£1.2 Billion for weather and climate supercomputer

All I can say to this news is – I wonder just where Boris is going to get the £1,200,000,000 (hopefully I’ve got the right number of zeroes) to splash out on this new piece of hardware for the Met Office? Most supercomputers have a very short shelf life, possibly not much more than five years, they’re almost obsolete before they’ve been installed, and I’ve never heard of a second hand market for them, because surely you just can’t drive down to the recycling centre with one. The Met Office had obviously been haranging the Government for a new supercomputer for sometime, and have even had a computer centre built to house it in on the Science Park in Exeter. Perhaps they used the departure of the ECMWF and their supercomputer to convince the Government, that if we are to compete with the best in the world after Brexit, we will need an even more powerful supercomputer. I read in David Shukman article (see below) that even this building might be surplus to requirements as the whole thing might be run remotely and run from Iceland to negate the effects on our CO2 emissions!

Will greater accuracy help?

It strikes me that being more accurate in the future about warnings of severe weather will be of little use. Take the heavy rain for storm Dennis as an a prime example, yes we knew storm Dennis was coming five days ahead of time, and the Met Office dished out the appropriate amber warnings to the areas concerned (although the red one was a little too late), but how many people heeded them? The EA have mobile apps than can email and alert you when your postcode area is at risk of flooding, obviously it’s a much under used service. Perhaps the larger problem we have is one of poor communication rather than accuracy.

Extending the reach of their forecasts

If the Met Office are going to convince me that this new £1.3 billion investment will be a wise investment, they would have to plan to extend the reach of their new NWP models in the future, not just to day ten but out to day thirty. What we have at the moment in the way of a monthly forecast from them using their ‘old’ £97 million supercomputer is shockingly imprecise.

Courtesy of the Met Office

If this is all they can manage in the way of long-range forecasts at the moment, then I have little confidence in their predictions about what will happen to the climate in 30, 50 or a 100 years time. Interestingly they don’t say anything about extending their forecasting reach, which they know very well that they can’t beyond ten days with any accuracy, so they make a big play about the resolution of the model on the new supercomputer which will be down to one kilometre across the UK.

What’s in it for us?

My biggest gripe with the Met Office is that any NWP data produce – and they produce vast quantities of it – the taxpayer in the UK never sees. As far I can tell Meteogroup – who now have the contract for producing the weather forecast on the BBC – rarely use NWP data from the UKMO although they can access it (at a price?). It’s difficult to be certain – because they never badge which particular model they are using – but I would guess that they use the rival NWP data from the ECMWF in most of their broadcasts. So the argument seems to be – let’s spend £1.3 billion on a new supercomputer to improve our forecasts that we will never see, of course the Government might argue that they know full well the BBC is not going to be around much for longer in its present state, so why worry about that.

Free up our NWP Data

The only NWP data that I can access from the Met Office, is a number of fax (facsimile) charts that they still broadcast, and which I download from NOAA, the origin of these charts go back to just after the second world war, and why they are still produced is a mystery to me. This hoarding of NWP data is unlikely to change even after they’ve installed their £1.2 billion supercomputer in Iceland or wherever. If the Americans can publish their NWP data, as do the Canadians, Irish, French and Germans, so why is it the Met Office just likes to sit on ours, just like Smaug sat on his mountain of gold in the Hobbit, and not let us see our good the model is?

Courtesy of the Met Office

In conclusion

My advice to the government would be to merge the Met Office and Environment Agency (including SEPA), and rationalise them into one single entity that covered the whole of the United Kingdom, and would be responsible for warning and coordinating a national response to the threat of any severe weather event be it drought, pluvial or coastal flooding, heatwave, wildfire or snow.

These natural disasters will become even more common in the future as the world continues to warm. Just look how devastating the bushfires and the flooding has been in Australia in the last few months, it requires a joined up response to manage the risk, long gone are the days Denis Howell and a minister for drought.

I was told when I worked for the Met Office that they owned the raindrop while it was falling, but as soon as it hit the ground it became the property and responsibility of the EA. That’s why the EA have hundreds more automatic rain gauges across the country than do the Met Office – how ridiculous is that.

I would ask the Met Office to soldier on with their old £97 million supercomputer – which was only installed in 2017 – and at the same time free up their NWP data so that we could all see it. The money saved from the merger, combined with the money saved by not investing in a new £1.2 billion supercomputer, could then be spent on improving the old Victorian drainage infrastructure of our towns and villages to mitigate the effects of heavier rainfall events that we will experience across the UK in what’s left of the 21st century, at t least then we would see some tangible benefit from it.

And finally…

And for another angle on the story here is the BBC News item about it from David ‘global warming’ Shukman.

Already the Met Office is pulling in more than 200 billion observations from satellites, weather stations and buoys out in the ocean every single day, and that’s set to increase.

David Shukman

I would be interested to see a breakdown of the 200 billion observations that David Shukman says the Met Office to use each day. I realise that they collect an immense amount of data from satellite and aircraft, but two hundred billion is a very large number. I typically download ~8,000 SYNOP observations each hour from ships, buoys, and land observations both manual and automatic. Even if you reduced a SYNOP observation into its constituent elements such as visibility, wind speed, wind direction, temperature, dewpoint, pressure etc, this would still only give you ~3.8 million different weather elements in a day. Not only that I was told many years ago that todays NWP models are only interested in the MSLP and maybe the temperature these observations provide, the rest is discarded. I think someone needs to have a look at how they did their maths to get 200,000,000,000 on this one.

Scroll to Top