Measurement and simulation of the effect of compaction on the pore structure and saturated hydraulic conductivity of grassland and arable soil

Measurements have been made of the effect of compaction on water retention, saturated hydraulic conductivity, and porosity of two English soils: North Wyke (NW) grassland clay topsoil and Broadbalk silty topsoil, fertilized inorganically (PKMg) or with farmyard manure (FYM). As expected, the FYM topsoil had greater porosity and greater water retention than PKMg topsoil, and the NW clay topsoil retained more water at each matric potential than the silty topsoils. Compaction had a clear effect on water retention at matric potentials wetter than -10 kPa for the PKMg and FYM soils, corresponding to voids greater than 30 mu m cylindrical diameter, whereas smaller voids appeared to be unaffected. The Pore-Cor void network model has been improved by including a Euler beta distribution to describe the sizes of the narrow interconnections, termed throats. The model revealed a change from bimodal to unimodal throat size distributions on compaction, as well as a reduction in sizes overall. It also matched the water retention curves more closely than van Genuchten fits and correctly predicted changes in saturated hydraulic conductivity better than those predicted by a prior statistical approach. However, the changes in hydraulic conductivity were masked by the stochastic variability of the model. Also, an artifact of the model, namely its inability to pack small features close together, caused incorrect increases in pore sizes on compaction. These deficiencies in the model demonstrate the need for an explicitly dual porous network model to account for the effects of compaction in soil.