How weird was the Mediterranean Sea heatwave?

Summary

---

More or Less (BBC Radio 4):

"How weird was the Mediterranean Sea heatwave?"

Episode Date: August 16, 2025
Host: Lizzie McNeil
Guest: Jules Kaita, Physical Oceanographer, National Oceanography Centre (UK)

---

Episode Overview

This episode investigates a viral claim on social media that the 2025 Mediterranean Sea heatwave represented a "1-in-216-billion-years" anomaly. Host Lizzie McNeil, with the help of physical oceanographer Jules Kaita, explores how sea surface temperatures are measured, why the claim is mathematically questionable, and what the actual odds are for such extreme events—placing it in the context of climate change and statistical reasoning.

---

Key Discussion Points and Insights

1. Origins of the 216-Billion-Year Claim

• Listener Prompt: John O’Malley alerted the show to a tweet from "MetForecast UK" about the Mediterranean Sea’s record temperatures.
• Quote:

  “They stated that the Med is experiencing a 1 in 216 billion year sea temperature anomaly.”
  (A/Lizzie McNeil, 00:36)

• Observation: This figure exceeds not just Earth’s existence but the age of the universe, prompting skepticism.

2. Mediterranean Sea Heatwave by the Numbers

• Extreme but Real:

  “The Mediterranean wide average temperature during June and July, for example, were about 3 to 4 degrees hotter than the 1985 to 2005 mean.”
  (B/Jules Kaita, 01:20)

• Temperature Measurement: Discussion of how satellite data since the 1980s enabled precise tracking of sea surface temperatures.

  “The last four years, in fact, have also been the warmest on record in the Mediterranean on this satellite record.”
  (B/Jules Kaita, 02:37)

3. How Are Sea Temperatures Measured?

• Modern Methods: High-resolution, daily, satellite-based infrared readings, only available since the 1980s.
• Surface Data Limitation: Satellites only measure the top centimeters to meters—significant because the surface is warmer than deeper water.
  (01:48–02:26)

4. Parsing the "Anomaly" Calculation

• Local Vs. Global: The Twitter claim used figures from specific Mediterranean spots, not the whole sea.
  (A, 02:55; B, 03:46)
• Specific Observations:
  • Some locations saw anomalies of 5 to 8°C above the 1985–2005 average.
  • Example: South coast of France hit 27.7°C vs. an historic mean of ~20°C.

    “Around as hot as a heated swimming pool for adults.”
    (A, 04:03)

  • Absolute records (31°C+) were set in 2024, not 2025.
    (B, 04:26)

5. Statistics 101: Interpreting Extreme Events

• Attempting the Math: Jules used the standard method:
  • Analyzing daily temperature anomalies over the reference period, sorting them into bins (e.g., 2–3°C, 3–4°C), and fitting a curve (Gaussian/normal distribution).
    (B, 05:34–06:20)
• Quote:

  "One of the things that's really nice about a Gaussian distribution is that it can be described statistically with only two numbers... the mean... and the standard deviation."
  (B/Jules Kaita, 06:20)

• Catch: Real-world data didn’t fit a normal distribution—so the calculation significantly overestimated the rarity.

6. What’s the Real Probability?

• Refining the Stats:

  "I've done some calculations with a more refined statistical approach, and I would actually estimate that a better value puts the return period on these temperature anomalies at more like one to 2,000 years. So one in kind of 2,000 year events."
  (B/Jules Kaita, 07:40)

• Quote:

  "There's quite a big difference between 1 in 2,000 years, which I would have estimated, and 1 in 200 billion, but 1 in 1,000 or 2,000 years is still enormous."
  (B/Jules Kaita, 08:14)

7. Link to Climate Change

• Key Point: These extreme events are extremely unlikely—unless baseline climate is shifting.
• Quote:

  "If you assume that nothing is changing in the environment... then the kind of water temperatures we saw in the Mediterranean this summer are extremely improbable."
  (A/Lizzie McNeil, 07:55)

• Conclusion: Factoring in human-induced climate change, these events become much more plausible.
  (A, 08:31)

---

Notable Quotes & Highlights

• "The Med is experiencing a 1 in 216 billion year sea temperature anomaly."
  (MetForecast UK via Lizzie, 00:36)

• "The Mediterranean wide average temperature during June and July, for example, were about 3 to 4 degrees hotter than the 1985 to 2005 mean."
  (Jules Kaita, 01:20)

• "I've done some calculations with a more refined statistical approach... the return period on these temperature anomalies is more like one to 2,000 years."
  (Jules Kaita, 07:40)

• "If you assume that nothing is changing in the environment... the temperatures we saw this summer are extremely improbable."
  (Lizzie McNeil, 07:55)

• "One in 1,000 or 2,000 years is still enormous."
  (Jules Kaita, 08:14)

---

Key Timestamps

• 00:36 — Introduction of the viral 216-billion-year figure
• 01:20 — Discussion of the anomaly and recent sea temperatures
• 02:37 — Satellite record confirms consecutive record-hot years
• 03:46–04:26 — Local anomalies of up to 8°C, and specifics about sea temperatures in certain regions
• 05:34–06:20 — Describing the statistical approach (Gaussian distribution)
• 07:22 — Explains why the normal distribution fails here
• 07:40 — Jules computes the actual return period (1-in-1,000 to 2,000 years)
• 08:14–08:31 — Summary: Statistical unlikelihood without climate change

---

Summary

The episode dissects a sensational social media claim about Mediterranean sea heat, reveals how extreme but not unprecedented the event was in a meaningful statistical context, and demonstrates that, while the viral number is wildly inaccurate, the real story is still deeply significant: these marine heatwaves are extraordinarily rare events in a historical climate but growing less so as human-driven warming continues. The show emphasizes careful statistical reasoning while underscoring the dramatic shift underway in our global climate system.

Transcript

[A]

Hello and thanks for downloading the More or Less podcast. We're the program that looks at the numbers in the news and in life and in the sea. And I'm Lizzie McNeil. Listener John O' Malley emailed in to more or lessbc co.uk, which by the way, anyone can do to ask about a tweet that appeared to be making a somewhat staggering claim. The post seen by more than a million people are and or bots on the social media site was from an account called MetForecast UK talking about how hot the Mediterranean Sea got at the start of July.

[B]

They stated that the Med is experiencing a 1 in 216 billion year sea temperature anomaly.

[A]

This, as the post pointed out, was a significantly longer period of time than planet Earth has been in existence. Indeed it is many more times than our best estimates of the age of the entire universe. The massive scale of this anomaly, so the account claimed, proved that climate change was real. So does the statistical logic make sense?

[B]

There's no doubt that the temperatures in the Mediterranean this year have been extreme.

[A]

This is Jules Kaita, a physical oceanographer working at the National Oceanography Centre in the uk.

[B]

The Mediterranean wide average temperature during June and July, for example, were about 3 to 4 degrees hotter than the 1985 to 2005. Me.

[A]

Now we're talking about degrees centigrade here and Jools is comparing the temperature now against the average sea temperature at the same time of year in that period, from the 1980s to the noughties. But why start in the 1980s? Well, that's when the good quality data starts.

[B]

The most accurate way that we measure sea temperatures is at present is from satellites. They give us high resolution and daily observations at the sea temperature.

[A]

Satellites in space measuring the temperature of the sea might sound a bit balmy. Basically everything emits infrared radiation according to its temperature, which a sensor on the satellite can well sense.

[B]

But one of the limitations is that they've actually only been up there since the 1980s. Another limitation of satellite observations is that they only measure the temperature of the sea surface. It roughly represents that kind of top centimeters to meters kind of scale.

[A]

This top layer of the sea is warmer than the depths, particularly if it's not stirred up too much by waves and storms. This data set has the advantage of always measuring the same thing and the trend is clear.

[B]

The last four years in fact, have also been the warmest on record in the Mediterranean on this satellite record. And each of those four years have been about one and a half degrees hotter than this 1985-2005 mean. So it's been hot and it's getting hotter.

[A]

So that's the big picture. Let's get to the 1 in 216 billion year anomaly thing. As with Jules initial comparison, the temperature anomaly we think they're talking about in the Twitter post is worked out by figuring out the difference in temperature on certain dates this summer compared to those average temperatures between 1985 and 2005. I say we think, because much of this is not totally clear from the Twitter post. We've tried to get in touch with the account, but no one's got back to us. Anyway, the big difference is that the temperature anomalies are for specific locations in the Mediterranean Sea, not the whole thing. Technically, that means that the statistical calculations only apply to these specific, specific locations too, not the whole med.

[B]

The temperature anomalies that we have seen, particularly in early July, were about 5 or even up to 8 degrees in some places. So that means that 5 to 8 degrees warmer than at that time of year. Going back to the reference period in

[A]

the quiet waters off the south coast of France, between Spain and Italy, the sea temperatures reached 27.7 degrees centigrade in early July, when the average from 1985 to 2005 was more like 2028 degrees is hot. Around as hot as a heated swimming pool for adults. Although just to be clear, the med does get hotter.

[B]

Surprisingly, these anomalies are actually just short of breaking the record in terms of absolute temperatures. So because early July isn't actually the hottest time of year in the Mediterranean, even though they're the hottest anomalies, they don't correspond with a the hottest absolute temperature. But actually, the record temperatures in the Mediterranean, which were in excess of 31 degrees in some places in absolute terms, occurred during 2024. So also very recently.

[A]

Anyway, back to the sea temperature anomaly being a one in many, multiple times longer than the age of the universe event, Jules thinks he's managed to reverse engineer the calculations that ends up at that absurd conclusion.

[B]

So it does sound a little bit wild. But I've done some analysis of the recent Mediterranean temperatures to try to get some understanding of where this number might have come from. I had to make some guesses about the methodology that they used.

[A]

Okay, belt up. We're heading into Statsland. The first step in this calculation is to work out the probability of any one of these sea temperature anomalies taking place.

[B]

So. So if you were to take all of the temperature anomalies from every day over some period of Time, Maybe it's that 20 year reference period that I mentioned. And then you count up the anomalies that fit into particular Bruins, say between 0 and 1 degree, 1 degree and 2 degrees, 2 degrees and 3 degrees. And then on the negative side as well. So you can have negative anomalies between 0 and minus 1, minus 1 and minus 2, and so on. So if you put all of those anomalies into the relative bins and then count the number of anomalies that you have in each bin, you can then fit essentially a curve to that data.

[A]

The curve that fits the data points will tell you how likely any future events are if the context stays the same.

[B]

And one of the simplest ways that you can fit a curve to that, and actually a fit that works for a lot of data samples, is what's called a Gaussian. So one of the things that's really nice about a Gaussian distribution is that it can be described statistically with only two numbers. One is the mean, so that's where the peak of the Gaussian lies. And then the other is the standard deviation. And essentially the standard deviation describes the region within which about a third of the data lies.

[A]

A normal or Gaussian distribution follows specific mathematical rules which, which mean that if you collect new data, you can calculate how probable those events are by working out how many standard deviations they are from the mean or average. When he did that analysis with the Mediterranean heat anomaly in July, hey presto, Jules came out with similarly massive numbers. But there's a catch.

[B]

Turns out that actually the Mediterranean doesn't follow a normal distribution quite so well.

[A]

The method the Twitterer used to calculate this number doesn't actually match the likelihood of observing different sea temperatures between 1985 and 2005. When Joules applied some statistical techniques to better match the actual distribution, he came up with a very different number.

[B]

I've done some calculations with a more refined statistical approach, and I would actually estimate that a better value puts the return period on these temperature anomalies at more like one to 2,000 years. So one in kind of 2,000 year events.

[A]

However, while the stats were not very good in that Twitter post, the key point was not entirely wrong. If you assume that nothing is changing in the environment, that is zero climate change, then the kind of water temperatures we saw in the Mediterranean this summer are extremely improbable.

[B]

Clearly, there's quite a big difference between 1 in 2000 years, which I would have estimated, and 1 in 200 billion, but 1 in 1000 or 2000 years is still enormous. So what does that actually mean? They are statistically extremely unlikely unless we take into account the change in the climate.

[A]

Once you factor in human made climate change, the likelihood of these events completely changes. Although you probably don't need a ridiculous timescale to say that. Thanks to Jules Kiter. And that's it for this week. If you've seen a number you think we should take a look at, please email us at more or [e-mail] until next time. Goodbye.