Dr. Aston Chipanshi

Agriculture and Agri-Food Canada has a long history of delivering agriculture related information products operationally. Products such as the Canadian Soil Information System (CANSIS), Drought Watch, Agriculture Sustainability Indices and Crop Yield Forecasts are among the most frequently accessed products for research and assessing the present and future sustainability of agriculture across Canada. Little known are the crop density maps (https://agriculture.canada.ca/atlas/dcrops) which are produced from processing Annual Crop Inventory data where major crops are mapped using data from optical and SAR sensors, crop scouting and reported insured farm acres by Crop Insurance Companies. Crop density maps represent land units where specific crops are most likely to be found in a given year. Given a crop specific mask, crop yield predictor variables are generated based on the area covered by the crop. Preliminary results in the forecasting of crop yields across Canada has shown that the predictive skill gets better when crop specific masks are used instead of the generic agricultural land mask (Zhang et al, 2018). Crop density maps are also used by Agricultural Insurance Companies to assess trends in crop type cultivation; Land Use Planners to monitor land conversion, Agricultural Land use Planners to monitor rotations and land conversions. This poster presents changes in crop density maps of canola and wheat, the two most widely grown crops on the Canadian Prairies. While the maps represent averaged pixel values, the online application tool allows the user to zoom in and investigate changes year by year in more detail. As an extension of this work, Research Scientists at Agriculture and Agri-Food Canada are investing the generation of crop density maps within the growing seasons using Machine Learning Algorithms, expected to provide improved accuracy.