award

On 5 October 2021, the Royal Swedish Academy of Sciences announced that the Nobel Prize in Physics would be shared:

  • One half jointly to Syukuro Manabe and Klaus Hasselmann “for the physical modeling of Earth’s climate, quantifying variability and reliably predicting global warming.”
  • One half to Giorgio Parisi “for the discovery of the interplay of disorder and fluctuations in physical systems from atomic to planetary scales.”

Manabe’s Reaction

When told about the prize, Manabe responded with characteristic humor and humility:

  • “I was really happy and surprised.”
  • “I never dreamed I would win the Nobel physics prize. If you look at the list of past winners, they are amazing people who have done marvelous work. In contrast, what I have been doing looks trivial to me.”
  • After wondering whether his work deserved comparison with previous physics laureates, he considered that climate change is now a major crisis for humanity: “For that reason, I thought, maybe it’s ok!” – said with a laugh.
  • The “great fun” quote: “It is great fun to use a model to study how climate change over the last 400 million years has evolved.” This phrase became the headline of Princeton’s press release.

The Nobel Lecture

Manabe delivered his Nobel lecture, “Physical Modeling of Earth’s Climate,” on 8 December 2021 in Stockholm. The lecture traced the history of climate modeling from the early pioneers. He acknowledged Joseph Smagorinsky as the inaugural director of GFDL who made his career possible. The lecture is available as a PDF from NobelPrize.org and was published in Reviews of Modern Physics (2023).

Connection to Charney and the ENIAC

The Nobel Committee’s scientific background document traced the lineage from von Neumann’s vision of numerical weather prediction through the ENIAC forecast (1950) to Smagorinsky’s GCM work and Manabe’s climate modeling. Manabe’s predictions directly informed the 1979 Charney Report, which established the benchmark for climate sensitivity that has held ever since.

Sources:

12. Manabe’s Personality

Manabe is universally described as joyful, curious, humble, and self-deprecating. Key quotes:

  • “Curiosity is the thing which drives all my research activity.”
  • “I did these experiments out of pure scientific curiosity. I never realized that it would become a problem of such wide-ranging concern for all of human society.”
  • “I spent a major fraction of my life thinking about the same thing. I really didn’t do much else. Now at 90, I finally decided to stop doing active research.”
  • He urged students to “follow their curiosity and their joy, rather than trying to predict what research may prove impactful in future decades.”
  • As a child, he “usually stayed at home, laid down and thought about something endlessly.”

He became a U.S. citizen, explaining with disarming honesty that he found American directness more compatible with scientific work than Japanese social conventions. In 2022, he was named a “Great Immigrant” by the Carnegie Corporation of New York.

Sources:

13. The Chain from ENIAC to Nobel – Every Link with Dates

Date Event Key People
1946 Von Neumann proposes using computers for weather prediction at IAS Princeton John von Neumann
April 1950 First successful numerical weather forecast on ENIAC, Aberdeen, MD Charney, Fjortoft, Freeman, Platzman, Smagorinsky
1950 Publication of ENIAC results in Tellus Charney, Fjortoft, von Neumann
1953 Smagorinsky completes Ph.D. at NYU, joins U.S. Weather Bureau Smagorinsky
1955 Von Neumann instigates creation of General Circulation Research Section; Smagorinsky appointed director at age 31 von Neumann, Smagorinsky
1955 Norman Phillips publishes first GCM experiment Phillips
1957 Von Neumann dies (8 February) von Neumann
1958 Manabe completes D.Sc. at University of Tokyo; arrives in Washington, D.C. Manabe
1959 Smagorinsky formally recruits Manabe to the General Circulation Research Section Smagorinsky, Manabe
1961 Kirk Bryan recruited to GFDL for ocean modeling Bryan
1963 Smagorinsky publishes primitive-equation GCM paper on IBM Stretch; introduces subgrid turbulence model Smagorinsky
1963 Lab renamed Geophysical Fluid Dynamics Laboratory (GFDL) –
1965 Manabe, Smagorinsky, & Strickler publish GCM with hydrologic cycle Manabe, Smagorinsky, Strickler
1967 Manabe & Wetherald publish “Thermal Equilibrium” – first quantification of CO2 warming (2.36 degrees C per doubling) Manabe, Wetherald
1968 GFDL moves to Princeton’s Forrestal Campus –
1969 Manabe & Bryan publish first coupled ocean-atmosphere model Manabe, Bryan
1975 Manabe & Wetherald publish 3D CO2-doubling study showing Arctic amplification (3 degrees C per doubling) Manabe, Wetherald
1979 Charney Report confirms Manabe’s climate sensitivity estimate (~3 degrees C +/- 1.5 degrees C) Charney, Manabe, Hansen
1981 Charney dies (16 June) Charney
1983 Smagorinsky retires as GFDL director after 28 years Smagorinsky
2005 Smagorinsky dies (21 September, age 81) Smagorinsky
2011 Richard Wetherald dies (9 October, age 75) Wetherald
5 Oct 2021 Manabe awarded Nobel Prize in Physics – 71 years after the ENIAC forecast Manabe
8 Dec 2021 Manabe delivers Nobel lecture in Stockholm Manabe

The Chain in Words

Smagorinsky stood in front of the ENIAC in April 1950 and helped produce the first computer weather forecast. Five years later, von Neumann chose him to lead a new research section dedicated to simulating the global atmosphere. Smagorinsky recruited a 27-year-old Japanese meteorologist named Suki Manabe, gave him access to the world’s fastest computers, and shielded him from bureaucratic interference for decades. Manabe, working with Richard Wetherald, built the models that first quantified the greenhouse effect and predicted Arctic amplification. The Charney Report validated their work. Seventy-one years after the ENIAC forecast, Manabe stood in Stockholm to receive the Nobel Prize in Physics – the culmination of a chain that began with a room-sized vacuum-tube computer in a basement at Aberdeen.

14. Additional Primary Sources

15. Remaining Gaps / Items to Verify

  1. Exact mechanism of Manabe recruitment: Was there a specific letter? Did Smagorinsky visit Japan? The verified account says Smagorinsky read one of Manabe’s papers and invited him. The precise logistics remain unclear in public sources. The AIP and UCAR oral histories of Smagorinsky may contain this detail.

  2. Manabe’s “faculty meetings” quote: The exact phrasing “They never disagree with anything. But they never do what you ask” is a widely circulated paraphrase. The verified quotes are slightly different (see Section 7 above). The user should use the verified versions or mark it as paraphrased.

  3. Which computer ran the 1967 Manabe-Wetherald paper: The 1D radiative-convective model was computationally light. GFDL had the IBM Stretch (1961–64) and CDC 6600 (from 1964). The 1967 paper was likely computed on the CDC 6600, but this is not explicitly confirmed in public sources.

  4. Manabe’s Nobel lecture – specific mentions of ENIAC/Charney/von Neumann: The full text is in the PDF linked above. The lecture does acknowledge Smagorinsky explicitly. Whether it mentions the ENIAC by name should be verified by reading the PDF.

  5. Richard Wetherald’s education: No public source provides details about where he went to college or graduate school.