The ABMI Wall-to-Wall Land Cover 2000 & 2010 Datasets
Full digital classification of Alberta’s landscapes.
The ABMI Wall-to-Wall Land Cover 2000 and 2010 datasets provide Alberta-wide, polygon-based representations of provincial land cover circa the years 2000 and 2010, respectively. They are based on the digital classification of 30 metre resolution Landsat satellite images, and enhanced using GIS datasets provided by the Government of Alberta. The land cover products contain approximately 1 million polygons, each, and comprise 11 classes, including water, shrubland, grassland, agriculture, exposed land, developed land and different forest types.
Large-area land cover maps based on consistent remote sensing data can help land managers to gain knowledge and make informed decisions.
The Wall-to-Wall data sets are suitable for regional-level representation, for spatially explicit, long-term trend monitoring. They offer a baseline for land management, monitoring, land-use planning; or for regional-level environmental, ecological, biological and other studies.
The ABMI Wall-to-Wall Land Cover circa 2000 data set is based on the combination of two pre-existing raster (i.e., pixel-based) land cover products that were derived using the digital classification of Landsat satellite imagery dated circa the year 2000. These offer partial but complementary coverage of Alberta:
1) The Canadian Forest Service’s Earth Observation for Sustainable Development land cover classification, covering Alberta’s forested regions
2) Agriculture and Agri-Food Canada’s Land Cover for Agricultural Regions of Canada classification, covering the province’s southern agricultural and settled regions
The ABMI’s Geospatial Centre first created a vector, polygon-based segmentation using the original satellite imagery, to represent relatively homogeneous, consistent land cover parcels across Alberta’s various landscapes. Next, the GIS analyst team proceeded to assign land cover classes to the resulting polygons using a sophisticated combination of GIS and image analysis procedures involving the original imagery and the two pre-existing raster land cover layers. Enhancements included the insertion of additional Government of Alberta data layers representing hydrological features (water bodies, rivers, etc.) and particular access features (pipelines, powerlines, railways and roads), into the polygon land cover data set. The result is an 11-class, province-wide, polygon-based land cover data set. The ABMI has assessed the accuracy of this product, using an extensive validation dataset, to be 75%.
The more recent 2010 Wall-to-Wall Land Cover dataset was produced using a combination of satellite imagery from circa 2010, image change detection techniques and GIS analysis, and includes new landscape features like cutblocks, wildfires, surface mining, roads and urban areas. As these inputs are of greater quality than the original, raster land cover inputs into the 2000 land cover data set, the accuracy of the resulting 2010 product is estimated to meet or exceed 75%.
For full details, please see:
Castilla, G., Hird, J., Hall, R. J., Schieck, J. and McDermid, G. J. 2014.
Completion and Updating of a Landsat-Based Land Cover Polygon Layer for Alberta, Canada. Canadian Journal of Remote Sensing, vol. 40(2), pp.92-109.
More questions? Feel free to contact:
Geospatial Centre Director
ABMI Research Fellow