Prairie crop yield estimates from modelled phenological development and water use

The feasibility of estimating average yields of spring wheat, barley and canola for Manitoba, Saskatchewan and Alberta, individually and collectively, from crop-specific model output routinely published in the Winnipeg Climate Centre's Agrometeorological Outlooks was examined. Statistically significant correlations were found between crop district/census division yields for 1988–1992 and modelled end-of-season crop water-use, and between these yields and the end-of-season ratios of water-use to water-demand and modelled days-to-maturity. Up to 69% of the observed variance in spring wheat yields could be explained by modelled agrometeorological variables. The best R2 values for barley and canola were 67 and 64%, respectively. The regression equations so derived were applied to 1993 model output and the resultant average crop district yield estimates for wheat, barley and canola, aggregated up to the provincial and the Prairie-wide levels, were compared with Statistics Canada's preliminary (i.e., July) and post-harvest (i.e., November) crop survey estimates and with yield estimates from the Canadian Wheat Board's Western Canada weather model. From these comparisons it was concluded that the Agrometeorological Outlooks can provide timely provincial and Prairie-wide estimates of average yields for these important Prairie crops.