Satellite Monitoring of Urban Sprawl and Assessing the Impact of Land Cover Changes in the Greater Toronto Area

This research investigates the extent of land cover change, particularly urban sprawl, in the Greater Toronto Area (GTA) and the nature of the resulting landscape fragmentation, particularly with regard to the Oak Ridges Moraine (ORM), an ecologically important and sensitive area for the region. Three scenes of Landsat TM imagery were acquired from July/August 1985, July/August 1995 and July 2005. The TM images and TM5 textures were classified into seven land cover categories, i.e., low-density built-up, high-density built-up, construction sites, agriculture, forest, golf courses, parks/pasture and water using a maximum likelihood classifier. The overall accuracies were 90% for 1985, 93% for 1995 and 91% for 2005. Landscape fragmentation due to spatiotemporal changes was then evaluated using spatial metrics.The landscape composition results indicate that significant urban growth and sprawl occurred in the GTA. Comparison of the land cover classes over the 20 year period reveals a shift in dominance in the landscape through a clear increase in low-density built-up land cover categories coupled with a clear decrease in agricultural areas. The landscape configuration metrics calculated suggest increased fragmentation of both agricultural and low-density built-up areas during the two decades. Combined with the change in their proportions, this meant more isolation and attrition for the former, while the latter gained area through the appearance of more frequent and relatively smaller patches. In addition, the contagion and proximity indices indicate that agricultural areas were increasingly reduced in area, subdivided into smaller patches and became more dispersed over the given time period, while the opposite was true of suburban areas. The most extensive urban expansion onto the ORM occurred in municipalities within the region of York and therefore a thorough investigation of the impacts of this expansion is strongly recommended. ∗ Corresponding author, e-mail: [email protected]