|
The proposed species distribution model has been developed by following the principles of supervised machine learning methods. The high-resolution map layers of distribution determinants were generated by the AW3D data (the world’s first 5m-resolution 3D map of the Earth) and drone-based mapping. Drone surveys conducted over mixed-species forests and improvement of a hyperspectral transforming method aid a detailed identification of vegetation species. Camera traps were placed to validate the model at selected sites in a deep forest for three months. Biodiversity conservation solutions were discussed through meetings with local people. It is concluded that the targeted critically endangered mammal most likely hide in hard-to-reach areas close to the border between Central Vietnam and Laos.
|
