4caster

It is impossible in scattered showery situations (as over the North Sea today) to have any confidence in the forecast FOR A PARTICULAR PLACE for more than an hour or two ahead. One can predict the MOVEMENT of showers, but it's usually impossible to predict when and where each shower will decay, and when and where new ones will pop up (which is what they do, literally!).

I sympathise. For example the BBC forecast for Hull presently gives a 58% probability of rain at 1300 today. That's only 35 minutes from now, yet a glance at the radar tells me that the showers are all over the North Sea and coastal extremities like Flamborough Head and Cleethorpes. I prefer the Met Office site for short term forecasts: https://www.metoffice.gov.uk/public/weather/forecast/map/gcrzgymcy#?map=CloudAndRain&zoom=7

Sorry Phil is leaving. He is the best of the national BBC forecasters in my opinion.

I know of one who had a reputation for pushing his own ideas rather than those of the chief forecaster, but I don't think I should name him here.

I don't believe the forecasters can manipulate the graphical forecasts beyond choosing which elements to display and on what scale. They can't move rain areas further east or west, for example, or alter the track of a depression. Occasionally they'll explain that the precipitation is overdone, or it might get a little further north, but they wouldn't say things like that without being confident about it. The same applied when the Met Office had the BBC contract. But then they had regular conferences with the Met Office Chief Forecaster, who would try to persuade them to stick to the party line. Again they wouldn't diverge unless they were pretty sure about it.

Yes, a marked difference between the two models' solutions. I had recorded the bulletin at 26/2200 BST and saw it for myself. We'll see which model is correct (if either!).

Sorry, I thought I had chosen an ECMWF chart, but it was another GFS one, and I tried to edit it but the postman came to the door, and I was not in time to edit it! This is the ECMWF chart for Monday at 00Z. The track of the depression is closer to the BBC one. But 850 HPa temperatures of Zero to MS 04 over SE England are not cold enough for snow to settle on the Downs.

I agree with the BBC forecast. Lying snow is most unlikely even on the Downs. The depression is coming from the western Med. GFS and ECMWF place it slightly further east, and this zoomed-in GFS chart shows the maximum spread westwards, at 0900Z on Monday. The dotted lines are 1000-850 HPa thicknesses, which need to be below 130 DM for snow on the Downs. This is four days away, so the exact track of the depression is doubtful. The rain could extend further westwards, or it could miss England altogether. The wide area chart is from ECMWF for the same time. It takes the depression away northeastwards across Germany.

Thank you, Summer Sun, for posting this comparison. Essentially all three show the same thing. In the early days of weather radar the Met Office had "quality control", a tedious operation that involved checking suspect echoes against surface observations and removing "Anomalous Propagation (AnaProp)" echoes, such as ground returns, with a light pen. This of course delayed the issuance of the imagery. Nowadays I think AnaProp is removed automatically and the forecaster/presenters have no means of altering the imagery except for choosing which layers to display, such as clouds, fog, etc. The NetWeather display is the clearest in my opinion, but a coastline and place names as the top layer would help readers to locate where they are under the veil of precipitation.

I agree that the cloud looks too similar to the snow. But the BBC forecasters have the ability to remove the cloud overlay so as to leave the precipitation showing, and provided they tell viewers that they have done so, the white colour will represent only falling snow. My gripe is that, particularly on charts showing the whole of the British Isles, the place names obscure the meteorology. For example the B in Birmingham lies near Rosslare, Ireland. One way around that would be to remove the black "fill" of the rectangles surrounding the place names so that the rectangles are transparent, and change the place names from white to black. The names would still be legible because the background is never pure black.