Predicting space weather with AI

Space weather matters. Not just in space, where it can be a matter or life or death, but also on earth, where it can create amazing beauty — think auroras — or incredible destruction.

For example, the Carrington Event was the most powerful geomagnetic storm on record, occurring in September 1859. Caused by a massive coronal mass ejection (CME) from the Sun striking Earth’s magnetosphere, it knocked out telegraph systems across Europe and North America, created auroras that were seen as far south as the Caribbean and Hawaii, and lit up the night sky so brightly that people reportedly thought it was dawn.

If a similar event occurred today, it could disrupt satellites, power grids, and communication systems worldwide.

Predicting space weather with AI

In this episode of TechFirst, we chat with Alex Pospeckov, CEO of Mission Space about the critical role of space weather forecasting. Mission Space is deploying advanced AI and a planned constellation of 24 satellites to predict space weather and protect billions of dollars worth of assets … as well as astronomer’s lives.

Alex explains the science behind solar flares, coronal mass ejections, and their potential devastating impacts.

00:00 Introduction to Space Weather
01:16 Understanding Space Weather
03:26 Historical Impacts of Space Weather
07:38 Predicting Space Weather with AI
11:30 Challenges and Solutions in Space Weather Prediction
17:14 Future of Space Weather Prediction
24:44 Conclusion and Future Plans

 

Transcript: space weather

Note: this is AI-generated and not a perfect record

John Koetsier: Did you know that we’re now predicting weather in space with AI? Hello and welcome to TechFirst. My name is John Koetsier. Most people may not be aware of it, but space weather exists and it matters even if you’re never gonna get showers or snow out in space.

Today we’re chatting with the CEO of a space weather prediction service that impacts billions of dollars of assets both in space, and right here on Terra Firma also has a potential to save lives. His name is Alex and, oh, it’s not Alex Poe. It’s his name is Alex Pospeckov, and he’s the CEO of Mission Space.

Welcome, Alex.

Alex Pospeckov: Hello, nice to meet you.

John Koetsier: Nice to meet you too. You threw me off there with your LinkedIn and the Alex Poe, but I got it right now. let’s start here. most people aren’t aware of what space weather is, what it looks like, what impacts has, what is weather in space.

Alex Pospeckov: So the simple way to describe it, imagine you are traveling in space.

And then emits this invisible bursts of radiation energy charged particles, which can be absolutely destructible. This is what we call space weather. And when something like this happens, it affects us. It’s in space on earth, and of course, in danger Life of astronauts in space tourism.

John Koetsier: Exactly. So if you’re in the ISS, the International Space Station, then there’ll be a place where you can go, as protected as you can get, whether the walls have, thicker metal or maybe water around to try and protect from some of those energetic.

Particles, some of that radiation. it’s funny because on your website, you list as one of the danger factors is space tourists. I mean, that’s gotta be one of the newest things, right? I mean, I know it exists and we see it. you know, SpaceX is doing it. We know that blue Origin is doing it,

But, there’s also ground-based installations that are impacted by this. Talk about some of that.

Alex Pospeckov: Yeah. So with, with Isaid, they have special room filled with, lead and well, when they have space station, you can have it. It’s not luxury, you must have it, if you’re a private space company, it’ll cost you a fortune to put Mm-hmm.

This amount of fleet. in space, that’s when you have like four people, three, three people on orbit. It’s fine. You could always tell them, okay guys, now it’s time to go to this room. Then imagine in 10 years we’ll have 5, 10, 6 stations. How many people will be. At one moment in the orbit, it’ll be not so easy to manage safe resources for them.

John Koetsier: there have been impacts, right? we’ve had significant challenges. Talk about some of the significant challenges in the past and what you might be worried about in the future.

Alex Pospeckov: the great event for space weather, happened in the middle of the 19th century.

It’s called Carrington Event, by the name of the scientist who found it. It was, the magnitude was so high. Then, on, on the, in Cuba, you could see solar flares and Northern Lights. a lot of the telegraph cables, they were burning as as this time. So if something big like this will happen tomorrow, then our civilization will be back to 19th century at all because now we are absolutely interdependent of the electronic communication of the satellites of the, of the old world, the modern

John Koetsier: mm-Hmm mm-Hmm.

Alex Pospeckov: solutions for it. Communications, but, events happening, once per 200 years, 300 years. But more often we see events affecting us daily. In 1999 in Quebec, one of the t matter was burned due to the geomagnetic storm and a couple million people lived without electricity for a week or so because it’s simply you can buy it out of the shelf tomorrow, switch off to the.

some backup supplies In sixties, there is like legend, like in seventies, one of the missile detection systems was activated ’cause of the geomagnetic storm. And we were at the brink of nuclear war, but then scientists or the people in charge figured out? No, no, no. It was just qua event.

Everything will be fine. I don’t know if it’s true or not. This is some kind of a legend. You could find a couple of articles about it, but you can understand the challenges we have.

John Koetsier: So the mechanism here is we’ve got solar flares and coronal, mass ejections on the sun, right? if that happens and it’s aimed at us, then it’s coming our way.

if you’re in space, you’re largely unprotected by the earth’s magnetic field The earth’s magnetic field typically protects us from those streams of charged particles. But if it’s massive. Then some of it’s getting through Right.

And impacting us. I mean, Cuba, wow. They saw the Northern Lights. It’s almost the Southern lights by the time this is there. Yeah, yeah, yeah. Or Aurora Astralis, right? Yes. Versus in Australia.

Alex Pospeckov: Different, exactly. in, it’s interesting. Then one of their poly mission at the beginning of the, seventies was delayed for half of the year.

they figured out it was like the huge space weather went with the power of seven zero per hour to give you and the audience example. So what’s the seven zero per hour is when Chernobyl power plant exploded. It was only two. these brave people. Would stay there forever. And when we speak about this moons moon settlements, when, when we speak about SI economy, nobody speaks about radiation.

But if you remember, the old Mankind TV series, the second, second season, the first series start with the big solar flare when the astronauts were on the moon. But it’s not science fiction. This is how things are really gonna happen. Maybe, it was a bit dramatic.

But still it’s reality that we need to face.

John Koetsier: And on the moon, we’re not far from having stations or bases I’ve had people on the show who have been building landers and vehicles for it. we know that, NASA and SpaceX are working on landing on the moon.

The good thing about being on the moon is that you can dig. And if there’s going to be something that’s going on, you can shield yourself potentially. It’s not easy, right? Getting construction equipment, digging mining equipment isn’t easy to get to get it to the moon, but you can get under 10, 15 meters of regolith and hopefully be protected,

Not so much in space. and obviously with a high energy impact on the earth, really challenging. Okay, so you provide space, whether. Projections, predictions. how do you do that?

Alex Pospeckov: So now you have a lot of limitations with the current services.

to give you one example, if you are nasa, it’s very hard for you to use data from the European Space Agency or the data from non-public sources in Europe and vice versa. If you’re a European agency, it’s very hard for you from the legal point of view to use even public data from US counterparties.

first of all, we gather a lot of different public and non-public sources. Then we normalize this data in our data lake because you have huge data gaps sometimes. You could see the data was not uploaded for the last, week or even two weeks. Then you can feed your models with data gaps of this size.

Then, we have our own constellation. We’re launching it early next year, which will generate our own sources of data on the low earth orbit. we’ll have 24 satellites that we’ll be able in real time predict and provide us. With huge amount of data. So we can do from just one of one set of the sensors, we can do thousand measurements per second.

So you can imagine amount of data that the whole constellation will generate for us. And the most important thing here that was developing in-House are on scientific machine learning models. ’cause all of the modern, all, all of the models for data, they were developed at the late nineties, at the beginning of 2000.

Of course there are some exemptions. Generally they were developed 20 years ago. So it’s not you simply can’t put machine learning Mm-hmm. To this thing. Mm-Hmm. Because you don’t have data lake. You, you don’t have unified data, you don’t have unified data So it’s still a bit tricky, complicated question.

the whole system. with thousand times more data than we can have now with our own scientific models, allow us to deliver customers tailored, made mitigation strategies in in the size. Because by the end of the day, you as customer, you just don’t want to know anything about what’s going on in the sun.

Just want somebody to tell you, okay, what are you, what should I do and what will happen by the end of the day? Mm-Hmm.

John Koetsier: have you already raised the money that you need to ship your satellites into orbit?

Alex Pospeckov: Yes. we had our pre seed round, now we’re finalizing our seed round.

So yeah, we have the funding for the cantellation for the development, but seed round is still process. Okay.

John Koetsier: Okay. So this is amazing, right? Because, rewind 10 years, there’s zero chance that a startup like yours can get a seed, pre-seed, maybe even a Series A and boom, we’re launching 24 satellites, right?

impossible, you’re talking a hundred thousand dollars a kilo or something like that. Right now, SpaceX, maybe we’re below 10 KA kilo. I’m not totally sure where we are, but it’s within the realm of possibility. It’s orders of at least one. Order of magnitude, maybe a couple orders of magnitude cheaper than it was.

Alex Pospeckov: Yes. And it’s also very important to understand that the ization now allows us to put our sensors to the size of half of the unit. because my CTO worked for NASA space shuttle missions and the middle of nineties, in his days. The size of space, weather sensors, I dunno what I just, this. So it was like the huge thing.

And it’s also, it’s, it’s also possible that now we have, cheaper access to the orbit. And the option to miniaturize our payload as much as we can.

John Koetsier: So let’s talk about the mechanics of how it works a little bit, because some people might be saying, how do you know if something comes off the sun?

you know it when it’s here, but how can you predict, right? there’s a difference between the speed of light and the speed of these charged particles that are heavy radiation, right? I. Think it’s a couple of days notice in some cases.

Is that correct?

Alex Pospeckov: Well, well now we’re going to the very, very interesting question because what’s what the prediction is? So what’s the baseline? Yeah. What the accuracy of the u usual weather prediction. I think most of our audience will say no. There is like no accuracy at all. I’m hundred sure. Yes. The weatherman always wrong.

Yeah, I’m hundred percent sure about it. So something like this happening with the space weather or not. So what’s the, what current system can predict Very good. If something very big is happening now. The further you go, the lower the power of the event. The least accuracy you can get.

But what we do is not about direct measurements. It’s about complex models measuring a lot of different parameters, 30 parameters. using the data and correlations between these, parameters, we can give the forecast. for example, now you don’t have any localization options.

Mm-Hmm. So if you go to the NOAA SW PC website for the space weather forecasting. You will see a lot of the esis, gk. Your next question is, okay, is it like 10 in Australia and zero in the United States? Or maybe it’s 10 in Europe. The answer is it’s just 10.

John Koetsier: Hmm.

Alex Pospeckov: So this is the level of localization we have now, and it’s a huge problem because in, may one of the, power plants on the northern island of New Zealand, they switched off the power grids because it was, it, it was alarm about place for the events.

Gonna happen, but what they figure out, the solar storm went through. But they turned off the power plant and, spent four days without full capacity. So it’s also the problem about localization and a lot of false alarms now. You can hear from the current systems.

Let’s say if you calculate properly in the last six years, the accuracy of the midterm forecast was. 7.5%. Ouch. it’s public information. Of course, nobody put it in this way, but, if you just calculate how many times. The official forecast missed and how many times they’ve done the false alarm, they were able to think if everything is all right, if something is going to be very bad.

John Koetsier: Okay. So, I assume you’re, then also let’s talk about some of these parameters. ’cause that’s super interesting to me. I assume you’re training some telescopes or you will be training some telescopes on the sun. You can see an event, you can see a solar flare, you can see a coronal mass ejection.

where is it on the sun? is it gonna miss us completely? are we gonna get the tail end of it? are we gonna get a full hit Are those some of the parameters you’re looking at?

Alex Pospeckov: Correct. So we, 24 satellites, we’ll have 12, let’s say on, why we need 24 satellites.

Like the first question. ’cause we need to have one satellite every hour at the north as a South Pole. To measure what’s going on at these points. Plus, when we have satellites on dark and bright side of the earth, we can see the difference. this level of localization allows me to have real time map of what’s going on, and we’re analyzing the differences between all of these parameters in near real time.

So it’s not about just, okay, I see the sun, the flare is coming. No, it’s, it’s, it’s not simple. It’s not work like this. Sometimes you have a lot of alerts. Sometimes the, sometimes it’s missing. But, there are a lot of things when, result of the impact of the space weather. Was during the low solar cycle.

Was when the solar flare was at the very low of its magnet.

John Koetsier: Mm-hmm.

Alex Pospeckov: So you need to monitor many parameters about it. And it’s more about analytics and the modeling, and less about measurements itself.

John Koetsier: Interesting, interesting.

So super cool. You can tell some level of danger and then when it’s here, you can give specifics about where it’s hitting and how absolutely a chance it is. And so that power plant now could, be prepared to shut down, but not necessarily shut down. Or if they shut down at all, maybe shut down for an hour or a couple hours and then they hear from you, Hey, no, it’s fine.

It’s all good. No worries. You’re not in the danger zone. They can get back up and. Hundreds of thousands of homes now have power again.

Alex Pospeckov: Exactly. That’s what’s happening, especially with satellite operators, especially when you have critical emissions. When you can’t switch off your satellite very often because you really need the capacity of the satellite.

Now you’re dealing with the result of this space weather event.

John Koetsier: mm-Hmm.

Alex Pospeckov: Something happen. Then you need to react.

John Koetsier: Mm-Hmm.

Alex Pospeckov: The more precise forecasting you have, the more time in advance you have for the planning. You either can switch off. Use the backup systems switch off different, radio channels or, or ground stations.

then you have some level of control of the events that’s gonna happen.

John Koetsier: It’s interesting from what you said earlier, it sounds like you’ll have at least some of your constellation in a polar orbit as opposed to all in an equatorial.

Alex Pospeckov: It’s like fuller or fuller orbits, all of them or just some.

we have two different layers. So let’s say one of them will be like 450, another one or a hundred, or just a bit more because we’ll have additional layer so we can see the difference between the, the difference of data between these layers. I must admit, my engineers, like every week they come in with new ideas that we have new idea for the new.

For the new pilot, we have new idea for the orbits. We have, oh, this is a playground

John Koetsier: for these guys. Right? Yeah, absolutely. Absolutely. This is a dream. Yes, exactly. Totally understand. and I guess you’ve hardened your satellites or you are hardening your satellites, so that even if there’s a serious event, they can continue to function.

Alex Pospeckov: We have different approach, so we try to build them as cheap as we can. Absolutely. our radio that we can launch them if something happened to them, we can launch not tomorrow because it’s, it is not so easy. But let’s say for us, we can produce it, you know, in one, one month because latest it’s very interesting.

Latest European space agency gave Airbus and the consortium of uk, maybe some US companies, 500 million. For one satellite for weather that will be launched in 2032 and 2033. Wow. My question was, guys. What if something will happen with this one asset?

John Koetsier: So

Alex Pospeckov: I mean, you already spent 10 years eggs

John Koetsier: baskets

Alex Pospeckov: Yeah.

John Koetsier: All of them in one.

Alex Pospeckov: Yeah. I can imagine such level, let’s say development when, when we speak about James web telescope. Mm-Hmm. This like the pinnacle of the engineering. Engineering, everything that was developed for 15 years. Yeah. Fully understand why we need to spend 15 billion for such thing then.

Now you have a lot of different systems. So here we use just totally different approach. Let’s do it as cheap as we can and send it as much as we can.

John Koetsier: Super interesting. It sounds like the European Space Agency and maybe NASA should be partnering with you and, shoving you, you know, a couple rounding errors of cash, as a backup plan, if nothing else.

So that, there’s good space, weather forecasting.

Alex Pospeckov: In US we have, US ACT that was adopted four years ago, which is stimulate government by space weather data from the commercial companies. It was like separate act that we have now here.

John Koetsier: Awesome. So, let’s talk a little bit about your customers.

we’ve already talked about power plants, satellite communications companies. any aerospace company, frankly, maybe even airlines. if you’re traveling at, 10 kilometers, up in the atmosphere that might matter to you who else would be a customer for you?

Alex Pospeckov: First of all, it’s aerospace companies, satellite operators and space tour. For example, private space station companies, because they’re fully aware about this problem. But I must admit, not all of them are aware about this problem. I spoke three weeks ago with one of the guys from one of the private space station companies.

They said, oh, space radiation on the space station. It’s the same as the airlines and the airplanes.

John Koetsier: If you are thinking about putting something into do, you should know something about this. This is a potentially kill all your customers event.

Alex Pospeckov: Yeah, absolutely. sometimes you see just the different level of understanding

About this problem. Airlines, of course, last in October and November, Pacific boarded three planes. Per week because of the geomagnetic storm. Because when you have gas space, water event, you don’t have any radio signal.

John Koetsier: Mm-Hmm.

Alex Pospeckov: there are a lot of talks from the pilots that told me, on the long haul flights during this space water event, we have a lot of errors in the avionics, but this information is still different.

John Koetsier: that’s freaky.

Alex Pospeckov: Yeah.

John Koetsier: Yeah,

Alex Pospeckov: So airlines fully aware about this problem now. one of my, our, our scientific team attended space weather, week scientific event this year it was in Portugal a lot of airlines also attended this event trying to figure out what can be done together.

Three years ago, four years ago, spice Weather Aviation guys, are you crazy? It was like this now situation is really getting, getting better. Well, especially

John Koetsier: since almost every modern aircraft flies by wire, you do not have physical control of your rudder, your trim or anything.

if your avionics get fried, if other electronics start to get fried, you’re in a brick that is flying through the air at, high speed and high altitude

Alex Pospeckov: with airlines. It’s, it’s also very interesting, space Weather Action Group, which was under coordination of, Scientific office of the White House.

They, two months ago, they published research about different needs of space, weather users. And they said that in aviation, women pilot one of the group, one of the most, let’s say, important group. The most affected by the radiation in space during the first three, during the first weeks of the pregnancy.

Oh, wow. Nobody raising this question. Yeah. And for them it’s also like nobody answering it. So it’s interesting to see how I. Let’s say new industries and new ways of thinking of the space weather. Now, because we, we had a lot of research in the late eighties during the Concord era. We have a lot of data about the radiation.

airlines even gave, pilots and, steward this, them gigger counters. But what’s happened after, a month, they said, well, we don’t know what to do. is it risky? So people were just too, too much afraid about it. The project was closed. Now you have like another, let’s say another return of this idea of what we should do without it.

in May, farmers in the US lost more than 1 billion and losses because all of the modern John direct equipment and another, agriculture equipment equipped with GPS modules, ah, and they’re fully, and they’re now working. They’re fully. Autopilot mode. Yeah. And then during the storm they had different coordinates and they worked for a week in the different coordinates.

John Koetsier: Oh, so

Alex Pospeckov: It’s anima. It was unimaginable even like five years ago. I’m not speaking about 10 years ago. Imagine. What else we’ll have like in the upcoming 15 years.

John Koetsier: yeah, exactly. As we get more technological, we’re more and more vulnerable. Hopefully we can start to harden some of the key infrastructure, but you know how humans are.

We don’t, often, prepare in advance for things. It’s too expensive, it’s too costly. Will it ever happen? Who knows? excellent. Alex, this has been a ton of fun, and super interesting. what’s your timeline? when are you gonna launch?

Alex Pospeckov: So we have our first lunch earlier next year.

John Koetsier: Mm-Hmm,

Alex Pospeckov: we’ll have to launch, in fall, and maybe another one by the end of the year. So to launch 24 satellites. Until 2028. I think 2026 will be the year where we’ll have the updated period. ’cause every time you’re sending something into space and engineers return with more cool ideas.

And we’ll have the full constellation in let’s say four years.

John Koetsier: very cool. super interesting to hear about and, hope it all works out

Alex Pospeckov: It was my pleasure.