The first storm warnings

How the telegraph turned observation into obligation

Before the electric telegraph, weather moved faster than news about it. A gale that wrecked ships off the Cornish coast on a Tuesday was, at best, a Tuesday event — the information reached no one ahead of the wind. Storm warning was not a service because it was not yet possible. The telegraph changed that arithmetic entirely, and once it changed, someone had to decide what to do with the surplus time it created.

That someone, in Britain, was Robert FitzRoy.

FitzRoy had spent years collecting wind and barometer readings from ship captains — the early work of what became the Meteorological Department he established within the Board of Trade — and by the early 1860s, those readings were arriving by telegraph from a network of coastal stations. He could see, on a given morning, what the barometer was doing at Valentia in Ireland, at Nairn in Scotland, at Jersey in the Channel. The picture was hours old by the time it reached him, but hours old was still ahead of the weather.

FitzRoy called his warnings "cautionary signals," and the mechanism was simple enough that a harbour master with a black canvas cone and a mast could operate it. A telegram would arrive specifying the direction of expected danger; the cone went up pointing the right way; mariners could read it from the deck. The system went live in 1861. It was the first operationalised storm warning service in the world, and it worked often enough that its occasional failures became politically visible — which is to say, it worked well enough that people stopped expecting to be surprised by the sea, and started expecting not to die in it.

The tension at the heart of warning

Warning is prediction made actionable, and prediction carries the possibility of being wrong. FitzRoy understood this, and it cost him. The scientific establishment of his day — the Royal Society chief among them — objected that the warnings rested on empirical rules of thumb rather than theory, and that confident warnings based on imperfect rules were worse than no warnings at all. FitzRoy's own name for what he was doing, the word "forecast," was coined partly to distinguish the act from mere conjecture, but the distinction satisfied no one who had already decided he was overreaching.

The tension was real, not merely political. FitzRoy's cautionary signals depended on recognising patterns — a falling barometer at Valentia, a wind shift in the Southwest Approaches — without a physical theory that explained why those patterns should hold. He had no concept of what Vilhelm Bjerknes ↗ would later call the initial-state problem: that a forecast requires knowing the present state of the atmosphere fully enough to carry it forward in time by physical law. FitzRoy was running the warning service before the theory that would eventually underpin it existed. This is not a criticism; it is the correct order in which things happened.

What he had instead was a systematic procedure. Observations came in from the network, were plotted — a practice that eventually formalised into the station plot, the compact symbol at each location carrying wind direction, cloud cover, pressure and temperature — and then compared against known patterns of development. When the pattern looked dangerous, the warning went out. The logic was the same logic any experienced mariner used; FitzRoy had simply mechanised it over a national network.

What observation required of the people running it

A storm warning service is, structurally, an argument that observation creates obligation. Once you can see the weather coming, you have accepted a duty to tell people. FitzRoy felt this acutely, and the feeling was not wrong. The question was how far the obligation extended and how confident a forecaster had to be before sending a signal.

The World Meteorological Organization ↗ did not exist in FitzRoy's time; international coordination of weather observation was patchy and informal. His network was national, and even nationally it was thin — the stations were wherever telegraph cables reached, which was the coast, not the interior. The interior was where fronts developed before they reached the coast, though the word "front" would not appear in meteorology until Jacob Bjerknes introduced it in the Bergen school's work half a century later. FitzRoy was catching the storm as it arrived at the edge of the observation network, not before it entered the country. This constrained the lead time of every warning he could issue.

Even so, the lead time was enough to matter. A cone hoisted in the morning could keep a coastal vessel in harbour through an afternoon gale. The signal did not need to be perfectly precise to save a crew; it needed only to be right more often than chance and early enough to act on. Verification — the practice of scoring a forecast against what actually happened — was not formalised in FitzRoy's time, but the harbour masters who received the signals kept informal records, and those records were good enough to persuade the Board of Trade to keep funding the service.

The service was suspended after FitzRoy's death in 1865, following continued pressure from critics who believed prediction without theory was presumption. It was reinstated within two years, because the gap in warnings was immediately felt by the people who had relied on them. That reinstatement is its own kind of verification: the signal, it turned out, had been doing something measurable. The fishermen and the harbour masters knew before the theorists had caught up.

What FitzRoy established, and what the telegraph made possible, was the idea that a meteorological office has a public function distinct from scientific research. Observations exist to become warnings; warnings exist to reach people before the weather does. The machinery that now runs a modern numerical forecast — the radiosonde ascents, the assimilation of satellite data, the ensemble runs at the European Centre for Medium-Range Weather Forecasts in Reading — is vastly more sophisticated, but it is still pointed at the same moment: the one before the storm arrives, when there is still time to do something about it.