摘要：Precision agriculture (PA) in crop production refers to the management approach in which inputs are applied according to the site-specific demand and timing in order to improve efficiency of inputs and farm profitability. Among the variety of techniques under the concept of PA, variable rate application (VRA) of both phosphorus and nitrogen fertilizers, and temporal as well as spatial variability of corn crop response to these fertilizers were studied for the potential contribution of VRA of fertilizers to farm profitability. The crop response data were generated by an on-farm experiment in Windom (Cottonwood County, Minnesota) in 1997, 1999, and 2001, using a commercially cultivated field of 10 ha area, where corn and soybean were rotated. Using a fractional factorial design, three phosphorus fertilizer rates (0.0, 56.2, and 112.3 kg ha-1) and five N-fertilizer rates (0.0, 67.4, 112.3, 157.3 and 202.2 kg ha-1) were applied to the field before corn planting. The observed yields were regressed on these treatment rates and site-specific crop response functions (SS-CRFs) were estimated using maximum likelihood estimation. The SS-CRFs were employed for the optimization problem and the site-specific economically optimal rates (SS-EORs) of phosphorus and nitrogen were obtained as the solution of the profit maximization problem. The temporal as well as spatial variations in these SS-EORs were observed; the estimated SS-EORs of phosphorus and nitrogen fertilizers varied among 33 sites from 0.0 to 112.3 kg ha-1 and from 0.0 to 202.2 kg ha-1, respectively in each year. Moreover, the results indicate that applying the SS-EORs could have been more profitable than uniform rate. The total gains from changing to the SS-EORs from the uniform rates (P = 56.2 kg ha-1 and N = 157.3 kg ha-1) would have been $63.58 ha-1, $64.34 ha-1, and $117.42 ha-1 in 1997, 1999, and 2001, respectively. This study indicates that VRA of fertilizers potentially improve the profitability of corn production. In order for this technology to become practical, some obstacles need to be removed, such as a less expensive method to obtain SS-EORs than on-farm experiences, less expensive machinery, more accurate functional form for SS-CRFs than quadratic functions. Once VRA technologies advance and become inexpensive, the "potential" gain from VRA can be fully realized.
