https://edition.cnn.com/interactive/2026/07/weather/solar-storms-power-grid/ I don't think the fact that Carrington happened 150 years ago necessarily makes such an event a "150 or more" year storm; we could just have been lucky. Is it possible for you to give me an idea of what might happen in Cambodia, given latitude and other factors?
Yes. And I think your objection to the phrase “a 150-year storm” is exactly right. 1859 being 167 years ago does not make another Carrington event “due.” If these events were independent with a constant probability, the elapsed time since the last one would tell us nothing about whether the next one is imminent. The real question is the estimated underlying frequency, and that is quite uncertain because we have a very short instrumental record.
In fact, there is good evidence that Carrington-class events can occur much closer together than 150 years. The February 1872 storm, only 13 years later, appears to have been comparable in intensity to Carrington; its aurora was observed extraordinarily far toward the equator, including Shanghai. And a 2012 statistical study explicitly treats recurrence estimates as uncertain rather than as a clock ticking toward the next event.
Cambodia is an interesting case
Cambodia is actually much better positioned than North America or northern Europe with respect to the classic power-grid hazard, but not completely immune.
The important distinction is between geomagnetic latitude and ordinary latitude. Cambodia is only about 10–14° north geographically, and is also a very low geomagnetic-latitude region. NOAA's ordinary storm-impact thresholds become increasingly concerned with power systems as one moves toward 40–60° geomagnetic latitude; at Kp=9, the principal affected region is described as poleward of about 40°.
So if you transplanted the 1859 storm onto today's world, Cambodia would not be among the places I'd expect to suffer the worst transformer damage. The classic nightmare scenario is much more applicable to places such as Canada, the northern United States and Scandinavia.
But there is an interesting wrinkle: very low latitudes have the equatorial electrojet. Cambodia is close enough to the magnetic equator for this to matter. During the 2015 storm, enhanced geomagnetically induced currents were actually measured in the Southeast Asian equatorial region, and the researchers found evidence that the equatorial electrojet contributed to the unusually strong magnetic variations there. Research specifically on the equatorial electrojet warns that developing power grids in Southeast Asia can be susceptible to this mechanism.
So I would characterize Cambodia as low risk from the classic high-latitude transformer catastrophe, but not zero risk of electrical-grid disturbance.
What I would expect you actually to experience
Suppose, purely hypothetically, that tomorrow we got something genuinely Carrington-class.
1. The first thing you might notice is probably nothing.
You don't suddenly get an electromagnetic pulse frying everything. The geomagnetic disturbance develops over hours, and its consequences depend enormously on the rate at which Earth's magnetic field changes, local geology, and the particular electrical network. NOAA emphasizes that GIC effects depend on the actual current paths through transmission lines and the ground, not simply on the size of the storm.
2. The Cambodian grid could experience voltage disturbances and protective trips.
Cambodia's national grid is now a fairly substantial system, with 115-, 230- and 500-kV transmission infrastructure and dozens of substations. The 2024 system had roughly 3,900 km of high-voltage transmission lines and 70 substations supplying 25 provinces.
That's considerably more infrastructure than the Cambodia of a couple of decades ago—and consequently more opportunity for GICs—but it is also a relatively compact system compared with the enormous North American grids.
In your particular area around Battambang, there is a western transmission system with substations in Battambang, Pursat, Banteay Meanchey, Siem Reap, etc. So you are connected to a real high-voltage network rather than an isolated little village circuit.
My guess for an extreme storm would therefore be temporary outages, voltage problems, or protective equipment tripping rather than transformers exploding all over Cambodia.
3. Your mobile phone probably wouldn't suddenly become useless.
The radio-frequency environment and satellite navigation are another matter. Solar storms can disturb the ionosphere and cause GPS/GNSS errors or loss of accuracy; NOAA specifically identifies navigation degradation as an effect.
The cellular network itself is mostly terrestrial infrastructure, so the phone isn't inherently vulnerable merely because there is a solar storm. But if electricity or backhaul infrastructure went down, obviously the consequences would propagate into cellular service.
Cambodia's telecommunications system is quite heavily mobile-based—there were about 20.7 million mobile subscriptions in 2025, more than the country's population. So electricity failure would probably be a more immediate problem for you than the solar radiation itself.
4. Satellites would have a bad day.
This is one of the consequences I'd take seriously even in Cambodia. Severe storms heat and expand the upper atmosphere, increasing drag on low-Earth-orbit satellites, while radiation and charging can interfere with spacecraft electronics, attitude control and communications. NOAA lists all of those among severe-storm effects.
GPS could therefore become unreliable precisely when you were trying to use it.
5. The Internet is less frightening than the CNN article may make it sound.
This is an interesting one. Long submarine cables contain conductive elements and consequently can experience induced voltages, but research by Google scientists found that even a Carrington-scale event is unlikely to physically destroy modern submarine fiber-optic cables.
The weak points are more likely to be the powered equipment at the ends, terrestrial networks, electrical supply, satellites and routing infrastructure.
And then there is the fun part: the sky
You are much less likely to get a spectacular aurora than somebody in Canada.
But an extraordinarily large storm could produce aurora at astonishingly low magnetic latitudes. The 1872 event produced credible overhead aurora around 20° magnetic latitude and reports considerably farther toward the equator.
So I wouldn't promise you a Cambodian aurora during a Carrington event—but if something substantially beyond an ordinary G5 storm happened, looking north would be worthwhile. You might see something nobody in Cambodia normally gets to see.
There is also an amusing complication: because you're near the equatorial region, Cambodia doesn't simply sit in a magical “safe zone.” The equatorial electrojet produces unusually strong local magnetic effects, and studies have specifically identified Southeast Asia as a region where these effects can matter to power infrastructure.
My rough Cambodia risk ranking
If I were sitting in your house when a genuine Carrington-class CME arrived, I'd rank the things I'd worry about roughly like this:
Power outage → telecommunications/internet disruption → GPS/satellite problems → inconvenience from grid instability → actual destruction of your household electronics.
That last one is important. I would not expect your computer, television, refrigerator, etc. to be individually “fried” by the solar storm. The danger is that the electrical grid supplying them becomes unstable or goes down.
And there's a broader point behind your original observation: “Carrington hasn't happened for 167 years” is almost useless as a risk argument by itself. We have documented extreme storms in 1872, 1921, 1989 and others, and evidence from historical observations and indirect records suggests that the distribution has a very long tail. The scientifically honest statement is not “we're overdue”; it's “we don't know the recurrence interval very well, but we know events of extreme magnitude are physically possible.”
For Cambodia, I'd add: the probability of a catastrophic nationwide electrical collapse from one is probably considerably lower than the CNN scenario might leave you thinking—but a severe storm could certainly cause regional outages and communications/navigation problems. The country's low magnetic latitude is a substantial mitigating factor for the conventional GIC threat, while its proximity to the equatorial electrojet is a countervailing factor.
And given that you're in rural northwestern Cambodia rather than sitting beside one of the enormous North American transmission corridors, I'd be considerably more concerned about being without electricity for a day or two than about witnessing an electrical apocalypse.
l would think the authorities, knowing such a thing was developing, would proactively cut power.
Yes — to some extent, absolutely. And this is one reason the “the grid gets fried” scenario is somewhat misleading.
The operators don't have to wait for the geomagnetic storm to hit. A major solar eruption can be observed on the Sun, and once a CME is identified as Earth-directed, there is typically many hours, sometimes a day or more, of warning before its arrival. The really dangerous geomagnetic effects become apparent as the storm approaches and begins developing.
And power-grid operators already have procedures for this. For example, the North American reliability standards explicitly require operating plans and procedures for geomagnetic disturbances, including actions by transmission operators and generators.
But there's an important catch: simply switching the whole grid off isn't necessarily the best protection.
The GIC problem is caused by quasi-DC currents induced in long conductors. The danger is particularly associated with long, high-voltage transmission lines and grounded transformer neutrals. Operators can therefore do things such as:
- Reduce the load on vulnerable transformers.
- Reconfigure the transmission network, reducing particularly bad current paths.
- Take particularly vulnerable transmission lines or transformers out of service.
- Reduce generation/load in ways that give the system more margin.
- Monitor transformer behavior and GICs and progressively take additional protective action if necessary.
NERC's procedures specifically contemplate operators directing individual generators or transmission facilities to take mitigating actions rather than simply abandoning the whole system.
And there's a nice bit of physics working in their favor: the transformer damage isn't instantaneous. NERC's benchmark analysis notes that transformer hot-spot heating has thermal time constants of roughly 5–20 minutes, while the storm's dangerous magnetic variations can persist for hours. That gives operators an opportunity to react.
But your idea becomes particularly attractive for Cambodia
If Cambodia were facing a truly exceptional event and the authorities had good warning, I could imagine the rational strategy being something like:
“We're going to have an ugly night. Let's deliberately have a controlled blackout rather than allow the transmission system to destroy several very expensive transformers.”
That's a very different proposition from the CNN-style scenario in which civilization discovers afterward that its transformers have mysteriously died.
The tradeoff is that turning off power doesn't eliminate every GIC. A transformer can be exposed to geomagnetically induced currents while the grid is still partially energized, and some equipment may need to remain connected for stability or because shutting it down creates other problems. The network has to be managed as an integrated system.
Also, Cambodia has a much smaller and less redundant grid than North America. That cuts both ways: fewer enormous interconnected transmission corridors to act as gigantic antennas, but also fewer alternative routes if something does fail.
So I think your intuition is basically correct: if a Carrington-scale CME were spotted coming, the rational response would be to sacrifice availability in order to preserve infrastructure. You might experience a deliberate, possibly prolonged blackout—but that is vastly preferable to discovering afterward that the transformers needed to restore the grid have been damaged and replacements have to be manufactured and shipped.
And, amusingly, you personally might be rather well positioned for that sort of event: rural Cambodia, a generator/backup power arrangement if you have one, food and water on hand, and no particular dependence on commuter trains or elevators. Your biggest complaint might be that the Internet has disappeared and somebody has to figure out how to keep the refrigerator cold.
Yes — to some extent, absolutely. And this is one reason the “the grid gets fried” scenario is somewhat misleading.
The operators don't have to wait for the geomagnetic storm to hit. A major solar eruption can be observed on the Sun, and once a CME is identified as Earth-directed, there is typically many hours, sometimes a day or more, of warning before its arrival. The really dangerous geomagnetic effects become apparent as the storm approaches and begins developing.
And power-grid operators already have procedures for this. For example, the North American reliability standards explicitly require operating plans and procedures for geomagnetic disturbances, including actions by transmission operators and generators.
But there's an important catch: simply switching the whole grid off isn't necessarily the best protection.
The GIC problem is caused by quasi-DC currents induced in long conductors. The danger is particularly associated with long, high-voltage transmission lines and grounded transformer neutrals. Operators can therefore do things such as:
- Reduce the load on vulnerable transformers.
- Reconfigure the transmission network, reducing particularly bad current paths.
- Take particularly vulnerable transmission lines or transformers out of service.
- Reduce generation/load in ways that give the system more margin.
- Monitor transformer behavior and GICs and progressively take additional protective action if necessary.
NERC's procedures specifically contemplate operators directing individual generators or transmission facilities to take mitigating actions rather than simply abandoning the whole system.
And there's a nice bit of physics working in their favor: the transformer damage isn't instantaneous. NERC's benchmark analysis notes that transformer hot-spot heating has thermal time constants of roughly 5–20 minutes, while the storm's dangerous magnetic variations can persist for hours. That gives operators an opportunity to react.
But your idea becomes particularly attractive for Cambodia
If Cambodia were facing a truly exceptional event and the authorities had good warning, I could imagine the rational strategy being something like:
“We're going to have an ugly night. Let's deliberately have a controlled blackout rather than allow the transmission system to destroy several very expensive transformers.”
That's a very different proposition from the CNN-style scenario in which civilization discovers afterward that its transformers have mysteriously died.
The tradeoff is that turning off power doesn't eliminate every GIC. A transformer can be exposed to geomagnetically induced currents while the grid is still partially energized, and some equipment may need to remain connected for stability or because shutting it down creates other problems. The network has to be managed as an integrated system.
Also, Cambodia has a much smaller and less redundant grid than North America. That cuts both ways: fewer enormous interconnected transmission corridors to act as gigantic antennas, but also fewer alternative routes if something does fail.
So I think your intuition is basically correct: if a Carrington-scale CME were spotted coming, the rational response would be to sacrifice availability in order to preserve infrastructure. You might experience a deliberate, possibly prolonged blackout—but that is vastly preferable to discovering afterward that the transformers needed to restore the grid have been damaged and replacements have to be manufactured and shipped.
And, amusingly, you personally might be rather well positioned for that sort of event: rural Cambodia, a generator/backup power arrangement if you have one, food and water on hand, and no particular dependence on commuter trains or elevators. Your biggest complaint might be that the Internet has disappeared and somebody has to figure out how to keep the refrigerator cold.
The good news is that I would not expect a Carrington event to simply make the Internet disappear worldwide.
The actual international Internet is overwhelmingly carried by fiber-optic submarine cables—more than 99% of international data traffic. The fibers themselves aren't particularly susceptible to the geomagnetic disturbance. The cables do contain metallic conductors to supply power to their repeaters, however, so there is a mechanism by which a solar storm can induce voltage in them.
The reassuring part is that research using actual cable measurements found that even a Carrington-scale event should produce voltages substantially below the tolerance of modern submarine cable power systems. Google's analysis estimated roughly 800 V of additional voltage for a Carrington event versus tolerances around 6,000 V in the systems they examined. A 2026 peer-reviewed study is a little more cautious—it confirms that geomagnetic storms do induce voltages in the conductive portions of submarine cables—but that isn't the same thing as saying the cables will fail.
The vulnerable parts are more likely to be the things at either end of the cable.
If Cambodia's grid is deliberately shut down, your ISP's local equipment may lose commercial power. Your generator could keep your modem/router running, but that doesn't help if the ISP's equipment, regional fiber nodes, or the international gateway has lost power.
And Cambodia has another advantage here: terrestrial fiber can provide alternative routes. The Internet doesn't require one particular submarine cable to remain operational; traffic can be rerouted through other cables and neighboring countries. The problem is that Cambodia, like many less-connected countries, doesn't have the enormous redundancy of somewhere like the United States. The ITU specifically notes that dependence on a small number of cable systems is a vulnerability for less-developed and underserved regions.
Satellite Internet is the interesting backup
If you had something like Starlink, it would give you a genuinely different failure mode because it doesn't depend on the Cambodian terrestrial Internet connection.
But it isn't immune to a huge solar storm. Severe space weather can disrupt satellites and GPS/GNSS, and a sufficiently extreme event could interfere with satellite communications. The ITU's 2026 assessment explicitly includes satellite disruption in its solar-storm scenario.
So I wouldn't regard Starlink as a guaranteed escape hatch—but it would be a useful independent path if your ordinary Cambodian ISP went down.
There is also a rather mundane issue: Starlink itself needs electricity. Your generator handles that, obviously, but the dish and router would become another load during a period when you're trying to conserve diesel.
So, for your particular situation, I'd imagine the hierarchy during an extreme solar storm as:
Generator: probably fine.
Your house's Wi-Fi: fine as long as your networking equipment has power.
Local Cambodian fiber: probably fine physically, but dependent on powered network equipment.
International Internet: probably still functioning, but potentially degraded or interrupted by power failures and routing problems.
Submarine cables themselves: surprisingly likely to survive.
Satellites/GPS: potentially the most genuinely space-weather-sensitive part of the Internet ecosystem.
And there's one amusing possibility: you could have electricity but no Internet. You'd be sitting there in the air-conditioned house, diesel generator humming away, refrigerator operating, lights on—and Firefox just sitting there saying unable to connect. That would probably be considerably more annoying to you than the blackout itself. 😄
If I were preparing your particular house for a very-low-probability event like this, I wouldn't spend money on elaborate solar-storm protection. Your generator and two weeks of fuel already address the largest practical vulnerability. A modest UPS for the modem/router/network equipment would actually be more useful, because it would bridge the frequent short interruptions that occur during ordinary grid problems too.
Yes — you're right. As of now, Starlink is not authorized to operate in Cambodia. Cambodia's telecom regulator said in August 2025 that SpaceX had not yet been granted a license and that Starlink equipment had not received type approval.
There apparently has been some discussion between SpaceX and the Cambodian government, but I don't see evidence that authorization has happened since then. Interestingly, Vietnam only began accepting Starlink orders in August 2026, so Cambodia is not simply being left behind while all its neighbors have it.
And the Cambodian regulator has been quite explicit about the unofficial equipment being sold: it considers importing, selling, and using unauthorized Starlink equipment illegal.
So for our hypothetical solar storm, Starlink shouldn't really be part of your contingency plan at all. Your actual backup hierarchy is much simpler:
Cambodian grid → your generator → local ISP/international fiber if it survives.
If the last link fails, you're offline, and—as you said—you have plenty of material on the computer that doesn't require the Internet. That's actually a pretty robust arrangement for something as extraordinarily unlikely as a Carrington-class event.
Oh, yes. I took your point too narrowly. You mean the Internet goes away there for plenty of ordinary reasons already—not merely that a solar storm might knock it out.
And Cambodia's own regulator confirms that this is a recurring practical problem: in 2025 it specifically complained about ISP failures to restore service promptly after fiber-optic cable cuts. There are also multiple international cable routes serving Cambodia, including AAE-1, MCT, and the Sihanoukville–Hong Kong cable, so an outage doesn't necessarily mean the entire country's Internet is gone.
So your experience probably gives you a much better mental model than the dramatic CNN scenario:
Internet disappears → shrug → wait for it to come back.
You've already got experience with the Internet being an occasionally unreliable utility rather than something that must be available continuously. A solar storm would merely be an unusually exotic reason for an outage.
And there's a nice distinction here: your generator protects you against the failure of electricity at your end; it can't protect you against a severed fiber cable somewhere between Cambodia and the rest of the world. Cambodia's MCT cable, for example, provides a direct international route toward Singapore, and it has had detected incidents of its own.
So in your hypothetical Carrington event, I wouldn't even regard “no Internet” as a particularly extraordinary consequence for you. The extraordinary part would be why it stopped working.
Yes, and that's an important distinction. Cambodia's Internet infrastructure can be quite functional even though the last mile is decidedly hands-on.
What you're describing sounds like a classic rural-network situation: the backbone is working, but something at the local distribution level has gone wrong, and rather than having a remote automated system diagnose it, some young technician eventually arrives on a motorbike, fiddles with a box or a fiber connection, and—presto—the Internet exists again.
In some ways that's a surprisingly resilient arrangement. The failure may be technologically unsophisticated, but the repair mechanism is fast and cheap. You don't need a highly redundant automated network if a kid can get from the ISP's local office to your house and fix the problem in half an hour.
And it reinforces what you were saying about a solar storm. The Cambodian Internet doesn't need to be perfect to be useful. If it goes down, you've apparently already incorporated "wait for the kid" into the operating assumptions. A Carrington event would just make the kid's eventual visit considerably more complicated if the problem were 500 km away on an international connection. 😄
There's also a broader reason Cambodia may fare reasonably well in a solar storm: its comparatively low latitude means it isn't sitting under the regions where the largest geomagnetically induced currents normally occur. So even though Cambodia's infrastructure has ordinary vulnerabilities, those vulnerabilities aren't necessarily the same ones that make a Carrington event frightening in northern countries.
No comments:
Post a Comment