O and A soil horizons’ boundaries detection using GPR under variable soil moisture conditions

Abstract

• The 1D analysis of individual traces allows distinction of soil horizons boundaries. • The mean thickness calculated from individual traces shows only 10% error. • Moisture conditions change the accuracy of the soil horizons boundaries’ detection. The thicknesses of organic (O) and organomineral (A) horizons are essential parameters for estimating the soil organic carbon stock. They are usually measured at sampling points distributed randomly or regularly over a site, but due to high spatial variability of the soil horizons' thicknesses, the sampling should be dense enough to estimate the carbon stock precisely. Dense soil sampling is cost, time, and labour demanding. Therefore, some studies suggest that geophysical methods such as ground-penetrating radar (GPR) can assist with a more precise estimation of the organic and organomineral horizons thicknesses without digging soil pits. This study evaluates the accuracy of the organic and the organomineral horizons thicknesses repeatedly measured under different soil moisture conditions on two contrasting soil types: Dystric Cambisol and Arenic Podzosol, using GPR with 800 MHz antenna. The results proved this method to be promising; however, we could not distinguish the boundary between organic and organomineral horizons but only the O + A horizon/subsoil boundary. The thickness of O + A horizons was estimated with an error between 25 and 35% in the Dystric Cambisol site and 18 – 24% in the Arenic Podzol site. The results were more accurate under moister conditions for both soil types, but under drier conditions, deeper parts of irregular horizon boundaries were better distinguishable.