Somebody still draws the isobars

But it hasn't been replaced — it's been promoted.

The chart that refuses to die

Open any weather app and swipe past the sunshine icons. Behind them, in every major forecast office, somebody is still looking at a map covered in curved lines. The synoptic chart — pressure contours, fronts, the whole hand-drawn-looking apparatus — is older than numerical weather prediction by about a century. What surprises people is that it is still in daily operational use, not as a museum piece but as a working document.

The word synoptic means, roughly, seen together at one time. Robert FitzRoy was among the first to insist on that simultaneity: observations taken at the same hour, mapped in one place, so the atmosphere could be read as a system rather than a collection of local curiosities. His daily weather charts at the Met Office's predecessor showed pressure patterns that moved, and FitzRoy understood that movement as something that could, in principle, be anticipated.

The isobar — a line connecting points of equal atmospheric pressure — is the main structural element of such a chart. Draw enough of them and you can see the troughs, the ridges, the circular closed systems that produce the weather people actually notice. Vilhelm Bjerknes, working in Bergen, Norway in the early twentieth century, elevated this geometry into a physics problem: if you could describe the atmosphere's state precisely enough, you could calculate its next state from known equations. His 1904 paper set the intellectual agenda that eventually produced the numerical model. The map, for Bjerknes, was the starting point — the thing that had to be right before any calculation could mean anything.

What a forecaster reads in the lines

A synoptic chart is not produced by a single model run stamped onto paper. A forecaster reads several models — from ECMWF in Reading, England, from NOAA's American systems, from whatever their own national service runs — and then decides what the atmosphere is most likely doing. The chart they annotate or approve is their synthesis, their argument about which pressure centres are real and which are model artefacts. The spacing of the isobars encodes wind speed directly: pack them tight and the pressure gradient is steep, meaning fast air. The shape of the pattern — whether a front is sharp or diffuse, whether a depression is deepening or filling — carries information that a sequence of numbers does not make immediately visible.

Fronts are the other indispensable feature. Jacob Bjerknes and his colleagues in Bergen described the frontal model of cyclones in the 1910s and 1920s, naming the boundaries between air masses that had arrived from different source regions. A warm front on a chart tells a forecaster to expect a specific sequence: high cloud first, then thickening and lowering cloud, then steady rain, then a rise in temperature. The chart collapses that physical narrative into a symbol, and an experienced eye reads the narrative back out of it without conscious effort.

This is why the synoptic chart survived the computer, rather than being replaced by it. A model grid gives you numbers at every point — temperature, humidity, wind components, vertical velocity — and those numbers are precise and honest about their uncertainty through ensemble methods. What they do not give you is the story. The chart gives you the story. Forecasters at every major centre, including the ECMWF ↗, still produce and scrutinise synoptic analyses because pattern recognition is something trained humans do quickly and reliably in two dimensions.

There is also a verification argument. A chart drawn by a forecaster is an explicit, falsifiable claim about where the high is, where the front lies, how fast the low is moving. When the observations come in the next morning, that claim can be checked against reality. The World Meteorological Organization ↗ maintains international standards for how such charts are drawn and decoded precisely so that verification is comparable across national boundaries. The isobar is not a decoration. It is a quantitative assertion, expressed in a form that has been legible to trained eyes for more than a hundred and fifty years, and the fact that computers now generate the first draft does not change what the line is saying.