Forest loss in Ethiopia continues to increase annually and contributes to the global increase in greenhouse gas emissions. Determining the direct drivers contributing to Ethiopia’s deforestation is critical to effectively enforce targeted conservation and management policies and protect tropical forest resources.
Tropical forests are vital for providing clean air, biodiversity, maintaining quality water flow, prevention against erosion, and mitigating climate change. The increasing global trend of forest loss risks the continual supply of services provided by these forests. The direct drivers contributing to tropical forest loss include human-induced land uses such as commercial agriculture, small-scale agriculture, Mining, Pasture, construction, and plantation forests. Determining the extent to which these drivers contribute to forest loss will enable governments, NGOs, national forest monitoring systems to concentrate REDD+ mitigation efforts towards specific proximate deforestation drivers where they will have the greatest impact.
Ethiopia, as one of the tropical countries, faces increasing forest loss, which places it among the emitters of greenhouse gases in the tropics. However, limited data availability makes it challenging to identify and account for the contribution of direct drivers influencing forest loss and eventually greenhouse gas emissions. Recent releases of high-resolution planet satellite imagery supported with the Hansen forest loss dataset provide an opportunity to manually identify the direct drivers of forest loss in Ethiopia. Together with deep learning methods, they can be used to automate the classification of deforestation drivers that would enable us to identify hotspots and local patterns of land-use change.
- Identify and become familiar with deforestation drivers in Ethiopia. Manually generate a reference dataset of different types of drivers of deforestation (commercial agriculture, small scale agriculture, Mining, Pasture, construction, and plantation forests).
- Develop and apply a segmentation method (e.g., U-Net or LinkNet) to predict the deforestation drivers in Ethiopia based on Planet data and Hansen forest loss dataset.
- De Sy, V., Herold, M., Achard, F., Avitabile, V., Baccini, A., Carter, S., Clevers, J.G.P.W., Lindquist, E., Pereira, M., Verchot, L., 2019. Tropical deforestation drivers and associated carbon emission factors derived from remote sensing data. Environmental Research Letters
- Descals, A., Wich, S., Meijaard, E., Gaveau, D.L.A., Peedell, S., Szantoi, Z., . High-resolution global map of smallholder and industrial closed-canopy oil palm plantations
- Hansen, M.C., Stehman, S.V., Potapov, P.V., 2010a. Quantification of global gross forest cover loss. Proceedings of the National Academy of Sciences 107, 8650–8655.
- Irvin, J., Sheng, H., Ramachandran, N., Johnson-Yu, S., Zhou, S., Story, K., Rustowicz, R., Elsworth, C., Austin, K., Ng, A.Y., 2020. ForestNet: Classifying Drivers of Deforestation in Indonesia using Deep Learning on Satellite Imagery, in: 34th Conference on Neural Information Processing Systems, Vancouver. p. 10.
- Background in Earth Observation or machine learning
- Python, java script programming
- Knowledge in land use
Theme(s): Modelling & visualisation; Integrated Land Monitoring