Stochastic simulation of soil strength/compaction and assessment of corn yield risk using threshold probability patterns

Soil compaction can suppress crop yields. Thus, compaction indices such as cone penetration resistance (PR) can be useful to farmers and researchers i nterested in identifying areas in the field where soil strength/compaction, or governing soil properties reflected by PR measurements, might be proble matic to crop yield. The general aim of this paper was to apply stochastic simulation to spatially delimit critically important soil strength patterns , the spatial disposition and uncertainty of those patterns, and to determi ne if broad strength/compaction spatial units defined via PR threshold prob ability patterns could be used to reflect corn yield variability under diff erent rotation, tillage, and trafficking management treatments. Shallow (3- 15 cm depth) linear PR patterns were strongly anisotropic while deeper patt erns (25-40 cm depth) were more regionalized in character. Stochastic simul ation produced very strong global patterns of low ( < 2 MPa) and high ( > 2 MPa) PR values. Broad spatial units defined on the basis of these patterns were used to generate corn yield risk maps. It was found that yields, unde r the various treatments, reflected the risk patterns in an expected way 75 % of the time over 5 years of observation. That is, under similar managemen t treatments, lower risk areas (areas where there was lower probability of soil strength exceeding 2 MPa in the top 3 to 15 cm of the soil) produced h igher yields than higher risk areas (areas where there was higher probabili ty of soil strength exceeding 2 MPa in the top 3 to 15 cm of soil). (C) 200 1 Elsevier Science B.V. All rights reserved.