Spot Imagery for the Detection of Anomalous and Stressed Vegetation Over Mineral Deposits in Semi Arid Terrain

SPOT imagery acquired in May 1986 of the areas that include the Dugald River lead-zinc lode, northwest Queensland, Australia and the Ngwenalekau copper deposit, Botswana has been evaluated and compared with LANDSAT MSS imagery of the same areas. The data have been processed using band combinations, band ratioing, principal components analysis and classificatory techniques and the outputs compared with those of geobotanical, geochemical and biogeochemical data as well as geological information. For both study areas the image processing techniques cited have successfully distinguished discrete vegetation associations related to discrete surficial and bedrock geological units; and, notwithstanding the sharp contrasts between the nature of the geobotanical anomalies that in both areas occur over mineralised bedrock in Proterozoic sedimentary sequences, they have successfully identified the areas of anomalous vegetation over the mineral deposits. Greater vegetational and geological detail is revealed than on outputs generated from LANDSAT MSS data [see Cole, 1977, 1982(a) and (b)]. The additional information yielded by SPOT imagery compared with LANDSAT MSS imagery results from its greater spatial resolution - 20 metres as against 79 metres - and from ti fact that the three SPOT MSS spectral bands - 0.50 0.59um, 0.615-0.68um and 0.79-0.89um - respectivelc cover those sections of the electro-magnetic spectrum in which leaf pigments, notably chlorophyll, absorption for photosynthesis and leaf cell structure strongly influence reflectance. This accords opportunities for the detection of reflectance aberrations caused by stress due to drought and/or metal toxicity as well as for the identification of anomalous vegetation communities of moderate sizr These cannot be accurately discriminated on LANDSA I MSS imagery on which only very large geobotanical anomalies like that over the Lady Annie phosphate deposit in northwest Queensland [Cole, 1977, 1984; Cole and Owen-Jones, 19771 can be detected.