Noninvasively in situ 3D imaging of tree root systems

Knowledge of root systems and its distribution are imported e.g. for estimation of biomass as well as for the prevention of damages of subsurface distribution networks. Although several studies dealt with this topic, reliable 3D imaging of roots is still current research. Most studies used Ground Penetration Radar (GPR) for detecting root systems because it is non-invasive, easy to apply and capable to provide high resolution subsurface images up to several meters. The measured hyperbolas indicate the presence of potential roots. However, the GPR signal does not distinguish between roots and other reflecting inhomogeneities. Moreover, the attenuation of the GPR waves in clayey environments and in areas with changing soil moisture can complicate root detection. In particular in urban areas, the presence of heterogeneities in the subsurface that are potential diffractors can interfere with root reflections and hinder a reliable interpretation. In this study, preliminary 3D imaging results of tree roots are presented obtained by circular surveying around the tree trunk with a circle cross-line distance of 10 cm and lateral resolution of 2 cm. Measurements were made with a 500 MHz antenna in combination with a high precision prismatic theodolite. After identifying potential tree root hyperbolas within the GPR profile, the corresponding x, y and z coordinates were extracted and visualized in 2D and 3D plots to reveal inter-spatial connections. Due to distribution of the point scatterers, we could distinguish other non-linear objects from linear structures. These linear structures indicated possible tree roots which were marked on the corresponding surface for further investigation