5 ms·
Compare and contrast the American and the British weather services. In the USA it was the case that, if the taxpayer paid for it, then it was free. Meanwhile,
by Theodores 1y ago
Compare and contrast the American and the British weather services.
In the USA it was the case that, if the taxpayer paid for it, then it was free. Meanwhile, in the UK, the Thatcher government decided that it would be best if government agencies earned their keep. As a consequence, in the UK, weather data became chargeable. This meant that there was no growth of third party applications that built on Met Office data.
This became a market distortion with TV weather, where the BBC had their data from the Met Office whereas ITV had to use American data, with a specialist broadcaster - The Weather Department - providing the video inserts. The Weather Department eventually came into competition with the Met Office providing their own video inserts, so the marketplace was uncompetitive yet 'competitive'.
Compare with the BBC online news, which does not have to pay for itself in the same way that the mainstream news organisations such as The Guardian, The Telegraph et al. have to either go for paywalls or a smorgasbord of adverts. It is essentially impossible to compete against the BBC for news eyeballs online in the UK due to this competition model.
With weather you need two independent supercomputer outfits that crunch the numbers and come up with forecast data. Why two (or more)? Because forecast models can be wrong in ways that human forecasters can understand and work with. If there are two 'sources of truth' then you can use your gut to go with which one feels right.
Vast supercomputer outfits cost lots of money and, until recently, only a government of considerable size could pay for such an investment. The investment isn't just in the computers, you need a constant stream of meteorologists that can work with the data, so that means one or more universities with the specialism. In the UK we have Reading, that is the best place to go to if you want to get into forecasting, for TV, the military, farming or aviation.
Free weather data is a public good and also 'soft power'. In Ireland they don't have their own weather supercomputers or universities that are renowned for producing expert weather forecasters, hence they need British or American help.
I also forgot the satellites, which is in another league of expense, but necessary for all kinds of observation data.
Observations are interesting as the reliance has historically been on airfields. We missed a chance to get mobile phone base stations reporting in to have vastly more data. Given the amount of money involved in mobile, when the governments sold off the spectrum they could have used the opportunity to insist on weather observation data being collected, in order to refine the model.
Sometimes you get a town next to a lake with a massive mountain behind it where the forecast will consistently be wrong due to how the modelling works and a lack of observation data to make corrections. There are many scenarios such as this and the mobile phone masts could have been used to fill the gaps.
If the UK provided free data (and the USA didn't) then we wouldn't talking about AccuWeather, it would be some British company that the world would be relying on for their apps. This British company would be modest in size but still providing good jobs and keeping graduates from Reading employed. In turn they would be working with people developing specialist apps, for example, weather for race courses and there would be much tax paid to government to make it worth their while, plus national prestige in having 'the best weather service' going.
Another example of a British mistake is the Ordnance Survey, purveyor of paid for maps. Nobody uses Ordnance Survey, it is always Google Maps, Apple Maps or OpenStreetMap. They get pennies from developers and architects that need the official Ordnance Survey maps but nothing from day to day usage, in effect they have lost out.
All these things need to be for the public good, paid for out of public taxation and a genuine free market of private enterprise fostered, otherwise nothing happens.
- teamonkey 1y agoMet office weather data has a free tier, an API with a limit of 380 calls per day.
- Theodores 1y agoThanks for that, I just checked out their free tiers and there is enough there to get started with. This is very much for hobby or development purposes, which is fine. Imaginably I might want to write an app that needs weather data at scale, even for a simple thing. Imagine I sold chocolate teapots online and, in the delivery notes, wanted to advise customers to not leave their deliveries outside for too long if it was going to be sunny in their postcode area on the day of delivery. The free tier might work. If it was something more popular, for example milk or veg boxes, then I would need to start paying. This would be for a marginal value but of wording in a delivery note, but something that might differentiate the service. Or, imagine I sold outdoor gear, with a newsletter that goes out to 100,000 customers. Depending on their postcode I might want to promote deckchairs or umbrellas. Although this level of personalisation isn't going to win any awards for innovation, it is something that could be done, enabling a medium sized company to compete against the big companies that are optimising their promotions for the weather. There is also the research aspect, to verify a suspected phenomenon against data in an amateur way that you would never get a research grant for. All considered, I wish they did free with no restrictions.
- teamonkey 1y agoI feel that it’s fine that businesses are charged for heavier use of a tax-funded service, especially when it would increase the running cost of those services, and where there’s no obvious benefit to the taxpayer. I get the argument about using taxpayer services to leverage a more competitive business environment, and I agree that’s important, but I don’t think an advertising newsletter is a good example. And honestly the data’s fairly cheap for high-accuracy weather data. You get a lot of API calls for not very much money.
- Theodores 1y ago