摘要

Agricultural practices need to be adapted to variable field conditions to increase farmers%26apos; profitability and environmental protection, so contributing to sustainability of farm management. This study proposes a combined approach of multivariate geostatistics and non-parametric clustering to delineate homogeneous zones that could be potentially managed with the same strategy. In a durum wheat field of Southern Italy, in organic farming, some soil physical and chemical properties (electrical conductivity; pH; exchangeable bases; total nitrogen; total organic carbon; available phosphorous), elevation and the Normalized Difference Vegetation Index were determined and interpolated by using geostatistics. %26lt;br%26gt;The clustering approach, applied to the (co)kriged estimates of the variables, produced the delineation of four sub-field zones. A significant relation between soil fertility and yield was not found in such zones. Despite this, the proposed approach has the potential to be used in future applications of precision agriculture. Further work could focus on site-specific nitrogen fertilization with suited machinery.

  • 出版日期2013-9