When Russian troops seized control of the Chernobyl nuclear power plant last year, following the invasion of Ukraine, President Volodymyr Zelensky called it “a declaration of war” against Europe. Others warned that Russia’s reckless seizure of the plant could trigger a nuclear disaster to rival Chernobyl’s 1986 radiological accident.
Their fears seemed well-founded when, on the night of the invasion, sensors began reporting sudden spikes in radiation levels in the Chernobyl Exclusion Zone (CEZ)—a 1,000-square-mile forested zone around the plant where radioactive soil from the 1986 disaster had settled.
Forty-two sensors recorded spikes that night and the next morning—some at levels hundreds of times higher than normal. The State Nuclear Regulatory Inspectorate of Ukraine (SNRIU) eased concerns that nuclear material had leaked from the plant, however, when it said the spikes were likely due to “resuspension” of radioactive soil stirred up by Russian military vehicles—an explanation widely accepted by many nuclear experts and the media.
But a group of environmental radiation experts disputes this conclusion. In a paper published in June by the Journal of Environmental Radioactivity, they detail why there’s no way soil resuspension could have caused the spikes and speculate that interference from an electronic warfare weapon was behind the surge instead.
Now, in what is becoming a deepening mystery, noted cybersecurity researcher Ruben Santamarta says he believes something else was the cause—data manipulation, possibly through a cyberattack.
Based on patterns he found in the spikes—batches of sensors geographically distant from one another recorded spikes at the exact same moment, while sensors closer to them recorded no elevation—he thinks a remote hacker or someone with direct access to the server processing the data manipulated the numbers.
After an extensive review of the data and other materials, Santamarta says he finds it hard to believe the explanation about soil resuspension was ever considered plausible. And he is surprised that authorities never bothered to examine the data for patterns or, if they did, kept that information from the public. He thinks those patterns discount theories about interference from electronic weapons, and he plans to present his findings at the BlackHat security conference in Las Vegas next week.
“I have collected a significant amount of evidence by different means, including OSINT [open source intelligence], hardware and software reverse engineering, and data analysis of the radiation levels,” he says “I think it is enough to seriously consider the possibility that these radiation spikes were fabricated.”
If Santamarta is right, his finding could have far-reaching implications for radiation-monitoring systems around the world, says a former nuclear safety official who asked to remain anonymous in order to speak freely about the matter. If the data was manipulated, it could undermine trust in radiation-monitoring systems or change how data from them gets reported publicly. Data from radiation monitors is often distributed publicly in near real time so that governments and nuclear experts can actively monitor conditions in populated cities and around nuclear facilities. But this creates a risk that hackers or others could alter data to trigger public alarm before proper verification can occur.
Russian troops entered the CEZ early on the morning of February 24 last year because it’s the shortest and most direct route from Russia-friendly Belarus to Kyiv, Ukraine’s capital 80 miles south of the plant. But some feared Russia’s interest in Chernobyl was more than strategic. They worried the military could cause a disaster using radioactive waste at the plant or drum up false claims that Ukraine was building a dirty bomb there.
After a day-long battle with Ukrainian troops and three hours of negotiations to establish parameters for Russia’s occupation of the plant, Russia took control of Chernobyl’s facilities. At 8:40 pm local time, 10 minutes after the SNRIU indicated that Russia had formally taken control of the plant, seven monitoring stations in the CEZ suddenly began reporting elevated radiation levels. The readings ranged from two to five times the normal radiation rate each sensor had historically detected, but one station showed a level eight times higher than normal.
Ukraine has two networks of sensors to monitor radiation at Chernobyl. A set of 10 sensors inside the plant is operated by the state-owned nuclear energy company Energoatom. A second network, known as a radiation-monitoring and early-warning system (the Ukrainian acronym for it is ASKRS), consists of about 68 battery-powered GammaTracer detectors spread throughout the CEZ (with a few positioned outside it). This network is managed by the State Specialized Enterprise Ecocenter (Ecocenter, for short), under the State Agency for the Management of the Exclusion Zone.
These detectors continuously record ambient gamma radiation levels in the CEZ, process the readings to calculate an average, then transmit that figure once an hour (or every two minutes in an emergency) to a base station in the Ecocenter’s office in the town of Chernobyl, about 10 miles from the plant. The data is transmitted via radio over a dedicated channel using a SkyLink protocol.
The data then gets analyzed and processed with DataExpert software and a custom Ecocenter program before being posted to the Ecocenter’s website. It’s also distributed to the SNRIU, the International Atomic Energy Agency (IAEA)—the UN body that monitors nuclear programs around the world—and other governments.
The data can be difficult to find on Ecocenter’s site, so a Ukrainian nonprofit called SaveEcoBot scrapes it and republishes the data on its own site for easier access. It’s these two sites that many people around the world were using to track radiological conditions at Chernobyl in real time on the day of the invasion, and that triggered alarm when people began posting screenshots of them on Twitter.
Radiation levels at Chernobyl are measured as “ambient dose equivalent” rates—essentially the amount of energy, due to ionizing radiation, that the human body would absorb if exposed to the level of radiation a sensor detects. Dose rates are reported as microSieverts per hour (aka μSv/h).
Following the first spikes at 8:40 pm on February 24, 2022, the next cluster of spike occurred at 9:50 pm, when 10 different sensors reported elevated radiation levels, as well as one that had been in the earlier cluster.
More cluster spikes occurred at 10:20 pm, 11:30 pm, and 11:50 pm, involving nine, six, and five sensors, respectively, and then the pattern switched. From 12:01 am to 12:20 am on February 25, there were several spikes involving just one or two sensors each time. Then at 9:20 am, 10 sensors spiked simultaneously, including one that increased nearly 600 times its normal radiation reading. At 10:40 am, nine sensors spiked. And at 10:50 am, the last spike occurred with a single sensor. This sensor spiked three times in all. Called Pozharne Depo, its baseline reading of 1.75 μSv/h spiked to 8.79 (at 8:40 pm), 9.46 (at 9:50 pm), and 32.2 (at 10:50 am the next morning). The sensors may have continued spiking, but the Ecocenter website stopped updating its data.
Like other European countries, Finland carefully tracks Ukraine’s radiation levels. According to Tero Karhunen, a senior inspector with STUK, Finland’s radiation and nuclear safety authority, if ambient dose rates rose above 100 μSv/h for more than 48 hours, it would generally trigger an evacuation of affected regions.
Two sensors nearly reached that threshold at 93 μSv/h, but then they and all the other sensors stopped reporting updates—or at least the Ecocenter stopped posting updated data to its website. It’s not clear why this stopped. The invasion caused internet disruptions in Ukraine, but this would not have prevented the sensors from transmitting their data to the base station; it would only have prevented the Ecocenter from publishing new data to its website.
Yet the Ecocenter did continue to publish new data for some sensors. Shortly after the sensors spiked, online updates of data from 30 of them stopped; but data for the remaining ones continued until they stopped updating at different times. Most of the sensor data was updating online again the following Monday, February 28—at which point all the sensors were reporting normal radiation levels. But by March 3, Ecocenter had stopped posting data altogether.
This may be because Russian troops began stealing and damaging equipment at Chernobyl—including the server and software Ecocenter used to receive and process sensor data. In a French TV news interview after Russian troops left Chernobyl at the end of March, Mykola Bespalyi, head of the Ecocenter’s central analytical lab, showed an empty cabinet that had housed the server, explaining that the center had lost its connection to the radiation sensors in the CEZ. Data transmissions only restarted in June when the IAEA and others helped Ukraine restore the radiation-monitoring system.
The spikes were initially attributed to shelling. In a news story published about an hour after the spikes began, an unnamed Ukrainian official said Russian shelling had hit a “radioactive waste repository” and implied that radiation levels had risen as a result. Anton Herashenko, advisor to Ukraine’s Interior Ministry, then warned that the attack could send radioactive dust into Belarus and the EU.
But the next morning, on February 25, the SNRIU said the cause of the spikes couldn’t be determined. Later it released a statement saying Ecocenter experts attributed the spikes to topsoil stirred up—or “resuspended”—by military equipment. At that point, media attention turned to ongoing battles elsewhere in Ukraine, and talk of the spikes dropped.
Not everyone bought the explanation, however. Karine Herviou, deputy director general in charge of nuclear safety at France’s Institute for Radiation Protection and Nuclear Safety, told reporters there was no coherent explanation for the spikes, though her group later issued a statement saying they didn’t have any information “to confirm or refute” the SNRIU’s assertion about soil resuspension.
Bruno Chareyron, a nuclear physics engineer and laboratory director for France’s nongovernmental Commission for Independent Research and Information on Radioactivity, also scoffed at the soil explanation. Instead, he told reporters at the time that the spikes might be the result of interference—from a cyberattack.
If you look at a map showing the places where the results had been increasing, “there was no logic” to the soil suspension explanation, he now says. And given that Russia had been prolifically hacking Ukrainian systems at the time of the invasion, it was reasonable to wonder whether a cyberattack had resulted in false data.
Hours after the SNRIU’s statement, the IAEA released its own. Apparently accepting that the radiation spikes were real, the agency said the elevated radiation levels posed no threat to the public. Oddly, though, it referred to the spikes as topping out at 9.46 μSv/h—a figure it received from Ukraine’s nuclear regulator. But the agency only had to look at the Ecocenter’s website to see that some sensors were reporting levels magnitudes higher than this at 58 and 65 μSv/h.
Only some of the GammaTracers are considered “regulatory” sensors—meaning the SNRIU is required to submit data from them to the IAEA. At least three of these regulatory sensors were among those reporting exceptionally high data spikes. But it appears that the SNRIU didn’t provide data from those sensors to the IAEA. It’s not clear why; the SNRIU didn’t respond to inquiries from WIRED.
Notably, there was chatter on Twitter at the time, among nuclear and radiation experts, that the spike data might be erroneous. But if the sensor readings being reported by Ecocenter on its website were accurate, then contrary to the IAEA’s statement that the spikes posed no threat to the public, “this would have been a very dangerous situation for the people in the area,” Chareyron says.
Why did the IAEA only reference the lower radiation spikes in its statement and not the higher ones? The IAEA, after asking WIRED to submit questions in writing, didn’t respond to this or any of the other detailed questions submitted to it, including whether it attempted to investigate the veracity and cause of the spikes.
In the US, the National Security Council followed events in Chernobyl closely but did not respond to repeated requests to discuss the mystery around the radiation spikes.
Soil Suspension, Debunked
As noted, Mike Wood, a professor of applied ecology at the University of Salford in the United Kingdom who studies environmental radiation, including in the CEZ, examined the spike data with four colleagues and ruled out soil resuspension as the cause.
Wood says there isn’t enough radiation in the CEZ soil to cause the level of spikes that occurred—not to mention that military vehicles traveled primarily on asphalt and hardened dirt roads, not in places where loose soil would have been stirred up.
“Even with conservative assumptions, you cannot get anything like the rises that we’ve seen in those dose rate spikes,” he says.
What’s more, experts told WIRED that if soil resuspension were the cause, the radiation levels should have dropped gradually over days as the soil and dust resettled. Instead, many of the sensors were back to reporting normal levels within 30 minutes to a couple of hours after reporting a spike, despite heavy military traffic continuing in the region.
There was also no uniformity or expected patterns to the spikes. Instead of sensors spiking at different times as radiation levels near them rose, multiple sensors in different locations spiked at exactly the same time. Some sensors reported spikes 12 to 38 times their baseline level, others 300 times above baseline. The sensor that spiked nearly 600 times its base level was 18 miles southeast of the plant, located along what Wood calls a “minor” road where “there is no logical explanation as to why there would be significant military activity.”
Olegh Bondarenko, director of the Ecocenter until 2011, agrees with Wood’s conclusions and calls the air suspension explanation “fantastical.” But he doesn’t think Wood’s alternate theory—that the spikes were caused by interference from electronic warfare weapons—was the cause either.
Electronic warfare weapons are used as jamming devices to hinder or block an enemy’s communications and signals.
But Karhunen says Finnish researchers conducted limited tests on the effects of electronic warfare weapons on radiation sensors and found that they could also cause sensors to report false readings up to 30 μSv/h—300 times the base levels for the test systems.
William Radasky, former chair of an International Electrotechnical Commission subcommittee on electromagnetic weapons and their effects, says interference can cause data spikes and, depending on a weapon’s strength, permanently damage sensors. If a weapon were close to a radiation detector when it was fired, “they would probably kill the electronics used with the sensor,” says Radasky, who has conducted research for the US government and military on effects of electromagnetic pulses on defense systems and the electric grid. The pulses wouldn’t interfere with the sensors’ ability to detect radiation, but they would affect the electronics that are used to translate the radiation signals from the sensor into saved data and then transmit that data to the Ecocenter. It’s worth noting that the IAEA visited Chernobyl after Russian soldiers left and reported that many radiation-monitoring stations were damaged and out of service. The agency never identified which sensors or explained the nature of the damage, however.
But if such weapons can produce spikes in the sensor data, that still doesn’t explain the anomalous nature of the spikes, Radasky and Bondarenko say. There were no reports of other equipment in the CEZ being affected by electromagnetic weapons. And radiation sensors in other parts of Ukraine where fighting occurred—and presumably where electromagnetic weapons would likely have been used—did not experience spikes.
Most significantly, sensors that showed spikes in their data were near sensors that didn’t record spikes, and the geographical distance between sensors that spiked defied logic. Many of the sensors that showed simultaneous spikes in the CEZ were more than 30 kilometers apart. Radasky says it would be possible to have a single weapon affect sensors geographically apart, but only at limited distances.
“The most powerful [electromagnetic weapon] I know about could create a high field over [only] a kilometer,” he says. “There is no way to simultaneously affect widely dispersed sensors … from a single weapon.”
If troops dispersed throughout the CEZ carried portable electromagnetic weapons and fired them at the same time, it would be possible to affect sensors far apart, Radasky says. “But … the likelihood that they would have set off those weapons at the exact same moment, causing simultaneous spikes, seems highly unlikely,” he says, noting that a pulse generally lasts for one microsecond.
This, plus patterns that Santamarta found in the spikes, “really does sound like … this is a hacking attempt and not an electromagnetic weapon attack,” Radasky says.
Bondarenko agrees. Every other explanation is “practically implausible or impossible.” It would have been easy, he says, to “write a script that causes certain sensors to elevate at a certain time and then to go back to normal at a certain time.”
Jan Vande Putte, a lead radiation specialist at Greenpeace Belgium, led a team that visited Ukraine last July to measure radiation levels in one part of the CEZ. He agrees that none of the other given explanations are plausible. But he cautions that Santamarta’s theory of data manipulation, while convincing, is still speculation without a forensic investigation to support it.
“I have seen so many examples of coming to wrong conclusions,” he says.
Santamarta began looking at the issue last year, after a non-Ukrainian nuclear engineer who has done research in the CEZ asked him to consider whether the sensors could have been hacked. Santamarta specializes in critical infrastructure vulnerabilities and in 2017 found unpatched flaws in radiation-monitoring systems that would let someone falsify data with the aim of simulating a dangerous radiation leak.
He studied the type of sensors used in Chernobyl—a model made by the France-based company Saphymo (later purchased by Bertin Technologies)—and obtained raw sensor data Ecocenter posted to its website, which contained time stamps identifying when each sensor spiked.
He identified 42 radiation sensors that reported spikes in four different patterns—patterns that he says support his assertion that the radiation spikes were fabricated. In the first pattern, 18 sensors reported spikes before going offline. In the second pattern, 17 sensors each reported spikes two times. The second spike was always incrementally higher than the first. For example, a monitoring station called Buryakovka showed a moderate spike on the night of February 24, from 2.19 to 3.54 μSv/h. But at 9:20 am the next morning, it shot up to 52.7 μSv/h.
The third pattern involved two sensors that each spiked three times. The fourth pattern involved five sensors that experienced two spikes—the first spike occurred at 8:40 pm and the second at 11:30 pm the night of the invasion. In each case, the second spike involved a lower value than the first spike—in other words, the second reading was still higher than the baseline level, but lower than the earlier spike. For example, a sensor station called Gornostaypol normally reported a baseline of .092 μSv/h, but during the first spike at 8:40 pm it shot up to .308 μSv/h and at 11:30 pm it reported a level of .120 μSv/h—about midway between the two.
Santamarta believes the patterns strongly suggest that someone wrote code to inject false data into the Ecocenter’s DataExpert database at intervals. The code then posted the false data to the Ecocenter’s website while suppressing legitimate data that came in from the sensors.
“After characterizing the spikes, which are clearly structured, it’s difficult not to assume some kind of human intervention behind them,” he says.
WIRED sent the State Agency for the Management of the Exclusion Zone a list of detailed questions about the patterns Santamarta found and asked whether it had conducted any investigation to determine the cause of the spikes. The agency declined to answer most of the questions and said it was unable to answer others because the events occurred when Russian forces controlled the CEZ and Ecocenter personnel weren’t in a position to know what occurred or to carry out “any corrective actions on the systems.”
Because Ecocenter staff weren’t present, “there is almost no information about the situation around the sensors and servers of Ecocenter in the described period,” Maksym Shevchuk, deputy head of the state agency, said in an emailed statement that his agency translated into English. He noted that any data transmitted by the sensors during that time was automatically received and “automatically transmitted in ‘as-is’ mode” to “various departments outside the exclusion zone,” suggesting that any data posted to the Ecocenter website and transmitted to the IAEA was automatically processed and sent without involvement from Ecocenter staff.
With regard to Santamarta’s findings, Shevchuk said his agency’s “competence does not include the assessment of hypotheses and assumptions on this topic coming from third parties,” and he therefore can’t comment on them.
Who and Why
Santamarta doesn’t speculate in his presentation about who was behind the manipulation if it occurred—he wanted to focus on finding a sound technical and plausible explanation of the cause. But there are two obvious suspects—Russia and Ukraine—both of which have means and motive.
Russia has repeatedly threatened a nuclear event to assert dominion in the conflict and, some argue, to warn NATO against getting involved. And Russian authorities have made numerous claims before and after the invasion that Ukraine was developing a radioactive dirty bomb. A Russian scientist told state media that Russian troops seized Chernobyl to prevent Ukraine from creating a dirty bomb, and the radiation spikes could have been used as “evidence” of illicit nuclear activity on the part of Ukraine.
What’s more, in an April 2022 report about the war, Microsoft revealed that Russian hackers affiliated with the FSB intelligence service had breached a Ukrainian “nuclear safety organization” in December 2021 and stolen data for three months with the aim of obtaining data that would support Russia’s disinformation claims about Ukraine, including claims that it was building a dirty bomb. The report didn’t identify the organization, but the breach shows that Russia had a particular interest in hacking nuclear organizations in Ukraine with the aim of supporting its disinformation campaign.
These suggest a potential method and motive for Russia. But there’s a hitch in this theory. After Ukrainian officials cited the spikes in their denunciation of Russia’s “reckless” seizure of the plant, Russian Ministry of Defense spokesperson Igor Konashenkov denied that any spikes had occurred. He didn’t say how he knew this, but Bondarenko believes Russian troops likely carried handheld radiation meters with them into the CEZ, where they may have gotten very different readings from those the Ukrainian sensors were reporting. A Russian scientist also told Russian media that once data from the CEZ sensors started being posted online again “it would be clear that the radiation situation at Chernobyl was under control.”
If Russia planned to use the spikes to bolster claims that Ukraine was building a dirty bomb at Chernobyl, why didn’t it seize the opportunity to further that claim, instead of disputing that the spikes were real.
As for Ukraine’s potential motives, on the day of the invasion and for days after, Ukraine was struggling to secure timely financial and military aid from Europe. Radiation spikes could have helped underscore the potential nuclear threat to EU leaders if they didn’t act quickly to help Ukraine expel Russian troops.
But there’s another possible motive as well. A Chernobyl worker may have inadvertently revealed it in an interview with the Economist after Russian troops left Chernobyl at the end of March.
He told the publication that during the occupation of the plant, Chernobyl staff had “exaggerated the threat of radiation” to Russian troops, identifying “problematic areas” that they should avoid—all as part of a “cheeky plan” to control where the soldiers went. He didn’t mention the radiation spikes, but they could conceivably have been part of this plan.
After Russian troops left Chernobyl, workers also told reporters that some of the troops had exhibited signs of radiation poisoning—they “developed huge blisters and were vomiting after ignoring warnings about digging trenches in radioactive soil.” Reporters have not been able to independently verify the claims. The IAEA conducted tests after the soldiers left and determined that the digging would not have posed a radiation threat to the soldiers, raising questions about whether the reports of radiation sickness were falsified to instill fear in Russian soldiers.
It may be noteworthy that, in June of this year, as tensions around the Zaporizhzhia plant heated up during Russia’s occupation of that facility, Moscow ordered its troops to halt the automatic transmission of data from the plant’s radiation sensors. The sensors continued to monitor radiation levels, but the data was manually collected from the sensors by IAEA staff.
There were a number of ways to verify the veracity of the spikes when they occurred last year, but there’s no sign anyone in Ukraine questioned the integrity of the data or ordered an investigation. Vande Putte says this was never discussed when his Greenpeace group traveled to Ukraine.
More than two dozen wireless sensors in the CEZ have aerosol filters attached to them that detect radiation levels in the air. Karhunen says the filters are “a hundred million times more sensitive” to small changes in radiation than the digital sensors. Before the invasion, Chernobyl workers collected the filters once a week to test them in a lab. These could have been tested to see if they detected the same radiation levels as the digital sensors. But it seems this didn’t happen during the Russian occupation of the plant when the activities of Chernobyl workers were strictly curtailed, and it’s not clear if the filters were collected and tested afterward. Vande Putte says the Russians left land mines behind in the CEZ, and this may have made it too dangerous for workers to collect the filters after they left.
Once the occupation ended, it would also have been possible to conduct a forensic investigation. Even though the Ecocenter data server, and any digital evidence of manipulation it contained, was no longer available because the Russians stole the server, investigators could have extracted data from the memory of the digital sensors in the CEZ—land mines permitting—to see if data stored in the devices matched what was posted to the Ecocenter website. If it didn’t, this could have bolstered the theory that the data was manipulated on the Ecocenter server. It’s not clear anyone did this, however. Ukraine’s computer emergency response team would likely have been involved in such an investigation, but a source with knowledge of the CERT’s activities says the organization never received a request to investigate the Ecocenter’s systems, and the Ecocenter didn’t respond to WIRED’s inquiries.
It’s possible Ukraine and the IAEA didn’t investigate the spikes because they simply had other more pressing things to address—for example, an ongoing crisis at the Zaporizhzhia nuclear plant. The State Scientific and Technical Center for Nuclear and Radiation Safety told WIRED that it did do a radiological survey after Russian troops left Chernobyl, but this was to determine if Russian forces had absconded with any nuclear materials or planted them in occupied regions to leave behind a nuclear hazard, according to Yuliya Balashevska, who headed the survey. The survey focused only on Kyiv and southeast border regions, and Balashevska said her organization has no access to the Chernobyl sensors and could not have examined them if it wanted to do so.
It may be the case that authorities simply never questioned the authenticity of the spikes and therefore saw no reason to investigate.
Santamarta, however, believes the IAEA and Ecocenter didn’t investigate because of the potential geopolitical implications if such an investigation reached “an inconvenient conclusion” that risked adding “more complexity to an already extremely complicated situation.”
Either way, he, like everyone else WIRED interviewed, says the truth about what occurred is more important than whatever the findings might reveal about who was involved.
Although it’s not clear whether, 18 months after the invasion, evidence still exists that could resolve the mystery of the radiation spikes. Greenpeace’s Vande Putte says an investigation is merited, and that Santamarta’s research is high-quality and “very valuable” as a starting point.
The implications, if the cause was intentional data manipulation, are global, given the potential precedent for manipulating sensors in other regions.
“The truth in these matters is really important. Where did it go wrong? Was it purely technical? Was it deliberate?” he says. “We need to get to the bottom.”