Htet Thiha Zaw uses night time satellite images to confirm likely resettlement in Rohingya areas.
Despite a steady stream of accounts of murder, rape and forced evacuations of people, a full understanding of events in northern Rakhine is still limited by a critical problem: data availability. Buthidaung and Maungdaw townships, where most of the atrocities occurred, are closed off to international observers and few journalists are able to report on the updates (those who attempt to pay a high price for simply just doing their job), with neither the government nor the army showing any signal of change from their current stance. Given this, how do we extract frequent and credible information of events on the ground?
This is not to say that the accounts of violence from survivors are not credible; in fact, many of them likely are. But imagine the constraints: retaliation for speaking out, difficulties in providing concrete evidence, and general mistrust of the Rohingya people by the Burmese public. It is highly likely that most of the victims’ experiences remain untold.
Here I try to provide evidence for something that has been happening over the past year in Buthidaung and Maungdaw: replacing former Rohingya places with Rakhine and Bangladeshi Buddhists (articles here and here). I use monthly luminosity data from Visible Infrared Imaging Radiometer Suite (VIIRS). Using images taken at night, the satellite detects electric lighting on the earth’s surface. Given that electric lighting comes from human activity (such as lighting from buildings, highways, fishing boats), the images are a strong indication of human presence on the surface. When using these images to detect human activity, it is important to remove lighting from temporary non-human sources such as stray light, lightning, lunar illumination, and cloud cover, as they can contaminate the actual lighting coming from human activity. This is done by filtering out lights that are not observed for a long time and using satellite images of clouds. The images used here are controlled for stray light, lightning, lunar illumination, and cloud cover. Unfortunately, VIIRS does not control for fires (which have been rampant throughout the conflict) but averaging monthly readings offers some mitigation. These provide a stronger indication that there is a settlement, since the area has to be consistently bright throughout the month, meaning it is less likely to be a short-term source of light, such as fires. In this way, we can infer the changes in major settlements over the course of the year.
This animation shows what I found from the maps. You can see the settlements as bright spots here. The big bright spot near the southern border that appears throughout the year is Sittwe.
August 2017 shows an aftermath of the army’s “clearance operations” from October 2016 – July 2017, triggered after violence between Burmese security forces and Arakan Rohingya Salvation Army (ARSA). Violence over that year saw at least 1,000 Rohingya people killed and 87,000 Rohingya people fleeing to Bangladesh. If the satellite images can detect the shocks to human settlement, the night lights will be significantly reduced in the images. Compared to the latter images, you can see in August 2017 that there is no light whatsoever coming from three major settlement areas, Buthidaung, Maungdaw, and Rathedaung. We can see that the area overall becomes brighter between August 2017 and October 2017, which could be a result of fires that occur during the army operations. This, in fact, coincides with the exodus that occurred after clashes between armed Rohingya and the police in late August 2017.
What is more surprising is that the bright spots we observe starting from October 2017 remain throughout the rest of the timeline covered here. Again, these bright spots remain despite the forceful removal and fleeing of local villagers, in the hundreds of thousands. Many of them remain in refugee camps in Bangladesh, under poor living conditions. We can also see the appearance of a bright spot to the west bank of the river in April 2018. Like previously mentioned, given the monthly averaging, these are highly likely to be settlements.
So, what does this series of maps tell us? It lends support to the reported stories that former Rohingya areas have been repopulated with Buddhist populations, and it has been going on for at least a year. The implications are certainly worrying, given the repatriation agreement between Bangladesh and Myanmar. Will the displaced people eventually get to return to the places they come from? Will they be provided with any credible guarantee of security and protection from violence? How will these newly settled people possibly react to repatriation? In other words, the night lights do not lie; the former homes of Rohingya people have already been taken away.
This is not unique to Myanmar; from the Central African Republic to Sri Lanka, post-conflict property conflicts are always a problem when refugees attempt to return to their normal lives. However, this simple mapping exercise is a reminder that any arrangement for repatriation should consider the potential property conflicts that can arise between returnees and new residents. Potential solutions should be developed, in addition to guaranteeing security, provision of basic human rights, and a path to citizenship. This has been inadequately addressed by the government up to now. We need to ensure that returnees can eventually decide their own livelihoods. We need to ensure that property conflicts do not trigger another long episode of deadly violence.
Htet Thiha Zaw is a PhD student in Political Science at the University of Michigan – Ann Arbor. His research interests lie in quantitative analysis of state building, institutional choice, and distributive politics, with a regional focus in Southeast Asia. He can be contacted on Facebook (Htet Thiha Zaw) or on Twitter (@HtetThihaZaw). His personal website is https://sites.lsa.umich.edu/htzaw.