Electronic Thesis and Dissertation Repository


Master of Science




Dr. Irena Creed



While progress has been made in automating wetland identification, identifying lost and restorable wetlands remains a challenge. A suite of automated methods was developed and applied to the Nose Creek watershed near Calgary, Alberta to establish a historical wetland inventory and the proportion of permanently versus temporarily lost wetlands. A power-law function of wetland area vs. wetland frequency using wetlands derived from the fusion of a high resolution digital elevation model and near-infrared data identified permanent loss of 11.0% by number and 0.6% by area. The difference between historical and existing wetlands was used to estimate a further temporary loss of 61.1% by number and 78.3% by area. Historical wetlands lost to ditch drainage are easily restored by ditch plugging. Therefore, an algorithm was created using digital terrain analysis that distinguished drainage ditches intersecting wetlands using surface curvature. The 1,588 ditch-drained wetlands identified represent a potential recovery of 11.7% of the temporary loss by number and 12.5% by area. Automated techniques to estimate wetland loss and identify priority wetlands for restoration provide powerful tools for wetland management.