Mapping of penetrometer resistance in relation to tractor traffic using multivariate geostatistics

The traffic of agricultural machines can cause soil compaction and high variability of soil structure, both along normal lines and along those parallel to the field plane. The aim of this research is to investigate the potential of geostatistical techniques for understanding and evaluating the within-field spatial variability of soil compaction, caused by the traffic of agricultural machines and/or the action of tillage implements. In July 2003 soil cone penetrometer resistance was measured in a sandy-silt Cambisol of inland Sicily, where a three-year rotation wheat (Triticum durum Desf.) - wheat - tomato (Solanum lycopersicum L.) was adopted, along three parallel 3-m long transects, from the soil surface to a depth of 0.70 m. A multivariate geostatistical approach, including exploratory analysis, variography, stochastic simulation and post-processing of simulations was applied to produce thematic maps of penetrometer resistance and probability maps exceeding a critical value, corresponding to different examples of tractor movement. Penetrometer resistance variation was erratic at the surface but showed high spatial correlation between data measured at different depths. The maps of probabilistic compaction risk showed that the soil volume, exceeding the penetrometer resistance of 2.5 MPa, critical for root growth, increased from 20% to 40% after the tractor had passed through five times.