Classification of vineyard soil physicochemical zones using non-invasive frequency-domain electromagnetic induction and NDVI methods

In this work, a novel method for obtaining maps about the physiochemical nature of soils from electromagnetic induction (EMI) data is investigated for a French vineyard. Specifically, K means clustering is used to produce maps from both the quadrature and in-phase components of EMI measurements. The commonly discarded in-phase component of EMI measurements exhibits a higher correlation with plant vigor, as measured by NDVI. In addition to the non-intrusive characterization, extensive soil samples are used to explore the correlations and causations that link soil physiochemical properties with EMI and NDVI data. It is demonstrated that the in-phase component of EMI is related to metal concentrations at the site, these properties co-vary with crucial properties related to plant vigor, e.g. cation exchange capacity and organic matter content.