The Prairie Shelterbelt Program of Agriculture and Agri-Food Canada, has produced many benefits for farmers and the prairie landscape: reducing soil erosion, protecting crops, controlling snow drifting over highways/roads and providing wildlife habitat. Due to growing concerns about rising levels of carbon dioxide in the atmosphere, the ability of shelterbelts to sequester carbon dioxide may also be of importance. Although the Prairie Shelterbelt Program has been distributing tree and shrub seedlings for more than 100 years, and records of the numbers of trees shipped have been kept, an inventory of the number of trees that have been successfully planted and their locations on the landscape does not exist. When observed on Spot- 2.5 m. panchromatic satellite imagery, shelterbelts have distinct shapes, textures, and spatial relationships with other objects within the landscape. Object-based image classification methods are well suited to segmenting remotely sensed imagery based on these characteristics and have the potential to be used for delineating shelterbelts. In this thesis, Definiens object-based image classification software is shown to be an effective way to create a provincial-scale shelterbelt inventory. The primary objective of this research was to develop a process by which the spatial coverage of shelterbelts in the Province of Saskatchewan could be determined to facilitate the estimation of the amount of carbon being sequestered. The results show that due to the large diversity of agro-environmental conditions across the province, and seasonal and contrast inconsistencies of the panchromatic satellite data used, this was not possible. Nonetheless, the results show that object-oriented classification methods have the potential to detect the location of shelterbelts with an accuracy of over 80%. It is also possible to obtain a reasonable estimate of carbon sequestration. Future shelterbelt inventories should focus on the enhanced data potential found in high-resolution colour-infrared imagery.