Effects of the stress field induced by a running tyre on the soil pore system

This study investigated the impact of vehicle traffic on soil physical properties by systematically collecting samples in a transect running from the centreline to the outside of the wheel rut in a wheeling experiment conducted on a clay loam soil at Suberg near Bern, Switzerland, in 2010. Four repeated wheelings were performed by a forage harvester (wheel load 6100 kg; tyre width 80 cm). Mean normal and horizontal stresses were measured with Bolling probes (at 10, 20 and 40 cm depth) and load cells (at 40, 50, 60 cm lateral distance from the centreline of the wheel rut at 10, 30 and 50 cm depth), respectively. Intact soil cores of 100 cm3 sampled at 10, 30 and 50 cm depth in a soil transcet running from the centreline of the wheel rut to the unwheeled part of the field were used for measurements of water retention and air permeability (ka) at −30, −100 and −300 hPa matric potential. The complete stress state in the soil profile beneath the harvester tyre was calculated using the SoilFlex model. Pore continuity index (N) and blocked air-filled porosity (εb) were estimated from the relationship between ka and air-filled porosity (εa) for a range of matric potentials. Calculated and measured stresses agreed well at all depths. At −100 hPa, εa was consistently lower under the centreline of the wheel rut than at the lateral edge of the rut or outside the wheel rut, while ka was lowest at the lateral edge of the wheel rut and highest outside the wheel rut, with intermediate values under the centreline of the wheel rut. Simulations of the stress field in the soil beneath the tyre indicated that the trends in ka were determined by both the mean normal stress and the shear stress, while the trend in εa was determined by the mean normal stress only. At 10 cm depth, the index of pore continuity (N) supported the interpretation that soil pores under the centreline of the wheel rut are primarily reduced in size, while pore continuity is highly affected at the lateral edge of the wheel rut, as indicated by a higher value of εb than in other locations. These results indicate that sampling along the wheel track transect can provide better information about traffic-induced changes on soil physical properties than random sampling in lateral locations relative to the centreline of the wheel rut.