Premium
In situ characterization of forest litter using ground‐penetrating radar
Author(s) -
André Frédéric,
Jonard François,
Jonard Mathieu,
Lambot Sébastien
Publication year - 2016
Publication title -
journal of geophysical research: biogeosciences
Language(s) - English
Resource type - Journals
eISSN - 2169-8961
pISSN - 2169-8953
DOI - 10.1002/2015jg002952
Subject(s) - litter , ground penetrating radar , radar , humus , environmental science , soil science , dielectric permittivity , ground truth , remote sensing , ecosystem , permittivity , atmospheric sciences , dielectric , soil water , geology , ecology , materials science , biology , computer science , telecommunications , optoelectronics , machine learning
Abstract Decomposing litter accumulated on the soil surface in forests plays a major role in several ecosystem processes; its detailed characterization is therefore essential for thorough understanding of ecosystem functioning. In addition, litter is known to affect remote sensing radar data over forested areas and their proper processing requires accurate quantification of litter scattering properties. In the present study, ultrawideband (0.8–2.2 GHz) ground‐penetrating radar (GPR) data were collected in situ for a wide range of litter types to investigate the potential of the technique to reconstruct litter horizons in undisturbed natural conditions. Radar data were processed resorting to full‐wave inversion. Good agreement was generally found between estimated and measured litter layer thicknesses, with root‐mean‐square error values around 1 cm for recently fallen litter (OL layer) and around 2 cm for fragmented litter in partial decomposition (OF layer) and total litter (OL + OF). Nevertheless, significant correlations between estimated and measured thicknesses were found for total litter only. Inaccuracies in the reconstruction of the individual litter horizons were mainly attributed to weak dielectric contrasts amongst litter layers, with absolute differences in relative dielectric permittivity values often lower than 2 between humus horizons, and to uncertainties in the ground truth values. Radar signal inversions also provided reliable estimates of litter electromagnetic properties, with average relative dielectric permittivity values around 2.9 and 6.3 for OL and OF litters, respectively. These results are encouraging for the use of GPR for noninvasive characterization and mapping of forest litter. Perspectives for the application of the technique in biogeosciences are discussed.