Prediction of nitrogen, active carbon, and organic carbon-to-clay ratio in agricultural soils by in-situ spectroscopy

Visible and near-infrared (vis-NIR) spectroscopy is a promising technology for the analysis of different soil quality parameters. In this study, we used in-situ vis-NIR spectroscopy in association with partial least squares regression to predict the total and the mineral (nitrate + ammonium) nitrogen content, the permanganate oxidizable carbon (POXC), as well as the ratio of soil organic carbon-to-clay content in different agricultural soils in Switzerland. These parameters can indeed be used as indicators of soil quality in response to agronomic practices. To this goal, a total number of 134 soil samples were used for carbon-, total nitrogen- and clay-related parameters, whereas 69 soil samples were used for the mineral nitrogen-related parameters. We found that the partial least squares regression model can successfully predict the total nitrogen and the POXC content as well as the ratio of soil organic carbon-to-clay content (ratio of performance to interquartile range, RPIQ > 2.62, R-2 > 0.73, Lin's concordance correlation coefficient > 0.83). As concerns the mineral nitrogen, it was not possible to successfully predict this parameter by vis-NIR spectroscopy. By demonstrating the possibility to reliably predict POXC content and the soil organic carbon-to-clay ratio, we show that vis-NIR can be also used to analyse soil parameters associated with both the quality of organic carbon and the structural quality of agricultural soils.