Analysis of hyperspectral airborne HyMap data for vegetation mapping around Lahnaslampi talc mine, Finland

Hyperspectral (450-2500 nm) airborne HyMap imaging spectrometer data (spatial resolution 5 m) were processed for mapping of environmental effects on vegetation due to talc mining activities at Lahnaslampi, Sotkamo, Finland. As a part of the european MINEO project, the HyMap imagery was interpreted by applying vegetation indices, using standard classification and hyperspectral mapping approaches. Classification processes and accuracy assessment were based on forest inventory data, land use data, geochemistry and soil electrical conductivity data, as well as aerial photograph interpretation of tree species composition and recent disturbances due to forest management and mining operations. First, a general forest stand classification was performed with the maximum likelihood classifier, which resulted in an overall accuracy of 78.2%. Then, vegetation changes, such as decline in foliage, were revealed next to the mine by low NDVI and SAVI index values. Finally, the standard classification and hyperspectral mapping approaches gave more refined result in specifying distribution of stands dominated by mature Norway spruce (Picea abies), and sapling stands of downy birch (Betula pubenscens) at contaminated sites (dust and/or seepage water plumes). Image processing results and spectroradiometer measurements of the tree species indicate that the deterioration in plant vitality, especially that of Norway spruce and downy birch can be detected as slightly increased reflectance in visible (500-700 nm) and clearly decreased reflectance in near-infrared (700-1400 nm) wavelengths. The present study indicated that the impact of the mine on vegetation was minor and the contamination effects were found only in close vicinity of the mine.