Delineation of management zones using soil properties and multivariate geostatistical approach

Agricultural fields naturally have within-field variations that may cause spatial variability in crop yields. Delineating a field into a few management zones (MZ) is therefore a practical management approach for precision agriculture. In this study, some soil properties (clay content, organic matter content and water content measured at different dates) and a multivariate geostatistical approach were used to delineate management zones. The temporal stability of soil water content spatial patterns may be an additional approach in delineating management zones. The present study was conducted on a 120000 m2 durum wheat field in Southern Italy, where soil properties and soil water content were measured at different times over three crop seasons (2005-06, 2006-07, and 2007-08) at 100 locations. The multivariate spatial and temporal data were analysed by Factor Kriging Analysis (FKA). All simple and cross model variograms were calculated by fitting a linear model of coregionalization including the nugget effect and two spatial structures at shorter (70 m) and longer range (500 m). The application of factor kriging allowed us to identify one regionalized factor at longer scale and to partition the field from the point of view of temporal stability of soil water content, into contiguous zones to be submitted to site-specific management. The method showed a great flexibility to combine data from several sources in the delineation of management zones.