A brief history of the Met Office

1854 · Founding

The Met Office was created in August 1854 as a small department of the Board of Trade, charged with collecting and publishing marine weather observations so that British shipping might run with a little less risk. Its first head was Vice-Admiral Robert FitzRoy — by then best known as the captain of HMS Beagle, the survey ship that had carried Charles Darwin around the world two decades earlier. FitzRoy was given the somewhat unromantic title of Meteorological Statist to the Board of Trade. He started with an office in Parliament Street, London, a staff of two, and an ambitious belief that weather, if observed systematically enough, could be turned into something close to a science of prediction.

The remit was originally narrow. Standard-issue barometers and instruction booklets were distributed to merchant captains. Logs came back through customs houses. The data was tabulated by hand. There was no public forecast service, no warnings, no radio, no telegraphed map. The point, at first, was simply to understand what the sea was doing.

1859 — 1865 · FitzRoy and the first storm warnings

On the night of 25 October 1859 a deep depression came across the Irish Sea and tore the steam clipper Royal Charter onto the rocks off Anglesey. More than 450 people drowned within sight of land. The Royal Charter Gale, as it became known, was the catalyst FitzRoy needed. He argued that the same telegraph network the railways had been building could be used to flash warnings ahead of weather rather than behind it.

The first gale warning was issued in February 1861. By the following year a public daily forecast was appearing in the press. FitzRoy coined the word forecast deliberately — he wanted a term that conveyed careful inference rather than prophecy. The system worked imperfectly. False alarms drew scorn; missed events drew worse. FitzRoy was working at the very edge of what mid-Victorian observation and theory could support, and the public, the Admiralty and the Royal Society were not always patient with him. Worn down by the criticism, and by long-standing ill health, he took his own life in 1865.

1866 — 1913 · Slow consolidation

After FitzRoy's death the warning service was briefly withdrawn on the recommendation of a Royal Society committee that felt the science was not yet strong enough. The result was such a public outcry — particularly from fishing communities — that warnings were reinstated within a few years. The office passed between government departments and was placed under the new Meteorological Council, but the core work continued: more stations, more observers, more uniform instruments. Daily weather reports settled into the columns of The Times. Synoptic charts of pressure and wind began to appear across the British Isles. The discipline was acquiring, slowly, the look of a profession.

1914 — 1919 · The First World War

Aviation pulled meteorology out of the maritime corner it had occupied for sixty years. The Royal Flying Corps and then the Royal Air Force needed wind and visibility forecasts at altitude, for fields they were still in the process of building. The number of observing stations multiplied; radiosonde precursors were trialled; upper-air work began in earnest. In 1919 the Meteorological Office became part of the Air Ministry, formally tying its fortunes to military and civil aviation for the next several decades.

1922 · Lewis Fry Richardson's forecast factory

In 1922 the Quaker mathematician Lewis Fry Richardson published Weather Prediction by Numerical Process. Working largely outside the Met Office establishment, Richardson proposed that the weather could be forecast by writing down the equations of motion of the atmosphere and solving them, step by arithmetic step, on a grid laid over the globe. His own trial run, computed by hand from a real observation set, produced a spectacularly wrong six-hour forecast — but his diagnosis of why was correct, and the equations themselves were right.

Famously, Richardson estimated that to keep up with the actual weather one would need around 64,000 human computers — people with desk calculators — arranged in a great domed hall, each working a small region of the grid, coordinated by signal lamps from a central conductor. The forecast factory was a thought experiment. It was also an astonishingly accurate preview of the way digital weather prediction would work three decades later, once the human computers had been replaced by electronic ones.

1939 — 1945 · The Second World War and D-Day

The Second World War turned the Met Office into a military asset. Forecasters were posted to bomber stations, naval commands and convoy planning rooms. The most famous single forecast in its history came in early June 1944, when a small group of Met Office and US Army Air Forces meteorologists under Group Captain James Stagg advised General Eisenhower on the weather window for the Normandy landings. The original target date of 5 June looked unworkable. Stagg's team, reading a clearing pattern moving in from the Atlantic, recommended a 24-hour delay. The invasion sailed on the 6th into a brief, uncomfortable lull. It is hard to think of another forecast on which more turned.

1959 — 1965 · The computer arrives

Richardson's factory finally appeared, in silicon, around the end of the 1950s. In 1959 the Met Office began using a Ferranti Mercury — the first computer in the United Kingdom put to operational forecasting. By the mid 1960s numerical weather prediction was producing forecasts on a daily timetable, and the equations Richardson had written by hand were being integrated thousands of times faster than he could have done. In 1961 the headquarters moved from London to Bracknell in Berkshire, where it would remain for more than forty years. The new site was built around the needs of large computing machines and the people who tended them.

1980s · Television, and the Great Storm of 1987

Through the 1980s television became the dominant channel for public forecasts. Met Office staff presented the weather on BBC One in evening dress for a generation; the maps moved from magnetic pieces to electronic graphics. Public familiarity with the brand grew, and so did expectations.

Then, on the evening of 15 October 1987, Michael Fish told viewers that a caller had been worried about a hurricane on the way and that, although it would be very windy, there was not a hurricane coming. Overnight the deepest depression on record over southern England came across the Channel and killed eighteen people, felled an estimated fifteen million trees, and left the south of the country looking like it had been combed flat. The forecast had warned of severe weather but had badly underplayed where the worst of it would land and how strong it would be. The clip became one of the most replayed in British television. The harder result was a sustained Treasury commitment to better observations, better models, and better radar — a quiet investment whose dividends are still being drawn today.

1990s · The Unified Model

In 1991 the Met Office released what is still the spine of its operational stack: the Unified Model, a single dynamical core that could be configured to run anything from a short-range UK forecast at high resolution to a century-long climate integration at coarse grid spacing. Maintaining one model rather than several made every improvement in physics, data assimilation or numerical method automatically available to all timescales. Around the same period the United Kingdom became a full member state of the European Centre for Medium-Range Weather Forecasts in Reading, whose own model still sets the global benchmark for medium-range skill.

2003 · Exeter, and the Hadley Centre

In 2003 the Met Office moved its headquarters from Bracknell to a new purpose-built campus in Exeter. The Hadley Centre for Climate Prediction and Research, founded in 1990, had by then become one of the most influential climate-science institutions in the world, contributing heavily to every assessment report of the Intergovernmental Panel on Climate Change. The 2000s also saw operational model resolutions close on the kilometre scale, the UKV configuration of the Unified Model giving forecasters the first credible look at convective showers in something close to their real size.

2015 · Naming the storms

For most of the Met Office's history its severe weather messaging was technical: a system of named pressure centres and impact-based warnings the public was mostly happy to ignore. In 2015 it launched, jointly with the Irish service Met Éireann, a public naming scheme for windstorms affecting the British Isles, with Storm Abigail the first to carry a human-friendly label. The Dutch service KNMI joined the group a few years later. The change was social rather than scientific, but it was striking: warnings became easier to talk about, easier to remember and easier to take seriously.

2022 — present · The supercomputer era

In 2022 the Met Office began transitioning its operational forecast workload onto a new generation of supercomputing capacity hosted on Microsoft Azure, replacing the on-premises Cray machines that had served Exeter through the 2010s. The new capacity has supported finer grid spacing, larger ensembles and far heavier data-assimilation workloads. At the same time, a wave of machine-learning weather models has arrived from outside the traditional meteorological world — DeepMind's GraphCast and DGMR, Huawei's Pangu-Weather, Nvidia's FourCastNet — and the Met Office is one of the services actively integrating them into its production stack, particularly for nowcasting on the zero-to-six-hour horizon where convective forecasting lives.

Today

The modern Met Office employs around 2,000 people across Exeter, the Scottish College in Aberdeen, and a network of observing and radar sites. It operates the National Severe Weather Warning Service, the Met Office mobile and web products, the DataHub API platform, and a substantial climate-science and defence programme alongside the daily public forecast. Its work feeds — directly or indirectly — into almost every weather product British people see, from the morning bulletin to the radar overlay on their phones. You can find their official surface at metoffice.gov.uk.

Why it matters to chasers

Almost everything we look at in the morning — the radar composite, the UKV soundings, the warnings that frame our day, the climatology that tells us what is or is not unusual — exists because of the people and machines built up over the past 170 years. We benefit from a national service whose standards, calibration and continuity are inherited from FitzRoy's first instrument lists. It is worth knowing that, and worth a small amount of patience with it when the forecast is wrong.

For more on the data we use from the Met Office and other sources, see UK weather data sources and our tools. For the longer story of British weather observation and chasing in this country, see history, chasing in the UK and the wiki. If you're new here, about Chaseit is a good place to start.