Someone looked at the live map the other day and asked me a question I really should have asked myself sooner.
“Why are 74% of the assets earthquake exposed? We don’t really have earthquakes in the UK.”
Fair question.
In fact, it’s exactly the sort of question this project is supposed to answer.
The Numbers
At the moment the pipeline has processed around 3,450 real UK energy assets.
Here’s what it currently reports:
- 1.0% Flood Exposed
- 74.4% Earthquake Exposed
- 8.3% Wildfire Exposed
Two of those numbers look fairly believable.
One of them sticks out like a sore thumb.
Whenever I see a number that’s dramatically different from everything else, I stop trusting the summary and start looking at the data underneath it.
Does the UK Actually Have Earthquakes?
It turns out…
Yes.
Quite a lot of them.
The British Geological Survey records 1,476 earthquakes of magnitude 2.0 or above between 1970 and 2024.
That’s roughly 27 a year, which matches what the BGS says publicly. The UK experiences a couple of hundred earthquakes every year, but only around twenty or thirty are large enough to reach magnitude 2 or above.
The important bit?
Almost nobody notices them.
They’re simply too small.
So What’s Behind That 74%?
Once I stopped looking at the headline and started looking at the distribution, things made a lot more sense.
Of every asset currently marked as earthquake exposed:
- The median nearby earthquake is just Magnitude 2.3
- 87% are closest to an earthquake below Magnitude 3
- Only 66 assets are anywhere near a Magnitude 4+ event
- None are near a Magnitude 5+ event
That’s a completely different story.
A Magnitude 2 earthquake is basically background noise.
It’s something a seismometer notices, not something most people ever feel.
For comparison, the biggest UK earthquake in recent years was the 2008 Market Rasen earthquake, which reached Magnitude 5.2. Nothing in this dataset comes anywhere close.
Exposure Isn’t Risk
This is where definitions matter.
My current rule simply says:
Find the nearest recorded earthquake of Magnitude 2 or above within 25km.
Technically…
That works.
But Britain is a small country with low-level seismic activity scattered all over the place.
When you combine 54 years of recorded earthquakes with a 25km search radius, you’re almost guaranteed to find something near almost every asset.
So the pipeline isn’t really saying:
“This asset is at significant earthquake risk.”
It’s saying:
“At some point over the last five decades, there has been a small earthquake somewhere within 25 kilometres.”
Those are two very different statements.
Compare That With Flooding
Flood exposure currently sits at 1%.
Wildfire sits at 8.3%.
Those numbers are much lower because those hazards are genuinely localised.
Flood zones exist in specific places.
Wildfires happen in specific places.
Earthquakes in Britain are different.
Small tremors are scattered across the whole country, so the current threshold naturally produces a much higher exposure percentage.
The Data Is Still Useful
I don’t think the answer is to throw the earthquake data away.
Far from it.
The raw fields are actually useful.
Knowing an asset is 8km from a Magnitude 4.1 earthquake recorded in 2019 tells you something.
Knowing it’s simply marked:
Earthquake Exposed = True
…doesn’t tell you very much on its own.
Context matters.
So What’s Next?
I think the exposure model needs refining.
Maybe that’s:
- increasing the minimum earthquake magnitude,
- reducing the search radius,
- or introducing severity bands instead of a simple Yes/No flag.
That’s exactly why I built this project.
Not to produce pretty dashboards.
To challenge the numbers.
Because the biggest lesson here wasn’t about earthquakes.
It was this:
When a number looks strange…
Don’t argue with it.
Go and see what it’s actually measuring.
More often than not, the data is telling the truth.
It’s just answering a different question than the one you thought you asked.
Gareth Winterman