Title: Weather Effects on Yield and Using this Information to Make In–Season Fertilizer Decision

A linear statistical model was used to evaluate the interactions of nitrogen fertility, weather, soil, and irrigation on corn (Zea mays L.) yield from 1990 to 1995 in southeastern North Dakota. The fertility treatments were 0, 45, 90, 135, 180, and 225 kg N ha^-1 and 4 irrigation scheduling methods were applied. Also, there were a wide range of cumulative growing degree days (GDD) and evapotranspiration (ET) during this study period. It was found that the amount of GDD was a significant factor contributing to poor yields in 1992 and 1993. The optimum yield was defined as the yield at an optimum level of combined soil and fertilizer N, specific to cumulative ET or GDD from 1 May to 10 July for each site-year.  The multivariate model indicated that optimum rates of N increased with increasing yield. Weather induced variations in yield potential should enter into the decision making process when making fertilizer recommendations.

A regression model was developed that was used to predict the end-of-season corn grain yields on 10 July. This least-squares regression model was based on cumulative ET or GDD from 1 May to10 July and N fertility and corn grain yield. The 10 July prediction of this model corresponded well to individually observed end-of-season yields (r² = 0.80), and the predicted optimum yield (Y_OPT) was highly correlated to observed end-of-season Y_OPT (r² = 0.885). This model could provide a season-specific yield potential used to modify N application during the growing season, which could result in fertilizer savings in cool years or fertilizer increases to take advantage of optimum growing conditions and increase yields.