Mining activities and the waste products they produce can have significant effects on the surrounding environment, ranging from localized surface and ground water contamination, to the damaging effects of airborne pollutants on the regional ecosystem. The SDKI Sustainable Management and Rehabilitation of Mine Sites for Decision Support project is working collaboratively with federal departments, provincial governments and industry to develop tools to aid in the assessment of the environmental impacts of mining.

Figure 1. A false colour image CV580 polarimetric C-band SAR image over an abandoned gold mine near Seal Harbour, Nova Scotia. Wetland areas appear in blue, open water appears black.

The long-term monitoring of environmental impacts requires a cost-effective method to characterize land cover and land cover changes over time. The use of remote sensing technology to generate reliable land cover maps is a valuable asset to completing environmental assessments over mining-effected areas. Remote sensing uses the electromagnetic properties of the earth's surface to separate land cover types. Figure 1 shows an airborne radar image from an area with several abandoned gold mines along the eastern shore of Nova Scotia. At this site, researchers are exploring the transport of arsenic and mercury from mine waste disposal sites. Radar sensors transmit and receive microwave energy from the earth's surface, which can be used to derive information on topography, surface roughness and moisture status. Radar imagery is particularly useful for classifying wetlands, due to the unique geometric and moisture characteristics of these ecosystems. In this image, swamp wetland areas appear blue in colour, adjacent to open water areas, which appear black. By tracking field information on mine waste contaminants in a geographic context, this information can be integrated with land cover maps to derive more complex information on hydrology and contaminant movement in the environment.

Figure 2. Classified land and water maps derived from Landsat-7 satellite imagery over a section of southeastern Nova Scotia: land cover classification (left) and water classification (right). larger image [JPEG, 64.3 kb, 590 X 293, notice]

Optical sensors collect the sun's reflected radiation from the earth's surface. This information is particularly useful for determining vegetation cover types over large areas. With a sound knowledge of the vegetation cover types in the region, more complex environmental information on biodiversity and land recovery can be determined. Figure 2 shows land cover and water classification maps derived from Landsat-7 imagery over the same region of Nova Scotia. The maps were derived using a clustering algorithm that examines natural groupings in the data based on the spectral variability over six bands in the visible and infrared regions. The land cover map illustrates patterns in the vegetation cover that will be used in conjunction with ground-based environmental studies to gain a fuller understanding of the processes of contaminant transport, in collaboration with the Metals in the Environment Program (MITE) at the Geological Survey of Canada. The water map depicts patterns in the spectral opacity of the water, and is related to the depth and sediment loading of the water.

Future work will examine the complementarities of radar and optical data for land cover classification. In addition, the use of multispectral versus hyperspectral satellite imagery for classification will be explored to determine the advantages of using a greater number of spectral bands to derive indicators related to mine site reclamation and the accompanying improvements in accuracy. The resulting maps will be used for other environmental studies to aid in the development of models to predict the movement of mine waste through ecological systems.