Unmanned Aerial Vehicle for Peatland Degradation Assessment

South Africa is a water-scarce country, and groundwater-fed peatlands are susceptible to various degradation causes due to water stress. Peat degradation releases CO2 gas into the atmosphere, contributing to climate change. Subsurface peat fire detection using remote sensing data is difficult (i.e. conventional surface fire algorithms cannot detect and monitor peat fire incidences). Therefore, two sites have been considered for unmanned aerial vehicle application to detect thermal anomalies and wetting levels, i.e. Molopo (Mol) and Molemane (Mal). Two sensors we mounted on a hexacopter drone to capture multispectral (4 bands) and thermal data. The thermal data was captured early morning to remove the sun-heat background and detect any underground thermal emissions. The multispectral data was captured in the mid of the day. The multispectral images were mosaicked, orthorectified, and then resampled to 0.05 m resolution. The thermal data was calibrated using the ground control point temperature data. The thermal images were resampled to 0.08 m pixel size. Ground data, such as soil moisture using a dielectric sensor and soil thermal characteristics using KD2 Pro Thermal Properties Analyser, was also captured. Various indices were calculated, such as the Simple ratio (SR), Normalized Difference Vegetation Index (NDVI), and Vegetation Red Edge Index. The combined index was also calculated to determine the surface moisture (inundation) as the primary indicator for peat fire susceptibility.