Toward the Ultimate Soil Survey: Sensing Multiple Soil and Landscape Properties in One Pass

Abstract

Many studies have found the range of spatial dependence is shorter than the distances used in most grid sampling, and government soil surveys have limited utility in many precision applications. A new multi‐sensor platform was developed that records apparent soil electrical conductivity (ECa), optical reflectance readings with red and near‐infrared light‐emitting diodes and pH along with topography data. The objective of this study was to evaluate its performance for estimation of soil cation exchange capacity (CEC), organic matter (OM), and pH on eight fields in four states, comparing results with lab‐analyzed samples, USDA soil surveys, and 1‐ha grid maps. Proximal soil sensor measurements correlated well with lab‐analyzed soil samples. The OM calibrations using ECa, optical, and/or topographic data showed good performance with R2 of 0.8 or higher and ratio of prediction to deviation (RPD) of 2.5 or greater in all fields. In CEC calibrations, five of six fields had higher than 0.86 for R2 and greater than 2.8 for RPD. The pH calibration results showed RPD of 2.1 or greater and R2 of 0.76 or higher in seven fields. The sensor maps showed small‐scale variability not detected at conventional grid sample scales or with USDA soil surveys. Using the proximal soil sensors, the average root mean square error of prediction for OM was 2.78 g kg−1, CEC 1.20 cmolc kg−1, and pH 0.26 for the project fields. These values are significantly lower than the soil property ranges found in the soil surveys. This is a promising development for improving farm practices and management.