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.

ECMWF: founders, directors, and early scientists, 1973-1990

Notes for a long-form blog post on the European Centre for Medium-Range Weather Forecasts and the people who built it during its first two decades. Convention signed 1973; came into force 1 November 1975; first real-time forecasts June 1979; first operational forecasts 1 August 1979 on a Cray-1A; first ensemble forecast 24 November 1992. Eighteen founding member states. Headquarters at Shinfield Park, Reading, Berkshire, United Kingdom.

The ECMWF Convention was first drafted in December 1971; thirty-two senior representatives from fourteen of the eighteen founding states attended the meeting on 9 and 10 December 1971. The United Kingdom won the bid to host the Centre, in part because of the proximity of Shinfield Park to the UK Met Office at Bracknell and to the University of Reading’s Department of Meteorology. The eighteen founding states of 1975 were Austria, Belgium, Denmark, Finland, France, Germany, Greece, the Republic of Ireland, Italy, Luxembourg, the Netherlands, Norway, Portugal, Spain, Sweden, Switzerland, Turkey, and the United Kingdom – effectively all of non-communist continental Europe plus Britain and Ireland.1

This was a profoundly unusual institution to build in 1973. The first Cray-1 had not been delivered yet. The European Space Agency was still the European Space Research Organisation. The European Economic Community had nine member states. The atmospheric general circulation model was still mostly a North American technology. ECMWF was the first European intergovernmental scientific institution to commit, at its founding, to producing operational forecasts on the world’s most expensive supercomputers, with no operational mandate inside any single national meteorological service to fall back on if the experiment failed. The decision to do this was largely Aksel Wiin-Nielsen’s.

Aksel Wiin-Nielsen (1924-2010)

Aksel Christopher Wiin-Nielsen was born on 17 December 1924 in Denmark and died on 26 April 2010 in Copenhagen at the age of eighty-five. (Some sources note Frederiksberg as the place of birth; the Wikipedia entry simply says “Denmark.” The often-cited birth date “26 April 1924” is incorrect; he died on his eighty-fifth birthday plus four months, not on his birthday.) He was the first Director of the European Centre for Medium-Range Weather Forecasts and arguably the most consequential European meteorological administrator of the twentieth century.2

Wiin-Nielsen took his master’s degree at the University of Copenhagen in 1952 and became scientific assistant to Ragnar Fjørtoft – one of the original Charney-Fjørtoft-von Neumann ENIAC barotropic forecasting team of 1950 – in the Department of Geophysics. Through Fjørtoft he was drawn into the new International Meteorological Institute in Stockholm that Carl-Gustaf Rossby had set up in 1947, and at which Bert Bolin and Norman Phillips and a generation of young Scandinavian-American numerical meteorologists were working through the early 1950s. In 1955 Wiin-Nielsen participated in what is now considered the first numerical weather prediction that completed its computation ahead of the time for which the forecast was made – a milestone in operational NWP, achieved on the Stockholm BESK computer. In 1959 he moved to the United States, where for fifteen years he worked first at the Joint Numerical Weather Prediction Unit (the trilateral civilian-military joint operation of the U.S. Weather Bureau, U.S. Air Force, and U.S. Navy) at Suitland, Maryland, then in 1961 at the National Center for Atmospheric Research in Boulder, Colorado, where he worked on the general circulation of the atmosphere.3

Wiin-Nielsen left NCAR in 1963 – the suddenness of the move noted by NCAR colleagues at the time – to become the first Chair of the Department of Meteorology and Oceanography at the University of Michigan in Ann Arbor, where he taught and did research for the next ten years. By the late 1960s he was one of the most visible American-based atmospheric scientists in international meteorological politics, well-connected to the WMO, to ICSU, and to what would become the ECMWF planning committees. Through the European Cooperation in Science and Technology (COST) framework, the eighteen founding states selected Wiin-Nielsen as the first Director of the European Centre for Medium-Range Weather Forecasts in 1973, and he took office on 1 January 1974 – before the Convention had even come into force, before the building at Shinfield Park existed, before any computer had been ordered.

His five-year directorship – 1 January 1974 to 31 December 1979 – was the institutional substance of ECMWF. He hired the first staff. He selected the Cray-1A as the centre’s first supercomputer (a decision that committed ECMWF to the Cray vector philosophy from the outset and put the centre on the same procurement track as NCAR, GFDL, and the National Meteorological Center). He oversaw the construction of Shinfield Park and the first generation of the operational NWP system, which made its first real-time medium-range forecasts in June 1979 and entered operational service on 1 August 1979 – six months before he stepped down. The famous quote often attributed to him – some version of “I would consider it sound finance to spend ECMWF’s entire annual budget on a single forecast that proved useful” – has not been definitively traced to a primary source, but reflects the institutional culture he established.

In 1979 the eighth WMO Congress appointed Wiin-Nielsen as Secretary-General of the World Meteorological Organization, the third holder of that office, and he served from 1 January 1980 to 31 December 1983. After his WMO term he returned to Denmark, becoming Director of the Danish Meteorological Institute (DMI) from 1984 to 1987. In 1987 he was elected President of the ECMWF Council – effectively returning to oversee the institution he had founded. From 1987 he was Professor of Physics at the University of Copenhagen, and Professor Emeritus from 1995. He served as President of the European Geophysical Society from 1990 to 1992. He published more than one hundred peer-reviewed papers; his textbook Fundamentals of Atmospheric Energetics (with Tsing-Chang Chen, Oxford University Press, 1993) remains a standard graduate-level reference.4

His honours included the Buys Ballot Medal of the Royal Netherlands Academy of Arts and Sciences (1982), the Wihuri International Prize (1983), Fellow and Honorary Member of the American Meteorological Society, and Member of the Royal Swedish Academy of Sciences. He received the International Meteorological Organization Prize (the WMO’s highest honour) posthumously in 2011, the year after his death.

The obituary in Eos (the AGU’s weekly news magazine) for 26 October 2010 was written by Lennart Bengtsson – the man who had succeeded Wiin-Nielsen as ECMWF Director and worked alongside him for nearly forty years. The two had first met in Stockholm in the early 1960s when Bengtsson was a graduate student under Bert Bolin, and they had remained close professional friends across continents and decades.

Lennart Bengtsson (1935-)

Lennart Olof Bengtsson, born 5 July 1935 in Trollhättan, Älvsborgs län, Sweden, is alive at ninety years of age (his ninetieth birthday in July 2025 was marked by celebrations at the Max Planck Institute for Meteorology in Hamburg). He was the second Director of ECMWF – and arguably the scientist who turned the institution from Wiin-Nielsen’s promising start into the dominant operational NWP centre in the world.5

Bengtsson was educated at Uppsala University and completed his doctorate at Stockholm University in the late 1950s under the wing of Bert Bolin and the International Meteorological Institute that Rossby had founded a decade earlier. From 1961 to 1974 – thirteen years – he worked at the Swedish Meteorological and Hydrological Institute (SMHI), where he built the Swedish national operational NWP service almost from scratch. (The slight ambiguity in dates between sources – some give 1962, others 1961 – reflects the standard difficulty of pinning down post-doctoral start dates in the Nordic university system.) During this period he published influential work on data assimilation, four-dimensional analysis, and the use of variational methods in the initial-condition problem – ideas that would directly feed into ECMWF’s data-assimilation programme a decade later.

Bengtsson joined ECMWF on 1 January 1975, four months before the Convention came into force, as Head of Research. He served in that capacity until 1981, then was promoted to Director of ECMWF on 1 January 1982 – succeeding Wiin-Nielsen’s interim successor – and served as Director until 31 December 1990. Under his directorship, ECMWF achieved the operational milestones that defined its first generation: the transition from grid-point to spectral models in April 1983 (T63, replacing the original N48 grid-point model); the implementation of the first experimental reanalysis project, which would become the ERA-15 reanalysis; and the development of the predictability and ensemble-forecasting research that, by 1992 under Tim Palmer’s leadership, would yield the first operational ensemble prediction system.6

After leaving ECMWF in 1990 Bengtsson moved to Hamburg as Director of the Max Planck Institute for Meteorology from 1991 to 2000, where his most consequential decision was to take ECMWF’s Integrated Forecasting System – the operational ECMWF model – and use it as the foundation for the ECHAM climate model series. ECHAM became the dominant European climate model for the IPCC simulations of the 1990s and 2000s and the basis of much of European climate science to the present day. After Hamburg, from 2000 onwards, Bengtsson became Senior Research Fellow at the Environmental Systems Science Centre at the University of Reading; from 2008 he served as Director of the International Space Science Institute in Bern, Switzerland.

His major awards include the Milutin Milanković Medal of the European Geosciences Union (1996), the René Descartes Prize for Collaborative Research (2005), the WMO International Meteorological Organization Prize (the IMO Prize, 2006), the Alfred Wegener Medal & Honorary Membership of the European Geosciences Union (2009), and Honorary Fellow of the Royal Meteorological Society (2009). The often-cited “2014 Royal Astronomical Society Gold Medal” attribution appears to be incorrect; the 2014 RAS Gold Medal for Geophysics went to John Zarnecki, not to Bengtsson, and no record of a Bengtsson RAS Gold Medal has been confirmed.7

A controversial public episode in 2014 – when Bengtsson briefly joined the board of the Global Warming Policy Foundation, a UK climate-policy think tank that has been criticised as a vehicle for climate skepticism, and resigned shortly thereafter – briefly disrupted his public profile and produced a flurry of media attention. Bengtsson himself stated at the time that he did not believe in any “systematic cover-up” of climate research and that his views remained mainstream. The episode is mentioned here for completeness, not because it has any bearing on his foundational scientific contributions to NWP and climate modelling, which are universally acknowledged.

The Bengtsson-Wiin-Nielsen partnership at ECMWF – six years overlapping, with Bengtsson as Head of Research under Wiin-Nielsen 1975-1979 and the institutional baton passing in 1981-1982 – is the structural backbone of how ECMWF was built. They were friends, they were colleagues, they were both products of the Stockholm International Meteorological Institute and the Rossby-Bolin-Phillips lineage, and between them they ran ECMWF for the first sixteen years of its existence.

William (“Bill”) Bourke

William (“Bill”) Bourke was an Australian meteorologist whose work on spectral methods for global atmospheric models in the early 1970s was foundational to the entire global NWP enterprise – including, eventually, ECMWF’s transition to a spectral model in 1983. The full biographical details of Bourke – his birth date and place, his education at the University of Melbourne or elsewhere – are surprisingly difficult to extract from the primary literature, but his scientific contribution is canonical.8

In the early 1970s Bourke was a research scientist at the Commonwealth Meteorology Research Centre (CMRC) in Melbourne, Victoria. The CMRC was a joint Bureau of Meteorology / CSIRO research entity that, after a 1973 review, became the Australian Numerical Meteorology Research Centre (ANMRC) in 1974. Bourke worked at ANMRC throughout its operational life (1974-1984), serving as Acting Officer-in-Charge in 1983-1984. After ANMRC was reorganised in 1985 into the Bureau of Meteorology Research Centre (BMRC), Bourke continued there for the rest of his career, retiring in the early 2000s. He presented “History of NWP in Australia – 1970 to the Present” at the 16th BMRC Modelling Workshop in Melbourne in December 2004 – effectively his own retrospective on the field he had helped to define.

Bourke’s two foundational papers were both published in the U.S. Monthly Weather Review. The first – “An Efficient, One-Level, Primitive-Equation Spectral Model” (volume 100 issue 9, 1972, pages 683-689) – formulated the global primitive equations in spectral form using prognostic equations for vorticity and divergence, with a grid-transform technique to evaluate nonlinear terms. The model conserved energy, angular momentum, and squared potential vorticity to a high degree. The second – “A Multi-Level Spectral Model. I. Formulation and Hemispheric Integrations” (volume 102 issue 10, 1974, pages 687-701) – extended the work to a multi-level baroclinic model suitable for hemispheric simulation, using spectral-grid transform procedures and an efficient semi-implicit time integration scheme. Bourke and colleagues then collaborated on “Global Modeling of Atmospheric Flow by Spectral Methods” in the volume General Circulation Models of the Atmosphere, edited by Julius Chang (Academic Press, 1977), which became the canonical reference for the spectral approach.9

The technical importance of these papers is that they made global primitive-equation NWP practical at production resolution. Spectral methods – representing fields as truncated series of spherical harmonics rather than as values on a finite-difference grid – have two enormous advantages over grid-point methods on the sphere: (1) they handle the convergence of meridians at the poles automatically, eliminating one of the most stubborn classes of numerical instability in grid-point models, and (2) they make the dispersive properties of the discrete equations exactly match those of the continuous equations for resolved scales. The dispersion-error advantage in particular meant that for medium-range prediction (5-10 days), spectral models needed roughly half the resolution of grid-point models for the same forecast skill – a roughly four-times-cheaper computational cost.

The connection from Bourke’s 1972-1974 Australian work to ECMWF runs as follows. The first ECMWF operational model in 1979 was a grid-point model (N48 resolution, 15 vertical levels, 210 km grid spacing, running on the Cray-1A). By 1981-1982 the research department under Bengtsson was actively pursuing a transition to spectral methods, drawing directly on Bourke’s hemispheric work and on the related theoretical work of Steven Orszag, Eliasen-Machenhauer-Rasmussen, and Brian Hoskins. The T63 spectral model that ECMWF brought operational in April 1983 was the first European operational spectral NWP model, and was – in lineage if not in code – a direct descendant of Bourke’s 1974 Melbourne formulation. Joseph Sela’s parallel adoption of spectral methods at the National Meteorological Center in Suitland, Maryland from 1980 onwards followed the same lineage.

The Bourke-to-ECMWF connection illustrates a recurring theme in late-twentieth-century atmospheric science: a methodological breakthrough at a remote national institution (Melbourne, Australia) becomes, within a decade, the operational standard at an international centre half a world away. The technology transfer was effected through the standard apparatus of academic publication plus visiting-scientist exchanges plus WMO scientific conferences – not through any formal bilateral arrangement.

Anthony (“Tony”) Hollingsworth (1943-2007)

Anthony Hollingsworth, known universally as Tony, was born on 6 July 1943 in Dublin, Ireland, and died suddenly on 29 July 2007 at age sixty-four while on holiday in his native Ireland. He was buried in the old Franciscan Abbey cemetery in Claregalway, and is survived by his wife Breda and their children Cormac and Deirdre. Hollingsworth was the longest-serving staff member in ECMWF’s history and one of the institution’s most influential research scientists across more than three decades.10

Hollingsworth graduated with first-class honours in mathematics and mathematical physics from University College Cork in 1964 – not Trinity College Dublin as some sources mistakenly state. He joined what was then the Irish Meteorological Service (now Met Éireann) on a cadet scheme intended to develop professional meteorologists, and worked initially as a forecaster. From the late 1960s he took leave from Met Éireann to pursue PhD studies at the Massachusetts Institute of Technology – the standard graduate destination for atmospheric scientists of that generation, the home of Charney, Lorenz, and the new generation of MIT-trained dynamicists. After his PhD, Hollingsworth returned to Europe to take a research position in the UK Universities’ Atmospheric Modelling Group (UMAMG) at the University of Reading – the same institution physically adjacent to the Shinfield Park site that ECMWF was beginning to occupy in 1974-1975.

Hollingsworth joined ECMWF on 1 March 1975 – before the Convention had come into force, as one of the founding scientific staff. He stayed for thirty-two years, until his sudden death in 2007. During those decades he led, in turn, the Physical Aspects Section, the Data Division, and the Model Division of the ECMWF Research Department. He was appointed Head of Research in 1991 – succeeding Bengtsson, who had left for the Max Planck Institute – and Deputy Director in 1995. He stepped down from these executive positions on his sixtieth birthday in 2003, but rather than retire he led the Europe-wide GEMS (Global and Regional Earth-System Monitoring using Satellite and in-situ data) project until his death.

Hollingsworth’s most cited single technical contribution is the Hollingsworth-Lönnberg method of 1986 for analysing observation and analysis quality in NWP data assimilation systems. The method is based on the simple but powerful observation that the differences between observations and the background forecast are a combination of background-forecast error and observation error. By computing histograms of innovation covariances stratified by spatial separation – and assuming that observation errors are spatially uncorrelated and that observation and background errors are uncorrelated – one can fit a model of the background covariance directly from operational observation data. The two key 1986 papers were Hollingsworth, Shaw, Lönnberg, Illari, Arpe, and Simmons in the Monthly Weather Review (volume 114, pages 861-879) and Hollingsworth and Lönnberg in Tellus 38A, pages 111-136. The method became the canonical approach at ECMWF and at every other major NWP centre, and continues in modified form to the present day for diagnosing background-error statistics from observational data.11

Hollingsworth received the Jule G. Charney Award of the American Meteorological Society and an honorary D.Sc. from University College Cork. The fact that he died at sixty-four, suddenly, at the height of his second post-executive career, was a substantial blow to the European NWP community. His successor as Head of Research at ECMWF was Anthony Hollingsworth’s protégé and long-time collaborator, Adrian Simmons.

Adrian Simmons

Adrian Simmons – one of ECMWF’s most distinguished scientists and the recipient of the Vilhelm Bjerknes Medal of the European Geosciences Union in 2012 – joined ECMWF in 1978 and remained there for more than four decades. He has been described in EGU citations as a researcher whose work made “a crucial contribution to the great success of ECMWF” across forty years of dynamical-meteorology and numerical-weather-prediction research.12

Simmons began his career with PhD work at the University of Cambridge on dynamical studies of planetary-scale disturbances in the stratosphere. He moved to the University of Reading to work on the development of the spectral atmospheric model that would feed directly into ECMWF’s 1983 transition. His foundational research at Reading included fundamental studies of linear and non-linear baroclinic instability on the sphere – pure dynamical meteorology of the kind that Charney and Eady had defined in the late 1940s, but extended to spherical geometry – and he became, before he was thirty years old, one of the world’s authorities on the spectral formulation of atmospheric dynamics.

When ECMWF began hiring research staff in 1975-1978, Simmons joined the new Research Department in 1978 and immediately began work on the second-generation operational model that would replace the original N48 grid-point system. His technical contributions over four decades at ECMWF were broad and consequential: comparisons of spectral and finite-difference simulations; development of energy- and angular-momentum-conserving vertical finite-difference schemes; introduction of hybrid sigma-pressure vertical coordinates (the so-called “Simmons-Burridge hybrid coordinate,” now standard in essentially every operational NWP system in the world); the treatment of orography in spectral models; the study of error growth and predictability; developments of the semi-implicit time-integration method; the use of reduced Gaussian grids in spectral models. Each of these contributions individually would be enough to define a career in atmospheric dynamics; together they constitute the bulk of the technical foundation of post-1983 ECMWF.

After Hollingsworth’s death in 2007, Simmons led the MACC (Monitoring Atmospheric Composition and Climate) project as the European Union flagship effort in atmospheric composition monitoring, succeeding Hollingsworth’s GEMS programme. He was a co-author on the ERA-Interim reanalysis publication of 2011 and on the ERA5 global reanalysis of 2020 – arguably the two most influential atmospheric reanalysis products of the modern era. Hans Hersbach’s acknowledgements in the ERA5 paper specifically credit Simmons (alongside Dick Dee) with “expert help” in development. Simmons retired from ECMWF in approximately 2017-2018 and was honoured at a 2017 ECMWF Symposium attended by Brian Hoskins and other senior figures from the international NWP community.13

Tim Palmer (1952-)

Timothy Noel Palmer, born 31 December 1952 in Kingston upon Thames, Surrey, England, is the founding architect of the ECMWF Ensemble Prediction System – arguably the single most influential operational innovation in NWP since the original spectral models of the 1970s. He is now Royal Society Research Professor and Professor of Climate Physics at the University of Oxford’s Clarendon Laboratory.14

Palmer took a first-class joint honours degree in mathematics and physics at the University of Bristol, followed by a doctorate in General Relativity Theory at Oxford University – a curious starting point for a meteorologist. He had a job offer to work in Cambridge with Stephen Hawking after completing his PhD; the offer was for further work in fundamental general relativity. Palmer turned it down, on the grounds (as he later put it himself) that the work was very abstract and didn’t impact people’s lives at all. He chose instead to pursue climate physics at the UK Met Office, joining the High Atmosphere Branch in the late 1970s. The “After all, I decided I’d rather do something useful” line is the title of his autobiographical chapter in the 2025 Springer volume Tim Palmer: Mathematical Physicist, Climate Dynamicist, Poet, Band Leader.

At the Met Office, Palmer worked initially on the dynamics of monthly forecasting – a group there had been doing statistical-empirical monthly prediction, calculating probabilities of different weather patterns; Palmer’s mandate was to introduce dynamical models to the long-range forecasting problem. This reformulation – from statistical to dynamical monthly forecasting – was the conceptual foundation for everything Palmer would later do at ECMWF.

Palmer joined ECMWF in 1986 to lead the newly formed Predictability and Diagnostics Division. He spent twenty-five years at ECMWF, until 2011. His central contribution was the development – with Roberto Buizza, Franco Molteni, Robert Mureau, Stefano Tibaldi, Joe Tribbia, and (later) Jan Barkmeijer – of the medium-range Ensemble Prediction System that became operational at ECMWF on 24 November 1992. The technical foundation of the EPS is the singular vector method (drawing on Brian Farrell’s 1982 work): instead of running many forecasts from random perturbations of the initial state, the EPS computes the “singular vectors” – the directions in initial-condition space along which forecast errors grow most rapidly over a 48-hour optimisation window – and seeds the ensemble from those most-amplifying directions. The result is an ensemble that captures, with as few as 50 members, the bulk of the predictability uncertainty for the medium range. The 1992 ECMWF Technical Memorandum No. 188 (Palmer, Molteni, Mureau, Buizza, Chapelet, Tribbia, “Ensemble Prediction”) is the canonical reference; the Buizza et al. 1993 paper on primitive-equation singular vectors implemented the method in IFS code; the Molteni et al. 1996 paper documented the operational system.15

Palmer’s second major conceptual contribution was the introduction of stochastic parametrisation into ensemble forecasting, beginning with a 1995 talk and extended in his subsequent papers. The argument is that the parametrisation of sub-grid-scale processes in NWP models – convection, turbulence, radiation – is intrinsically stochastic rather than deterministic, because it represents the unresolved fast scales by their statistical effect on the resolved slow scales. Adding controlled noise to the parametrisation tendencies, calibrated against high-resolution simulations, captures part of the unresolved-scale uncertainty and reduces the systematic biases that conventional deterministic parametrisations introduce. The Palmer group’s “Stochastically Perturbed Parametrisation Tendencies” (SPPT) scheme, in operational use at ECMWF since the early 2000s, is the practical realisation of this idea.

Palmer left ECMWF in 2011 to take up a Royal Society Research Professorship at Oxford as one of the Royal Society’s 350th Anniversary Research Professors. He is a Professorial Fellow of Jesus College, Oxford, and co-director of the Oxford Martin Programme on Modelling and Predicting Climate. He was elected Fellow of the Royal Society in 2003 and appointed Commander of the Order of the British Empire (CBE) in the 2015 New Year Honours. His major awards include the Buys Ballot Medal of the Royal Netherlands Academy (2003), the WMO International Meteorological Organization Prize (2007 – in some sources this is given as 2014), the Carl-Gustaf Rossby Research Medal of the AMS (2010), the Dirac Gold Medal (2014), and the Royal Astronomical Society Gold Medal for Geophysics (2023). His 2022 trade book The Primacy of Doubt: From Quantum Physics to Climate Change, How the Science of Uncertainty Can Help Us Understand Our Chaotic World (Oxford University Press) brings the philosophical lessons of ensemble forecasting to a general audience.16

Mike Cullen

Michael (“Mike”) Cullen has spent most of his career at the UK Met Office, with extended secondments to ECMWF and the University of Reading. He is the architect of much of the modern Met Office Unified Model and the leading mathematical theorist of the semi-geostrophic equations in atmospheric dynamics. His role in the ECMWF story is adjacent rather than central, but the technical and mathematical bridge between the two institutions runs largely through him.17

Cullen joined the Met Office as a research scientist in 1970 and through the period to 1988 was responsible for developing the numerical methods used in the operational weather forecasting models – the predecessors of the Unified Model. From 1987 to 1991 he designed both the software formulation and the integration scheme for the UK Met Office Unified Model, the integrated weather-and-climate prediction system that since 1991 has been the operational backbone of UK forecasting. The Unified Model’s dynamical core uses semi-implicit semi-Lagrangian time-stepping – a scheme that eliminates the stability constraints on numerical model integration and dramatically improves computational efficiency. The 2003 update to a fully compressible non-hydrostatic dynamical core, and the later ENDGame core, all rest on Cullen’s foundational work.

Cullen’s purely mathematical work on semi-geostrophic theory – the simplified version of the primitive equations valid at large horizontal scales, which exhibits singular fronts that can be analysed using ideas from Monge-Kantorovich optimal transport – is summarised in his 2006 monograph A Mathematical Theory of Large-Scale Atmospheric Flow (Imperial College Press). The convexity principle due to Cullen and R. James Purser (the U.S. NCEP scientist) requires that simplified solutions describing large-scale phenomena must be stable to parcel displacements; this convexity property is the basis of proofs of existence of solutions to the semi-geostrophic equations. Cullen is a Visiting Professor at the University of Reading and supervises PhD students at Imperial College London, the University of Bath, and Reading. He is the Met Office representative on the ECMWF Scientific Advisory Committee for much of the 1990s and 2000s.

David Burridge (third Director, 1991-2004)

David Martin Burridge succeeded Bengtsson as the third Director of ECMWF on 1 January 1991 and served until 2004 – a thirteen-year directorship that bridged the centre’s transition from a primarily-European research institution to a global operational prediction centre. Burridge graduated from Bristol University with a first-class degree in mathematics and completed a PhD in fluid dynamics. After a year teaching at Florida University, he joined the UK Met Office in 1970 as a research scientist. He transferred to ECMWF in 1975 – one of the founding scientific staff alongside Hollingsworth and Bengtsson – and worked there for the next twenty-nine years. The “Simmons-Burridge hybrid coordinate” mentioned earlier is the joint product of his collaboration with Adrian Simmons. After his retirement in 2004, Burridge led the international THORPEX programme of WMO – the global research programme on improving high-impact weather forecasting. He received the European Meteorological Society Silver Medal in 2010.18

Erland Källén (Director of Research, 2009-2017)

Erland Källén obtained his PhD from Stockholm University in 1980 and has been Professor of Dynamic Meteorology at Stockholm University since 1996 (Professor Emeritus from 2022). He held an early position at ECMWF before returning to Stockholm, and joined ECMWF for a second time in July 2009 as Director of Research, succeeding the late Tony Hollingsworth and Hollingsworth’s interim successor. Källén served as Director of Research at ECMWF from 2009 to 2017 – eight years – before resuming his Stockholm University position in July 2017. He served as Director of the Centre for Climate Research Singapore from 2018 to 2020. The achievement Källén himself most often cites from his ECMWF tenure is the centre’s nine-day forecast of Hurricane Sandy in 2012, which gave US authorities unprecedented advance warning of the late-October landfall.19

Florence Rabier (Director-General, 2016-2025)

Florence Rabier, born 1964 in France, was the eighth Director-General of ECMWF, serving from 1 January 2016 to mid-2025 – a nine-year tenure that made her the longest-serving female director of any major European intergovernmental scientific institution to date.20

Rabier joined Météo-France as an undergraduate engineer at the École nationale de la météorologie. She took a master’s degree in meteorology from the Toulouse III - Paul Sabatier University and a doctorate from the University of Paris (Pierre and Marie Curie University) in 1992 – her dissertation was on variational assimilation of meteorological observations in the presence of baroclinic instabilities. From 1992 to 1998 she worked at ECMWF for the first time, where she developed the four-dimensional variational data assimilation method (4D-Var) that has been at the core of every modern operational data assimilation system since. Her 1997 implementation of operational 4D-Var at ECMWF was the first worldwide – a milestone she frequently cites as the highlight of her career. From 1998 to 2013 she worked at Météo-France in Toulouse, progressing from Head of the Observations team to Deputy Director of GMAP (the French data-assimilation group) and specialising in satellite-observation assimilation.

Rabier returned to ECMWF in 2013 as Director of Forecasts, and in 2016 was appointed Director-General. She was appointed Knight of the Legion of Honour in 2014; elected Honorary Member of the American Meteorological Society in 2022; elected to the French Academy of Technologies in 2024; received the European Meteorological Society Silver Medal in 2025; and was elected President of the European Meteorological Society in 2026. From 2021 to 2023 she chaired the European Space Agency Advisory Committee on Earth Observations.

How the people fit together

The institutional structure of ECMWF’s first sixteen years can be sketched as a sequence of overlapping handovers, all rooted in two intellectual lineages: the Stockholm International Meteorological Institute under Bert Bolin (Wiin-Nielsen 1955; Bengtsson 1960s; Källén 1980), and the Cambridge-Reading-MIT axis of UK and Irish atmospheric dynamics (Hollingsworth at MIT then Reading; Simmons at Cambridge then Reading; Cullen at Met Office Bracknell; Burridge at Met Office). The original 1975 founding cohort – Wiin-Nielsen, Bengtsson, Hollingsworth, Burridge, joined by Simmons in 1978 – ran ECMWF from 1975 through approximately 2004 with extraordinary continuity. The next generation – Palmer joining in 1986, Källén returning in 2009, Rabier first in 1992 – inherited the institution and built on it.

The geographical pattern is striking: the founding generation was overwhelmingly Scandinavian (Wiin-Nielsen Danish, Bengtsson Swedish, Källén Swedish, with Bolin in the background) and Anglo-Irish (Hollingsworth Irish, Burridge English, Simmons English, Cullen English, Palmer English). The Mediterranean countries that founded ECMWF – France, Italy, Spain, Greece – supplied much of the day-to-day operational and political leadership but, until Rabier’s elevation in 2016, none of the directors. The pattern of “Scandinavian intellectual stewardship of a British-hosted European institution” is one of the more distinctive features of late-twentieth-century European meteorology, and its roots reach back through Bert Bolin to Carl-Gustaf Rossby’s 1947 Stockholm institute, and through Ragnar Fjørtoft to Vilhelm Bjerknes’s 1917 Bergen School.

Footnotes

Selected references

  • Bengtsson, L. “Aksel Wiin-Nielsen (1924-2010),” Eos, Transactions of the American Geophysical Union, vol. 91 no. 43, 26 October 2010, p. 396.
  • Bourke, W. “An Efficient, One-Level, Primitive-Equation Spectral Model,” Monthly Weather Review, 100, 683-689, 1972.
  • Bourke, W. “A Multi-Level Spectral Model. I. Formulation and Hemispheric Integrations,” Monthly Weather Review, 102, 687-701, 1974.
  • Bourke, W., B. McAvaney, K. Puri, and R. Thurling. “Global Modeling of Atmospheric Flow by Spectral Methods,” in J. Chang (ed.), General Circulation Models of the Atmosphere, Methods in Computational Physics vol. 17, Academic Press, 1977, pp. 267-324.
  • Cullen, M.J.P. A Mathematical Theory of Large-Scale Atmospheric Flow, Imperial College Press, 2006.
  • Dee, D.P., S.M. Uppala, A.J. Simmons, et al. “The ERA-Interim reanalysis: Configuration and performance of the data assimilation system,” Quarterly Journal of the Royal Meteorological Society, 137, 553-597, 2011.
  • Hersbach, H., et al. “The ERA5 global reanalysis,” Quarterly Journal of the Royal Meteorological Society, 146, 1999-2049, 2020.
  • Hollingsworth, A., D.B. Shaw, P. Lönnberg, L. Illari, K. Arpe, and A.J. Simmons. “Monitoring of observation and analysis quality by a data assimilation system,” Monthly Weather Review, 114, 861-879, 1986.
  • Hollingsworth, A. and P. Lönnberg. “The statistical structure of short range forecast errors as determined from radiosonde data. Part I: The wind field,” Tellus, 38A, 111-136, 1986.
  • Molteni, F., R. Buizza, T.N. Palmer, and T. Petroliagis. “The ECMWF Ensemble Prediction System: Methodology and validation,” Quarterly Journal of the Royal Meteorological Society, 122, 73-119, 1996.
  • Palmer, T.N. The Primacy of Doubt: From Quantum Physics to Climate Change, How the Science of Uncertainty Can Help Us Understand Our Chaotic World, Oxford University Press, 2022.
  • Palmer, T.N. “The ECMWF ensemble prediction system: Looking back (more than) 25 years and projecting forward 25 years,” Quarterly Journal of the Royal Meteorological Society, doi:10.1002/qj.3383, 2019.
  • Palmer, T.N., F. Molteni, R. Mureau, R. Buizza, P. Chapelet, and J. Tribbia. “Ensemble Prediction,” ECMWF Technical Memorandum No. 188, 1992.
  • Simmons, A.J. and D.M. Burridge. “An energy and angular-momentum conserving vertical finite-difference scheme and hybrid vertical coordinates,” Monthly Weather Review, 109, 758-766, 1981.
  • Wiin-Nielsen, A. and T.-C. Chen. Fundamentals of Atmospheric Energetics, Oxford University Press, 1993, 400 pp.

  1. ECMWF founding history. Convention signed 1973, came into force 1 November 1975. First operational forecasts 1 August 1979 on Cray-1A. First ensemble forecast 24 November 1992. Sources: ECMWF “History” page https://www.ecmwf.int/en/about/who-we-are/history; ECMWF “Key facts and figures” https://www.ecmwf.int/en/about/media-centre/key-facts-and-figures; Wikipedia “European Centre for Medium-Range Weather Forecasts” https://en.wikipedia.org/wiki/European_Centre_for_Medium-Range_Weather_Forecasts. ↩

  2. Wiin-Nielsen biographical: died 26 April 2010 in Copenhagen, age 85. Born 17 December 1924 in Denmark (Frederiksberg per some sources, simply “Denmark” in Wikipedia). Wikipedia “Aksel C. Wiin-Nielsen” https://en.wikipedia.org/wiki/Aksel_C._Wiin-Nielsen; ECMWF obituary 2010 https://www.ecmwf.int/en/about/media-centre/news/2010/aksel-wiin-nielsen-passed-away. ↩

  3. Wiin-Nielsen at Stockholm 1952-1959 under Fjørtoft; first NWP completed ahead of forecast time at Stockholm BESK 1955; Joint Numerical Weather Prediction Unit Suitland MD 1959-1961; NCAR 1961-1963; University of Michigan 1963-1973 as first Chair of Department of Meteorology and Oceanography. Sources: as above. ↩

  4. Wiin-Nielsen ECMWF 1974-1979; WMO Secretary-General 1980-1983; Danish Meteorological Institute Director 1984-1987; ECMWF Council President 1987; University of Copenhagen Professor of Physics 1987-1995. Awards: Buys Ballot Medal 1982, Wihuri International Prize 1983, IMO Prize 2011 (posthumous). Bengtsson, L., “Aksel Wiin-Nielsen (1924-2010),” Eos, vol. 91 no. 43, 26 October 2010, p. 396. ↩

  5. Bengtsson biographical: born 5 July 1935, Trollhättan, Älvsborgs län, Sweden. 90th birthday celebrated July 2025 at Max Planck Institute for Meteorology Hamburg. Sources: Wikipedia “Lennart Bengtsson” https://en.wikipedia.org/wiki/Lennart_Bengtsson; Max Planck Institute “Lennart Bengtsson is celebrating his 90th birthday” https://mpimet.mpg.de/en/communication/news/lennart-bengtsson-is-celebrating-his-90th-birthday; Prabook biographical entry https://prabook.com/web/lennart_olof.bengtsson/40704. ↩

  6. Bengtsson at SMHI 1961-1974 (some sources 1962); ECMWF Head of Research 1975-1981; ECMWF Director 1982-1990. Source: ECMWF “History” and EGU citations https://www.egu.eu/awards-medals/alfred-wegener/2009/lennart-bengtsson/. ↩

  7. Bengtsson awards: Milutin Milanković Medal 1996, René Descartes Prize 2005, IMO Prize 2006, Wegener Medal 2009, Honorary Fellow RMS 2009. The “2014 RAS Gold Medal” attribution is unconfirmed; the 2014 RAS Gold Medal for Geophysics went to John Zarnecki, see RAS award archive https://ras.ac.uk/sites/default/files/2019-04/Gold_medallists.pdf. Sources: EGU and ECMWF citation pages cited above. ↩

  8. William Bourke biographical detail is sparse in publicly available sources. Bourke was at the Commonwealth Meteorology Research Centre (CMRC) Melbourne, Victoria, in 1972; this became the Australian Numerical Meteorology Research Centre (ANMRC) in 1974, and the Bureau of Meteorology Research Centre (BMRC) in 1985. Bourke served as Acting Officer-in-Charge of ANMRC in 1983-1984 (CSIRO Annual Report). He presented “History of NWP in Australia – 1970 to the Present” at the 16th BMRC Modelling Workshop, Melbourne, 6-9 December 2004. Sources: https://www.eoas.info/biogs/A000912b.htm; https://www.eoas.info/biogs/A000913b.htm. ↩

  9. Bourke 1972 paper: “An Efficient, One-Level, Primitive-Equation Spectral Model,” Monthly Weather Review, vol. 100 no. 9, pp. 683-689, September 1972. Bourke 1974 paper: “A Multi-Level Spectral Model. I. Formulation and Hemispheric Integrations,” Monthly Weather Review, vol. 102 no. 10, pp. 687-701, October 1974. Bourke et al. 1977 chapter: “Global Modeling of Atmospheric Flow by Spectral Methods,” in J. Chang (ed.), General Circulation Models of the Atmosphere, Methods in Computational Physics vol. 17, Academic Press, 1977. ↩

  10. Hollingsworth biographical: born 6 July 1943 Dublin Ireland; died 29 July 2007 in Ireland age 64. UCC mathematics 1964; Met Éireann forecaster; PhD MIT; Reading UAMG; ECMWF from 1 March 1975. Source: ECMWF obituary 2007 https://www.ecmwf.int/en/about/media-centre/news/2007/dr-anthony-tony-hollingsworth-1943-2007; Irish Times obituary https://www.irishtimes.com/news/irish-meteorologist-of-international-repute-1.954580. ↩

  11. Hollingsworth-Lönnberg method 1986 papers: Hollingsworth, A., D.B. Shaw, P. Lönnberg, L. Illari, K. Arpe, and A.J. Simmons, “Monitoring of observation and analysis quality by a data assimilation system,” Monthly Weather Review, 114, 861-879, 1986. Hollingsworth, A. and P. Lönnberg, “The statistical structure of short range forecast errors as determined from radiosonde data. Part I: The wind field,” Tellus 38A, 111-136, 1986. ↩

  12. Adrian Simmons EGU Vilhelm Bjerknes Medal 2012 citation: https://www.egu.eu/awards-medals/vilhelm-bjerknes/2012/adrian-simmons/. Simmons “Cambridge then Reading then ECMWF 1978” career path: ECMWF “Adrian Simmons’ early academic career” https://www.ecmwf.int/en/elibrary/80428-adrian-simmons-early-academic-career; ECMWF “Symposium for Adrian Simmons” 2017 https://www.ecmwf.int/en/learning/seminars/symposium-adrian-simmons. ↩

  13. Simmons-Burridge hybrid coordinate: standard reference is Simmons and Burridge 1981, “An energy and angular-momentum conserving vertical finite-difference scheme and hybrid vertical coordinates,” Monthly Weather Review, 109, 758-766. ERA-Interim reanalysis: Dee et al. 2011, Quarterly Journal of the Royal Meteorological Society, 137, 553-597. ERA5: Hersbach et al. 2020, QJRMS 146, 1999-2049 – Hersbach acknowledges Simmons and Dee. ↩

  14. Tim Palmer biographical: born 31 December 1952 Kingston upon Thames Surrey England. Bristol University maths/physics; Oxford D.Phil. General Relativity Theory. Sources: Wikipedia “Tim Palmer (physicist)” https://en.wikipedia.org/wiki/Tim_Palmer_(physicist); Royal Society profile https://royalsociety.org/people/tim-palmer-12036/; Oxford Physics profile https://www.physics.ox.ac.uk/our-people/palmer. ↩

  15. ECMWF EPS founding paper: Palmer, T.N., F. Molteni, R. Mureau, R. Buizza, P. Chapelet, and J. Tribbia, “Ensemble Prediction,” ECMWF Technical Memorandum No. 188, 1992. Operational implementation of EPS: 24 November 1992. Buizza et al. 1993 first IFS singular-vector implementation. Molteni et al. 1996 documents the operational system. Palmer 2019 retrospective: “The ECMWF ensemble prediction system: Looking back (more than) 25 years and projecting forward 25 years,” QJRMS, doi:10.1002/qj.3383. ↩

  16. Tim Palmer awards: FRS 2003, CBE 2015, Buys Ballot Medal 2003, IMO Prize circa 2007 (varies by source: 2007 vs 2014), Carl-Gustaf Rossby Research Medal AMS 2010, Dirac Gold Medal 2014, RAS Gold Medal Geophysics 2023. “After All, I Decided I’d Rather Do Something Useful”: chapter in Springer 2025 volume “Tim Palmer: Mathematical Physicist, Climate Dynamicist, Poet, Band Leader” https://link.springer.com/chapter/10.1007/978-3-031-81650-5_5. ↩

  17. Mike Cullen at Met Office from 1970; Unified Model dynamical core 1987-1991. A Mathematical Theory of Large-Scale Atmospheric Flow, Imperial College Press, 2006. Met Office profile https://www.metoffice.gov.uk/research/people/mike-cullen. ↩

  18. David Burridge ECMWF Director 1991-2004; Bristol mathematics; Florida University one year teaching; Met Office 1970; ECMWF from 1975. EMS Silver Medal 2010. Source: EMS award page https://www.emetsoc.org/awards/award/david-burridge/. ↩

  19. Erland Källén PhD Stockholm 1980; Stockholm University Professor of Dynamic Meteorology from 1996; ECMWF Director of Research 2009-2017; Centre for Climate Research Singapore 2018-2020; retired 2022. ECMWF newsletter article 2009 https://www.ecmwf.int/en/newsletter/152/news/new-director-research-takes-his-post; Academy of Europe page https://www.ae-info.org/ae/Member/Källén_Erland. ↩

  20. Florence Rabier born 1964 France; ENM/Toulouse III/Paris VI; ECMWF 1992-1998 (4D-Var); Météo-France 1998-2013; ECMWF Director of Forecasts 2013-2016; Director-General 2016-2025. Knight of the Legion of Honour 2014. Sources: ECMWF profile https://www.ecmwf.int/en/about/who-we-are/staff-profiles/florence-rabier; Wikipedia “Florence Rabier” https://en.wikipedia.org/wiki/Florence_Rabier. ↩