Projected climate memory and inherited warm-tail risk in accelerated European summer warming

The Weather's Long Memory

Imagine the Earth's climate not as a series of random daily weather reports, but as a giant, complex orchestra. In this orchestra, the "weather" is the loud, fast-playing violin section that changes every minute. But underneath that, there are deep, slow-moving cellos and basses—the oceans, the soil moisture, and the massive air currents—that take a long time to change their tune. Scientists call these slow-moving parts "climate memory." Just like a person who has been running a marathon might still feel the effects of that run days later, the Earth's climate can carry the "echo" of past conditions, like a warm ocean or a dry summer, into the next year.

Why does this matter? Because Europe has been getting hotter, and heatwaves are becoming more intense and dangerous. For a long time, scientists tried to predict these extremes by looking at simple trends (like "it's getting warmer every year") or by assuming that next summer will look a lot like this one. But this new research asks a deeper question: Does the Earth's "memory" of past slow changes, like a lingering warmth in the Mediterranean Sea, actually load the dice for future heatwaves? If the planet is carrying a heavy backpack of past heat, does that make it much more likely to trip and fall into a dangerous heat extreme, even if the weather that day seems normal?

---

The Paper's Story: The Earth's Heavy Backpack

In this study, researchers Mauricio Herrera-Marín, Alex Godoy-Faúndez, and Diego Rivera decided to investigate this "climate memory" across 28 different regions of Europe. They treated the climate like a giant machine where the current summer temperature is made of three ingredients: a smooth, slow background warming (like a rising tide), a "memory" of what happened in previous years (the slow cello section), and a brand-new "innovation" or surprise that happens right now (the fast violin).

To figure this out, they looked at data from 1950 to 2024, using a super-detailed digital map of the atmosphere called ERA5. They built a special kind of math model to separate the "memory" from the "surprise." Think of it like trying to guess how hot next summer will be. You could just guess based on the trend (it's getting hotter), or you could guess based on last summer (it was hot, so maybe this one will be too). But this team asked: What if we look at the "slow state" of the Earth—like how warm the Mediterranean Sea has been over the last few years? Does that warm ocean act like a heavy backpack that makes the whole system more likely to overheat?

The Main Discovery: The "Warm Tail" Effect

The paper found that yes, this "memory" is real and useful, but not in the way you might expect. It turns out that knowing the Earth's memory doesn't help you predict the average temperature of next summer very much better than just looking at the trend. If you just want to know if next summer will be "a bit warmer than usual," the memory doesn't add a huge amount of extra clarity.

However, the story changes completely when we look at the extremes—the really dangerous, record-breaking heatwaves. Here, the "memory" acts like a risk amplifier. The researchers found that when the accumulated memory of the Mediterranean region is high (meaning the ocean and land have been holding onto heat for a while), the chance of a "warm-tail" event—a super-hot day, a long heatwave, or a very high number of hot days—jumps up significantly.

Specifically, in years with high accumulated memory, the model predicts that the probability of these extreme heat events increases by about 8 to 11 percentage points compared to years with low memory. This holds true for looking 1, 3, and even 5 years into the future. It's as if the Earth's "backpack" of past heat doesn't make the average day hotter, but it makes the system so sensitive that when a heatwave does try to happen, it hits much harder.

What the Paper Rules Out and Clarifies

The authors are very careful not to overpromise. They explicitly state that they are not finding a single, perfect "memory kernel" (a specific mathematical shape that describes exactly how the memory works). They tried different ways to measure this memory—like a simple average of the last few years, a weighted average that cares more about recent years, and a more complex "tempered" filter. Surprisingly, all of these methods worked about the same. This suggests that while we know the memory exists and is useful, the data is too short to tell us the exact "shape" of that memory. It's like knowing a ghost is in the house because the lights flicker, but not being able to describe exactly what the ghost looks like.

They also clarify that their model is not a crystal ball for predicting the exact temperature of a specific day. In fact, if you try to use their model to predict the current state by adding the "memory" and the "surprise" together, you just get the actual temperature back (which is a tautology, or a circular statement). The real value is in using the "memory" part alone to identify which years are loaded with risk.

How Sure Are They?

The researchers are confident about the direction of their findings but cautious about the strength of the proof. They used a clever test called a "circular-shift placebo." Imagine taking the history of the Mediterranean's warmth and shuffling the years around randomly, so the warm years no longer line up with the hot summers. When they did this, the link between memory and heatwaves got weaker, but it didn't disappear entirely.

The statistical evidence is "moderate" rather than "decisive." The chance that this result happened by random luck is low (around 5% to 14% in their tests), but it's not low enough to be considered a slam-dunk, iron-clad proof in the strictest scientific sense. They also found that their model tends to underestimate how big the effect actually is in the real world. The model says the risk goes up by 8–11%, but the real data shows the difference between high-memory and low-memory years is actually much bigger (around 18–38 percentage points). This suggests that the "memory" loads the system into a state where it is ready to explode, but the final explosion depends on other things happening that same summer, like wind patterns or dry soil, which the simple memory model doesn't fully capture.

The Takeaway

So, what does this mean for us? The paper suggests that we should stop looking at climate memory just as a way to predict the average temperature. Instead, we should see it as a "risk-loading" variable. It's like checking the fuel gauge on a car. A full tank of gas (high memory) doesn't mean the car will drive faster on its own, but it means that if you step on the gas pedal (a heatwave event), the car has the potential to go much faster and cause more damage.

The authors conclude that while we can't perfectly predict the exact shape of the Earth's memory from just 70 years of data, we can definitely see that when the Mediterranean and other slow parts of the climate system are "loaded" with heat, Europe becomes much more susceptible to extreme heat. It's a warning system: if the memory is high, the dice are loaded, and we should be on high alert for the worst-case scenarios, even if the average forecast looks just a little warm.