Elsevier

Soil and Tillage Research

Volume 144, December 2014, Pages 53-71
Soil and Tillage Research

Toward a tool aimed to quantify soil compaction risks at a regional scale: Application to Wallonia (Belgium)

https://doi.org/10.1016/j.still.2014.06.008Get rights and content

Highlights

  • We developed a model of the precompression stress (Pc) at the regional level and applied it to Wallonia (Belgium).

  • The model takes into account the spatial structure of the data either through geostatistical methods or from the knowledge of the statistical distribution of parameters.

  • A higher uncertainty was found at pF 2.5 than at pF 1.8. Uncertainty was also found higher for clay and clayed loess than for other textural classes present in Wallonia.

  • We developed a model of the risk corresponding to the application of loads on the soil.

  • Subsoil compaction risks exist mainly in loamy forest soils with small coarse fragments that support loads similar to that existing on logging machines.

Abstract

The spatial analysis of the soil compaction risk has been developed at the regional level and applied to Wallonia (Belgium). The methodology is based on the estimation of the probability of exceeding the preconsolidation stress due to the application of loads on the soil.

Preconsolidation stresses (Pc) are computed from the pedotransfer functions of Horn and Fleige (2003) at pF 1.8 and 2.5 and classified into 6 categories ranging from very low Pc (<30 kPa) to extremely high Pc (>150 kPa). The computation requires the knowledge of pedological (texture, organic content), mechanical (bulk density, cohesion, internal friction angle), and hydraulic variables (water content available, non-available water content, air capacity, saturated hydraulic conductivity). These variables are obtained from databases like HYPRES or AARDEWERK or from pedotransfer functions. The computation of Pc takes into account the spatial structure of the data: in some cases, data are abundant (e.g., texture data) and spatial variability is taken into account through geostatistical methods. In other cases, the data is sparse but uncertainty information can be extracted from the knowledge of the statistical distribution. Maps of the most probable Pc class are produced. Uncertainty is computed as the classification error probability. Implementation of these methods in Wallonia showed that Pc values higher than 120 kPa are reached either on 64% of the territory at pF 2.5 or on 55% at pF 1.8. A higher uncertainty was found at pF 2.5 than at pF 1.8. Uncertainty was also found higher for clay and clayed loess than for other textural classes present in Wallonia.

The risk of compaction is defined as the probability that Pc is exceeded by the stress created by a load applied to the soil at a depth of 40 cm, the loads being similar to those induced by agricultural or forestry tires. It appeared that subsoil compaction risks exist mainly in loamy forest soils with small coarse fragments supporting loads similar to that existing on logging machines.

In the zones where the uncertainty is low, the developed tool could be used as a basis for providing policy measures in order to promote soil-friendly farming and forest practices.

Introduction

Compaction concerns agricultural and forestry crops and results from the passage of heavy machines on sensitive soils, mainly during harvest operations and harvest transport. The detrimental effects of soil compaction on the crop production have been reported in many studies on both agricultural and forest soils (Hakansson and Reeder, 1994, Hamza and Anderson, 2005, Greacen and Sands, 1980, Goutal, 2012). Compaction causes a decrease in porosity and an increase in soil strength that may restrict root growth and affect the density and diversity of soil mesofauna and bacterial communities (Soane and van Ouwerkerk, 1995, Batey and McKenzie, 2006, Frey et al., 2009, Lipiec et al., 2012). Soil compaction not only reduces crop and forest production, but has also negative environmental effects (Jones et al., 2003). Indeed, saturated hydraulic conductivity is reduced, increasing the risk of runoff of water and pollutants toward surface waters, and the movement of nitrate and pesticides into ground waters. The volume of soil available to act as a buffer for pollutants is reduced. The risk of soil erosion increases through the presence of excess water above compacted layers. Because of the reduction of soil aeration, production of greenhouse gases through denitrification may occur by anaerobic processes (Jarvis, 2007, Hoefer, 2010).

Considering the detrimental effects of soil compaction, the proposal of the EU Commission for a Soil Framework Directive mentions soil compaction as one of the major threats to a sustained quality of soils in Europe (COM, 2006). The compaction of the subsoil, defined as ‘subsurface soil material that lies below the normal cultivation depth or pedological A horizon’, is particularly problematic since it is difficult and expensive to alleviate (Hakansson and Reeder, 1994, Spoor et al., 2003). Subsoil compaction risks are increasing with growth in farm size, increased mechanisation and equipment size, and the drive for greater productivity (Jones et al., 2003).

In analysing soil compaction, a distinction has to be made between the susceptibility of soils to compaction and their vulnerability. Susceptibility is the likelihood that compaction occurs if subjected to factors that are known to cause compaction (Louwagie et al., 2009). Susceptibility to compaction depends on quasi-permanent characteristics such as texture and carbon content and on short-term changing characteristics such as soil moisture condition. It ranges from sand (least susceptible) – loamy sand – sandy loam – loam – clayed loam – loamy clay to clay soils (Woods et al., 1944, cited by Louwagie et al., 2009). Medium- and fine-textured loam and clay soils are resistant to mechanical pressure at low water contents but they are highly susceptible to severe compaction at high water contents (Horn et al., 1995).

The soil’s vulnerability to a given threat is determined taking into account the inherent soil susceptibility and an exposure estimate based on an evaluation of the stresses inflicted by land management and climate (Troldborg et al., 2013). Jones et al. (2003) propose a simple classification system for subsoil vulnerability to compaction using a two-stage process. First, the inherent susceptibility of the soil to compaction is estimated on the basis of the relatively stable soil properties, such as soil texture, nature of clay, bulk density, organic matter content, structure, soil moisture content and soil moisture potential. Second, the susceptibility class is converted into a vulnerability class through consideration of the likely soil moisture status at the time of critical loadings. The authors conclude that some improvements could be brought to the method, including namely the use of pedotransfer functions. Another method for estimating the soil’s vulnerability or the risks of soils being further compacted is obtained by comparing calculated soil strengths with vertical stresses created by a given wheel. The soil strength is usually expressed by the precompression stresses evaluated from pedotransfer functions (Van den Akker, 2004, Horn and Fleige, 2003). More recently, Troldborg et al. (2013) developed Bayesian belief networks for assessing the risk of soil compaction, allowing the combination of available data from standard soil surveys and land use databases with qualitative expert knowledge.

In order to face the challenge of the Soil Framework Directive, if implemented, the Governments of the European Union wish to identify areas of risk and develop relevant policy measures suited to provide soil-friendly farming practices. The report of the SoCo (Sustainable Agriculture and Soil Conservation) project presents a European map of natural soil susceptibility to compaction (Louwagie et al., 2009). Based on soil properties, it gives an idea of the geographic spread of compaction susceptibility. Unfortunately, this map does not provide sufficiently accurate information to determine the extent of actual and potential problems at a Regional Scale and to bring responses to the regions in Europe who have been asked to develop environmental plans.

In Wallonia (South of Belgium), political discussion on the problem of compaction is namely taken into account by the Walloon Forest Code which is in application since 13th September 2008 (Décret relatif au Code forestier wallon, 2008) and prohibits explicitly damages on the ground that could have long-term consequences on the forests vitality. Wallonia occupies around 17,000 km2. Forest areas represent 530,600 ha while agricultural areas represent 756,000 ha. Forest soils are mainly Cambisols, while agricultural soils are mainly Luvisols.

The aim of the paper is thus to develop a methodology that would be suitable to help the policy-makers in order to limit soil compaction. The methodology concerns the subsoil (40 cm depth) because compaction in this horizon is generally considered as particularly serious because of its persistence. As far as possible, the methodology should involve the use of existing databases. The main challenge concerns the structure of the information. In some cases, the information is abundant while in other cases, there is a lack of data. At the same time, the uncertainties relative to the data and to the modelling process have to be taken into account and their contributions in terms of overall uncertainty on the results have to be quantified.

Section snippets

Material and methods

The methodology comprises two stages. Firstly, the susceptibility of soils to compaction is assessed by computing the soil strength, this latter being expressed by the precompression stress (Pc). Secondly, the vulnerability of soils is analysed by computing the vertical stresses created in the soils by a load similar to that applied by a wheel and comparing it to the precompression stress.

Measurements analysis

On the experimental sites, as expected, the measurements performed at 40 cm depth revealed significant variability (Table 5). The soil texture variability was characterized by high values of the coefficient of variation (CV) of clay and sand contents in all sites, and high CV of loess content in forest sites. Bulk density (BD) was higher in agricultural soils than in forest soils. Overall, the variability of BD was small, excepted for one forest site (Fauvillers).

The measured values were

Discussion

The proposed methodology provides estimates of the modal class of Pc (1–6) and the compaction risk at regional scale. As acquiring direct Pc measurements is prohibitive in time and costs, there is a need for a model able to valorise already existing data. Of course, the accuracy of the results depends on the quality of those data. The level of accuracy of the estimations can be discussed at the light of two types of error, namely, random and systematic errors.

The computation of modal class Pc

Conclusion

The basis of management tool for assessing soil compaction were developed and applied to Wallonia (Belgium). This tool evaluated the susceptibility of subsoil compaction at a regional scale by estimating the modal class of precompression stress (Pc) on basis of the pedotransfer functions (PTFs) of Horn and Fleige (2003). The uncertainty was quantified by estimating the classification error probability on the modal class. Simulations showed that uncertainty was higher at pF 2.5 than at pF 1.8.

Acknowledgments

This work was funded by the SPW (Service Public de Wallonie), DGARNE (Direction générale opérationnelle de l'Agriculture, des Ressources naturelles et de l'Environnement).

References (48)

  • T. Keller

    A model for the prediction of the contact area and the distribution of vertical stress below agricultural tyres from readily available tyre parameters

    Biosyst. Eng.

    (2005)
  • M. Lebert et al.

    A method to predict the mechanical strenght of agricultural soils

    Soil Till. Res.

    (1991)
  • J. Lipiec et al.

    Effects of soil compaction on root elongation and anatomy of different cereal plant species

    Soil Till. Res.

    (2012)
  • J.C. Refsgaard et al.

    Uncertainty in the environmental modelling process – a framework and guidance

    Environ. Modell. Softw.

    (2007)
  • P. Schjønning et al.

    Modelling effects of tyre inflation pressure on the stress distribution near the soil–tyre interface

    Biosyst. Eng.

    (2008)
  • G. Spoor et al.

    Subsoil compaction: risk, avoidance, identification and alleviation

    Soil Till. Res.

    (2003)
  • M. Troldborg et al.

    Application of Bayesian belief networks to quantify and map areas at risk to soil threats: using soil compaction as an example

    Soil Till. Res.

    (2013)
  • J.J.H. Van den Akker

    SOCOMO: a soil compaction model to calculate soil stresses and the subsoil carrying capacity

    Soil Till. Res.

    (2004)
  • J. Wösten et al.

    Development and use of a database of hydraulic properties of European soils

    Geoderma

    (1999)
  • E. Alonso et al.

    A constitutive model for partially saturated soils

    Géotechnique

    (1990)
  • T. Batey et al.

    Soil compaction: identification directly in the field

    Soil Use Manage.

    (2006)
  • P. Bogaert et al.

    Estimating soil properties from thematic soil maps: the Bayesian maximum entropy approach

    Soil Sci. Soc. Am. J.

    (2002)
  • J. Boussinesq

    Application des Potentiels à L'étude de L'équilibre et du Mouvement des Solides Élastiques

    (1885)
  • S. Brais et al.

    Soil compaction induced by careful logging in the claybelt region of northwestern Quebec (Canada)

    Can. J. Soil Sci.

    (1998)
  • Cited by (13)

    • Forest stand susceptibility mapping during harvesting using logistic regression and boosted regression tree machine learning models

      2020, Global Ecology and Conservation
      Citation Excerpt :

      SSM predicts the spatial probability of stand damages, which can be practical in forest harvesting management. Thus, the study of SSM is useful for forest management planning (Baskent and Keles, 2005; Legues et al., 2007), harvest scheduling (Hifi and M’Hallah, 2005; Mok et al., 2007), land use policy (Bremer et al., 2010), and major scientific research (D’Or and Destain, 2014) throughout the world. Nowadays, different models are used in the natural sciences and natural hazards including, LR (De Abreu et al., 2017), generalized additive models (Pecchi et al., 2019), neuro-fuzzy systems (Dedić et al., 2019; Ghorbanzadeh et al., 2018), frequency ratio (Nourani et al., 2014; Pirasteh and Li, 2017), machine learning techniques (Ghorbanzadeh et al., 2019a,b,c; Pourghasemi et al., 2019a, b; Shahabi et al., 2019; Pourghasemi et al., 2020), and tree-ensembled techniques (Shearman et al., 2019), but this is a fact that a few studies reported the SSM for forest harvesting (Reeves et al., 2012).

    • Soil research challenges in response to emerging agricultural soil management practices

      2020, Advances in Agronomy
      Citation Excerpt :

      The translation into real conditions can lead to misinterpretations. Most soil compaction modeling studies are available for the 1D (e.g., Stettler et al., 2014), for 2D (e.g., D'Or and Destain, 2014) and 2.5D (e.g., Keller et al., 2007; Schjønning et al., 2008, 2015a) layer for different depths, but not real 3D connections; 3D modeling is rare (e.g., Duttmann et al., 2014), and 4D is not available. Most soil compaction research is done in the laboratory (e.g., Canillas and Salokhe, 2002) or on single plots during wheeling experiments (e.g., Keller et al., 2014).

    • Mapping soil compaction – A review

      2018, Current Opinion in Environmental Science and Health
      Citation Excerpt :

      Troldborg et al. [45] developed Bayesian belief networks for assessing the risk of soil compaction by combining available data from standard soil surveys and land use databases with qualitative expert knowledge. D’Or and Destain [46] computed the pre-consolidation stress from pedotransfer functions [27] based on knowledge of pedological, mechanical and hydraulic conductivity obtained from available databases. Spatial variability may be taken into account through geostatistical methods if data are abundant and uncertainty information can be extracted from the knowledge of the statistical distribution in the case of sparse data.

    • Does soil compaction increase floods? A review

      2018, Journal of Hydrology
      Citation Excerpt :

      Lebert (2010) estimated the soil compaction risk for the arable land in Germany using results of Houšková (2008), while Troldborg et al. (2013) used Bayesian belief networks by combining data from standard soil surveys, land use and expert judgement. D’Or and Destain (2014) computed the preconsolidation stress (determined based on the stress-strain relationship of soil obtained during laboratory compression according to Casagrande (1936)) from pedotransfer functions (Horn and Fleige, 2003), based on pedological, mechanical and hydraulic characteristics within a geostatistical framework. Recently, geophysical techniques for investigating subsurface compaction in agricultural soils have been advanced, such as ground penetrating radar (Lane et al., 2016, Wang et al., 2016, André et al., 2012), although these are usually limited to rather small scales.

    View all citing articles on Scopus
    View full text