Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
6348454 | International Journal of Applied Earth Observation and Geoinformation | 2016 | 14 Pages |
Abstract
Prosopis spp. is a fast and aggressive invader threatening many arid and semi-arid areas globally. The species is native to the American dry zones and was introduced in Somaliland for dune stabilization and fuel wood production in the 1970â¿¿s and 1980â¿¿s. Its deep rooting system is capable of tapping into the groundwater table thereby reducing its reliance on infrequent rainfalls and near-surface water. The competitive advantage of Prosopis is further fuelled by the hybridization of the many introduced subspecies that made the plant capable of adapting to the new environment and replacing endemic species. This study aimed to test the mapping accuracy achievable with Landsat 8 data acquired during the wet and the dry seasons within a Random Forest (RF) classifier, using both pixel- and object-based approaches. Maps are produced for the Hargeisa area (Somaliland), where reference data was collected during the dry season of 2015. Results were assessed through a 10-fold cross-validation procedure. In our study, the highest overall accuracy (74%) was achieved when applying a pixel-based classification using a combination of the wet and dry season Earth observation data. Object-based mapping were less reliable due to the limitations in spatial resolution of the Landsat data (15â¿¿30Â m) and problems in finding an appropriate segmentation scale.
Related Topics
Physical Sciences and Engineering
Earth and Planetary Sciences
Computers in Earth Sciences
Authors
Wai-Tim Ng, Michele Meroni, Markus Immitzer, Sebastian Böck, Ugo Leonardi, Felix Rembold, Hussein Gadain, Clement Atzberger,