The weather conditions of the past few days have posed testing problems for the forecasters. They have made some errors, but on the whole seem to have made as good a job of it as could have been hoped.

A letter published in The Daily Telegraph yesterday accused the Meteorological Office of producing a forecast on Saturday to surpass Michael Fish's infamous there-isn't-going-to-be-a- hurricane report of 10 years ago. Quoting the forecast for SE and central S England: "Early mist clearing to give sunny spells & scattered showers. A light south-easterly breeze", the writer from Surrey said: "This I managed to read by candlelight during a six-hour power failure and amid the sound of tiles being ripped from our roof."

In fact, the Met Office did not do badly in the circumstances. Everything was proceeding in an orderly fashion until Friday. There was a neat queue of depressions performing their winter migration across the Atlantic, passing south of Iceland, then north of Scotland before disappearing somewhere above Scandinavia. At least, that is what they should have been doing. Last week's predictions tracked them accurately until everything became confused as the one they called Low J overtook Low I, speeding beneath it and throwing everything into confusion.

Yesterday's Atlantic chart was a complete mess, with no fewer than five distinct occluded fronts marked. Four of them had the British Isles surrounded, and they seemed all to be travelling in different directions. The presumptuous Low J, which was meant to pass more or less harmlessly north-eastwards over Scandinavia, had changed its mind when it reached Scotland, and drifted back to Iceland and the north Atlantic.

An occlusion is the result of a wedge of warm air being lifted above the ground by cold air creeping beneath it. The ascent of warm air may produce heavy rain or thunderstorms, but more often when a depression occludes, the warm air spreads out and its ascent is considerably diminished. The low-pressure area then fills out and is gradually dissipated.

In the present case, the successive depressions seem all to have been battling for space in the sky, creating masses of air that have been pushing each other out of the way in a manner that has been too complicated to predict more than a short time in advance.

Looking back at the forecast in the Telegraph that provoked the critical letter, it is noticeable that its "General situation" summary was typical of a January day with low pressure coming in from the west: "Eastern parts will have a mostly dry day ... pleasant sunny spells ... only a small chance of a shower ... Scotland will have occasional showers ... the wind will be light or moderate and southerly ... fresh in the north." On the front page of the same paper, however, the weather summary said: "Widespread rain, heavy in places, windy." In the few hours between putting the main forecast in the paper and compiling the front page, everything had changed.

Perhaps every forecast should come with a sell-by time, beyond which its predictions should be taken, if not with a pinch of salt, at least with a heavy coat and umbrella, just in case.