Elsevier

Pedosphere

Volume 32, Issue 4, August 2022, Pages 602-615
Pedosphere

Using pXRF and vis-NIR spectra for predicting properties of soils developed in loess

https://doi.org/10.1016/S1002-0160(21)60092-9Get rights and content

ABSTRACT

Visible near-infrared (vis-NIR) and portable X-ray fluorescence (pXRF) spectrometers have been increasingly utilized for predicting soil properties worldwide. However, only a few studies have focused on splitting the predictive models by horizons to evaluate prediction performance and systematically compare prediction performance for A, B, and combined A+B horizons. Therefore, we investigated the performance of pXRF and vis-NIR spectra, as individual or combined, for predicting the clay, silt, sand, total carbon (TC), and pH of soils developed in loess, and compared their prediction performance for A, B, and A+B horizons. Soil samples (176 in A horizon and 172 in B horizon) were taken from Mollisols and Alfisols in 136 pedons in Wisconsin, USA and analyzed for clay, silt, sand, pH, and TC. The pXRF and vis-NIR spectrometers were used to measure the pXRF and vis-NIR soil spectra. Data were separated into calibration (n = 244, 70%) and validation (n = 104, 30%) datasets. The Savitzky-Golay filter was applied to preprocess the pXRF and vis-NIR spectra, and the first 10 principal components (PCs) were selected through principal component analysis (PCA). Five types of predictor, i.e., PCs from vis-NIR spectra, pXRF of beams at 0–40 and 0–10 keV (XRF40 and XRF10, respectively) spectra, combined XRF40 and XRF10 (XRF40+XRF10) spectra, and combined XRF40, XRF10, and vis-NIR (XRF40+XRF10+vis-NIR) spectra, were compared for predicting soil properties using a machine learning algorithm (Cubist model). A multiple linear regression (MLR) model was applied to predict clay, silt, sand, pH, and TC using pXRF elements. The results suggested that pXRF spectra had better prediction performance for clay, silt, and sand, whereas vis-NIR spectra produced better TC and pH predictions. The best prediction performance for sand (R2= 0.97), silt (R2= 0.95), and clay (R2= 0.84) was achieved using vis-NIR+XRF40+XRF10 spectra in B horizon, whereas the best prediction performance for TC (R2= 0.93) and pH (R2= 0.79) was achieved using vis-NIR+XRF40+XRF10 spectra in A+B horizon. For all soil properties, the best MLR model had a lower prediction accuracy than the Cubist model. It was concluded that pXRF and vis-NIR spectra can be successfully applied for predicting clay, silt, sand, pH, and TC with high accuracy for soils developed in loess, and that spectral models should be developed for different horizons to achieve high prediction accuracy.

Section snippets

INTRODUCTION

Proximal soil sensing techniques have been widely used to provide information on soils. Visible near-infrared (vis-NIR) and portable X-ray fluorescence (pXRF) spectra have been used as rapid, environment-friendly, and low-cost techniques to measure or predict soil properties with high accuracy, and in a non-destructive way compared to traditional chemical methods (Taylor et al., 2004; Waiser et al., 2007; Rossel et al., 2010, 2011; Rouillon and Taylor, 2016). Vis-NIR and pXRF spectrometers have

Soil sampling and analysis

The study was conducted in the Driftless area, Dane County, Wisconsin, USA. The study area was approximately 7 000 ha and was mainly agricultural land, with maize, soybean, and alfalfa as the most prevalent field crops. The climate in this region is characterized by warm, moist summer and cold, dry winter. The mean annual precipitation and mean annual temperature are 857 mm and 7.3 °C, respectively. Loess materials cover much of the landscape in the study area. The soils were described and

Characterization of soil horizons

The distributions of soil particle size fractions (clay, silt, and sand), TC, pH, and elements of the 348 soil samples are shown in Fig. 2 for A, B, and A+B horizons. The A horizon had an average of 12% sand, 66% silt, and 22% clay. The Bt horizon had an average of 15% sand, 58% silt, and 27% clay. In the soil profiles, clay, Mg, Al, Fe, and Sr increased from A to Bt horizon, whereas TC, TN, pH, Ca, Mn, and Zn noticeably decreased from A to Bt horizon. The soil samples had higher concentrations

Model performance using pXRF and vis-NIR spectra for prediction of soil properties

In our study, good results were obtained for predicting clay, silt, sand, pH, and TC from vis-NIR and pXRF spectra. The PCA+Cubist method predicted clay, silt, and sand from vis-NIR spectra with good accuracy (validation R2 = 0.84, 0.60, and 0.71, respectively). Clay and sand were more accurately predicted than silt, which may be due to the special characteristics of clay (O–H in water, FeOH–, AlOH–, and MgOH– in the structure of mineral crystal lattice) and corresponding spectral features in

CONCLUSIONS

The Cubist model was used with individual or combined pXRF and vis-NIR spectra through PCA to investigate its performance in predicting clay, silt, sand, TC, and pH of soils developed in loess. Its prediction performance was evaluated for data from A, B, and A+B horizons. We found that the models using individual pXRF spectra produced better predictions for clay, silt, and sand, whereas those using individual vis-NIR spectra performed better in TC and pH predictions. Combining vis-NIR and pXRF

ACKNOWLEDGEMENTS

The authors are grateful to Mr. Dave Evans, who sampled these pedons for his master project. The first author is supported by the Scientific Research Projects (BAP) (No. 2019-2757) of Eskisehir Osmangazi University for postdoc research at the Department of Soil Science, University of Wisconsin-Madison.

SUPPLEMENTARY MATERIAL

Supplementary material for this article can be found in the online version.

References (54)

  • A Rawal et al.

    Determination of base saturation percentage in agricultural soils via portable X-ray fluorescence spectrometer

    Geoderma

    (2019)
  • M Rouillon et al.

    Can field portable X-ray fluorescence (pXRF) produce high quality data for application in environmental contamination research?

    Environ Pollut

    (2016)
  • A Sharma et al.

    Characterizing soils via portable X-ray fluorescence spectrometer: 3. Soil reaction (pH)

    Geoderma

    (2014)
  • A Sharma et al.

    Characterizing soils via portable X-ray fluorescence spectrometer: 4. Cation exchange capacity (CEC)

    Geoderma

    (2015)
  • U Stockmann et al.

    Utilizing portable X-ray fluorescence spectrometry for in-field investigation of pedogenesis

    Catena

    (2016)
  • T Tóth et al.

    XRF-measured rubidium concentration is the best predictor variable for estimating the soil clay content and salinity of semi-humid soils in two catenas

    Geoderma

    (2019)
  • D D Wang et al.

    Synthesized use of VisNIR DRS and PXRF for soil characterization: Total carbon and total nitrogen

    Geoderma

    (2015)
  • D C Weindorf et al.

    Advances in portable X-ray fluorescence (PXRF) for environmental, pedological, and agronomic applications

    Adv Agron

    (2014)
  • D Y Xu et al.

    X-ray fluorescence and visible near infrared sensor fusion for predicting soil chromium content

    Geoderma

    (2019)
  • Y K Zhang et al.

    Soil horizon delineation using vis-NIR and pXRF data

    Catena

    (2019)
  • Y D Zhu et al.

    Characterizing soils using a portable X-ray fluorescence spectrometer: 1. Soil texture

    Geoderma

    (2011)
  • R Zornoza et al.

    Near infrared spectroscopy for determination of various physical, chemical and biochemical properties in Mediterranean soils

    Soil Biol Biochem

    (2008)
  • R Andrade et al.

    Prediction of soil fertility via portable X-ray fluorescence (pXRF) spectrometry and soil texture in the Brazilian Coastal Plains

    Geoderma

    (2020)
  • R Andrade et al.

    Assessing models for prediction of some soil chemical properties from portable X-ray fluorescence (pXRF) spectrometry data in Brazilian Coastal Plains

    Geoderma

    (2020)
  • R J Barnes et al.

    Standard normal variate transformation and de-trending of near-infrared diffuse reflectance spectra

    Appl Spectrosc

    (1989)
  • E Ben-Dor et al.

    Near-infrared analysis as a rapid method to simultaneously evaluate several soil properties

    Soil Sci Soc Am J

    (1995)
  • L Benedet et al.

    Soil subgroup prediction via portable X-ray fluorescence and visible near-infrared spectroscopy

    Geoderma

    (2020)
  • Cited by (9)

    • Assessing the effect of soil to water ratios and sampling strategies on the prediction of EC and pH using pXRF and Vis-NIR spectra

      2022, Computers and Electronics in Agriculture
      Citation Excerpt :

      In addition, limited studies have explored sampling strategies' effect on prediction accuracy (Zhang and Hartemink, 2017; 2021; Chatterjee et al., 2021; Gozukara et al., 2022c) of soil properties and none of the studies to date have focused on the effect of soil to water ratios on prediction accuracy of EC and pH using individual or combined pXRF and Vis-NIR spectra. However, some studies have focused on the effect of individual or combined pXRF and Vis-NIR spectra on prediction performance of EC and pH (Weindorf et al., 2016; O’Rourke et al., 2016a; Wan et al., 2019; Zhang and Hartemink, 2020; Gozukara et al., 2022a-b-c), whereas some studies have focused on vertical and horizontal variation across the soil profile wall (Adhikari et al., 2016; Grauer-Gray and Hartemink, 2018; Mancini et al., 2021; Gozukara et al., 2022b), sampling strategies (Zhang and Hartemink, 2017), and different machine learning algorithms (Yang et al., 2019) on the prediction performance of EC and pH. Although each of these factors can significantly affect prediction performances, no studies have evaluated these factors together. In particular, there is an increasing need for quantitative soil data and information using spatial analysis, which has been utilized by many studies compared to traditional soil sampling based on high labor and extensive soil sampling.

    • Soil particle size prediction using Vis-NIR and pXRF spectra in a semiarid agricultural ecosystem in Central Anatolia of Türkiye

      2022, Catena
      Citation Excerpt :

      It is well known that soil surface often contains higher organic materials than subsurface (Jiang et al., 2017; Yost et al., 2019; Gozukara et al., 2021b, 2022a) and soil mineralogy varies along with the soil profile compared to the soil surface (Yost et al., 2019; Saricaoğlu et al., 2021; Kowalska et al., 2022). These differences between topsoils and subsoils layers affect the prediction accuracy for sand, silt, and clay content (Silva et al., 2020; Gozukara et al., 2022a). However, no studies have been conducted in which differences such as individual or combined soil spectra in the soil surface, soil profile, and soil surface and profile wall datasets for prediction of sand, silt, and clay distribution in semiarid regions.

    View all citing articles on Scopus
    View full text