Coarse Roots Biomass Estimation Using GPR: Forward Simulation and Field Controlled Experiment.

Ground penetrating radar (GPR) has been tested for coarse roots detection and quantification (including the estimation of root diameter, biomass, and fresh weight) since the late 1990s. Current GPR-based coarse root biomass estimation method relies on the regression model between root biomass and reflectance indexes extracted from GPR radargrams (e.g., pixels within threshold range). Because of the high permittivity of water, the reflectance indexes controlled by the permittivity contrast between different media have been proved to be sensitive to root moisture content and soil moisture content. However, up to now, now research has been aware of the impact from root moisture content on the estimation of coarse root biomass (which is independent of root moisture content) by reflectance indexes (which are closely related to root moisture content). In this study, the impacts of root moisture content on current GPR-based coarse root biomass estimation are quantitatively evaluated by both forward simulation and controlled field experiment. In all three GPR antenna frequency levels (i.e., 0.5GHz, 0.9GHz, and 2GHz), the result shows: (1) the correlation between GPR indexes and coarse root biomass decreases when root moisture content range expands; (2) the correlation between root fresh weight and GPR indexes is stronger than that between root biomass and GPR indexes; and (3) the biomass of root with lower moisture content is underestimated and that of root with higher moisture content is overestimated by the current GPR-based root biomass quantification method. Then, a new GPR-based coarse root biomass estimation approach is proposed which takes into account root moisture content.