El Niño Is Back
El Niño is back.
And not just in the slightly boring, technically-it-counts sense either.
The World Meteorological Organization expects the current event to strengthen considerably through the end of 2026 and continue affecting global weather into 2027.
Closer to home, the Met Office says it could increase the chances of wetter and stormier conditions across north-western Europe during autumn and early winter.
That includes the UK.
And, more importantly for this project, the North Sea.
Predictably, the headlines have already started doing what headlines do.
“Unprecedented El Niño.”
“Stormy winter ahead.”
“Extreme weather heading for Britain.”
Which all sounds suitably alarming.
But once you get past the headline, what does it actually mean?
The Pacific Is Quite a Long Way Away from Nottingham
El Niño begins with unusually warm water in the tropical Pacific Ocean.
The North Sea is, rather obviously, nowhere near the tropical Pacific Ocean. Nottingham isn’t either, for that matter.
Its influence reaches us indirectly by changing large-scale atmospheric circulation, which can then affect the jet stream and alter the likelihood of particular weather patterns.
The important word there is:
Likelihood.
El Niño does not directly create one particular storm over the North Sea.
It does not mean every storm this winter will be stronger.
It does not even guarantee that the UK will have an unusually stormy winter.
The Met Office itself is clear that the relationship between El Niño and European weather is complicated and far from guaranteed.
It changes the odds.
That’s different from predicting the outcome.
So What Does It Mean for Offshore Energy?
The North Sea contains an enormous amount of energy infrastructure.
Offshore wind farms.
Oil and gas platforms.
Subsea cables.
Pipelines.
Electrical substations and grid connections.
All of it has been designed to operate in difficult conditions.
But difficult conditions still have consequences.
High winds can stop turbines from generating.
Large waves can prevent maintenance crews from reaching an asset.
Equipment faults can take longer to repair because nobody can safely get near them.
A relatively minor failure during calm weather might be fixed within hours.
The same failure during a major storm could potentially leave an asset unavailable for days.
That creates a fairly obvious question:
As North Sea storms become more severe, do offshore outages become longer or more frequent?
It feels like something somebody should already know.
And, to be fair, parts of the industry probably do.
But that isn’t quite the question I’m interested in. The question is whether I can demonstrate it using public data, without pretending the data says more than it does.
Headlines Aren’t Operational Data
This is where it stops being a weather story and turns into a data problem. Which is, naturally, the bit I find interesting.
A headline can tell me that a storm was severe.
An outage record can tell me that generation dropped or an asset became unavailable.
But neither one necessarily proves that the storm caused the outage.
An offshore wind farm may already have been undergoing planned maintenance.
A turbine may have stopped because wind speeds exceeded its operating limit. That is technically an interruption to generation, but it is also the machine behaving exactly as designed.
A grid connection problem could make an entire site appear unavailable even though the turbines themselves are fine.
And sometimes the public record may simply say:
Unavailable.
That is not enough to declare the storm responsible.
Timing alone is not causation. Two things happening on Tuesday does not mean one caused the other.
But timing is still a perfectly sensible place to start looking.
The Data Question
So I’m not going to treat El Niño as the cause of anything. I’m using it as the topical excuse to pull apart the wider question properly.
For each significant North Sea storm, I want to collect measurements such as:
- maximum wind speed
- minimum atmospheric pressure
- wave height
- storm duration
- storm track
- distance from offshore assets
I can then compare those storm windows against publicly available outage and incident records.
That should let me ask:
- Were more assets unavailable during severe storms?
- Did outages last longer?
- Were particular wind farms repeatedly affected?
- Did assets nearer the storm track show a stronger effect?
- Was the outage genuinely unplanned?
- Did bad weather delay the restoration rather than cause the original fault?
That last distinction matters.
The storm might not break the equipment.
It might simply stop anyone getting out there to fix it.
From an operational and insurance point of view, that can still be very expensive.
Reusing the Global Shock Pattern
I have already built something similar for the Global Shock project.
That pipeline separates:
- the underlying time-series data
- the real-world event
- the phases that followed it
Those datasets are then connected using a stable event_id.
The subject was different.
Oil prices instead of offshore outages.
Wars and supply disruptions instead of storms.
But the underlying data problem is remarkably similar.
Something happens in the real world.
A time series moves.
The difficult bit is working out what happened during the event, what happened afterwards, and whether the two things are genuinely related.
For this project, each storm becomes an event.
It gets an event window, severity measurements and the offshore assets likely to have been exposed to it.
Outage records can then be attached to the same event without mixing the storm data, operational data and chart logic into one enormous file.
Same pattern.
Different risk.
The Answer?
I don’t know.
Severe storms might clearly correlate with longer offshore outages.
They might only affect access and repair times.
The public outage data might be too vague to produce a reliable answer at all.
That last result would be mildly irritating after doing all the work.
But it would still be a result — and a much better one than forcing the data into a dramatic answer it can’t support.
Because this project is not about forcing a dramatic conclusion from an interesting chart.
It is about building a defensible way to test the question.
El Niño gives me the topical starting point.
The storms give me the measurable events.
The outages give me the operational consequence.
Now I need to see whether the data can genuinely connect them.
Gareth Winterman