Forward simulation of root's ground penetrating radar signal: Simulator development and validation

Background and aims: It remains unclear how the limiting factors (e.g., root size, root water content, spacing between roots, and soil water content) affect root investigation using ground penetrating radar (GPR). The objective of this study is to develop a theoretical forward simulation protocol of synthesizing root's GPR signal and test the feasibility of our proposed simulation protocol in evaluating the impacts of limiting factors on GPR-based root detection and quantification. Methods: The proposed forward simulation protocol was developed by integrating several existing numerical models, such as the Root Composition Model, the Root Dielectric Constant Model, the Root Electrical Conductivity Model, the Soil Dielectric Constant Model, the Soil Electrical Conductivity Model, and a newly-established model (Root Length-Biomass Model). Resolution and GPR index obtained from both field collected radargrams and corresponding simulations were compared to validate the accuracy of simulation. Results: Simulated radargrams exhibit similar resolution with that of the in situ collected. The same trends of root radar signals against different levels of root size, root water content, interval between roots, root depth, and antenna frequency were observed on both in situ radargrams and simulated radargrams. Strong correlations (correlation coefficients ranging from 0.87 to 0.96) were found between GPR indices extracted from the simulated data and those from the field collected data. Conclusions: Our proposed forward simulation is effective for assessing the impacts of limiting factors on root detection and quantification using GPR. This forward simulation protocol can be used to provide guidance for in situ GPR root investigation and can predict the accuracy of GPR-based root quantification under site-specific conditions.