Phenological stages-based NDVI in spring wheat yield estimation for the Canadian prairies

Spring wheat is a major export crop of the Canadian prairies (Alberta, Manitoba, Saskatchewan). Approximately 75 percent of its total production is exported through a grain marketing agency. In order to ensure optimum profits from the exports, pre-harvest wheat yield estimates are required as a component to the export marketing strategy and to drought planning. The more accurate the yield estimates, the better the strategy. Currently a regression model is employed, namely, the Western Canada Wheat Yield (WCWY) model to obtain the yield estimates. The model employs monthly weather data (average temperature and precipitation) and determines a drought index for most weather stations (about 120) across the prairies. The indices are then averaged and regressed against the average wheat yield for the prairies, to develop the WCWY model. The objective of this paper is to improve the performance of the WCWY model by utilizing daily, in place of monthly weather data and, further, by incorporating a Normalized Difference Vegetation Index (NDVI) based variable into the model