Accuracy of infrared (IR) models to measure soil particle-size distribution (PSD) depends on soil preparation, methodology (sedimentation, laser), settling times and relevant soil features. Compositional soil data may require log ratio (ilr) transformation to avoid numerical biases. Machine learning can relate numerous independent variables that may impact on NIR spectra to assess particle-size distribution. Our objective was to reach high IRS prediction accuracy across a large range of PSD methods and soil properties. A total of 1298 soil samples from eastern Canada were IR-scanned. Spectra were processed by Stochastic Gradient Boosting (SGB) to predict sand, silt, clay and carbon. Slope and intercept of the log-log relationships between settling time and suspension density function (SDF) (R2 = 0.84–0.92) performed similarly to NIR spectra using either ilr-transformed (R2 = 0.81–0.93) or raw percentages (R2 = 0.76–0.94). Settling times of 0.67-min and 2-h were the most accurate for NIR predictions (R2 = 0.49–0.79). The NIR prediction of sand sieving method (R2 = 0.66) was more accurate than sedimentation method(R2 = 0.53). The NIR 2X gain was less accurate (R2 = 0.69–0.92) than 4X (R2 = 0.87–0.95). The MIR (R2 = 0.45–0.80) performed better than NIR (R2 = 0.40–0.71) spectra. Adding soil carbon, reconstituted bulk density, pH, red-green-blue color, oxalate and Mehlich3 extracts returned R2 value of 0.86–0.91 for texture prediction. In addition to slope and intercept of the SDF, 4X gain, method and pre-treatment classes, soil carbon and color appeared to be promising features for routine SGB-processed NIR particle-size analysis. Machine learning methods support cost-effective soil texture NIR analysis.
Determining soil particle-size distribution from infrared spectra using machine learning predictions: Methodology and modeling
