Multi-Temporal Electrical Conductivity and NDVI Analysis for Vineyard Functional Zone Mapping

Precise soil spatial identification and characterization is crucial for optimizing vineyard management and enhancing grape quality. Various approaches exist for characterizing spatial soil variability, all aimed at zoning and identifying areas that, despite experiencing the same climate, exhibit different crop responses and therefore require differentiated management. However, the complexity of soil-plant interactions and the dynamic nature of soil properties over time necessitates the optimization of existing zoning methodologies. For instance, electrical conductivity (EC) mapping is a common technique, but relying on single-date measurements often fails to capture the full extent of spatial and temporal soil variability, even within a single growing season. Furthermore, commonly used electromagnetic induction (EMI) instruments operate at multiple frequencies to analyze different soil depths, making it challenging to directly relate these measurements to the specific soil volume explored by plant roots. Focusing on a well-defined soil depth, even if coarsely related to the root zone, would be more relevant for plant-soil interaction studies. Identifying the optimal period for characterizing soil spatial variability is therefore a key objective.

In this context, within the Agritech National Research Center project (https://agritechcenter.it/it/),  we study the use of multi-temporal EC data, acquired with a GF Instruments CMD MiniExplorer 6L, for delineating functional homogeneous zones within an Aglianico DOC vineyard at Tenuta Donna Elvira, Grottoni (AV), Italy. The CMD MiniExplorer 6L, capable of measuring EC at up to nine depths within 2 meters by combining its horizontal and vertical dipole configurations, provided detailed soil information.

EC data were collected over five distinct days spanning from April to late August, capturing seasonal soil moisture variations. Concurrently, multi-spectral imagery was acquired using a DJI Phantom 4 RTK drone across a broader timeframe from April to late October. Normalized Difference Vegetation Index (NDVI) values were derived from the drone imagery to assess canopy vigor and variability.

A k-means clustering approach was applied to the daily EC datasets, exploring various depth combinations to generate 36 distinct clustering outputs for each acquisition date. This multi-depth approach allowed for a comprehensive assessment of soil variability at different scales. The resulting EC-derived clusters were then compared with the mean NDVI values extracted for each cluster. This comparison aimed to evaluate the relationship between soil electrical properties and vine vigor, as reflected by NDVI.

The analysis revealed a strong correlation between EC-derived clusters and NDVI, demonstrating the effectiveness of EMI measurements for differentiating soil properties relevant to vineyard performance. The study also highlighted the influence of acquisition timing on the efficacy of soil classification, identifying optimal periods and depth configurations for maximizing the differentiation of functional zones. This multi-temporal, multi-depth approach provides valuable insights for precision viticulture, enabling targeted management practices based on spatially explicit soil and canopy information. The results contribute to a better understanding of soil-vine interactions and offer a practical methodology for efficient vineyard zoning.