[00:00]
The butterfly effect is much stranger than we thought. A group of researchers just showed that if you take into account quantum physics, then it doesn't just make predictions difficult, it makes
[00:12]
predictions impossible after a certain time. They worked this out concretely for one of the most complicated chaotic systems we know, Earth's atmosphere. Does this mean that we can't predict
[00:26]
the climate either? Let's have a look. Stress can damage your health both physically and mentally. Ask me how I know. If you struggle with mental health problems or know someone who does, today's
[00:40]
paid partner BetterHelp can help. BetterHelp matches you online with a credential therapist based on your needs. You can talk by video, phone, or messaging, whatever works for you. I
[00:52]
think BetterHelp is a service that the world needs because especially with mental health problems, it shouldn't eat up someone's last remaining energy to find a therapist. For me, this was a real obstacle. Yes, I know they've seen criticism in the past, but they've made changes and now have
[01:09]
a 4.8 out of 5 rating in the App Store. More than 6 million people have gotten BetterHelp this way. Use betterhelp.com/sabine or my code during sign up for 10% off your first month. And now,
[01:25]
back to the science news. The name butterfly effect goes back to meteorologist Edward Lorenz who found out that the weather on our planet is extremely sensitive to small perturbations.
[01:38]
The flap of a butterfly's wings in Brazil could set off a tornado in Texas. More formally, if you start with two initial states that are just a tiny little bit different, then over time, the difference between them can grow dramatically. This part of the butterfly effect is quite well
[01:55]
known, but few people knew that this actually wasn't Lorenz's point. His point was what climate scientist Tim Palmer has called the real butterfly effect. If you read Lorenz's original paper,
[02:08]
you'll see that he was really arguing that weather prediction suffers from a sort of Zeno effect. If you improve the precision by which you know the initial conditions by a factor of say 2, then
[02:20]
you can increase the time over which you can make an accurate prediction by one half. Improve the precision by another factor of 2, and you add another one half squared, then one half cubed, and so on. This means that even if you could measure the initial conditions to infinite precision,
[02:37]
you'd not be able to make predictions beyond a certain time horizon, even though the system's fully deterministic. Whether this real butterfly effect is actually "real" in the sense of happening
[02:49]
in nature has never been settled, because Lorenz didn't, strictly speaking, derive it. He just made a compelling argument. I found it both interesting and somewhat concerning that this real butterfly
[03:01]
effect has gotten almost entirely lost in the literature, because if it happens in nature, it doesn't just happen for the weather forecast. It'd be a general property of chaotic systems
[03:13]
relevant for everything from nuclear fusion to quantum computing to black holes. In any case, the authors of the new paper now looked at this question, whether there is such a finite horizon for predictions for the weather, and they found that yes, there is, and it's 129 days.
[03:31]
After that time, they say weather forecast isn't just practically, but even theoretically impossible. Even if you knew the initial state perfectly, had a perfect model, and unlimited computing power, you still wouldn't be able to get past the 129 days. Their argument,
[03:49]
however, is rather different from the argument that Lorenz made. Lorenz made a general argument which said that inevitable changes on small scales like molecules would spread to larger and larger scales, eventually making predictions impossible. The authors of the new paper have sidestepped the
[04:07]
question of just what happens at the limit to infinitely small scales by instead asking what the physical causes for unavoidable disturbances are and what we know about them. Their answer is kind of obvious. Earth constantly receives quanta of light from the sun,
[04:24]
but we don't know the exact quantum state of the sunlight. We can't know it. Quantum physics forbids it. This quantum uncertainty first affects very small scales, but from there it spreads to larger and larger scales until it has spread through the entire atmosphere.
[04:39]
They calculate how long it take for the energy of the quanta to spread, and that gives them the 129 days. What does this mean for climate models that make predictions over years or hundreds of years?
[04:52]
The maybe somewhat surprising answer is pretty much nothing because climate models already deal with averages. They'll tell you the average ocean temperature as a function of average carbon dioxide or the probability of certain extreme events. They won't tell you whether it rains
[05:08]
in Paris on July 14th, 2030. The calculation for the butterfly effect tells you how long it takes until average prediction is the best you can do. It doesn't make the average prediction
[05:20]
any better or worse. I give this paper a 5 out of 10 on the [ __ ] meter. That's because I'm not super convinced by their argument that quantum effects entirely erase the information of the initial state. Still, I think that this paper's conceptually an important contribution
[05:37]
to the literature and I wish that physicists in other disciplines would pay more attention to this. The intersection of quantum physics and chaos is maybe the most underrated research
[05:49]
direction in the foundations of physics today. When you want to know if there's a fundamental limit to the precision your quantum computer can reach before spending a few billion dollars on it, maybe that's just me. If you try to predict what someone will say from what
[06:05]
they've previously said, how many words does it take until you lose all predictive power? Broccoli. Do you think ChatGPT would have made this joke? Thanks for watching. See you tomorrow.