More thoughts on the new BBC weather graphics

[loganberrybunny.livejournal.com]

crossposted to my own LJ

I've now seen some more of the BBC's imminently-arriving new graphics, partly via the TWO forum and partly thanks to enteirah telling me about Weatherscape's own page on the new system, which hosts several MP4 videos of it in action. On the basis of what I've now seen, I'm still reserving judgement, since an awful lot depends on how the forecasters themselves use the new graphics, but overall I'm not terribly impressed. Some things that came to mind:

• The projection used for the all-UK view is terrible, making northern Scotland look tiny; even on the zoomed-in Scottish map the Northern Isles are hard to pick out. It's entirely possible that this problem will become all too obvious very soon, since on Tuesday and Wednesday the distance between the north coast of Scotland and the Central Highlands could easily make the difference between snow and rain, yet there's hardly any space between the two places on the map.


• The contrast between cloudy and clear areas of the country at night isn't nearly high enough, and again is hard to pick out at higher latitudes. And I have near-perfect colour vision; people with some types of colour-blindness are going to have a very hard time with this. I don't actually mind the brown colour of the land too much in itself, though.


• As I said in my last post about this, it's a bad idea to imply (as the second video on Weatherscape's page does) that rainfall can be predicted accurately enough to forecast rain in Leeds but not in Doncaster. By its very nature meteorology is an inexact science and this is just going to lead to more complaints from people who assume, reasonably but wrongly, that the greater precision shown on the map implies a greater precision in the underlying forecast models.


• Leading on from the above, a much more serious problem: the rain and cloud areas don't match up properly! On this graphic, for example, you can clearly see rain falling from sunny skies in parts of central Europe. This is something that is entirely unacceptable for even the most basic of weather forecasts, so I'm astonished to see it on a shiny new system like this.


• I'm not at all happy with the simplistic distinction between "cloudy" and "not cloudy" areas on the map. With the current 2D symbols, heavy and overcast weather can be shown by a black cloud symbol and fair yet cloudy weather with a white one. As I said in a post a while back, this is not a particularly wet country on the whole, but is a very cloudy one, and it's important to know what sort of cloud is expected. And what happens if there's a large depression of the kind Britain often gets in the autumn, when the whole country is overcast? How do we know?


• The satellite-derived cloud pictures (of past weather) are fine, and very clear. But then the maps used for those are more 2D than the main forecast, and I don't think the two facts are unrelated. I haven't yet seen a synoptic North Atlantic chart, so am holding fire on that, but we do need a proper one with pressure centres marked in millibars and warm, cold and occluded fronts distinguished. I really fear a change to hopeless ITV-style "Met Office windflow chart" territory.


• The animations aren't as annoying as I thought they'd be, but really aren't necessary and are a little distracting. There also needs to be a clear distinction between light and heavy rain, and between sleet and snow. The gradual change in shading of the temperature boxes is fine, but we do also need to retain the temperature contour maps which are excellent for showing at a glance how the overall feel of the coming days is expected to change.


• There will apparently be more live forecasts on News 24. This is very much to be welcomed, so long as they're not hopeless 30-second affairs, since (despite having a list somewhere) I can never remember which News 24 forecasts are live at the moment! Hopefully forecasters will manage to restrain themselves from showing off too much, and to stick to giving out actual information.


• Finally, text: the absurdly dumbed-down summary at the start of the forecast is an utter waste of time. "Cool with some sun" is so vague they might as well not bother with it at all; I could probably work that out for myself! The whole point of watching a professionally produced forecast is to get detailed information from experts, not to get one sentence in three seconds. Also, the text saying things like "3PM FRIDAY" at the bottom is, if not terrible, not very well integrated either.

Overall? I'm afraid I don't expect to be happy once this comes in, and I wouldn't be at all surprised if some changes were made within the first month as a result of viewers' feedback. It was certainly interesting to note that Everton Fox didn't sound hugely enthusiastic about the new graphics on the radio. Even if it proves a disaster, scrapping the system altogether would be just too big a loss of face (and money) to be contemplated, but I would hope that at least some of the problems that I've listed will be addressed. The coming week should actually prove a good test for the new system, in that at least some parts of the country are likely to see rain, sun, snow, sleet, frost and thunder at various times, with quite widely varying temperatures. So, we'll see on Monday...