In a very basic sense, the essence of weather forecasting was captured by William Shakespeare in Richard II: Men judge by the complexion of the sky, The state and inclination of the day.
And for forecasts of an hour or so ahead, this method works quite well. Within such a short timescale, it is rarely necessary to worry much about what is happening beyond the visible horizon, since events occurring more than 50 miles away are unlikely to impinge upon the weather of the immediate future. But when forecasting conditions for longer periods into the future, the area of interest becomes significantly larger. In these circumstances, as John Ruskin, the 19th-century art critic and amateur meteorologist once explained it over 150 years ago: "The meteorologist is impotent if alone; his observations are useless, for they are made upon a point, while the speculations to be derived from them must be on space."
Observations are needed from places perhaps several hundred miles away, and from these a three-dimensional picture of the atmosphere must be assembled, the volume of which depends on the range of the forecast. As would be expected, the longer the forecast period, the greater the amount of information needed at the outset. In the case of Ireland, forecasts for a day ahead require observations from Europe, the North Atlantic and the eastern seaboard of the United States. To produce a two-day prediction meteorologists need information over most of the northern hemisphere, and in the vertical, data from the ground to the base of the stratosphere, 50,000 feet or more above the ground. Tiny disturbances, barely visible and thousands of miles away, may affect Irish weather in 48 hours' time. As the period of interest gets even longer, the amount of information involved becomes immense, and there is no question of handling it in any other way than by computer. For a forecast of five or six days ahead, it is necessary to know the initial conditions over the entire globe, and to take into account exchanges of energy and water vapour between the atmosphere and the surface of the Earth over both land and sea. And when it comes to the latest kinds of predictions, still at the experimental stage - seasonal forecasts, where attempts are made to gauge the general character of the weather several months in advance - not only is it necessary to analyse the whole atmosphere in great detail, but the developing thermal structure of the ocean must also be predicted to a depth of several tens of metres.