Remote mine site rehabilitation monitoring using airborne hyperspectral imaging and landscape function analysis (LFA)

Landscape Function Analysis (LFA) is an environmental monitoring technique which is internationally recognized as a method of measuring and monitoring ecosystem function and rehabilitation progress. Airborne hyperspectral data were used to generate maps of the three Landscape Function Analysis (LFA) Indices over a 5 × 20 km area of a rehabilitated bauxite mine in southwestern Australia. The technique involved measuring field spectra of 35 LFA sites and generating Final Regression Coefficients using partial least squares analysis for the three LFA Indices, Stability, Infiltration and Nutrient Cycling. Both field and airborne spectral data were used in the calibration dataset, as the ground based sampling technique was not able to capture the information on the crowns of the tall eucalypts in rehabilitation older than eight years of age. The Final Regression Coefficients were applied to the airborne hyperspectral imagery to produce LFA Maps with good correlations to the field LFA measurements (Stability r²=0.76, Infiltration r²=0.67 and Nutrient Cycling r²=0.71 for n=15). The results demonstrated that airborne hyperspectral data could be reliably used to derive spatially continuous LFA maps over forest rehabilitated after bauxite mining in the Darling Range of Western Australia.