A shipwreck in 1859 gave us the phrase weather forecast
When a storm sank the Royal Charter in October 1859, Robert FitzRoy responded by drawing charts to predict gales, and he called the practice forecasting the weather. Within two years he ran a telegraph warning service for ships, and The Times printed the first daily forecasts in 1861, despite mockery in the press.
Forecasting uses measurements of the present atmosphere, land and ocean, plus the physics of meteorology, to project conditions at a given place and time. It was once done by hand from barometer readings and the look of the sky; today computer models juggle countless variables. Humans still choose which model to trust, drawing on pattern recognition and knowledge of each model's biases.
People have read the sky for millennia. Babylonians around 650 BC combined cloud patterns with astrology, Aristotle wrote Meteorologica about 350 BC, and Theophrastus compiled a Book of Signs. Chinese and Indian methods date to around 300 BC, and in 904 AD Ibn Wahshiyya's Nabatean Agriculture discussed forecasting from winds and the moon. Folk rules such as a red sunset promising fair weather accumulated over generations, though many fail rigorous statistical tests.
The real turning point was the electric telegraph of 1835. Before it, weather news travelled perhaps 60 to 120 kilometres a day; by the late 1840s reports from far upwind arrived almost at once. Francis Beaufort, who devised the wind force scale and expanded record-keeping at 200 coastguard stations, and his protege FitzRoy are credited with turning forecasting into a science. In 1854 FitzRoy took charge of a Board of Trade office collecting weather data at sea, the forerunner of the Met Office, and fifteen telegraph-linked land stations fed him daily reports.
Forecasts still grow shakier the further ahead they look. The atmosphere is chaotic, the equations need enormous computing power, starting measurements carry errors and some processes remain poorly understood, so forecasters run ensembles and compare models to gauge confidence. The payoff is large: in 2009 the United States spent roughly 5.8 billion dollars on forecasting and reaped benefits estimated at six times that, from storm warnings to crop planning and power demand.
Source: Weather forecasting
