Evaluating the Impact of Water Content on Belowground Biomass Quantification Using Ground Penetrating Radar

Abstract:

One of the main objectives of the natural sciences is to describe the composition of the biosphere. A major difficulty in determining the amount of subsurface biomass in a system is to avoid destructive and laborious methods commonly used. Since the last decade, different authors have shown the feasibility and assessment of applying ground-penetrating radar (GPR) in root detection and imaging, a geophysical non-invasive and non-destructive method used to detect changes in physical properties on the subsurface. However, the relationship between the detection capacity and the relative dielectric permittivity εr of the roots has been minor explored. In this study, we combine direct measurements of the electrical properties of Eucalyptus grandis trunk samples with the responses in radargrams. We used a 2.6 GHz antenna for acquisitions in a controlled environment, a 2 x 2 x 1 meters sand-box. After acquiring and processing 2D profiles, we extracted indexes obtained from the generated radargrams, such as the amplitude of the reflected wave and an estimated diameter. Correlations between these indexes show challenges in calculating root dimensions. Wet roots in dry soils present the best results in GPR imaging, due to the difference between relative dielectric permittivity of the soil and the target. Dry samples are detectable only after applying artificial gains to data, but could easily be confused with other objects as small rocks. Many factors affect the accuracy of GPR root detection: target diameter, depth, moisture, soil conditions and antenna frequency. In this study, we give new insights for future development of site-specific methods based on εr variation.