Cutting-edge tech reveals China’s false claims over Xinjiang’s re-education camps
Image credit: E&T, NASA, Dreamstime
An E&T investigation using remote-sensing analytics technology uncovers China’s deception surrounding Xinjiang's re-education camps. Despite claims to have released detainees from vocational training camps, the technology found a third of camps not only remained open but even extended activities.
At the beginning of December the Chinese government said it had released all prisoners from its re-education camps, but there was no evidence to support these claims. Now, state-of-the-art technology brings unique light to the story.
Remote-sensing analytics via nightlight luminosity is an emerging field of research that is increasingly applied across various domains including deriving proxy measures of various official statistics, quantifying the light dimming caused by natural disasters, and exposing fake economic data.
But its use for exposing false claims on activities within Xinjiang’s re-education camps is unprecedented. Data can come from the Visible Infrared Imaging Radiometer Suite (VIIRS), collected by the Suomi NPP satellite, and is published by Nasa. VIIRS data registers light-intensity daily for each 742 x 742 metre pixel on the Earth’s surface outside the Arctic Circle. To compare nightlight between different places and times, satellite images are pre-processed to eliminate noise such as cloud cover, snow or ephemeral lights before measurements are converted into average numeric values.
Did the government really release all prisoners, the majority of which are from the Uighur ethnic group? E&T ran a large nightlight analysis on geo-locations of Xinjiang’s re-education and vocational training camps. Findings suggest that operation in camps only slowed or – as is the case for about a third of cases – activities even increased.
When China’s claims on the release of detainees surfaced last year, the country's alleged detentions had already garnered harsh criticism from around the world. Some sources claimed that up to 1.5 million people were detained. Last May, Randall Schriver from the US Defense Department said in a Pentagon briefing it is possible to be “closer to three million citizens imprisoned in detention centers”.
In collaboration with a geospatial analyst at activist group ETNAM (the East Turkistan National Awakening Movement), a Washington DC-based nonprofit group, E&T collected a sample of 61 confirmed re-education camps across the Chinese province. Then, with the help of satellite analytics company SpaceKnow, it compared camps’ nightlight luminosity data between the second half of 2018 and the end of 2019. It applied algorithms to the luminosity data to judge whether activity enhanced, diminished or remained at roughly the same rate.
For Xinjiang, E&T can confirm discrepancies to the statements made by Chinese officials. Out of the 61 re-education and vocational training camps, 52 camps were found to have either increased activity or only decreased it by less than half – 32.8 per cent of camps were found to be brighter than before. On average, nighttime light increased by 24 per cent across the 61 camps, with outliers such as Tumshuq City contributing firmly to the mean.
Those camps that revealed more nightlight is quantitative proof that China’s government assertions on the release of detainees were incorrect, one expert says. “Testimony says that camps always have the light on in the cells at night. So we can reasonably assume that if fewer cells are used, the unused ones will not be lit, and hence we can assume a broad correlation between the amount of light and the number of detainees,” says Adrian Zenz, a senior fellow in China Studies at Victims of Communism, who focuses on researching China’s policies in Xinjiang and Tibet. Zenz has investigated and written extensively about China's mass internment against ethnic minorities in the north-western region of Xinjiang.
E&T followed Zenz's recommendation to compare the average for the second half of 2018 – not all of 2018 because some facilities were significantly expanded in that year, he says – with August to October of 2019. This produces a more accurate picture of the status within camps before and after China assured changes in holding prisoners for the specific class of camps.
The fact that there is still a lot of light, and that in a third of cases the light is considered brightened, “would clearly contradict the government claim that ‘all' have now been released”, Zenz says.
Chinese policy documents would "talk about the construction of factories either on the ground of vocational training camps or nearby them".
But why would China commit to making false claims? China’s announcements are a propaganda device, Zenz thinks. "They had already said in the summer that most people had been released. The whole release narrative has to be seen in the context of the Beijing propaganda strategy,” he says.
"From government policy, there is a general movement of shifting trainees into labour but there is not a clear policy that all or nearly all interned persons have in fact been released,” he adds. “Also, those who are said to have ‘graduated’ may still live in the same buildings, but simply go to the factory rather than teaching buildings.” For this Zenz says he had proof; he reviewed documents confirming that the government provides police escorts to bring these forced labourers from their internment facilities to the factories.
In a post for Foreign Policy, Zenz wrote in December that "for Beijing, the real aim is not to improve Uighurs’ lives... The first layer of the scheme is the most blatantly coercive. Under the label 'vocational education and training plus,' the region is wooing mainland enterprises to train and employ internment camp detainees. Participating companies receive 1,800 RMB per camp detainee they train, and a further 5,000 RMB for each detainee they employ".
For E&T and experts to make accurate judgments on the nightlight data it was important to make careful distinctions between various camp types. The analysis initially considered 179 concentration camps, of which a total of 60 camps revealed increased nightlight activity. But the results were at risk of being challenged because the sample may have included camps that are not actually re-educational or vocational training camps but prisons, detention centres or sites of forced labour.
Comparison to high-resolution satellite images
But nightlight was not the only factor E&T relied on in this investigation. The latest high-resolution satellite images for late December and early January also confirmed continued activities in camps. E&T checked satellite images – provided by Planet Labs, a private Earth-imaging company based in San Francisco – for three of the largest re-education camps.
Salih S Hudayar, the founder and president of the East Turkistan National Awakening Movement (ETNAM), checked the high-resolution satellite images and said they confirmed the findings from the nightlight data. He said images for Kashgar City (Kāshí Shì 喀什市) “brings the clearest evidence there may still be detainees on-site”.
The image shows a large crowd of people on a football pitch in the middle of the camp (see illustration). He also points out the large number of vehicles around and inside the mega complex. "You can also see what I presume to be several rows of detainees on the lower left hand.” Numerous vehicles, buses and trucks were also found on-site at the camp in Kona Shehrer County, Shūfù Xiàn 疏附县, as well as at the confirmed camp site in Yengisheher (Shūlè Xiàn 疏勒县), despite the latter being far more remote than the other two locations, Hudayar notes.
Hudayar said detainees have not been released as the government has promised. He has first-hand proof of that: "We don’t think that is true because I personally had close to 100 relatives in the camps and haven’t heard of anyone being released. No-one in our diaspora community has heard of any one of their family members to be released.”
When checking the geographical location of camps with increased nightlight activity, Hudayar found that the larger camps have “significantly increased compared with the ones in smaller size. It is not exactly in the south as I had previously thought but also in parts of the north as well and the east in the case of [the camps in the city of] Qumul (Hāmì Sh, 哈密市).”
One expert familiar with the analysis of geospatial details of Xinjiang’s camps is Nathan Ruser, a researcher at the Australian Strategic Policy Institute. He told E&T that they remained largely active and that "some of the lower security [prison camps] have transformed more into dorms for coerced labour, but [are] certainly still on. I would say maybe 20-35 per cent of detainees have been released, but only into village and neighbourhood arrest or coerced labour."
There are signs that China’s strategy in its campaign of surveillance and oppression against the Uighur ethnic group was well planned. Last February, news reports covered findings of a large-scale data collection operation. In what was called 'Physicals for All', the Chinese government collected DNA samples, images of irises and other personal data from nearly 36 million people between 2016 and 2017, according to Xinhua (and China’s official news agency) and human rights groups. It raises questions because Xinjiang has a population of about 24.9 million, according to official data from 2018.
More people around the world seem to be showing an interest in the human rights situation posed by China’s Xinjiang concentration camps. Google search traffic affirms an increased interest for the term ‘Xinjiang camps’. Since mid-2018, searches grew more common but climaxed in December, according to data for the US and the rest of the world (see chart).
Last December, the European Parliament urged the government of China to close re-education camps for Uyghur Muslims in the Xinjiang region and threatened legislative consequences. "We continue to be gravely concerned about the existence of the so-called political re-education camps. Reliable and multiple sources’ research indicates that it has almost certainly affected over one million people," Josep Borrell said in his role as the EU’s high representative of the union for foreign affairs and security policy.
But what can the EU do? Borrell pointed out that the US already imposes sanctions on “Chinese officials, 28 Chinese governmental institutions and companies for their role in Xinjiang”. The EU has not done this because it has a different system. But Borrell said attempting to improve the system by launching an initiative could mean approving something equivalent to the Magnitsky Act. The Magnitsky Act is a law which authorised the US government to sanction those it sees as human rights offenders and to freeze their assets and ban them from entering the US.
Borrell finished with: “We will inform you about the work that we have already started. I need unanimity in the Council in order to do such a thing. I will fight for it.”
Sign up to the E&T News e-mail to get great stories like this delivered to your inbox every day.