Elsevier

Geoderma

Volume 271, 1 June 2016, Pages 225-233
Geoderma

Comparison of infrared spectroscopy and laser granulometry as alternative methods to estimate soil aggregate stability in Mediterranean badlands

https://doi.org/10.1016/j.geoderma.2016.02.025Get rights and content

Highlights

  • IR spectra outperformed LG-data to predict aggregate stability values and classes.

  • NIR-DR and MIR-ATR data provided best prediction (RPD = 3.55 & 2.88; R2 = 0.93 & 0.90).

  • Spectral prediction relied on a tight link between SOC and aggregate stability.

  • The combination of IR and LG data did not yield a better prediction model.

Abstract

Soil aggregate stability is a key indicator of soil resistance to erosion, but its measurement remains fastidious for large scale uses. Alternative time and cost-effective methods are thus needed. Our objective was to assess and compare the efficiency of laser granulometry (LG) and soil mid- and near-infrared spectroscopy (MIR/NIR) as alternative methods to assess soil aggregate stability in Mediterranean badland soils. A collection of 75 badland soil samples was used, showing wide variations in soil aggregate stability. Three different categories of measurements were performed: (i) the aggregate breakdown kinetics of the [< 1 mm] size fraction under stirring and sonication, tracked by repeated particle size distribution measurements, using LG, (ii) mid-(diffuse-MIR-DR and attenuate transmitted reflectance — MIR-ATR) and near-(NIR-DR) infrared spectra of the fine soil fraction [< 2 mm] and (iii) the soil aggregate [3–5 mm] stability, using the standardized method (ISO/FDIS 10930, 2012). Partial least squares regression models were used to predict soil aggregate stability using LG data and infrared spectra. Results showed that NIR-DR and MIR-ATR data provided the best prediction model for soil aggregate stability values (RPD = 2.61 & 2.74; R2 = 0.85 & 0.87), followed by MIR-DR data (RPD = 2.24; R2 = 0.89) and finally LG data (RPD = 2.12; R2 = 0.80). For a quantitative use of the models to assign soil samples to standardized soil aggregate stability classes (ISO/FDIS 10930, 2012), infrared spectra also provided the best accuracy, with a misclassification rate below 30% for NIR-DR and MIR-ATR models, while it reached 43% with the LG-based model. The combination of IR and LG data did not yield a better prediction model for soil aggregate stability values and classes. Infrared-based method also provided best results in terms of time-saving strategy, reducing the measurement time to 8 min only. To conclude, infrared spectra (NIR-DR and MIR-ATR) outperformed LG-data to predict soil aggregate stability. Further development of this technique would require calibrating a set of soil-type specific prediction models for a wide range of soil types.

Introduction

Soil aggregation is a key process in soil structuration, influencing a wide range of soil properties, such as water holding capacity, water availability to plant roots and organic carbon sequestration (Bronick and Lal, 2005, Zhao et al., 2007). The stability of soil aggregates, referring to their resistance to breakdown under disruptive forces, is also a good indicator of soil resistance to erosion (Barthès and Roose, 2002). Surface erosion by water is important in Mediterranean regions (García-Ruiz et al., 2013), exhibiting frequent intense rainfall events (Borga et al., 2014) and the occurrence of badland areas. Badlands refer to patches of the landscape characterized by a dense network of ephemeral ravines incising soft erodible rocks with poorly cohesive soils and minimal vegetation (Gallart et al., 2013). In such context, soil aggregate stability offers a possibility to assess soil erodibility without running expensive and time-consuming field runoff experiments (Le Bissonnais, 1996).

Several methods have been developed to measure soil aggregate stability (Chepil, 1952, Kemper and Rosenau, 1986, Le Bissonnais, 1996, Amezketa, 1999). To allow comparison between studies, a standardized protocol, composed by three disruptive tests accounting for slaking, mechanical breakdown and physico-chemical dispersion was recently proposed (ISO/FDIS, 10930, 2012), from the work of Le Bissonnais (1996) and Le Bissonnais and Arrouays (1997). However this standardized method remains time-consuming. Labor-intensive steps of this method mainly relate to the numerous manual steps, including wetting and several sieving steps (Rawlins et al., 2013). In soils containing more than 10% of coarse soil particles, such as gravels or marly platelets, an additional step of manual removal of stones further increases the labor investment (Erktan et al., 2016). This actually limits the deployment of this method to a high number of field sites, while there is a growing interest in quantifying the variation of aggregate stability and related soil properties both in time (Abiven et al., 2009, Algayer et al., 2014) and space (Rawlins et al., 2015).

To reduce time investment in soil aggregate stability measurements, two main alternative strategies have been explored: the use of (i) laser granulometry (e.g. Bieganowski and Witkowska, 2010) and (ii) visible, near and mid infrared spectroscopies (e.g. Chang et al., 2001, Madari et al., 2006, Gomez et al., 2013).

Laser granulometry (LG) offers the possibility to run quick, easy and reproducible measurements of soil effective particle size distribution (PSD) of a soil sample in circulating water (Arriaga et al., 2006, Ryzak and Bieganowski, 2011, Sochan et al., 2012). Several alternative methods for time-effective soil aggregate stability measurements were developed using this technique (Beuselinck et al., 1999, Bieganowski and Witkowska, 2010, Schomakers et al., 2011, Grangeon et al., 2012, Grangeon et al., 2014). The most recent developments of the LG-based technique for rapid and reproducible soil aggregate stability assessment were conducted by Rawlins et al., 2013, Rawlins et al., 2015. Running continuous PSD measurements, they tracked mean weight diameter (MWD) variations of soil aggregate samples (1–2 mm), progressively disintegrating in circulating water due to hydrodynamic forces and sonication. Soil aggregate stability was assessed by the difference in the continuous size distribution of water stable aggregates and the disaggregated material. The main advantages of the LG-based technique for soil aggregate assessment rely on (i) accounting for a wide range of soil fractions, (ii) tracking soil disintegration by continuous measurements, (iii) limiting the bias related to the operator, and (iv) being less labor-intensive.

Visible, near and mid infrared (Vis, NIR and MIR) spectroscopies have proved to be effective alternative methods to conventional laboratory analyses for assessing several soil properties (Viscarra Rossel et al., 2006), especially biochemical characteristics (O’Rourke and Holden, 2012, Cécillon et al., 2012, Yang et al., 2012). By contrast, studies investigating the prediction of soil structural characteristics are still scant and contradictory (Cécillon et al., 2009, Askari et al., 2015). Nevertheless, recent studies found promising, but still relatively inaccurate results for the prediction of soil aggregate stability using infrared spectroscopy (Madari et al., 2006, Cañasveras et al., 2010, Gomez et al., 2013). Soil aggregate stability is influenced by several soil components, such as soil carbon concentration, calcium carbonates, clays and iron oxydes, which are all well predicted by infrared spectroscopy (Malley et al., 2004, O’Rourke and Holden, 2012, Gomez et al., 2013) and may explain the high potential of this technique to predict soil aggregate stability accurately.

Despite increasing effort in the parallel development of LG and infrared spectroscopy as time and cost-effective methods for determining soil aggregate stability, no study attempted to compare directly these two methods.

The objective of this study was to assess and compare the efficiency of LG and soil infrared spectroscopy (NIR/MIR) as alternative methods to assess soil macro-aggregate stability in Mediterranean badland soils. It was hypothesized that (i) NIR and MIR spectra are suitable for prediction of soil macro-aggregate stability due to their ability to detect key drivers of soil aggregate stabilization, such as soil organic carbon (SOC), clays and carbonates, and that (ii) the disintegration of the [< 1 mm] soil aggregates under stirring and sonication within the LG is correlated to the disintegration of the [3–5 mm] aggregates after rapid immersion in water, constituting the standardized method of soil macro-aggregate stability measurement.

Section snippets

Study site and soil sampling

The soils used in this study originate from Mediterranean badlands in the Southern French Alps, near Digne-les-Bains at the Draix-Bléone Environmental Research Observatory (http://oredraixbleone.irstea.fr/; 44°08′N, 6°20′E). The site is characterized by a Mountainous and Mediterranean climate (Vallauri, 1997), with an annual mean air temperature of 10.3 °C and an annual mean precipitation of 900 mm (Mathys, 2006). Precipitation regime is heterogeneous, with most rainfall concentrating in May and

Results and discussion

Overall, the 75 soil samples collected covered the entire range of soil stability classes, as defined by the international norm ISO/FDIS 10930 (2012), with (a) very unstable class for 8 samples, (b) unstable class for 14 samples, (c) moderately stable class for 15 samples, (d) stable class for 13 samples and (e) very stable class for 15 samples (Table 1; Erktan et al., 2016).

Conclusions

Among the five datasets used for prediction, NIR-DR and MIR-ATR spectral data offered the best prediction model for soil aggregate stability values (RPD = 2.61 and 2.74; R2 = 0.85 and 0.87; misclassification rates = 22% and 28%), and also the best prediction of soil aggregate stability classes. LG-data also provided a satisfying prediction model (RPD = 2.12; R2 = 0.80), but error prediction still needs to be reduced for a routine use. The combination of IR and LG data did not yield satisfying predictive

Acknowledgments

We thank Electricité de France (EDF), Agence de l'eau Rhône, Méditerranée et Corse, Région Provence-Alpes-Côte-d'Azur and European Union (FEDER program “L'Europe s'engage en PACA avec le Fonds Européen de Développement Régional”) for the financial support. We also thank Astrid Morel, Fabien Candy, Sophie Labonne and Thomas Goître for their help in the fieldwork and/or in the laboratory. Finally, this study was made possible by a research grant provided by the University of Grenoble Alpes.

References (56)

  • J.M. García-Ruiz et al.

    Erosion in Mediterranean landscapes: changes and future challenges

    Geomorphology

    (2013)
  • C. Gomez et al.

    Laboratory Vis–NIR spectroscopy as an alternative method for estimating the soil aggregate stability indexes of Mediterranean soils

    Geoderma

    (2013)
  • B.E. Madari et al.

    Mid- and near-infrared spectroscopic assessment of soil compositional parameters and structural indices in two Ferralsols

    Geoderma

    (2006)
  • O. Maquaire et al.

    Characterisation of alteration profiles using dynamic penetrometry with variable energy. application to weathered black marls, Draix (Alpes-de-Haute-Provence, France)

    C. R. Géosci.

    (2002)
  • N. Mathys et al.

    Erosion quantification in the small marly experimental catchments of Draix (Alpes de Haute Provence, France). Calibration of the ETC rainfall–runoff–erosion model

    Catena

    (2003)
  • B. Stenberg et al.

    Visible and near infrared spectroscopy in soil science

  • R.A. Viscarra Rossel et al.

    Visible, near-infrared, mid-infrared or combined diffuse reflectance spectroscopy for simultaneous assessment of various soil properties

    Geoderma

    (2006)
  • H.L. Zhao et al.

    Soil properties, crop productivity and irrigation effects on five croplands of Inner Mongolia

    Soil Tillage Res.

    (2007)
  • B. Algayer et al.

    Short-term dynamics of soil aggregate stability in the field

    Soil Sci. Soc. Am. J.

    (2014)
  • E. Amezketa

    Soil aggregate stability: a review

    J. Sustain. Agric.

    (1999)
  • F.J. Arriaga et al.

    A fast method for determining soil particle size distribution using a laser instrument

    Soil Sci.

    (2006)
  • C. Beleites et al.

    hyperSpec: a package to handle hyperspectral data sets in R, R package version 0.98-20150304

  • L. Beuselinck et al.

    Assessment of micro-aggregation using laser diffractometry

    Earth Surf. Process. Landf.

    (1999)
  • A.M.R. Bieganowski et al.

    Determination of soil aggregate disintegration dynamics using laser diffraction

    Clay Miner.

    (2010)
  • L. Cécillon et al.

    Assessment and monitoring of soil quality using near infrared reflectance spectroscopy (NIRS)

    Eur. J. Soil Sci.

    (2009)
  • C.W. Chang et al.

    Near-infrared reflectance spectroscopy–principal components regression analyses of soil properties

    Soil Sci. Soc. Am. J.

    (2001)
  • W.S. Chepil

    Improved rotary sieve for measuring state and stability of dry soil

    Soil Sci. Soc. Am. Proc.

    (1952)
  • R. Core Team

    R: A Language and Environment for Statistical Computing

    (2015)
  • Cited by (15)

    • Data fusion of visible near-infrared and mid-infrared spectroscopy for rapid estimation of soil aggregate stability indices

      2021, Computers and Electronics in Agriculture
      Citation Excerpt :

      In respect of the current study as well as that of Shi et al. (2020), a rather strong control of OC on aggregate stability index was observed (0.75 and 0.84 respectively at p > 0.01) with correspondingly good vis-NIR prediction accuracies. MIR model for FW in the present study is considered better than those reported in the literature [i.e. Cañasveras et al. (2010) (R2 = 0.59 and SEV of 0.4 mm for cross-validation) and Erktan et al. (2016) (R2 of 0.45, RMSE of 0.62 mm for cross validation]. The results for the data fusion as well as that of SW for vis-NIR and MIR could not be put into context with other studies as no such studies exist in published literature.

    • Hyperspectral response and quantitative estimation on soil aggregate characters

      2021, Catena
      Citation Excerpt :

      According to previous publications, the relatively good prediction on MWD, MA, MIA and OMA were achieved with Rv2 value of 0.72 (MWD), 0.62 (MA), 0.65 (MIA), 0.43 (OMA) (n = 83) (Cañasveras et al., 2010; Shi et al., 2020), Rv2 = 0.56 (MWD) (n = 113) (Cañasveras et al., 2010; Gomez et al., 2013). Some other studies also indicated the R2 of validation models for MWD, MA, MIA, and OMA were entirely lower than 0.59 (n = 150) (Erktan et al., 2016; Henaka Arachchi et al., 2016; Le Bissonnais, 1996; Yang et al., 2017). The predictive models in our study (n = 199) outperformed above results, and the number and variability of soil samples were probably the main factors to affect model accuracy.

    View all citing articles on Scopus
    View full text