«Medium-Range Weather Prediction Austin Woods Medium-Range Weather Prediction The European Approach The story of the European Centre for Medium-Range ...»
Labrousse always envisaged going back to France after a short time, but in fact he stayed at the Centre for close to eight years, before returning to become Director-General of Météorologie Nationale, the French Meteorological Service. Perhaps we can look ahead to a party in December 1981, when the Council bade farewell to Labrousse. The Council President Dr E. Linglebach from Germany, having recognised Jean Labrousse’s “great skill and ability” in recognising the important problems, noted: “you have always found workable solutions”, and further: “j’ai admiré votre logique française et votre humeur gallic!” Bengtsson and Wiin-Nielsen were working on getting the experimental forecasting up and running. In line with his objective to start operational forecasting soon, Wiin-Nielsen contacted two groups in the USA, who were well advanced in terms of model building. One was at the University of California, Los Angeles (UCLA), led by Professor Yale Harvard Mintz, “the only person I know” said Wiin-Nielsen “who was named after two universities!” The other was at the Geophysical Fluid Dynamics Laboratory (GFDL), under Dr Joseph Smagorinsky. Both of them agreed to make their modelling and other software available, on condition that the Centre sent a scientist to work with their groups for a few months, to gain a full understanding of the complex software.
In the meantime, Labrousse was working on getting temporary use of a computer for installation at Bracknell. A Service Agreement with Control Data Limited came into effect on 26 August 1975. The hired CDC 6600 was slow, and although far from satisfactory for the requirements, it had enough capacity to allow trial forecasts. It was installed in John Scott House, a building close to Fitzwilliam House. In December the Service The first Director 11 Agreement was changed to a Lease Agreement, giving unlimited access to the computer. In addition, time was purchased on the IBM 360/195 at the Meteorological Office.
Recruitment of staff from Member States for the Research and Operations Departments continued. All were conscious that the spin-up time allowed for the entire complex system to get to fully operational forecasting was very short — too short, in the opinion of some.
When it came to appointing the Head of Administration, Germany strongly supported Dr Wolfgang Dieter von Noorden for the post. He replaced Mr Clark, who if given the choice would have liked to continue.
It is fair to say that the working styles of von Noorden and Wiin-Nielsen were very different. Wiin-Nielsen needed to make a myriad of decisions large and small in a rather short period and while under pressure to produce results quickly. Von Noorden’s background in the larger and more bureaucratic administration of the Federal Republic of Germany did not match well with Wiin-Nielsen’s requirements at the time. Discussions on administrative and legal matters were at times difficult, even heated. After a relatively short time, von Noorden left the Centre, to take up an appointment with INMARSAT in London.
Committee meetings moved from Belgium to Britain. Conference rooms of sufficient size and with the required facilities for simultaneous interpretation were unavailable in Bracknell. Suitable premises were found at the Headquarters of the International Coffee and Cocoa Organisation in London. Those who attended the meetings remembered them for the four different kinds of excellent coffee, always provided for free! Centre staff gradually gained more experience with meetings. The underlying papers got shorter and better, thanks largely to the precision and brevity of the original English documents, whose preparation was handled by Ernest Knighting (normally referred to simply as “K”), a consultant who had recently retired from the Meteorological Office. K did a “marvellous job” of introducing Wiin-Nielsen, Bengtsson and Labrousse to the sometimes subtle nuances of the British system. Labrousse later referred to him as “une figure, très intelligent, très fin, avec un esprit critique très acerbe et au final très constructif.” At an early stage, an estimate was needed of how many members of staff would finally be required. A surprisingly small number — just over 30 — was allowed for the Administration Department. An international organisation has heavy requirements for administrative personnel including recruitment of international staff, and for translation, as well as general services, building maintenance and liaison with the authorities of the host
12 Chapter 1country. The Operations Department became the largest; it was clear that the Centre would work round the clock. Many technical staff would be required to supervise and maintain the computer and telecommunications installation, the software and other technical equipment. These were estimated to total about 65. Around 35 scientists would be required for the Research Department. The total was thus taken to be around 130. The architect assigned to the building project, Mr Kidby, needed these numbers, even though they were a shot in the dark at that early stage.
Kidby also needed an estimate of how many square metres would be needed for computing equipment and other technical installations. That was more difficult, as the planning staff still had no idea what computers might be acquired in the years to come. The most pessimistic assumption had to be made that the largest machines then available would be installed. This proved to be wrong, as the Centre’s choice, a CRAY computer, was highly compact. On the other hand, the more usual problem of the building being too small was avoided; later there was adequate space for replacement mainframe computers, which would run in parallel with those already installed.
Furthermore, space was available for a large archive and for the many magnetic tapes used by the computer system in the 1970s. It was not until more than 30 years later that the Computer Hall would need to be extended; a contract for this extension was signed in July 2004.
The architect also needed to know how many of the staff would be men, and how many women; this would affect the number of toilets required.
Wiin-Nielsen looked him in the eye and told him that there would be equal numbers of each. Kidby proceeded accordingly.
Working with Kidby went well on the whole. Kidby said that it was good working with precise people, but there was one point of serious disagreement. There was an energy crisis at the time in the UK. As the electricity supply might fail, it was important for the Centre to have two large diesel generators, which could provide the Centre with the backup supply required, and some large batteries to ensure that computing would continue uninterrupted if the power supply failed. This was absolutely essential, as it would take up to 30 minutes to get the diesel generators up and running. Data could therefore be lost, and the programs running adversely affected. Kidby agreed to all this, but when Wiin-Nielsen said the batteries should be in the basement below the computer room, Kidby disagreed: “We don’t do basements in Britain”. The reason for this was that they were always damp and hence unusable. Wiin-Nielsen explained there were basements in the Netherlands and Denmark in areas below sea level. But the answer was the same: “We do not do basements”. There was a deadlock. One weekend The first Director 13 Wiin-Nielsen, Bengtsson and Labrousse visited the site and visualised the finished building in drawing form. They realised that if the whole complex was rotated through a few degrees, the computer room would be on a sloping section of the site, so there would be room for two floors on the low side and one on the high side. Wiin-Nielsen suggested this at the next meeting. Agreement was reached, and the batteries were installed on the ground floor under the computer room, which was strictly speaking no longer a basement.
A separate wing held an excellent lecture theatre seating 126, and a large conference room for the Council, its Committees and other groups, containing an oval table large enough to accommodate the Chairman, 42 delegates and 40 advisers. Five interpreters’ booths allowed for simultaneous interpretation to and from the five official languages of the Centre.
There were also smaller meeting rooms. The final wing contained the offices, with the library on the top floor.
It was necessary to have discussions with the UK government on matters concerning the Centre, such as negotiating the Headquarters Agreement between the Centre and the UK, which laid down the rights and obligations of the Centre; Wiin-Nielsen was given a contact at the UK Foreign Office, Miss Phyllis Smith. She helped greatly with many issues raised, and wrote the first draft of the Headquarters Agreement. This was based on similar agreements with other organisations, but contained one perhaps rather unusual provision. The Centre was granted a 999-year lease on the land free of charge, with the condition that, when the land and buildings reverted to the UK, the buildings had to be in the same condition as received. WiinNielsen was intrigued; he asked Kidby how long he thought the building would last. The answer was that they “didn’t build for centuries any more, only perhaps for 60-70 years”. After a little discussion, he and Wiin-Nielsen agreed that this would be a problem for others to worry about! Wiin-Nielsen signed the Agreement for the Centre.
In the two weeks 1–12 September 1975, the first of what was to become an annual series of ECMWF Seminars was held at the Met Office College in Shinfield Park. Prof Pierre Morel from Laboratoire de Météorologie Dynamique (LMD) France dealt with data and its assimilation in numerical models, Dr Kiku Miyakoda from GFDL reviewed how physical processes were modelled, as well as numerical methods. Dr Cecil Leith from NCAR described progress in understanding uncertainties in the initial state and in the representation of physical processes. More than forty participants attended from the Member States. This was the beginning of the Centre’s major programme of advanced training. Each year since, the Centre has organised
14 Chapter 1well-attended Training Courses in meteorology and computing, as well as Seminars and Workshops.
At the first Council session on 4–6 November 1975, Wiin-Nielsen presented his first report to Council. (The role of the Council and its Committees is outlined in Annex 2.) Contracts with the Centre staff had expired the preceding Saturday, but had been extended to cover the period of the Council session! However, he noted that with the Convention coming into force, and the adoption of staff regulations and financial regulations, the days of improvisation were over; the Centre was now on a sound footing. He noted the importance of the forthcoming major First GARP Global Experiment (FGGE) exercise, planned for about the time that the Centre would be ready to begin operational forecasting.
The Centre’s headquarters building was opened on 15 June 1979 with speeches from His Royal Highness Prince Charles, Prof Lauri Vuorela of Finland, who was Council President at the time, and Wiin-Nielsen. Dr E.
Süssenberger, first Council President, and as we shall see later a key figure in planning the Centre from the beginning, was among the guests invited to attend the opening ceremony.
While the contract for the Centre’s computer was put out to tender, in reality there was no credible competitor; this was a one-horse race. The contract was negotiated and signed with Cray Inc. Such a major purchase had to be approved by the Council, taking into account the opinions and recommendations of the Finance Committee. Labrousse was outstanding in presenting the issue to the Committee and Council. He had considered all the possible clauses of the long and complicated contract and answered questions clearly. The representative in Europe of Cray Inc, Mr Peter Appleton Jones, was also of great help. The Centre had the first prototype CRAY-1, later replaced by a completely new machine. It — and the same was true for its successors — was surprisingly reliable for such complicated hardware and software. Before the start of operations, foreseeing the absence of a backup mainframe computer, Member States were advised to plan for the loss of perhaps one forecast per week, or two or three a month, to allow for unexpected hardware or software problems. In the event, only a handful of forecasts were partially or completely lost in the first operational year from 1 August 1979. These were later re-run to maintain a full archive. Operational forecasting seven days per week began on 1 August 1980; none of the forecasts were lost after that date and delays were few.
Wiin-Nielsen left the development of the science to Bengtsson and his staff in the Research Department. They made rapid and substantial progress in creating the Centre’s own forecasting model. Studies of the model software obtained from the USA, and the experience gathered from other institutions, The first Director 15 as well as their own substantial stock of experience, all contributed. The task, quite simply, was to put together a model consisting of the best components from the scientific literature or created in-house. Bengtsson was a driver; he demanded, and demanded again, more and more of his staff. He was impatient with doubters. He never accepted “luck” as an explanation for success, or “bad luck” as an explanation for failure. He peppered his staff with questions, constantly raising the level of expectation. He had “the vision thing”.
Perhaps more important, he had the staff who were able and willing to carry out the necessary research. It was common to find Centre staff working late into the evening, and at weekends and holidays. Years later, when Bengtsson was Director, the prospect was raised by the Administration Department of keeping account of staff hours worked. Bengtsson vetoed this rapidly. He knew that if staff realised just how much time they were putting in, this would likely have resulted in a reduction of the hours worked!