Optimum spring N fertilizer rates for winter wheat in Virginia vary widely from field to field, but traditionally spring N is applied at a uniform rate to all fields. A recently-developed tissue test procedure provides a field-specific evaluation of crop N status and predicts optimum N rate for the second spring N application in a split spring application management system. However, this procedure is based on a small number of researcher-planted experiments utilizing a single cultivar; it fails to provide field-specific rate recommendations for the first spring N application; and it is not accessible to farmers who are unwilling to split their spring N applications. Our objectives were: to evaluate the reliability of the tissue test procedure in a large number of farmer fields; to develop a method for making field-specific N rate recommendations for the first spring application in a split-application management system; and to develop a method for making field-specific N rate recommendations in a single-application management system. Forty-five spring N rate experiments were established in farmer fields over a five-year period. A range of N fertilizer rates was applied in early spring (Zadoks growth stage 25) and again in mid-spring (Zadoks growth stage 30) in all possible combinations. Yield data were used to calculate economic optimum N rates at growth stage (GS) 25 and GS 30 with split-application management, and at GS 25 with single-application management. These optima were regressed against a variety of predictor variables measured in the same fields. The reliability of the previously-developed tissue test procedure for making GS 30 N rate recommendations was confirmed. Tiller density was the best predictor of optimum GS 25 N rate with split-application management, while soil nitrate to 90 cm was the best predictor of optimum GS 25 N rate with single-application management. These three relationships fit together to form a flexible and powerful system for making spring N rate recommendations for winter wheat. This system increased estimated profit and apparent fertilizer efficiency in these experiments. / Ph. D.

The loss of nitrogen from agricultural land to ground and surface waters is currently a major concern in the Commonwealth of Virginia. Farmers use nutrients on plants to maximize profit from the selling of a crop. For them losses of nutrients through inappropriate nutrient application are undesirable. Thus more effective nutrient management is beneficial for both the farmers and the remainder of society. Achieving environmental quality goals while taking into account farmers' income risk is essential when making fertilization recommendations. This study on Richmond county, uses the EPIC model and stochastic dominance analysis to compare four different fertilization regimes at the field and farm level; and then uses a sensitivity analysis to examine how the ranking of different regimes are affected by changes in crop prices. Results suggest that regime 4 which is the EPIC automatic fertilization regime has a better performance than the other regimes. However potential costs not accounted by EPIC involved at the farm level when using regime 4 may negatively affect its adoption by farmers. Thus further studies need to be done to understand and assess the performance of regime 4. The results do not show any significant difference between farm and soil type levels of analysis. The sensitivity analysis mainly affects the less risk averse decision makers, and change only the ranking of the three fertilizer regimes specified by the researcher as opposed to the automatic fertilizer option of EPIC. / Ph. D.