I’m going to kick things off by taking you behind the scenes. I’m writing this in my office space at home, looking out of the window at a sky that’s largely covered by thick, white clouds, with a patch of grey here and there. So naturally, I’m wondering if it’s going to rain when I pop out to get my steps in later.
That acts as proof to my opening point... us Brits are obsessed with the weather. When was the last time you went a day without at least one of the conversations you had including at least a passing mention to the temperature, the lack of summer sunshine or how cold it’s getting now autumn is upon us?
That may be because the weather impacts all of us to a varying degree. It’s a common denominator that makes for easy small talk. We all find ourselves getting annoyed when the weather forecast isn’t right and people remember Michael Fish and the hurricane.
So why are we focusing on the weather now, when it tends to be a constant discussion point? The topic of weather went beyond small talk and onto the front pages last week when parts of the UK experienced tropical style downpours, which caused the creation of a “new river” that submerged cars in Bedfordshire, while Luton experienced a “small tornado”.
What does the data say?
- The flood waters in Bedfordshire, which was seemingly the worst hit region of the UK, were 8 feet deep. This equated to enough water to fill two Olympic-sized swimming pools being pumped from the A421.
- Research shows Brits spend nine minutes a day (or FIVE MONTHS of their lives) talking about the weather, check forecasts 11 times a week and post weather-related social posts once a day on average.
- The UK has at least 100 words and phrases for rain. From the quite literal (downpour, torrential) through to the more visual (the heavens have opened, nice weather for ducks).
While our obsession with the weather is easy to ridicule at times, the recent flooding highlights a necessity to accurately predict extreme weather as early as possible.This is even more important in places where extreme weather is more persistent. East Africa, Asia and South America regularly experience hurricanes and heavy rains that destroy villages and cities.
Is this where AI comes in?
Google DeepMind’s MacRobert Award-winning GraphCast software was launched last November and it’s reported that it out-performed the European Centre for Medium Range Weather Forecasting across 90% of more than 1,300 atmospheric variables, including humidity and temperature. It also correctly predicted the landfall of Hurricane Lee much further north than traditional weather forecasting methods, 10 days in advance.
While meteorologists working in more traditional weather forecasting may question the value of AI over approaches they’ve been using for decades, there is clearly value in the ability to crunch 40+ years of weather data to predict outcomes and trends. Especially if they allow for additional warning when floods or extreme weather fronts are on the horizon.
Our client the Royal Academy of Engineering is hosting a panel discussion on exactly this on October 17th, chaired by BBC broadcaster and meteorologist Tomasz Shafernaker and representatives from the likes of the Met Office and Google DeepMind. You can register interest in attending in-person or online here.
Come back for next week’s Your Weekly Antidote, another dose of data-driven news analysis on one of the biggest stories of the week from your favourite comms tech agency.
