Dr Catherine Ross has been working on the extensive archives of the Met Office for more than a decade and is a mine of information on its history, role and contributions to many aspects of our nation and individuals.
Catherine traces the use of pre-digital IT back to the employment of the Victorian telegraph to transmit readings and broadcast predictions and storm warnings. Between then and the invention of the stored program digital computer, Dr Ross charts the use of other technologies, including “computors” and the early vision of how armies of people with mechanical calculators might have presaged the use of super computing.
The Met Office was one of the first users of digital computers as we know them in 1951, running its programs on the Leo at Cadby Hall. Its first own computer was a Ferranti Mercury. The Met Office’s ever more sophisticated numerical modelling of the atmosphere has created a continually expanding demand for computing power. That has made it one of the most demanding users of processing power, leading it to use larger and larger supercomputers from Control Data, Cray and IBM.
- 2004 – gains a Master of Arts in Ancient History at Durham University
- 2004-08 – goes on to attain a Doctor of Philosophy at Durham
- 2008 – graduates with Master of Archives and Records Management from the University of Liverpool
- 2008 – becomes an archivist at Army Medical Services Museum
- 2012 – joins the Met Office as Archivist
Interview conducted by Richard Sharpe on 22 September 2022 on Teams.
Catherine Ross is the Met Office Archivist and has been doing the role for ten years. She explains how she became an archivist, saying: “The general route into being an archivist generally starts with a history degree. I studied ancient, medieval and modern history at Durham, then ancient history through my Masters and PhD. At the same time I worked in my college archive; St Mary’s in Durham. From that, in order to stay involved in history, I decided that archives would be a good career path for me. I qualified at Liverpool.” Catherine says that being organised is one of the key skills needed, adding, “Archives is one of the fields of knowledge and information management. You are responsible for the collection, preservation and access of materials, whatever the archive might be. You need to decide what’s coming in, make sure that it comes in correctly, that you know where it is, that you can always access it for somebody else, and when you get it out, you need to make sure that you remember where it goes back to so that it never gets lost.” Speaking about the Met Office archive, she adds: “Within the Met Office we have our analogue archives, and we have a digital library and archive where we have a number of born digital collections; the things that are being produced digitally now, things like space weather, national severe weather warnings, some of the atmospheric pollution type forecasting. We’ve digitised some of our most important series. That’s still a drop in the ocean in the size of the archive, but it is a significant part of the most useful information that’s now digitised (scanned, and made available as PDFs), so that it’s much more accessible around the world for people wanting to access our data.” Having done a rough calculation, Catherine says that the Met Office’s archive would be “approximately three times the height of Ben Nevis.” The English and Welsh archives are stored in Exeter, the archives for Scotland are held in Scotland, and the archives for Northern Ireland are held in Northern Ireland, Catherine explains: “We’ve done that because we believe that they should be held in their country of origin, although they’re all owned by the Met Office, they belong to their country of origin.” Catherine has overarching responsibility for the maintenance and safekeeping of all of the collections, adding: “Here at Exeter, not only have we got England and Wales, we’ve got all of the international data, private weather diaries, rare books, all sorts of other things that sit around that core data collection, and that’s all my responsibility as well.” A career as an archivist
The idea for meteorological offices around the world originated at an international conference in Brussels in 1853. Catherine explains: “It all comes down to Lieutenant Matthew Fontaine Maury who was in the US Navy. He was injured and placed in charge of the depot of ship logs. Most people would have thought was something of a backend job, however, he realised that he now had a mine of information, which, if he could put all of this information into one place, i.e. as a series of charts, would be really useful to seafarers. He then took that a step further and realised that everybody else’s logs would add more information, and if everybody got together on this then we could really create a useful series of charts for the whole world, not just the Atlantic and some of the Pacific. He wrote to governments around the world suggesting an international conference, and the UK was one of many who sent representatives. Out of that was the agreement to found the Met Office and then to start collecting and sharing data. It’s probably the earliest international data sharing agreement in existence.” As a result of the conference, on 1 August, 1854, the Department of Trade launched the Meteorological Department, based in Parliament Street, Westminster. Robert FitzRoy, captain of the Beagle and a surveyor, was appointed as Meteorological Statist to the Board of Trade. FitzRoy came to the attention of Admiral Beaufort, who saw FitzRoy as a protégé, and as well as introducing him to Darwin, also recommended FitzRoy as the head of the fledgling Meteorological Office. Catherine adds: “His job was to set up an observing network. At this point the interest was marine. The Met Office was established for the protection of life and property at sea, but in the Victorian period one would suspect that the stress may more have been on property. FitzRoy’s job was to start putting in place the mechanisms to collect data, things like designing a log with columns, dates, etc, elements that all Royal Naval vessels would use, and then moving this out to merchant marine as well, so that you had as many ships as possible observing. Then thinking about where he could put land observation stations, our very earliest anemometers that record windspeed and direction. We have very early anemometers from Bermuda and also from Ascension Island.” All of FitzRoy’s early information collection was done by post. Catherine says: “The origins of the Met Office were in data collection, climatology as we would call it now, the concept of forecasting was not there. What the weather would do was considered an act of God and so you couldn’t possibly forecast it.” Origins of the Met Office
A great storm, known as the Royal Charter storm, which took place on 25-26 October 1859, changed attitudes to forecasting the weather. Over 800 people lost their lives at sea during the storm which saw 133 boats sink, including the Royal Charter which was wrecked off the coast of Anglesey. Catherine explains: “The storm hit the headlines and FitzRoy approached Parliament to say that having taken observations for years, having gained the knowledge and understanding of how weather systems progress, the storm could have been foreseen. He put together the Board of Trade Report Number 10, which went into Parliament requesting that he be allowed to start a gale warning service. There was quite a lot of doubt that this was really possible, but in light of the disaster it was a case of give it a go because what have we got to lose. FitzRoy set that up.” A critical factor in this was the beginning of electronic communication. Catherine adds: “It was critical because we couldn’t possibly have done that without the birth of weather telegraphy. You cannot give a forecast if you cannot get data from one site to another and back again in a timely fashion, and that was impossible without the telegraph network. All of the founding of forecasting depended on the birth of the telegraph network. Fitzroy founded the storm warning service, it’s now known as the shipping forecast and it’s believed to be the longest running national forecasting service in history. It’s an icon of Britain now.” The service operated through a telegraph station located next door to numbers 1 and 2 Parliament Street. Telegrams were delivered straight to FitzRoy and his team who examined the observations and decided if warnings needed to be sent to the harbours via telegraph. The harbour masters would then physically send a message to ships via a visual system also invented by FitzRoy. Catherine adds: “FitzRoy designed a system of cones and drums based on large canvas shapes that would look the same whichever angle you saw them from, so there was no risk of confusion, which would be hung from a mast or other tall point in sight of the harbour and the shore for coastal shipping.” The system provided ships with information about which direction the gale was potentially coming from and an estimation of when the storm would arrive. Catherine continues: “The first warning on the Tyne in February 1861 was ignored and there was significant loss of life as a result. It was not ignored after that.” Storm Warning Service
Catherine explains that the shipping forecast we know today were not officially laid out until the 1920s. She says: “They grow and they change over time. The area is widened and the actual shipping forecast areas are made smaller and more precise. In the 1920s a group of meteorologists from the UK and from the North Sea coasts got together and decided what they thought the areas should be and named them. Most of them had geographic terms, for example; Forties is because it’s normally forty fathoms deep, Dogger because it’s Dogger Bank. The furthest south used to be called Finisterre, but during the twentieth century there was increasing confusion between Cape Finisterre, which was a Spanish meteorological service forecasting point, and the shipping forecast area Finisterre, and so Finisterre was renamed FitzRoy.” Shipping Forecast
Having established the Met Office, used the telegraph, and introduced the storm warning, FitzRoy, then introduced the public weather forecast. Catherine explains: “FitzRoy believed that the shipping forecast was important to mariners. He also understood that mariners not just at the large ports needed it, so he designed a barometer for the fishermen, which he personally financed. He also thought that the British public would find it useful to have a weather forecast. He didn’t have permission for it but he went ahead anyway and published his first forecast in The Times on 1st August 1861. It’s a very simple, very general forecast. He didn’t actually include the east on the first day. In our books, we have a note that says ‘from tomorrow we will include the east’. The first forecast was accurate, but he was chancing his arm. Marine forecasting was slightly easier because they had more data, more understanding, more experience of forecasting. Anyone that’s been in the navy naturally becomes a forecaster by means of having been on board a ship for thirty years, it’s not the same for land forecasting. They had data which came through the telegraph network but they didn’t necessarily have the depth of scientific knowledge to understand more complex synoptic situations and the forecasts were wrong on occasion, possibly more than on occasion. He did receive a lot of criticism, particularly from members of the Royal Society. He also nearly bankrupted himself in putting out those barometers to the small fishing ports, and it is believed to be one of the reasons why he took his life in 1856.” Catherine says that FitzRoy was known as being a bit of a depressive, adding: “I think people would refer to it now probably as manic depression. He did suffer with bouts of depression.” FitzRoy wrote the first textbook on weather forecasting in 1863/64 or ’64, Catherine says: “In it, he coins terms like ‘forecasting’ and ‘synoptic’, which are FitzRoy’s terms. He talks about forecasters needing to talk in terms of likelihood rather than certainty, which you’ll still very much hear to this day. He started putting forward theories such as air masses. There’s an image in the back of his book where you can see warm and cold air masses mixing. He was starting to understand circulatory theories and those concepts of meteorology. But there was a lot more that didn’t have a solid theory behind it, and that’s one of the reasons why the Members of the Royal Society reacted so negatively to his approach, because it wasn’t based on a solid proven theory, because that takes time.” After FitzRoy’s death, there was a review of the Met Office. The Royal Society called for it to be shut down completely, instead, management of the Met Office was given to the Royal Society and all forecasting was stopped with a focus on building the scientific theory. Catherine says: “Very swiftly afterwards the gale warnings were restarted because of the increased loss of life, a significant public outcry, and questions in Parliament, etc. However, land forecasting and the public weather forecasts didn’t restart until the 1870s when they felt they had sufficient theoretical underpinning to be more accurate more of the time. “The basics of meteorology are developed in those twenty, thirty years at the end of the Victorian period, including the understanding of isobaric theory, pressure, isobars, how winds flow along isobars, all those kind of things that are very critical. Many other things came much later in the mid-twentieth century; the understanding of how the upper air affects our own meteorology, the tropopause, the jet stream, etc.” Public weather forecast
James Glaisher was an English meteorologist, aeronaut and astronomer, who discovered various weather effects through his experimentation with ballooning. Catherine explains: “Glaisher was definitely forward thinking, or certainly upward thinking. He had instruments tied all over the cockpit and we believe he encountered the tropopause point for the first time. This is the point where temperature decreases with height and then levels off very briefly before decreasing again. He started to understand that things change as you go up and winds can be in different directions or can be stronger. Glaisher also established a rainfall observing network which involved ordinary men and women all over the country taking rainfall observations and sending them in. That was the start of what is a very, very dense and very important climate network that we have still to this day. It was called the British Rainfall Organisation, and later became part of the Met Office network. It was one of the earliest sets of data to be put onto Hollerith cards and analysed because it was so important.” James Glaisher
The Met Office started using Hollerith cards from 1920 and international data sharing started from 1926. Catherine says: “The Met Office received marine observations from the Royal Meteorological Institute of the Netherlands. Later on data is being shared with India or coming in from the US and Hollerith cards became the de facto means of moving data around the world.” Hollerith Cards
The early technology used to record findings, included ready reckoners, pen and paper etc. Catherine explains: “They had two additional things to help them. One of them, designed by Galton, was a trace calculator, otherwise known as a pantograph. It was used in the 1870s. When meteorological instruments took an observation it created a line on a piece of paper, known as curves. Although, they weren’t necessarily entirely curves, they were up and down and all over the place according to temperature, pressure, etc. These needed to be turned into numbers, so Galton developed a trace calculator or trace computer. … It was manual but it was a way of turning graphs to numbers. Later on Kelvin harmonic analysis was brought in for statistical analysis of pressure and temperature, again, from the traces.” The next phase of technology use in the Met Office was telegraphy. Catherine adds: “The focus was on weather telegraphy rather than the data analysis. Data analysis was being done by people, who were known as computers, I have managed to trace that from the 1860s onwards.” Early technology
With the introduction of Marconi’s wireless telegraphy, the first weather forecast from a vessel, the SS Caucasian, was sent on 10th January 1909. Catherine says: “After Marconi introduced the wireless telegraphy, we start to see a shift in the number of marine observations collected following the Titanic disaster which marked a key step change. There were Marconi Rooms in ships before then because we know that the Titanic had a wireless room, it was sending out signals. But with the loss of the Titanic, there is suddenly a huge interest in marine information, sea ice information, and so the US launched a ship to go out and research. The UK focussed on ship-to-shore wireless, and it established ship-to-shore communications for sending marine observations, post 1912. So that’s when we definitely see that coming into play in the Office.” Telegraphy
In 1922 Lewis Fry Richardson pioneered numerical weather prediction. He had been involved in meteorology since 1913 when he was appointed the Superintendent of Eskdale Muir Observatory. Catherine says: “He had no knowledge of meteorology at all, he was a physicist and he was brought in to bring a more theoretical approach to the understanding of meteorology, to start working out some of the underlying equations that might help to calculate numerical weather prediction, as it would become. He started working on this concept and produced the first draft of his book, Weather Prediction by Arithmetical Finite Differences, not long afterwards.” Unfortunately, his work was cut short by the start of the first world war when he took a stance as a devout Quaker that none of his work could be used in any way to further the war and he resigned from the Met Office. Catherine continues: “He took his work with him and worked on it privately and produce a worked up forecast based on a set of observations. It was quite inaccurate and he was disappointed, but Peter Lynch, a meteorologist, has gone back and looked at it and the only thing that Richardson didn’t understand at that point was the concept of data smoothing. If he’d applied data smoothing, a concept that didn’t exist at that point, he would have had an accurate forecast. His calculations, which were carried out in a war zone, were bang on, which is incredible.” Following the war, Richardson returned to the Met Office, and published his book; Weather Prediction by Numerical Process, in 1922. Catherine adds: “In his book, he is describing how a computer could work, seventy years before the advent of a computer … he’s describing the master-slave relationship of a supercomputer with multiple nodes. He’s got the idea in his head long before anybody had got a computer that could even get close to what he was trying to do. His basic equations are still there at the bottom of the models today, because they’re the basic equations of thermodynamics.” Numerical weather prediction
As the BBC celebrates its hundredth anniversary in 2022, it is also the hundredth anniversary of the very first weather broadcast transmitted from Savoy Hill, London. Catherine says: “It went out on 2LO two hours after the BBC first crackled into life. It didn’t reach necessarily all that many people, but it was a text produced by the Met Office that a presenter then read out.” Television broadcasting of the weather forecast was due to start in 1936, however, two factors prevented its success at that time; the second world war and the fact that not many people had access to a television. Catherine says: “The war certainly didn’t help, but the main reason was that simply not enough people had a television, so you just weren’t reaching enough of an audience. The proof of concept was there, you could read out or even put a chart on a screen and describe what was happening through reading a text – we haven’t got as far as an actual forecaster in front of a television at this point – but there simply weren’t enough televisions, the network was not big enough. So they paused that until such times as the television network was large enough to make it worthwhile. There were no forecasts at all of any variety during the First or Second World War, because that data could be picked up and used by the enemy. The only thing you had eventually was a few coded forecasts which were put out for the farmers to enable them to get the crops in before a storm, or warnings of frosts for the apple crops etc, that’s the only forecasts that you had during the wartime periods.” Broadcasting
The forecasts for the D-Day landings played an important part in selecting the date to attack and are recorded in the Met Office archives. Most of those working at the Met Office were seconded to the RAF and there was collaboration between allied meteorologists. James Stagg, a member of Met Office staff, was the forecaster at the time and it was his responsibility to decide what the forecast was going to be and to get agreement, or not, from the allies. Catherine says: “Ideally they would all agree it, but if they didn’t, he had to decide which way he was going to go. Early June was extremely unsettled and so there was a very high degree of discontinuity between the forecasters. The plan had been to invade on 5th June. On 4th June Stagg had to give a forecast saying that a front was expected to come in, sit across the English Channel and make it impossible for the armada to cross in any degree of safety. The American meteorologists didn’t agree with that and it was a tough decision to call, but it was the correct decision. Stagg took the decision to the joint chiefs, and Eisenhower made the decision not to go. However, at 3am on 5th June, Stagg returned with the next forecast which advised that ‘it’s your call, it’s marginal but we think it’s good enough’. You can see this in Stagg’s diary; the final and irrevocable decision, the invasion is on. Clearly he was not convinced it was going to be okay. He knew just how marginal that forecast really was, because he notes that Admiral Creasy and various others of the joint chiefs congratulated him and he said they need to wait and see what happens.” Had the decision not been taken, the landings would have been delayed until later in June coinciding with a storm blew in which had not been predicted with accuracy and caused damage to the Mulberry harbours, Catherine adds: “Stagg wrote a brief report afterwards which notes the weather of both periods. … Eisenhower sent back a copy with a note across the top saying thanks, and ‘thank the Gods of war we went when we did’.” D-Day Forecasting
After the war, which saw the development of the Colossus and the Bombe by Turing at Bletchley, the growth of National Physical Laboratory, computers in Cambridge and Manchester, and in the US, Lyons catering company decided to build its own computer based on the Cambridge computer. This move caught the attention of the Met Office and in May 1948 a meeting was held jointly by the Met Office and Imperial College to discuss ‘the possibilities of using electronic computing machines in meteorology’. Catherine explains what happened next. “A Forecast Research Division was established at the Met Office and after one member of staff went on a training course to use the Cambridge computer, two members of Met Office staff, Fred Bushby and his assistant, Mavis Hinds, started using the Lyons LEO at Cadby Hall, Lyons headquarters, to do the very, very first explorations into the application of computing to meteorology. We’re not really looking at NWP at this point, it’s a question of whether we can produce charts, can we put data in and get some kind of model out. The first atmospheric model calculations were put together by Sawyer and Bushby, that was also happening in ‘51/52. It was very, very early days, trying to see what they could produce with a computer.” It was after this, in 1959 that the Met Office decided to invest in its own Ferranti computer for its forecasting division located at the Met Office head office in Dunstable. Catherine says: “It was ordered in 1955, delivered in 1959. The first programming course was in 1956, so by the time it’s delivered in 1959 you have staff who are capable of using the machine. In addition, they’d identified the maintenance engineer and he went to the Ferranti factory and helped to build a Mercury computer so that he could learn how it worked.” The Met Office uses its computers for general office work, forecasting and climatology. Catherine says: “We have general computer requirements, forecasting and climatology which is dealing with the data, quality control, research, testing out new models, new equations etc. The first machine was called Meteor which very quickly ran out of power to be able to deal with all of these things going on at the same time.” With the move to Bracknell, the Met Office purchased a new English Electric machine; KDF9. Catherine adds: “They took the Ferranti with them initially and then they built the KDF9 which was called Comet. The decision is made that the forecasts are good enough, they are useful, they are at least as useful as a synoptic forecaster, if not more so, they’re dealing with the data fast enough, we’re going to go operational, we’re going to start using numerical weather prediction to produce our forecasts. At that point it’s got to be fast enough, it can’t fall over, it’s now got to produce a forecast twice a day on a timetable forever more. 2nd November 1965 was when the Met Office started the operational use of numerical weather prediction, with great fanfare and many reporters. All the meteorologists wanted it much more low key in case it went wrong. It didn’t and the forecast was accurate.” In 1971 the Met Office bought an IBM 360 195. The investment was made to allow it to shift from its three-level model to the ten-level model of forecasting (levels in the atmosphere) which they had developed off site at using the Science Research Council’s Ferranti Atlas. Catherine says: “The IBM 360 195, called Cosmos, was powerful enough to run the ten-level model. It came in 1971 and on 1st August that was when the three-level model was replaced by the ten-level model operationally. So that’s a big step change in forecasting. … It was at this stage that Fortran makes its first appearance in Met Office coding language. By the end of 1971 we had over 160 staff trained in Fortran IV. The first Met Office Computers
At around the same time, the Met Office also started using the Manchester Computer. Catherine explains: “The Lyons computer was good but it wasn’t powerful enough. It’s a running theme, the computers are not powerful enough. They needed to move on, so they started using the Ferranti Mark 1 (the Manchester Computer) to carry out calculations as a precursor to a system of mathematical forecasting. It was testing and experimentation again, but it proved that proof of concept and there might be a future for the concept of NWP.”
In 1978, the Met Office introduces the national weather radar network. Radar had been used for a variety of purposes since the second world war, the shift in 1978, saw the use of radar to look for rainfall and the national weather radar network was established. Catherine adds: “That was the start of a significant amount of additional data starting to come in into the Office that needed to be processed and fed into the models.” National weather radar network
By 1982, the Met Office, despite having added an additional IBM to the Cosmos to allow it to do some research in addition to the ten-level modelling, started to run out of computational power. The decision was made to purchase its first supercomputer. Catherine says: “The reason for it came out of growing international interest in climatology; the start of the concept of climate change has arrived by now. It was agreed that the UK would play a significant role in the World Climate Programme. Obviously that required going back to our database known as Midas, and being able to process all of that data for which they needed a machine that was an order of magnitude more powerful than the IBM. So that’s when they moved to supercomputers. The intention was it would be attached to the IBM mainframe, but the mainframe started to creak because it couldn’t quite keep up with what was required. From there, there was continuous progress.” By 1991, with the need for yet more power, the Met Office introduced a Cray C-90. This was capable of implementing the Unified Model operationally. Catherine says: “The Unified Model is that rather than having different models running so that one is giving you today’s forecast, one is looking a week out, and another one is looking at your climate model, you build them all into one so that your Unified Model is running everything from what’s going to happen in twenty-four hours to what might happen in fifty years, all at the same time. That mesh is not only getting finer, it’s also stretching out, so all of your models run together, which requires an awful lot of computing power.” Two years later, the Met Office introduced its second Cray to provide yet more power. This was followed in 2015 by the Cray CX40 and the building of a ‘supercomputer’ building. Catherine explains: “When we brought in the new Cray, part of the concept was around having capacity not only to do our own forecasting and research, but to be able to make supercomputing facilities available for other organisations involved in climate research. The Met Office maintained its two computer halls within its HQ, but built a second building where a third supercomputer was located. … They’re all connected up so the data can flow wherever it needs to flow, but it was focussing on enabling research, not just for the Met Office, but for other organisations involved in climate research.” Supercomputers
The Met Office moved to Exeter in 2002. Catherine says of the operation: “It was a huge relocation project, which was very successful if you consider that about eighty per cent of the staff moved with the Met Office. That’s a fairly significant uptake, to move your life, your children, your spouse, to the other side of the country. The machines were moved down individually. You’ve always got to have your back-up, so they were moved down on the backs of low loaders with guards. We went operational virtually seamlessly between Bracknell being switched off and Exeter being switched on, and then the second computer was moved down.” Exeter
The Met Office has kept pace with social media. Catherine says: “The purpose almost goes back to FitzRoy’s original concept of reach. It’s not about research, it’s about reaching audiences. FitzRoy first reached out with the telegraph network, then the newspapers, then of course there was the radio, then there was television. So it’s about reaching your audience wherever your audience are. It started off with YouTube in 2007 and moved onto other social media; Twitter in 2009, Facebook in 2010, Instagram 2013, Snap Chat 2017, TikTok 2019.” Social Media
Interviewed by Richard Sharpe Transcribed by Susan Nicholls Abstracted by Lynda Feeley Interview Data