Practical insights for ECa-based soil mapping: Case studies in croplands and vineyards

Abstract

Mapping the within-field variability of soil properties is essential for precision agriculture. This work aims to develop an efficient approach to mapping soil properties using proximal geophysical methods, minimizing the number of soil sampling points required for the site-specific calibration of the soil parameter estimation models.Ten experimental fields in northern Italy were surveyed using geoelectric and electromagnetic induction methods to obtain maps of soil apparent electrical conductivity (ECa). Soil samples (topsoil and subsoil) were collected following a dense regular grid and analysed to determine chemical and physical properties.For each field, the relationships between ECa and soil parameters were studied. The possibility of obtaining detailed maps of some soil parameters by using a reduced number of sampling points, selected on the basis of the ECa map, was then investigated. The Mean Absolute Percentage Error (MAPE) was used to assess the estimation error.The results showed that in our study sites ECa is mainly correlated with textural parameters. The MAPE trends indicated a low variability in the minimum number of sampling points across the study sites. At each field, the calibration model for estimating soil properties could be built with a limited number of sampling points, that are representative of its soil characteristics. Thus, the soil information collected at the few selected sampling points provides insights into the within-field variability of each soil parameter, giving an indication of whether it is worth continuing the process of detailed mapping, which occurs when the coefficient of variation of the parameter is medium to high.

Keywords

• Soil properties
• Proximal geophysical techniques
• Soil sampling optimization
• Soil spatial variability
• Precision agriculture