Reliably Mapping Clay Content Coregionalized with Electrical Conductivity

Core Ideas Surface clay content was derived through a combination of clay measurements and a proximal sensor. Apparent electrical conductivity (ECₐ) supported reliable clay content estimation. Predictions of clay coregionalized with ECₐ performed well, even for one soil sample per two hectares. Understanding the spatial variability of soil texture within field soils is important due to its influence on a large number of soil and plant related processes and for site‐specific application of inputs that are crucial to crop production. It remains a problem to obtain a reliable clay content map based on a limited number of sampling locations. The objective of this study was to identify spatial variability of soil clay content and the behavior of the estimation result for different spatial resolutions of measured clay content (0–20 cm depth) in combination with a coregionalization approach using apparent electrical conductivity (ECₐ). In a silty loam soil, soil clay content was measured at 96 points in a 50‐m by 50‐m grid within an agricultural field. ECₐ was measured using a contact sensor Veris 3150. Data were analyzed with ordinary kriging and cokriging while using ECₐ at a shallow depth. We analyzed different sampling scenarios based on clay subsamples of 48, 24, and 12 data points distributed over the 27‐ha field. In all scenarios investigated here, the RMSE stayed in the range of 3 to 4% by using different validations, with cokriging performing constantly better than ordinary kriging. Clay content maps estimated with cokriging maintained a satisfactory precision when the sampling density was reduced to one sample per two hectares, a result that leads to the conclusion that electrical conductivity in combination with spatial coregionalization demonstrated to be a promising tool to estimate the spatial variation of clay content even at a low clay sampling density.