I still can’t fully understand how poorly the Met Office dealt with their warning of heavy rain and thunderstorms across the south for last night, but with the help of the BBC Spotlight Southwest forecast at 2230 BST, I think I can show how both the BBC and the Met Office got it wrong.
Figure 1 – Courtesy of the BBC
The first frame that David Braine shows is the actual weather radar image for 22 BST (good on you David), which shows a band of thunderstorms along the English Channel, that he says are heading north which is what they were doing, well a bit east of north (fig 1). He must realise that these storms will affect Devon, but he’s in a dilemma, because he now’s stuck with showing forecast graphics that are out of step with the radar.
Figure 2 – Courtesy of the BBC
The next frame he shows is the classic cut to the NWP forecast frame for 23 BST, and it’s far from seamless, with just two active cell aligned SW-NE, rather than numerous cells aligned E-W in the weather radar. Cornwall and Devon are now clear of any activity, either from rain or thunderstorms (fig 2).
Figure 3 – Courtesy of the BBC
The next frame at midnight swiftly takes the cells NE and shrinking them, especially the one in Lyme bay (fig 3).
Figure 4 – Courtesy of the BBC
And hey presto in the next frame for 01 BST, the Lyme bay storm has vanished (fig 4).
So the NWP forecast weather was way out of synch with the weather at 22 BST, and then the model proceeds to ignore the observational radar weather data as if it wasn’t there. Not only does the model dismiss the rain, but the forecaster dismisses it as well, with a flick of the wrist and a classic David Braine comment of “so we end up with a largely dry night“.
I saw David Braine interviewed a little later on the rolling BBC news channel, and he did emphasise that the storm in Coverack was very localised, but he then spoils it all by saying that we can’t say how much rainfall fell because we have so few rain gauges, but he forgets that we do happen to have a state-of-the-art weather radar and lightning detection networks that’s sorely under used in situations like this. We can estimate quite accurately just how much rain fell, anywhere yesterday (fig 7), I know because I do it on a regular basis and blogged about the ‘hot-spot’ off the Lizard yesterday before the story had been widely picked up by the media.
The growing problem with weather forecasting since the 1980’s, is I believe is its over reliance on NWP, especially in the short-term, and at times it’s almost total disregard of observational data, especially if it conflicts with what the model is saying. That kind of thinking explains why the Met Office were so reluctant to extend yesterdays warning for heavy rain beyond 2355 BST, which in my opinion should have been done by 2200 BST at the latest on the evidence of the weather radar, especially so in light of the earlier flash flooding that occurred at Coverack. Curiously, the Chief was finally persuaded to update the existing warning area for today to include the south (fig 5), but it didn’t happen till 0127 BST (the time I received their email), and the previous warning had been out of force for around 90 minutes.
The Thunderstorms in mid-Devon put on a spectacular lightning show from 2330 to 0230 BST, with sporadic heavy rain at times, I recorded 23.8 mm of rain in the 24 hours at Bradninch, Exeter airport reported 10.2 mm, but Dunkeswell just to the east, received only 1.2 mm of rain which underlies the localised nature of the storms. Here’s the weather radar sequence for last night from 21 UTC to 02 UTC (fig 6).
And just to emphasise why the warning area required extending after 2355 BST to include all southern coastal counties from Devon eastward, here’s my estimated rainfall accumulations from the weather radar for 18-06 UTC. As you can see, a swathe of areas across Hampshire had rainfall totals in excess of 50 mm in that period, it’s also interesting to see the track of the thunderstorms as they moved NNE up the Exe Valley in Devon (fig 7).
Most of the gauges that reported 24 hour rainfall totals in the 06 UTC SYNOPs missed the heaviest of the rain, which is understandable given the localised nature of thunderstorms and how they tracked (fig 8). I know that there are many more gauges out there than this, and I would love to get my hands on the totals from the automatic gauges and fill the gaps.
I think the Met Office had basically got the wording in the yellow alert for heavy rain for yesterday correct. Where they got it wrong was not extending the area of the warning for the next day to include southern coastal counties in good time. What if a Coverack type deluge had happened to a more populated location such as Portsmouth or Brighton just after midnight when there was no warning in force at all? I can almost detect a kind of arrogance in how they delayed updating the warning, as if they know best, and wouldn’t be hurried into doing it, the flash flooding at Coverack was a warning (excuse the pun) which they failed to see, but fortunately they got away with it, this time.