Tested over a seven-year period at the Canada Glacier, a GeoAI framework developed by the University of Wollongong combines field surveys, drone imagery and satellite data to monitor polar vegetation.
Researchers from the University of Wollongong, associated with the Australian Research Council's Securing Antarctica's Environmental Future program, publish in the ISPRS Journal of Photogrammetry and Remote Sensing a geospatial AI framework to monitor tiny Antarctic ecosystems: mosses, lichens and cyanobacteria. These organisms, sensitive to the warming and depletion of the ozone layer, are dispersed in small spots on isolated terrain, too large for field work alone and too thin for satellite resolution.
A three-storey chain
The team, led by Narmilan Amarasingam, built a hierarchical chain: ground observations calibrate drone mapping, which in turn provides training and validation data for models applied to satellite images. The device was tested for seven years at the Canada Glacier in the McMurdo Dry Valleys, one of the most vegetated areas on the continent. It produces consistent maps on all scales, from biological crusts of a few centimetres to the entire glacier, and helps distinguish a transient fluctuation from a sustainable climate-related transformation — without pretending to replace the on-site missions, whose use it increases.
Read the full article on the original source
Go to the source →