Challenges in the development of analytical soil compaction models

Soil compaction can cause a number of environmental and agronomic problems (e.g. flooding, erosion, leaching of agrochemicals to recipient waters, emission of greenhouse gases to the atmosphere, crop yield losses), resulting in significant economic damage to society and agriculture. Strategies and recommendations for the prevention of soil compaction often rely on simulation models. This paper highlights some issues that need further consideration in order to improve soil compaction modelling, with the focus on analytical models. We discuss the different issues based on comparisons between experimental data and model simulations. The upper model boundary condition (i.e. contact area and stresses at the tyre–soil interface) is highly influential in stress propagation, but knowledge on the effects of loading and soil conditions on the upper model boundary condition is inadequate. The accuracy of stress transducers and therefore of stress measurements is not well known, despite numerous studies on stress in the soil profile below agricultural tyres. Although arable soils are characterised by distinct soil layers with different mechanical properties, analytical models rely on a one-layer approach with regard to stress propagation, an anomaly that needs further attention. We found large differences between soil stress–strain behaviour obtained from in situ measurements during wheeling experiments and those measured on cylindrical soil samples in standard laboratory tests. We concluded that the main reason was differences in loading time, and suggest that future research should concentrate on in situ stress–strain behaviour during short time, dynamic loading.

Research highlights▶ Accurate estimation of the upper model boundary condition is crucial. ▶ Accuracy of stress transducers and therefore stress measurements is not well known. ▶ One-layer analytical approach for stress propagation may prevent accurate estimation of soil stress. ▶ In situ soil stress-strain behaviour significantly differs from stress-strain behaviour measured on soil samples in laboratory tests.