Protocol for multivariate homogeneous zone delineation in precision agriculture. (March 2016)

• Record Type: Journal Article
• Authors: Córdoba, Mariano A.; Bruno, Cecilia I.; Costa, José L.; Peralta, Nahuel R.; Balzarini, Mónica G.
• Is Part Of: Biosystems engineering. Volume 143(2016:Mar.)
• Page Start: 95
• Page End: 107
• Publication Date: 2016-03
• Subjects: Spatial data -- Spatial autocorrelation -- Multivariate classification -- Site-specific management -- Software specifications; Bioengineering -- Periodicals; Agricultural engineering -- Periodicals; Biological systems -- Periodicals; Génie rural -- Périodiques; Systèmes biologiques -- Périodiques; 631
• DOI: 10.1016/j.biosystemseng.2015.12.008
• Languages: English
• ISSNs: 1537-5110

Abstract

Abstract : Uniform management of agricultural fields has been increasingly replaced with environmentally based management, which is benefited by the identification of homogeneous zones within crop fields. Such zones are based on the classification of field sites into groups of homogeneous features. Multiple causative agents of variability and the response of agricultural crops should be considered for zoning. Several correlated variables are usually measured and georeferenced for this purpose at multiple sites within the field. This paper presents an approach to promoting the integration of different statistical tools for identifying homogeneous zones based on site covariates. The methodological innovation of this work involves cleaning and re-scaling of spatial data, as well as multivariate and geostatistical analyses in a logical sequence (protocol). Statistical topics for further improvement and protocol applications are noted. The analytical process has been illustrated using a rain-fed wheat crop (60 ha) from the Argentine Pampas, with apparent electrical conductivity, elevation and soil depth as master variables for zoning, and yield, soil organic matter and clay to validate the created management zones; however, it may be applied to other production systems using georeferenced data. The R scripts and the sample file to run the proposed protocol are provided as electronic supplementary material. Highlights: A protocol to delineating multivariate homogeneous zones is presented. We include data clean-up, multivariate classification and filtering for spatial data. On line access to R scripts for implementing the whole protocol is provided.