Application of multivariate geostatistics in delineating management zones within a gravelly vineyard using geo-electrical sensors

In gravelly soils, surveys are generally time-consuming, labour-intensive and costly. This limits the possibility of adopting an appropriate sampling to determine within-field spatial variability. The potential use of electro-magnetic induction scans (EMI) to measure apparent electrical conductivity (EC) and improve the estimate accuracy of sparsely sampled primary variables was assessed in a 5-ha gravelly soil vineyard in Valpolicella, north-eastern Italy. EC was measured using a Geonics EM38DD operating in both horizontal and vertical mode. Geo-electrical investigations were also done in 18 positions with the electrical resistivity tomography (ERT) method to obtain high-resolution images of the soil profile, The spatial variability of soil properties and their relationships with EC in horizontal and vertical mode was estimated using multivariate geostatistical techniques. Spatial dependence between EC and physical soil properties (particle-size distribution) was explored with factorial kriging analysis (FKA) that could isolate and display sources of variation acting at different spatial scales, expressed as regionalised factors, which was followed by fuzzy c-means classification for zoning the vineyard. There was a generally close relationship between EC and the measured physical properties: EC was negatively correlated with the coarser texture components (gravel and sand) and positively with the finer ones (clay and silt). EC measurements were also consistent with ERT profiles, evidencing the presence of gravelly parent material, with low electrical conductivity, variably distributed in the 3 dimensions and affecting vine rooting depth. FKA isolated two significant regionalised factors which, with an acceptable loss of information, give a concise description of the soil physical variability at the different selected spatial scales. These factors, used in fuzzy c-means classification, allowed the delineation of zones to be managed separately. The results prove that EM38DD could be advantageously used to map soil spatial variability in gravelly soils, even if ground-truth soil samples are obligatory to understand and interpret the EC measurements. (C) 2009 Elsevier B.V. All rights reserved.