Patrick Filippi*1, Edward J. Jones1, Bradley J. Ginns1, Brett M. Whelan1, Guy W. Roth1, Thomas F.A. Bishop1
1 The University of Sydney, Sydney Institute of Agriculture, Sydney, New South Wales, Australia, email@example.com
Subsoil alkalinity is a common issue in the alluvial cotton-growing valleys of northern NSW, Australia. This causes nutrient deficiencies, toxicity, and inhibits root growth, which can have a damaging impact on crops. The depth at which a soil constraint is reached is important information for farmers, however, this is hard to measure spatially. This study predicted the depth in which a pH constraint (pH > 9) was reached to a 1 cm vertical resolution for a 1 m soil profile on a dryland cropping farm in northern NSW, Australia. Equal-area quadratic smoothing splines were used to resample vertical soil profile data, and a random forest model was used to produce the depth-to-pH-constraint map. The model to spatially predict soil pH across the farm was accurate, with an LCCC of 0.63, and an RMSE of 0.47 when testing with leave-one-site-out-cross-validation. About 77% of the area was constrained by a pH greater than 9 within the top 1 m of soil. The relationship between the predicted depth-to-pH-constraint map and cotton and grain (wheat, canola, and chickpea) yield monitor data was analysed for individual fields. The deeper in the soil profile a pH constraint was reached, the greater the crop yield. A strong relationship was found for wheat, canola, and chickpea (Spearman’s correlation (rs) of 0.75, 0.66, and 0.58, respectively), and a moderate relationship for cotton (rs = 0.37). The modelling approach presented could be used to identify the depth to other soil constraints, such as soil sodicity. The outputs are a promising opportunity to understand crop yield variability, which could lead to improvements in management practices.