FitzRoy, and the word forecast

The man who named an idea — and was hounded for it.

A sailor who read the sky differently

Robert FitzRoy had already done the thing most people remember him for — carrying Charles Darwin around the world on HMS Beagle — before he turned his considerable intelligence to meteorology. The voyage, completed in 1836, left him with a navigator's instinct for pattern and a surveyor's respect for precision, and both would prove useful when, in 1854, the British Board of Trade created the Meteorological Department and put FitzRoy in charge of it. He was forty-nine, a vice-admiral in waiting, and about to invent something that had never existed: a public warning service for the weather.

The department's original mandate was modest — collect data from ships to improve charts for mariners. FitzRoy did that, and then kept going. He standardised the barometers being issued to harbour stations around the British coast, insisted on a common protocol for reading them, and began receiving the data by electric telegraph, which was new enough to still feel miraculous. What the data told him, read in aggregate, was that weather moved in recognisable ways. A deep low to the west of Ireland behaved predictably enough, often enough, to be worth acting on.

By 1861 he had established a network of fifteen coastal stations reporting daily by wire to his office in London, and he began issuing gale warnings whenever the pattern looked threatening. He also began publishing daily weather statements in The Times — not descriptions of the day's conditions as they already existed, but statements about what the following day would bring. He called them forecasts, a compound he coined deliberately: fore, meaning beforehand, and cast in the older sense of a calculation or estimate. The word forecast, in this meteorological sense, appears in the historical record as FitzRoy's coinage from around 1861.

The argument about presumption

The forecasts worked often enough to be useful and failed often enough to provoke fury on both sides. Fishermen and shipowners who had sailed in conditions FitzRoy had warned against found themselves angry; scientists who considered the atmosphere far too complex to be predicted at all found the whole enterprise embarrassing. The Meteorological Magazine, in the tone that technical journals manage particularly well, doubted that any rational basis for such predictions existed.

FitzRoy was not, by the standards of later science, working from a rational basis in the formal sense — he had no equations, no physics of fronts, no understanding of the three-dimensional structure of a storm. What he had was an empirically trained eye and a growing faith in the telegraph-fed synoptic map, which let him see the atmosphere across a thousand miles in something close to real time. Synoptic means simultaneous: the word distinguishes the kind of map where all measurements are taken at roughly the same hour from any other kind of geographical display. The synoptic chart was a new instrument, and FitzRoy was among the first to use it operationally.

He published The Weather Book in 1863, the first attempt at a practical guide to forecasting for a general audience. It argued that a trained observer with a barometer, a knowledge of wind direction and a synoptic map could make useful predictions for the following twelve to twenty-four hours. The science behind it was incomplete. The intuition behind it was sound.

What the pressure cost him

FitzRoy's position became impossible in two directions at once. Scientists considered him a charlatan for claiming to predict the unpredictable. Sailors and the press considered him criminally negligent on the occasions when his forecasts missed — when gales struck coasts he had cleared, or warnings paralysed fleets in harbours while the sea stayed calm. He ran his department on a budget too small for the ambition, argued constantly with the Board of Trade over resources, and bore the editorial attacks of The Times, which had initially been glad to publish his columns and later glad to criticise them.

He died in 1865, by his own hand, in circumstances the period recorded with the usual careful vagueness. He was fifty-nine years old and had not resolved the question of whether public weather prediction was science, service, or hubris.

Within two years of his death, the storm warning service he had built was quietly suspended, then reinstated under pressure from fishing communities who turned out to have found it useful after all. The Met Office, which traces its institutional lineage directly to FitzRoy's Meteorological Department, still operates storm warnings today from its headquarters in Exeter.

The forecast after FitzRoy

The conceptual revolution that FitzRoy's intuition needed but never had came a generation later from Bergen, Norway, where Vilhelm Bjerknes and his son Jacob Bjerknes developed the theory of fronts — the boundaries between air masses of different temperature and moisture — in the years around the First World War. This gave forecasters a physical framework to hang the barometric patterns on. The triangles and semicircles that appear on synoptic charts today are their notation, formally standardised in the decades that followed.

The mathematical turn came from Lewis Fry Richardson ↗, who reasoned in the early 1920s that if the atmosphere obeyed equations, a forecast was simply the numerical solution of those equations run forward in time. His attempt — a hand-calculated six-hour forecast for a single point over central Europe — took weeks of arithmetic and came out badly wrong, but the method was correct. What it required was computing power that would not exist for another three decades. When Jule Charney and his collaborators ran the first successful numerical forecast at the Institute for Advanced Study in Princeton in 1950, it was on one of the first electronic computers, and FitzRoy's empirical pattern-matching was on its way to being replaced by physics.

None of that erases what he did. He looked at a barometer reading arriving by wire from a coastal station on a November morning and wrote down what he thought would happen next. He was the first person in an official capacity to do that, and then to publish it, and then to stand behind it when the ships came back. That is still what a forecast is — a statement about a physical system's future state, issued with enough confidence to be useful and enough honesty to be checked.