Quantitative, non-destructive estimates of forest coarse root biomass using 3-D ground-penetrating radar (GPR)

Abstract

<p>We evaluated 3-D imaging of coarse root structure and biomass using ground-penetrating radar (GPR). GPR surveys were conducted in a white pine forest in southern Ontario, Canada. GPR profiles were obtained across two test plots (6 and 17 m² area), using 1-GHz GPR and a MEMS (micro-electro-mechanical systems) accelerometer. Test plot surveys evaluated the effects of micro-topography, soil moisture content, and root diameter and spacing. In addition, with the aid of the outcome of the control test plots two other plots (25 and 400 m²area) were surveyed with varying line sample spacing to investigate the restraints on resolution brought about by line sampling density.</p> <p>Accounting for antenna tilt is necessary to determine an accurate and more precise position of root mass. The antenna tilt was >45^o pitch, >28^o roll and up to 10^o yaw due to surface micro-topography of the forest floor. Vector 3-D imaging enhanced the diffraction amplitude (15.5% increase) and centralized the position of the root. Radial surveys provided root continuity and produced better root imaging.</p> <p>GPR largely underestimates coarse root biomass when a line spacing of 25 cm is used. However similar results are found with smaller line spacing (12.5 cm). A maximum line spacing of 10 cm provided continuous root structure and differentiation of roots spaced 10 cm apart and greater. A sampling line spacing of 5 cm and an inline sampling interval of 0.5 cm in low soil moisture conditions provided the detection of roots that were a minimum of 1.4 cm in diameter.</p>