Evaluation of Spectral Pretreatments, Partial Least Squares, Least Squares Support Vector Machines and Locally Weighted Regression for Quantitative Spectroscopic Analysis of Soils

Abstract

Soil testing requires the analysis of large numbers of samples in the laboratory that is often time consuming and expensive. Mid-infrared spectroscopy (mid-IR) and near infrared (NIR) spectroscopy are fast, nondestructive and inexpensive analytical methods that have been used for soil analysis, in the laboratory and in the field, to reduce the need for measurements using complex chemical/physical analyses. A comparison of the use of spectral pretreatment as well as the implementation of linear and non-linear regression methods was performed. This study presents an overview of the use of infrared spectroscopy for the prediction of five physical (sand, silt and clay) and chemical (total carbon and total nitrogen) soil parameters with near and mid-infrared units in bench top and field setups. Even though no significant differences existed among pretreatment methods, models using second derivatives performed better. The implementation of partial least squares (PLS), least squares support vector machines (LS-SVM) and locally weighted regression (LWR) for the development of the calibration models showed that the LS-SVM did not out-perform linear methods for most components while LWR that creates simpler models performed well. The present results tend to show that soil models are quite sensitive to the complexity of the model. The ability of LWR to select only the appropriate samples did help in the development of robust models. Results also proved that field units performed as well as bench-top instruments. This was true for both near infrared and mid-infrared technology. Finally, analysis of field moist samples was not as satisfactory as using dried-ground samples regardless of the chemometrics methods applied.

Long‐term organic carbon sequestration in tidal marsh sediments is dominated by old‐aged allochthonous inputs in a macrotidal estuary

“…Before the OC concentration was estimated using partial least squares regression (PLSR) with 4 latent variables, the soil spectra were preprocessed using the standard normal variate procedure (SNV, see e.g. Igne et al., ; Viscarra Rossel, McBratney, & Minasny, ). Total SOC stocks were calculated down to a depth of 0.6 m for each site by combining depth profiles of OC concentration and soil bulk density.…”

“…Before the OC concentration was estimated using partial least squares regression (PLSR) with 4 latent variables, the soil spectra were preprocessed using the standard normal variate procedure (SNV, see e.g. Igne et al, 2010;Viscarra Rossel, McBratney, & Minasny, 2010 | 2501 depth at every sampling location, with the exception of the low marshes at salinity 27 and 30.…”

Major Issues of Diffuse Reflectance NIR Spectroscopy in the Specific Context of Soil Carbon Content Estimation

“…Despite these obstacles, local methods generally give good results (Igne et al (2010) and Gogé et al (2012) with LWR andFernández Pierna andDardenne (2008) with LOCAL). In their study, Gogé et al (2012) applied the LWR method on 2500 soil samples representative of the entire French territory.…”

“…Once a tree has been grown and possibly pruned, it will have some nonpartitioned nodes called terminal nodes. Computational times are huge (Igne et al, 2010;Vohland et al, 2011) and this method has not yet much applied on soil data. The CART (Breiman, 1984) underpin these methods, which differ in the way of calculating the final output.…”

“…Furthermore, selection of the best preprocessing method is generally based on the model performance and not in order to optimize the sample subset selection for local methods (Igne et al, 2010). Sample preparation before spectral measurement is a way to control the effects of the physical influence factors such as moisture, particle size, and bulk density .…”