Key points are not available for this paper at this time.
Abstract Management zone (MZ) or variability zone delineation is a critical component of precision agriculture (PA), enabling site‐specific management to optimize crop production and resource efficiency in response to within‐field variability. This study evaluated whether digital soil maps (DSM or continuous soil property prediction maps) can serve as a superior information layer compared to the Soil Survey Geographic Database (SSURGO) for soil MZ delineation. High‐resolution DSM data of four farmer fields in northeast Oklahoma, including soil features such as macro‐ and micronutrients, soil texture, and chemical properties at multiple depths, were used in two clustering techniques, k ‐means and fuzzy c ‐means (FCM), to delineate DSM‐based MZs. Performances of DSM‐ and SSURGO‐based MZs were evaluated using the variance reduction (VR) index based on yield monitor data from four fields between 2014 and 2020. In a baseline comparison (i.e., same number of MZs), k ‐means and FCM achieved a relative VR increase of 78% on average across all fields compared to SSURGO (with an absolute VR difference of 4%). When the number of MZs increased, VR was further improved by DSM‐based clustering, particularly with four to five MZs (VR increased by 236% with five MZs, with an absolute VR difference of 13%). Our results showed that DSM‐based clustering outperformed SSURGO‐based zoning in reducing the within‐zone yield variability. The leverage of DSM and clustering techniques enabled finer‐scale on‐farm yield variability detection and therefore enhances MZ precision. The insights from this study can inform future site‐specific management strategies, ultimately supporting sustainable resource allocation, optimizing inputs, and minimizing environmental impacts.
Abbasi et al. (Sat,) studied this question.