In October of 2017 the Myanmar military began an ethnic cleansing of the persecuted Rohingya population in the north of the country. Villages were burned and an estimated 6,300 people were killed at the hands of the Myanmar army. To escape the violence, 1.1 million Rohingya fled across the Northern border of the country into Bangladesh, which is already one of the most population dense countries in the world. In an effort to house hundreds of thousands of refugees, the government of Bangladesh opened up a massive refugee camp known as Kutapalong in a previously forested section of the country.
To stay warm, cook food, and earn income, many of the 800K+ refugees at the camp harvest lumber from the surrounding forests, causing an immense amount of deforestation. However on-the-ground aid organizations have not had the time or resources to really assess the magnitude of the deforestation caused by the camp, as seen in Figure 1.
In order to better quantify the environmental degradation that is ongoing at Kutapalong and the potential impacts it could have on human health, our project leverages Planet’s high-resolution satellite imagery to A) objectively map the extent and rate of deforestation in the camps, and B) evaluate the impacts of this land cover change on water quality in the Naf River, which may be affected by increased sediment runoff in the absence of vegetation on the hilly terrain. Deforestation, which is clearly visible in the Planet imagery, can have drastic effects on local ecosystem services including reduced natural hazard protection, loss of food and extractive resources, and reduced climate regulation. Sedimentation in the adjacent Naf River, also detectable by multispectral satellite data, can also be correlated with water-borne ailments and waterway pollution.
We stated by calculating the water drainage basin that the main set of refugee camp is in, and used as this Area of interest to determine deforestation rates over the past year. We then calculated an NDVI for the region to see the change in forestry coverage during the time frame. We expanded the NDVI by using a machine learning algorithm to train an AI to detect the change in forest cover, a tool that could possibly be used to more closely track not just fires in forest area but also deforestation due to logging.
Planet’s very high spatial resolution also allowed us to examine relative suspended sediment loads at different times of the year even in narrow parts of the river. Preliminary analysis shows that the sediment loads since November 2017 have far exceeded the natural cycle of flood-driven sedimentation, and have remained high even through the dry season in early 2018 (see Fig. 2).
This weekend serves as the starting point for our group to work more closely alongside the UN and the IRI Institute at Columbia to assess the environmental consequences of human migration and come up with novel tools for assessing the environmental degradation caused by a large influx of migrants into an area. We see this project not so much as a “solution,” but more as tools for decision-makers on the ground to effectively assess priorities for mitigation strategies and preventative care.