Ground-penetrating Radar to Detect and Quantify Residual Root Fragments Following Peach Orchard Clearing

Consecutive replanting of peach (Prunus persica) trees on the same orchard site can result in various replant problems and diseases, including armillaria root disease (Armillaria spp.), which develops upon contact between the roots of newly planted trees and infested residual root pieces in the soil. There is little information regarding the quantity of roots remaining in stone fruit orchards following tree removal and land clearing. We investigated the utility of ground-penetrating radar (GPR) to characterize reflector signals from peach root fragments in a controlled burial experiment and to quantify the amount of residual roots remaining after typical commercial orchard clearing. In the former experiment, roots ranging from 2.5 to 8.2 cm in diameter and buried at depths of 11 to 114 cm produced characteristic parabolic reflector signals in radar profiles. Image analysis of high-amplitude reflector area indicated significant linear relationships between signal strength (mean pixel intensity) and root diameter (r = -0.517; P = 0.0097; n = 24) or the combined effects of root diameter and burial depth, expressed though a depth × diameter term (r = -0.630; P = 0.0010; n = 24). In a peach orchard in which trees and roots had been removed following typical commercial practice (i.e., trees were pushed over, burned, and tree rows subsoiled), a GPR survey of six 4 × 8-m plots revealed that the majority of reflector signals indicative of root fragments were located in the upper 30 to 40 cm of soil. Based on ground-truth excavation of selected sites within plots, reflectors showing a strong parabolic curvature in the radar profiles corresponded to residual root fragments with 100% accuracy, whereas those displaying a high amplitude area represented roots in 86.1% of the cases. By contrast, reflectors with both poor curvature and low amplitude yielded roots for less than 10% of the excavated sites, whereas randomly selected sites lacking reflector signals were devoid of any roots or other subsurface objects. A high level of variability in the number of residual roots was inferred from the radar profiles of the six plots, indicating an aggregated distribution of root fragments throughout the field. The data further indicated that at least one residual root fragment would be present per cubic meter of soil, and that many of these fragments have diameters corresponding to good to excellent inoculum potential for armillaria root disease. Further GPR surveys involving different levels of land clearing, combined with long-term monitoring of armillaria root disease incidence in replanted trees, will be necessary to ascertain the disease threat posed by the levels of residual root biomass observed in this study.