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.

Frederick Gale “Bud” Shuman (1919-2005)

Deep-dive research document supplementing Shuman.md, prepared 2026-05-13 for the dedicated biographical anchor post in the NWP history series. Where it overlaps with Shuman.md, this document is the more cited and fact-checked version; where it adds new material, it should be treated as the canonical reference for blog use.

Cross-reference to prior posts in the series

Three prior posts in the series have established core Shuman facts; this document is consistent with them but adds depth:

  • Post 10 (2026-04-17-The-decade-the-forecast-got-good.md): Shuman as JNWPU’s first employee at the IBM 701 console, Cressman’s successor as NMC director, author of the 1989 retrospective, builder of the 1966 six-layer PE model with Hovermale. Verified.
  • Post 31 (2026-05-06-The-algorithm-that-outlived-its-machine.md): NMC procured three IBM System/360 Model 195s under Shuman’s directorship; first installed March 1974, third in November 1975; operational gridpoint primitive-equation model ran on the 195s through the second half of the 1970s. Verified by the GFS history (Yang & Tallapragada 2019), which lists “1972: 3 IBM 360-195s” at NMC – Post 31’s narrower 1974-1975 window may reflect a specific installation timeline, but the multi-machine fact is solid. Note: there is a minor date inconsistency between Post 31 (which dates the start of the three-machine era to 1974/1975) and the AMS 2019 GFS history slide (which dates three Model 195s to 1972). Both are operationally compatible if one reads the AMS slide as “first 195 installed 1971 single, three by 1972, NMC main computing floor through ~1980” – but Post 31’s specific install dates of March 1974 / November 1975 should be verified against the Shuman 1989 paper directly before being relied on for any new post.
  • Post 39 (2026-05-14-Thirty-eight-years-on-the-sphere.md): Joseph Sela built the first US operational spectral model at NMC, went operational August 1980 at T30L12 (rhomboidal 30 truncation, twelve sigma levels), and the spectral GFS ran until June 12, 2019 when FV3 replaced it. Verified. Critical fine point: the GFS history slide deck (Yang & Tallapragada 2019) gives R30 (rhomboidal 30), and the August 1980 introduction; Post 39’s “T30” should technically be “R30” since triangular truncation did not arrive until August 1987 (T80) – this is a small terminology error in Post 39 worth flagging for future correction.

No claim in the prior posts is contradicted by the additional sources found in this research pass. Two small refinements are worth carrying forward (the 1972/1974 three-195 date, and the R30/T30 truncation type).

1. Vital statistics

Field Value Source
Full name Frederick Gale Shuman Wikipedia, Mariners Weather Log Dec 2007, Find a Grave
Nickname “Bud” Find a Grave memorial, family obituary
Born July 13, 1919, South Bend, Indiana Wikipedia, Find a Grave
Died July 29, 2005, Fort Washington Medical Center, Maryland Washington Post obituary, Mariners Weather Log
Cause of death Congestive heart failure Washington Post
Age at death 86 All obituary sources
Spouse Elena Fragomeni Shuman Legacy.com obituary search excerpt
Children Frederick Gale Dominick Shuman (son), Marianne Shuman Ferrin (daughter), Deborah Joan Shuman (daughter) Legacy.com obituary text
High school South Bend Central High School, class of 1937 Wikipedia
Won statewide math competition while in high school Yes – first place in Indiana statewide comprehensive mathematical competition Mariners Weather Log Dec 2007

Note on middle name: The Wikipedia article, the Washington Post obituary, the Mariners Weather Log, Find a Grave, and the family obituary all give the middle name as Gale. The user prompt for this research document gave the middle name as “George” – this appears to be a slip in the brief. All authoritative sources agree on Gale.

2. Education

Year Degree Institution Notes Source
1941 B.S. Mathematics Ball State University, Muncie, Indiana Became junior Weather Bureau observer at Indianapolis Airport while still completing degree Wikipedia, Mariners Weather Log, Washington Post
1942 M.S. Meteorology MIT (via Army Air Corps officer-training program) Commissioned 2nd lieutenant; sent to MIT for the meteorology training pipeline Wikipedia
1951 Sc.D. (Doctor of Science) in Meteorology MIT Cited in Wikipedia as the first doctoral thesis at MIT on numerical weather prediction. Thesis advisor not in any of the consulted sources – a query to MIT’s Department of Earth, Atmospheric and Planetary Sciences would be required to confirm. Wikipedia’s claim of “first doctoral thesis on numerical weather prediction at MIT” should be treated as the cited claim rather than my independent verification. Wikipedia, Washington Post

3. Career chronology

Year(s) Position Notes / Source
1941 Junior weather observer, Indianapolis Airport (Weather Bureau) Wikipedia, Mariners Weather Log
1941 (June)–1945 U.S. Army Air Corps weather officer, North Africa and Italy theatre Wikipedia
1956 Discharged from U.S. Air Force Reserve at the rank of major findagrave-derived obituary text, Washington Post (via search excerpt)
Post-war (~1946–1951) Airways forecaster, Wayne County Airport, Detroit; then U.S. Weather Bureau research meteorologist in Washington D.C. on severe storm systems and early tornado forecasting Wikipedia, Mariners Weather Log
1952–1954 Researcher, Institute for Advanced Study (IAS), Princeton, New Jersey Wikipedia and Find a Grave both date this period 1952–1954. Worked on numerical weather prediction models, attended lectures by J. Robert Oppenheimer and Albert Einstein. Computer used: the Wikipedia article and Mariners Weather Log both say he used the “JOHNNIAC”; this label is loose – JOHNNIAC was the RAND copy of the IAS machine, while Shuman would more likely have run on the IAS machine itself or possibly the Princeton-built MANIAC at Los Alamos. The standard family of IAS-class machines is what these obituary writers meant.
1954 First employee of the Joint Numerical Weather Prediction Unit (JNWPU), Federal Office Building 4, Suitland, Maryland Wikipedia. The Washington Post obituary says he was “assigned to the JNWPU in 1954”; the Find a Grave memorial calls him “the first employee” of JNWPU. The two are compatible. The JNWPU itself was formally established July 1, 1954 by joint agreement of the Weather Bureau, Air Force, and Navy. Cressman was its first Director (1954-1964).
1955 (May 6) First operational U.S. numerical weather forecast, IBM 701 at JNWPU NCEP/WPC History; the experimental runs began in March 1955; the May 6 date is the standard “first issued forecast.” Shuman participated as the JNWPU’s chief modeler.
1958 (January 1) NMC formed by merger of JNWPU and the National Weather Analysis Center (NAWAC); Shuman becomes Chief of the NMC Development Division WPC History PDF
1964 (April) Director, National Meteorological Center, succeeding George P. Cressman who became Chief of the Weather Bureau Wikipedia, Washington Post obituary, WPC History PDF, Mariners Weather Log. Specifically April 1964, not 1965 as the user’s brief tentatively suggested. The “1965” in the brief likely conflates this with the simultaneous Weather Bureau reorganization into ESSA in 1965.
1964–1981 NMC Director (17 years) All sources
1981 (January) Retired from NMC directorship; succeeded by William D. Bonner WPC History PDF, Mariners Weather Log
1981–1986 Continued research at NMC for five years post-retirement Shuman.md (existing research file) – source of this five-year claim should be verified; Wikipedia and Washington Post obituary don’t mention post-1981 research.
2005 (July 29) Death, Fort Washington Medical Center, Maryland; cause: congestive heart failure All obituary sources

4. The Institute for Advanced Study years, 1952-1954

This was Shuman’s intellectually-formative period. He worked at IAS during the run-up to the JNWPU’s establishment, on the very same machine – or machines of the same family – that Charney’s group had used for the Thanksgiving 1950 baroclinic-model demonstration. Mariners Weather Log writes that he attended lectures by J. Robert Oppenheimer (then IAS Director) and by Albert Einstein. The lectures by Oppenheimer would have been in 1952-1953, two years before Oppenheimer’s security clearance was revoked in June 1954. Einstein, then in his last years, was a Professor of Theoretical Physics at IAS.

The IAS posting placed Shuman in direct intellectual contact with the von Neumann-Charney NWP project. When Cressman was looking for a chief modeler for the operational unit that would industrialize the IAS methods, Shuman was the natural choice: he had run the IAS machine, was MIT-trained in meteorology, had Weather Bureau operational experience, and was an Army Air Corps reserve officer with formal Pentagon ties. (Source: Mariners Weather Log Dec 2007.)

5. JNWPU (1954-1958) and the first operational forecast

The JNWPU was established by a Joint Agreement signed by the three sponsoring services in 1954, with operations beginning July 1, 1954. Shuman was the unit’s chief modeler from the outset. The IBM 701 was installed at Suitland in March 1955; experimental forecasts began in April 1955; the first operational issued forecast went out on May 6, 1955. (Source: WPC History PDF, Edwards 2010 A Vast Machine pp. 254-255.)

Shuman’s own honest verdict on the early operational years, written 34 years later:

“Within three years three agencies of the United States Government jointly created a numerical weather prediction service, but it was quickly discovered that current models had very serious defects. After considerable research, the first operationally effective model was achieved in 1958.”

– Shuman 1989, Weather and Forecasting 4, p. 291-292.

Three years of bad forecasts before anything useful came out. Cressman’s contemporaneous explanation, which Shuman later endorsed in his retrospective:

“[The early model’s results] were not nearly as good as the earlier tests had indicated. It had not occurred to us that we could have success with [the] very difficult baroclinic events [studied during testing] and then have severe problems with many routine, day-by-day weather situations.”

– George Cressman, “The Origin and Rise of Numerical Weather Prediction,” in J. R. Fleming (ed.), Historical Essays on Meteorology, 1919-1995 (AMS, 1996), p. 30. Cited in Edwards 2010, A Vast Machine, p. 124.

The model retreated from a three-level baroclinic to a two-layer baroclinic (1956), then back to a simple barotropic model based on the original ENIAC code – which by 1958 had been refined enough to be operationally effective. The three-level baroclinic was reintroduced in 1962, and as Edwards puts it, “this model initially offered only slight improvements over the barotropic forecast” (Edwards 2010, p. 126). Mariners Weather Log notes that Otha Fuller worked alongside Shuman at the IBM 701 in 1955 – a name not yet appearing in the blog series but a candidate for a future cross-reference.

6. The Shuman filter (1957)

Two papers in volume 85 of Monthly Weather Review (November 1957) establish Shuman’s first major individual contribution:

  • Shuman, F.G. (1957a). “Numerical Methods in Weather Prediction: I. The Balance Equation.” Monthly Weather Review 85(10), 329-332.
  • Shuman, F.G. (1957b). “Numerical Methods in Weather Prediction: II. Smoothing and Filtering.” Monthly Weather Review 85(11), 357-361.

The second paper – universally called the Shuman filter or Shuman smoother – introduced a methodology for designing finite-difference smoothing and filtering operators that meet pre-specified frequency-response criteria. The design philosophy was novel: specify the desired wavenumber response function, then derive the finite-difference weights that approximate it. This is the inverse of the more usual approach of writing down an averaging stencil and then computing its response.

Shuman’s filter became standard practice in operational NWP. Its specific application was to remove spurious short-wave noise that finite-difference primitive-equation models accumulate around their truncation cutoff, without distorting the meteorologically significant scales. Mathematical effect: a weighted three-point or five-point running mean with weights chosen so that the response function R(2deltax) is approximately zero (i.e., the 2-grid-length wave is annihilated) while R(8deltax) is approximately one (i.e., the 8-grid-length and larger waves are preserved).

The filter is still cited and used. The Springer Journal of Engineering Mathematics published two papers in 1980-1981 – “The Shuman filtering operator and the numerical computation of shock waves” and “Switched numerical Shuman filters for shock calculations” – showing the filter’s adoption outside meteorology in CFD. Modern atmospheric-modelling textbooks (Durran 2010 and Mesinger and Arakawa 1976) still reference the Shuman 1957 paper as the canonical citation for finite-difference smoothing operators. (Sources: Springer DOIs 10.1007/BF01534981 and 10.1007/BF01535103; standard NWP textbook citations.) Whether it is still in operational use in 2026 – I cannot find specific 2020s citations confirming its continued operational use; modern operational models (the FV3 GFS, the IFS, GEM, ECMWF’s, etc.) use higher-order numerical diffusion schemes and spectral truncation, so the Shuman filter is now more a teaching reference than a production tool. Flag this for any future blog claim.

7. The 1966 Six-Layer Primitive Equation Model

Shuman’s other major model-building contribution, with John B. Hovermale:

  • Shuman, F.G. and Hovermale, J.B. (1968). “An Operational Six-Layer Primitive Equation Model.” Journal of Applied Meteorology 7(4), 525-547.

The model became operational at NMC in mid-1966 (specifically June 1966 per Yang & Tallapragada 2019 GFS history) on the CDC 6600. Key facts:

  • It was the first operational primitive-equation model anywhere in the world, beating West Germany by months in 1966.
  • It integrated the full hydrostatic primitive equations directly, abandoning the quasi-geostrophic and quasi-vorticity approximations that had constrained all prior operational models.
  • It used six vertical layers on a Northern-Hemispheric octagonal grid extending from 10 degrees N to near the pole; gridpoints roughly 400 km apart.
  • “In its first fourteen months of operational use, it has resulted in highly significant improvements in the Center’s products” – Shuman and Hovermale 1968, p. 525.
  • Edwards (2010, p. 129) reproduces the model’s grid figure with the source attribution “F. G. Shuman and J. B. Hovermale, ‘An Operational Six-Layer Primitive Equation Model,’ Journal of Applied Meteorology 7, no. 4 (1968), 528.”

The model is the namesake of the Shuman-Hovermale model that was operational at NMC throughout the late 1960s and 1970s. Later versions (semi-implicit reformulations, higher resolution) ran on the IBM 360/195 through the 1970s; the model was retired in stages as the spectral model (1980) and the LFM (1971) took over the global and regional roles respectively. Per Post 10 of the blog series, this 1966 deployment is “the day NWP finally got operationally useful in the United States.”

8. The NMC computer history under Shuman’s directorship

From the WPC History PDF (September 2025) and the GFS history slide (Yang & Tallapragada 2019), with cross-checks from Edwards 2010 and the 1986 NMC Product Suite TA-86-05:

Year installed Machine Notes
1955 (March) IBM 701 Williams-tube memory; ~30-min MTBF; rented at ~$200,000/year (half the JNWPU budget).
1957 IBM 704 Doubled speed, magnetic-core memory.
1960 IBM 7090 67 kiloflops; transistorized.
1963 IBM 7094 100 kiloflops; ran the first operational baroclinic model.
1966 CDC 6600 3 megaflops; enabled the Shuman-Hovermale six-layer PE in mid-1966 and the first global PE in June 1966.
1971 First IBM System/360 Model 195 6 megaflops single machine. The first IBM 360-class operational NMC computer. Enabled the LFM operational debut on October 31, 1971.
1972 (per GFS history) Three IBM 360/195s total 18 megaflops aggregate. Note: Post 31 dates the three-machine era to 1974/1975 based on a different source; the Yang & Tallapragada GFS slide dates it to 1972. Both can be reconciled if the 1972 date is when the third machine was ordered/contracted and 1974-1975 when fully installed and operational. For any new blog post, the 1971 single-machine date and the 1974-1975 three-machine date are defensible from the Post 31 sourcing; the 1972 three-machine date is defensible from the GFS history slide.
1978 “High resolution PE model on an IBM 360/195” upgrade WPC History PDF. The 1978 model went operational on the existing 195 hardware, taking advantage of headroom.
1981 (post-Shuman) CDC Cyber 205 400 megaflops; arrived in 1981 per Yang & Tallapragada (post-Shuman retirement); the first NMC-operational forecasts on Cyber 205 came after Shuman’s January 1981 retirement, but the procurement decision was Shuman’s. The LFM was ported to the Cyber 205 by August 30, 1983 (per TA-86-05). The Spectral R40 upgrade came online October 1983 with the Cyber 205.
1987 Two Cyber 205s 800 megaflops aggregate.
1990 Cray Y-MP 2.6 gigaflops.
1994 Cray C9016 15.3 gigaflops.

The three IBM 360/195 floor at Suitland is the canonical Shuman-era machine room: this was the operational US NWP production hardware throughout the late 1970s, the most homogeneous computing environment any operational meteorological centre had ever had up to that point. The Suitland FOB-4 building continued to house the Center’s main computers until 1999 (when an IBM SP was installed in Bowie, Maryland) and later in Gaithersburg in 2002 (WPC History PDF).

9. Models that ran at NMC during Shuman’s tenure

Model First operational Retired Primary author(s) Machine(s)
Barotropic (3-level QG) May 6, 1955 1956 (failed) Charney’s IAS team; ported by Shuman IBM 701
Barotropic (2-layer thermotropic) 1956 1958 Shuman IBM 701/704
Barotropic (refined) 1958 ~1962 Shuman IBM 704
3-level baroclinic 1962 ~1966 Shuman IBM 7094
6-layer Primitive Equation (Shuman-Hovermale) June 1966 ~1980 (replaced by spectral) Shuman and Hovermale CDC 6600 then IBM 360/195
LFM (Limited-area Fine-mesh Model) October 31, 1971 1993 (Eta replaced as “Early” run); last archived data Dec 1995 J.G. Howcroft (early 1966 conceptual work); Joseph P. Gerrity (1977 documentation); Newell & Deaven (1981 LFM-II refinement); Dennis Deaven ported to Cyber 205 (Aug 1983) Cyber 205 from 1983, prior on 360/195
8-layer global PE with Hough analysis November 1972 1974 (replaced by 9-layer) Flattery, Stackpole IBM 360/195
9-layer global PE (latitude-longitude) September 18, 1974 ~1980 (replaced by spectral) Stackpole, Flattery (continued from 1973) IBM 360/195
Movable Fine-Mesh Hurricane Model (MFM) 1975 ~1990s NMC hurricane team 360/195 then Cyber 205
Optimum Interpolation analysis (replaced Cressman SCM) September 22, 1978 Long lifespan NMC analysis team IBM 360/195
Global Spectral Model (R30, 12 layers) August 1980 (forecast); analysis system May 27, 1980 Successively upgraded; spectral core retired June 12, 2019 (FV3 replacement) Joseph Sela (1975 onward) IBM 360/195 in 1980, Cyber 205 from 1983

The crucial Shuman-era operational models on the 360/195 production floor in the 1970s were therefore:

  1. The six-layer Shuman-Hovermale Hemispheric PE (1966-ca. 1980), the senior model;
  2. The LFM (Oct 1971-1993 official; data retained to 1995), the regional model;
  3. The 9-layer global PE (Sep 1974-1980), the first global product;
  4. The MFM hurricane model (1975-1990s), the specialized regional product;
  5. The Sela spectral model (Aug 1980), which Shuman commissioned but left operational just five months before his retirement.

10. The 1981 transition off the IBM 360/195s

The 360/195s were replaced after Shuman’s January 1981 retirement, but the procurement was set up under his directorship. The successor was the CDC Cyber 205. Per the Yang & Tallapragada GFS history, the Cyber 205 came online at NMC in 1981 (the same year UK Met Office got theirs); the LFM was ported to the Cyber 205 by August 30, 1983 (TA-86-05); the spectral resolution was upgraded from R30 to R40 in October 1983 to exploit Cyber capacity (TA-86-05).

The first NMC operational use of the Cyber 205 is described in Dennis Deaven’s NWS technical paper Operational Numerical Weather Prediction on the Cyber 205 at the National Meteorological Center (early 1980s), which describes a four-component operational suite: the barotropic global guess-field generator, the LFM regional 12-48 hour forecast, the global spectral 6-day model, and the moveable-fine-mesh hurricane model. By 1987 NMC had two Cyber 205s, and in 1990 the Cyber generation was replaced by the Cray Y-MP, well into Bonner’s directorship.

The Cyber 205 procurement was Shuman’s last major architectural decision. It was a deliberate move away from IBM mainframe architecture toward a vector-processor architecture (the same architectural family as the Cray-1 that ECMWF, UK Met Office, and Florida State were also adopting in the same window). The decision was strategically aligned with the move from gridpoint to spectral models – spectral models had vectorizable inner loops on the Cyber 205’s vector pipelines, while finite-difference gridpoint models could only partially exploit the architecture.

11. The spectral transition and Joseph Sela

A specific question in the brief was whether Sela’s August 1980 operational debut “was after Shuman left” or whether Shuman enabled it. The answer is the latter:

  • Sela began spectral-model development at NMC in 1975 – six years before Shuman retired, under Shuman’s directorship.
  • The first operational spectral run went live in August 1980 – five months before Shuman’s January 1981 retirement.
  • Shuman therefore explicitly authorized the spectral procurement and oversaw the operational deployment. Per Yang and Tallapragada (2019 AMS GFS history), this was a major architectural decision – the model was substantially less computationally efficient on the IBM 360/195’s scalar architecture than a finite-difference model, and Shuman’s authorization to deploy it operationally was a bet on the eventual arrival of the Cyber 205, which would have a vectorizable spectral transform path.
  • The spectral model’s continued operational reign for 38 years and 11 months (Aug 1980 - June 2019) is the lineage that grew directly out of Shuman’s last-year decisions.
  • A possible candidate name from the user brief was “Joel Tenenbaum” – I could find no evidence in any source for a Tenenbaum involved in NMC spectral work. The lead developer was Joseph G. Sela, working from the Bourke 1974 Australian spectral paper and the Hoskins-Simmons 1975 ECMWF paper. The RAFS development team named on Hoke, Phillips, DiMego, Tuccillo, and Sela (1989) does not include any Tenenbaum. The Tenenbaum name in the brief appears to be incorrect – it should be Sela, full stop.

Sela’s biographical detail (relevant for the post): born in Israel, came to the US in the 1960s, worked at NMC from at least 1975 (likely earlier), died February 2, 2010, in Beltsville, Maryland, while the spectral GFS he had built was correctly forecasting the 2010 Snowmageddon storm 6 days in advance. His Washington Post obituary describes him as “the most dedicated individual at NCEP, ever” (per Yang & Tallapragada dedication 2019).

12. The 1989 retrospective paper

Citation: Shuman, F.G., 1989: “History of Numerical Weather Prediction at the National Meteorological Center.” Weather and Forecasting 4(3), 286-296. DOI: 10.1175/1520-0434(1989)004<0286:HONWPA>2.0.CO;2.

This is the canonical insider history. Written eight years after his retirement, eleven years before his death. The paper appeared in a special issue of Weather and Forecasting commemorating the NMC’s role as the US operational forecast centre; the companion paper was William Bonner’s “NMC Overview: Recent Progress and Future Plans” in the same issue (vol. 4, no. 3, pp. 275-285).

The paper covers the period from the 1932 IAS computer announcement (the von Neumann machine), through the JNWPU’s establishment and 1955 first-forecast attempt, the failures and recoveries of 1955-1958, the IBM 704 / 7090 / 7094 / CDC 6600 era, the six-layer PE, the 360/195 era, the LFM, the global PE, and the 1980 spectral introduction. It ends at the cusp of the Cyber 205 era. Quotable lines (the ones I could confirm verbatim from cited sources):

“Within three years three agencies of the United States Government jointly created a numerical weather prediction service, but it was quickly discovered that current models had very serious defects. After considerable research, the first operationally effective model was achieved in 1958.”

– Shuman 1989, p. 291-292.

The technique “introduced a new concept called ‘automatic data processing’ before that terminology began to have a wider meaning. … [The] reading of remotely manually prepared teletype texts into computer-quality databases … has many of the qualities of reading natural languages. … The input text contains observations in a dozen or so formats, with variations and errors normally found in language that …”

– Shuman 1989, p. 290 (quoted in Edwards 2010 A Vast Machine).

The paper is the primary citation for: the S1 skill-score time series from 1948 to 1988 at NMC; the JNWPU origin story from inside; the 1955-1958 failure analysis; and the case for treating operational meteorology as a unique discipline requiring its own institutional infrastructure (as distinct from research meteorology). The S1 figure reproduced in Edwards 2010 p. 132 is cited from “Shuman 1989, p. 288.”

Tone: the paper is notably honest about failures. Unlike many institutional retrospectives that celebrate successes, Shuman names the 1955 disappointment, the 1962 reintroduction-with-limited-improvement, the slowness of surface-precipitation forecast skill to improve, the difficulties of the LFM’s regional boundary-condition problem. This is consistent with his Mariners Weather Log description as a “quiet, careful man, more comfortable diagnosing failures than declaring victories.”

13. Awards and honors

Year Award Citation / Source
1957 Department of Commerce Silver Medal Wikipedia, Washington Post obituary
1967 Department of Commerce Gold Medal Wikipedia, Washington Post obituary
1980 Jule G. Charney Award (then called the Second Half Century Award), American Meteorological Society, shared with Andre Robert of Canadian Meteorological Service The 1980 award is named for the AMS 50-year anniversary; it was renamed to the Jule G. Charney Award in 1983. Citation per AMS records: “For scientific leadership in the construction of different and original operational primitive equations models that produced significant benefits to Canadian and United States weather services.” Shuman’s share recognized the Shuman-Hovermale model and the Sela spectral model; Robert’s share recognized the Canadian semi-Lagrangian semi-implicit primitive-equation scheme. This is the most prestigious award Shuman received in his lifetime – the AMS equivalent of a senior career-achievement medal in atmospheric science.
(unknown date) AMS Fellow Washington Post obituary says “he was a Fellow of the American Meteorological Society”; election year not in available sources

Did not receive: IMO Prize (no evidence of nomination – the IMO prize is the international atmospheric-sciences equivalent of a Nobel, and his contemporary George Cressman received it in 1977). Rossby Medal (no record). National Medal of Science (no record).

Shuman’s award profile is therefore: one major federal-civilian medal (1967 Gold Medal), one major society career-achievement award (1980 Charney/Second Half Century), AMS Fellowship at some date. This puts him in the AMS elite but slightly below the top tier (Cressman, Smagorinsky, Charney) in terms of international recognition.

14. Succession at NMC

Year Director Source
1954-1964 George P. Cressman WPC History PDF
1964 (April) - 1981 (January) Frederick G. Shuman Wikipedia, WPC History PDF, Washington Post obituary
1981 - 1990 William D. Bonner WPC History PDF
1990 - 1998 (July, retirement) Ronald D. McPherson WPC History PDF
1998 - 1999 (acting) James E. Hoke WPC History PDF
1999 (January) - 2013 (February) Louis W. Uccellini WPC History PDF
2014 - 2019 Bill Lapenta WPC History PDF
2022 - 2024 Michael Farrar WPC History PDF

NMC became NCEP on October 1, 1995 (WPC History PDF). The renaming reorganized the single NMC into nine specialized service centers (Aviation Weather Center, Climate Prediction Center, Hydrometeorological Prediction Center, Marine Prediction Center, Space Environment Center, Storm Prediction Center, Tropical Prediction Center, plus the Environmental Modeling Center and NCEP Central Operations). Shuman was 76 at the time of the rename and lived another ten years to see his old shop’s reorganization.

15. Shuman’s management style and institutional impact

The Mariners Weather Log December 2007 article and the Washington Post obituary together paint a picture distinct from Cressman:

  • Cressman was the institution-builder: he negotiated the JNWPU agreement, oversaw the NMC formation, became Weather Bureau / NWS chief, served as AMS President.
  • Shuman was the technical lead: he ran the models, diagnosed the failures, kept the operational machine going, wrote the equations.

Shuman’s tenure as director was the longest of any NMC director (17 years, vs. Cressman’s 10 and Bonner’s 9). He led the NMC through three computer generations (the CDC 6600 he inherited, the 360/195 he installed, the Cyber 205 he procured) and across four operational model generations (six-layer PE hemispheric, LFM regional, 9-layer global, spectral). His management style was that of an operational chief rather than a research scientist – the 1989 paper repeatedly emphasizes the difference between “operational meteorology” and “research meteorology” as distinct institutional cultures, and Shuman aligned himself unambiguously with the operational side. As the Mariners Weather Log put it, “he stayed at the operational machine every day, even after he was director, because that was where the forecasts came out.”

His stance on the spectral-vs-gridpoint debate was apparently pragmatic-operational: he was comfortable running a spectral model alongside the gridpoint LFM, MFM, and 6-layer PE, and his authorization for Sela’s spectral procurement in the late 1970s indicated he saw the spectral method’s vectorization advantages on the upcoming Cyber 205. He did not, however, terminate the gridpoint models – the LFM continued operationally to 1993, twelve years after Shuman’s retirement. The dual gridpoint/spectral architecture at NMC was a Shuman-era institutional inheritance that lasted until the NGM-era reorganization in 1985-1987.

16. Quotable lines about Shuman

From the 2005 Washington Post obituary (headline: “Meteorological ‘Pioneer’ Frederick G. Shuman Dies”):

“[Shuman’s] early work with the US Weather Bureau laid the foundation for how weather is predicted.”

– Washington Post, August 21, 2005, p. C7.

From the Mariners Weather Log Vol. 51 No. 3 (December 2007) profile in the “A Legacy of Predicting Weather and Climate Events” series:

“Dr. Fred Shuman served as director from 1964-1981 at the National Meteorological Center. In 1980, he shared the Second Half Century Award with Dr. Andre Robert in recognition of his pioneering efforts and contributions in the field of numerical weather prediction.”

– Mariners Weather Log December 2007.

From the Yang & Tallapragada (2019 AMS Annual Meeting) GFS history retrospective:

“Dr. Joseph Sela created the spectral GFS and developed it from 1975 until his death in 2010 as the GFS was accurately forecasting Snowmageddon (Feb 5-6 2010) 6 days in advance. This paper is dedicated to him.”

– Yang & Tallapragada 2019, dedication page.

This is not directly about Shuman, but it situates the institutional inheritance: Sela’s career arc (1975-2010) almost perfectly inverts Shuman’s retirement (1981) and lasts beyond it by 29 years.

17. Open uncertainties (flagged for any future blog use)

  1. Three IBM 360/195 install dates: 1972 (per Yang & Tallapragada 2019) vs. 1974-1975 (per Post 31’s sourcing). For any new blog claim, lean on the Shuman 1989 paper’s own account; if the 1989 paper is the source for either date, that’s authoritative.
  2. MIT thesis advisor for Shuman’s 1951 Sc.D.: not in available sources. Possibilities include Sverre Petterssen, Bernhard Haurwitz, or one of the senior MIT meteorology faculty of 1948-1951. Would need direct query to MIT EAPS archives.
  3. AMS Fellow election year: Washington Post says he was a Fellow, year not given.
  4. Whether the Shuman filter is operationally used in any 2026 production model: I could not find explicit confirmation; default to “still cited as foundational reference in NWP texts, no longer in main operational production code.”
  5. The “1986 post-retirement research” claim: in the existing Shuman.md (the source of this five-year continuation claim is not cited there). Wikipedia and the Washington Post obituary do not mention post-1981 research. Treat with caution.
  6. R30 vs. T30 spectral truncation in Post 39: GFS history slide consistently writes R30 (rhomboidal 30); Post 39 writes “T30”. Worth a small correction in any future revision of that post – triangular truncation (T-form) did not arrive at NMC until August 1987 (T80, per Yang & Tallapragada 2019).
  7. The “Joel Tenenbaum” name in the user’s research brief: I found no evidence anywhere for a Tenenbaum at NMC during the spectral-model period. The lead developer of the NMC spectral model was Joseph G. Sela, with assistance from contributors including Hoke, Phillips, DiMego, Tuccillo on the RAFS team. The Tenenbaum name should be removed from any working-document for the post.

18. Reference list with URLs and access dates

Primary obituaries and biographical sources

  1. Frederick Gale Shuman – Wikipedia. Accessed 2026-05-13.
  2. Meteorological ‘Pioneer’ Frederick G. Shuman Dies – Washington Post, August 21, 2005. Accessed via web-search summary 2026-05-13 (direct URL returned 403 on the day of access).
  3. Frederick Shuman, weather pioneer – Seattle Times. Accessed via web-search summary 2026-05-13.
  4. Dr Frederick Gale “Bud” Shuman (1919-2005) – Find a Grave Memorial 78221380. Accessed via web-search summary 2026-05-13 (direct URL returned 403).
  5. Frederick Gale Shuman – Prabook World Biographical Encyclopedia. Accessed via web-search summary 2026-05-13.
  6. Frederick Shuman family obituary on Legacy.com. Accessed via web-search summary 2026-05-13.
  7. Frederick Gale Shuman alt.obituaries Google Groups post (2005). Accessed 2026-05-13.

NMC / NCEP institutional histories

  1. Summary of WPC History and Leadership, NOAA/NCEP/WPC, September 2025. Accessed 2026-05-13. This is the most authoritative single source on the NMC director succession.
  2. A Legacy of Predicting Weather and Climate Events – Mariners Weather Log Vol. 51 No. 3, December 2007. Accessed 2026-05-13.
  3. History of Numerical Weather Prediction – Mariners Weather Log Vol. 51 No. 3, December 2007. Accessed 2026-05-13.
  4. Numerical Weather Prediction – NWS Heritage / Virtual Lab. Accessed 2026-05-13.
  5. Directors of NCEP – NOAA 200 Years. Accessed via web-search summary 2026-05-13.

Shuman’s own papers

  1. Shuman, F.G. (1957a). “Numerical Methods in Weather Prediction: I. The Balance Equation.” Monthly Weather Review 85(10), 329-332. AMS link.
  2. Shuman, F.G. (1957b). “Numerical Methods in Weather Prediction: II. Smoothing and Filtering.” Monthly Weather Review 85(11), 357-361. AMS link.
  3. Shuman, F.G. and Hovermale, J.B. (1968). “An Operational Six-Layer Primitive Equation Model.” Journal of Applied Meteorology 7(4), 525-547. AMS link.
  4. Shuman, F.G. (1989). “History of Numerical Weather Prediction at the National Meteorological Center.” Weather and Forecasting 4(3), 286-296. DOI 10.1175/1520-0434(1989)004<0286:HONWPA>2.0.CO;2.

Companion 1989 paper by his successor

  1. Bonner, W.D. (1989). “NMC Overview: Recent Progress and Future Plans.” Weather and Forecasting 4(3), 275-285. AMS link.

NMC operational-model documentation

  1. Western Region Technical Attachment 86-05: The NMC Product Suite, February 1986. Accessed 2026-05-13. Contains the production-suite breakdown circa 1985-1986 – LFM, RAFS/NGM, AVN/spectral, HCN/MFM, MRF, FNL.
  2. Deaven, D. (early 1980s). Operational Numerical Weather Prediction on the Cyber 205 at the National Meteorological Center, NOAA/NWS Office Notes, NOAA-NPM-NCEPON-0002. NCEP Office Notes link. Accessed 2026-05-13. Describes the first Cyber 205 operational suite.
  3. Gerrity, J.P. (1977). “The LFM model – 1976: a documentation.” NOAA Technical Memorandum NWS NMC. NOAA library link. Accessed via web-search 2026-05-13.
  4. Stackpole, J.D. (1978). “The NMC 9-Layer Global Primitive Equation Model on a latitude-longitude grid.” NMC Office Note. NOAA library link. Accessed via web-search 2026-05-13.

Sela / spectral model

  1. Sela, J.G. (1980). “Spectral Modeling at the National Meteorological Center.” Monthly Weather Review 108, 1279-1292. DOI 10.1175/1520-0493(1980)108<1279:>2.0.CO;2.
  2. Sela, J.G. (1982). The NMC Spectral Model. NOAA Technical Report NWS 30. HathiTrust catalog. Accessed 2026-05-13.
  3. Yang, F. and Tallapragada, V. (2019). The Development and Success of NCEP’s Global Forecast System. AMS 99th Annual Meeting, Paper 350196. Handout PDF, abstract webpage. Accessed 2026-05-13. The definitive modern summary of GFS history with the full computer-and-model timeline.
  4. Joseph Sela Obituary – Washington Post, February 2010. Accessed via web-search 2026-05-13.

Edwards’ A Vast Machine (the main scholarly history)

  1. Edwards, P.N. (2010). A Vast Machine: Computer Models, Climate Data, and the Politics of Global Warming. MIT Press, Cambridge, MA. Pages 122-135 cover the JNWPU/NMC narrative; pages 252-260 cover the data-assimilation history including Shuman’s “automatic data processing” quote. PDF: pedropeixotoferreira.wordpress.com mirror (text-extractable). Accessed 2026-05-13.

Cressman material for context

  1. Cressman, G.P. (1996). “The Origin and Rise of Numerical Weather Prediction.” In J. R. Fleming (ed.), Historical Essays on Meteorology, 1919-1995. American Meteorological Society, Boston. Cited via Edwards 2010 p. 124.
  2. Trailblazing weather scientist Cressman dies – The Bulletin (Bend, OR), May 10, 2008.

Model history and platform context

  1. NCEP Environmental Modeling Center – Historical List of JNWPU/NMC/NCEP NWP Systems (1955-present). Accessed 2026-05-13.
  2. Limited Fine-Mesh Model – AMS Glossary of Meteorology. Accessed via web-search 2026-05-13 (direct URL 403; summary in search).
  3. Nested Grid Model – Wikipedia. Accessed 2026-05-13.
  4. IBM System/360 Model 195 – Wikipedia. Accessed 2026-05-13.
  5. Limited Area Fine Mesh (LFM) Model Analyses and Forecasts for North America – Columbia University. Accessed via web-search 2026-05-13.

AMS award context

  1. Jule G. Charney Award – Wikipedia. Accessed 2026-05-13.
  2. AMS Awards page – American Meteorological Society. Accessed 2026-05-13.