Mapping Queensland land cover according to FAO LCCS using multi-source spatial data

摘要

Recent developments in the application of high-resolution satellite data for extracting spatial information have encouraged land cover mapping activities throughout Australia. Together with the increase of these mapping activities, the need for standardizing land cover classification schemes for maps has been emphasized. In 2000, FAO published a widely accepted land cover classification system based on priori (pre-decided) approach, which can be applied to any area of the world. This study examined land cover features of the geographically diverse state of Queensland, Australia, with a special emphasis on dynamic land cover differences, and applied fundamentals of FAO LCCS to classify land cover of two different regions. These two regions; 1) the arid Mt. Isa region in northern Queensland; and 2) the urbanized southeast region including Brisbane and the Gold Coast, cover a diversity of landscape types. The combination of these regions is representative of a large portion of other regions of the Australian continent. Improving the resolution of national land cover maps is of a national priority as it required to address a wide range of environmental and natural resource issues. Classifications were conducted on SPOT 10m satellite data, supported by 2.5m high resolution SPOT true colour images, ASTER, and some Landsat scenes. Other GIS data layers including SLATS 2001-2003 (Statewide Land cover and Trees Study) data and extensive field survey information collected from both areas were also utilized. Both regions were classified initially into three levels (dichotomous phase) of the FAO LCCS and then spectral features, field investigations, and other image attributes were used to generate 4th level (Modular-Hierarchical phase) land cover categories. Results show a very satisfactory land cover map at 10m resolution compared to existing land cover products. The standard FAO classification scheme provides a standardised system of classification that can be used to analyse spatial and temporal land cover variability throughout the country. This approach also has the advantage of facilitating the integration of Australian land cover mapping products with global land cover datasets.