Agricultural Data Analytics for Environmental Monitoring in Canada
Huffman, T., Olesen, M., Green, M., Leckie, D., Liu, J., Shang, J. 2017. Agricultural Data Analytics for Environmental Monitoring in Canada. In: Federal Data Science, Transforming Government and Agricultural Policy Using Artificial Intelligence. Batarseh, F.A., Yang, R. Edits. Academic Press.
The development of accurate international reports and domestic policies related to environmental sustainability requires high-accuracy, high-resolution and multi-temporal national resource maps. Although a number of land cover and vegetation maps have been produced for Canada over the past 30 years, the classification schemes often differ, spatial resolution and coverage varies between products and classification accuracy generally remains at or below 82-86%. This chapter outlines an approach that uses appropriate analytical procedures to integrate a wide variety of raster and vector spatial products in order to generate a series of land use maps that reflect the best estimate of the class at each of over 6 billion 30m pixels in Canada. The premise of the approach was that careful evaluation of inputs and rigorous development of rules regarding accuracy and preponderance of evidence should bring out the best of each input and allow the development of output maps with higher class and overall accuracies than any of the individual input products. The project resulted in maps for 1990, 2000 and 2010, all at the same scale and categorized according to the 6 classes of the Intergovernmental Panel on Climate Change (IPCC); Forest, Cropland, Grassland, Wetland, Settlement and Otherland.
Report a problem on this page
- Date modified: