Visible Near-Infrared Spectroscopy and Pedotransfer Function Well Predict Soil Sorption Coefficient of Glyphosate

Abstract

The soil sorption coefficient (Kd) of glyphosate mainly controls its transport and fate in the environment. Laboratory-based analysis of Kd is laborious and expensive. This study aimed to test the feasibility of visible near-infrared spectroscopy (vis–NIRS) as an alternative method for glyphosate Kd estimation at a country scale and compare its accuracy against pedotransfer function (PTF). A total of 439 soils with a wide range of Kd values (37–2409 L kg−1) were collected from Denmark (DK) and southwest Greenland (GR). Two modeling scenarios were considered to predict Kd: a combined model developed on DK and GR samples and individual models developed on either DK or GR samples. Partial least squares regression (PLSR) and artificial neural network (ANN) techniques were applied to develop vis–NIRS models. Results from the best technique were validated using a prediction set and compared with PTF for each scenario. The PTFs were built with soil texture, OC, pH, Feox, and Pox. The ratio of performance to interquartile distance (RPIQ) was 1.88, 1.70, and 1.50 for the combined (ANN), DK (ANN), and GR (PLSR) validation models, respectively. vis–NIRS obtained higher predictive ability for Kd than PTFs for the combined dataset, whereas PTF resulted in slightly better estimations of Kd on the DK and GR samples. However, the differences in prediction accuracy between vis–NIRS and PTF were statistically insignificant. Considering the multiple advantages of vis–NIRS, e.g., being rapid and non-destructive, it can provide a faster and easier alternative to PTF for estimating glyphosate Kd.

Integration of Vis–NIR Spectroscopy and Machine Learning Techniques to Predict Eight Soil Parameters in Alpine Regions

“…Vis-NIR can reflect the overtones and combinations of basic molecular vibrations, such as clear responses to functional groups like C=O, N=H, and O=H. Many scholars have explored the prediction of soil parameters using Vis-NIR technology and achieved good results [20][21][22], for instance, the physical and chemical properties of soil and the composition of minerals, including TN [23], SOM [14], soil moisture [24], organic carbon [25], soil exchangeable cations [26], and the soil adsorption coefficient of glyphosate [27]. However, there are differences in Vis-NIR prediction models in different regions due to regional differences in soil types and physicochemical properties, surface cover, and climate.…”

“…However, different modeling methods and parameters have different accuracies. In general, the modeling accuracy is deep learning models > machine learning models > linear models but may vary depending on the parameters or region [27,48,49]. Deep learning often requires a large amount of training data to improve the accuracy and robustness of the model, and a small amount of data suffers from many problems, such as difficulty in convergence and overfitting.…”