Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
6458863 | Computers and Electronics in Agriculture | 2017 | 10 Pages |
â¢The Algorithm is based on a well-established agronomical concept.â¢Site-specific N-fertilization trials were used for calibration.â¢The algorithm was tested on commercial farms on usual field-scale.â¢Canopy N uptake was estimated using a tractor mounted remote sensing device.â¢N-balance surpluses were reduced significantly when applying N site-specifically.
Winter oilseed rape (WOSR) is a major crop in Germany, combining economic benefits with a high value in crop rotation, but it still lacks agronomically sound concepts for site-specific nitrogen (N) fertilization. Since ecological challenges resulting from high optimal N rates and a low N harvest index are approaching on WOSR cropping systems, optimizing N fertilization becomes crucial. Recent studies showed the importance of taking autumnal N uptake into account when estimating optimal N rates for WOSR, thus autumnal N is pivotal in the algorithm that is introduced in this study. The algorithm was parameterized by using data from site-specific N fertilization trials and optimized to reduce N fertilizer amounts. Afterwards it was tested on different commercial farms in northern Germany. The autumnal N uptake was estimated using hyperspectral reflection measurements gained from tractor-mounted devices, and the data was processed to N application maps used for the N application in spring. In addition, a uniform optimal fertilization and a uniform application of average N rates calculated by the algorithm were applied to provide control treatments. Yield, N balance and economic net-revenue were evaluated for each treatment. Yields from site-specific fertilization were slightly lower (0.06Â t/ha) than from uniform optimal treatment but 0.22Â t/ha higher than from the uniform application of the site-specific N amount (not significant in both cases). The N balance was significantly lower when fertilizing site-specifically instead of applying uniform optimal N rate, while the net-revenues were slightly higher.