Elsevier

Waste Management

Volume 55, September 2016, Pages 176-190
Waste Management

Understanding leachate flow in municipal solid waste landfills by combining time-lapse ERT and subsurface flow modelling – Part II: Constraint methodology of hydrodynamic models

https://doi.org/10.1016/j.wasman.2016.04.005Get rights and content

Highlights

  • A new constraint methodology of hydrodynamic models.

  • Single and dual continuum approaches.

  • Hydrodynamic parameters.

Abstract

Leachate recirculation is a key process in the operation of municipal solid waste landfills as bioreactors. To ensure optimal water content distribution, bioreactor operators need tools to design leachate injection systems. Prediction of leachate flow by subsurface flow modelling could provide useful information for the design of such systems. However, hydrodynamic models require additional data to constrain them and to assess hydrodynamic parameters. Electrical resistivity tomography (ERT) is a suitable method to study leachate infiltration at the landfill scale. It can provide spatially distributed information which is useful for constraining hydrodynamic models. However, this geophysical method does not allow ERT users to directly measure water content in waste. The MICS (multiple inversions and clustering strategy) methodology was proposed to delineate the infiltration area precisely during time-lapse ERT survey in order to avoid the use of empirical petrophysical relationships, which are not adapted to a heterogeneous medium such as waste.

The infiltration shapes and hydrodynamic information extracted with MICS were used to constrain hydrodynamic models in assessing parameters. The constraint methodology developed in this paper was tested on two hydrodynamic models: an equilibrium model where, flow within the waste medium is estimated using a single continuum approach and a non-equilibrium model where flow is estimated using a dual continuum approach. The latter represents leachate flows into fractures. Finally, this methodology provides insight to identify the advantages and limitations of hydrodynamic models. Furthermore, we suggest an explanation for the large volume detected by MICS when a small volume of leachate is injected.

Introduction

This is the second of two related papers that attempt to improve the understanding of leachate flow in municipal solid waste (MSW) landfills by time-lapse ERT and subsurface flow modelling.

To reduce the impact of the landfill on the environment, the bioreactor concept has been developed for more than a decade in Europe (Reinhart and Townsend, 1998). This concept is based on leachate recirculation through a leachate injection system (LIS).

However, controlling the quantity of injected leachate remains a challenge (El-Fadel et al., 1996, Rosqvist and Destouni, 2000, Zeiss, 1997). Indeed, bioreactor operators need to design LIS to ensure optimal water content distribution into the waste deposit cell, as recommended by Reinhart and Townsend (1998). In the best cases, operators use hydraulic empirical laws, which generally are not based on experimental observation of the leachate flow inside the waste.

In recent decades, several hydrodynamic models for landfills have been developed and could help LIS design by predicting the volume of waste wetted by the injection process. The prevailing approach for modelling leachate flow in solid waste media is based on the assumption of a homogenous porous media (Demetracopoulos et al., 1986, Khire and Mukherjee, 2007, Korfiatis et al., 1984, Straub and Lynch, 1982). In these studies, a single continuum model represents waste and leachate flow is simulated by solving the Richards equation (Richards, 1931).

Because of the large heterogeneities observed in MSW, the assumption of an equilibrium model can be challenged. In addition, preferential flows are believed to be the reason why the single continuum approach considering homogeneous medium may not be in agreement with field observations (Ehrig, 1983, Uguccioni and Zeiss, 1997). A classical way is to represent preferential flows by distributed model parameters with the resolution of a single continuum model.

Another possibility, which is often described in soil science, consists in representing preferential flows by: (i) a dual continuum approach to represent the medium and (ii) the resolution of dual porosity or dual permeability models for water flow simulation (Gerke and Van Genuchten, 1993). This approach assumes that the porous medium is divided into two interacting pore domains, the first one corresponding to fractures (or macropores) with rapid water flow and the second one corresponding to the surrounding matrix domain (or micropores) with slow water movement. While dual porosity models assume that water in the matrix is stagnant, dual permeability models allow water flow in the matrix as well. To our knowledge, the dual porosity model was only used by Tinet et al. (2011) to simulate 1D leachate flow into the unsaturated waste column while the dual permeability model was used by Han et al. (2011) in columns filled with newspapers. In both cases, a dual continuum model was introduced because the single continuum approach (the Richards equation) failed to describe the experiments. A similar concept was used by Rosqvist and Destouni, 2000, Beaven et al., 2003, Rosqvist et al., 2005, Woodman et al., 2005 and Woodman et al. (2011) to model the contaminant transport in saturated conditions, with the so-called Mobile-ImMobile (MIM) transport model, corresponding to the dual porosity model or the BIModal advection model (BIM), corresponding to the dual permeability model. The MSW medium has therefore been described as a system composed of a matrix linked to fractures (macropores), which enables complex exchanges of contaminant during leachate flow. It allowed reproducing experimental observations (at the laboratory scale), while the single continuum model (advection dispersion equation) did not.

The main drawback of the dual continuum approach is the definition of the hydrodynamic parameters (porosity and permeability) of each continuum (fractures and matrix) and of the parameters ruling the fluxes exchanged between the two different continua.

Therefore, additional information is required to calibrate hydrodynamic models with an appropriate range of hydrodynamic parameters.

The determination of hydrodynamic parameters (i.e. moisture retention properties, porosity, hydraulic conductivity, density) is somewhat difficult in landfills because of the heterogeneity of the porous waste medium (Bendz et al., 1998, McCreanor and Reinhart, 2000). Many researchers have assessed hydrodynamic parameters on waste samples at the laboratory scale (Benson and Wang, 1998, Breitmeyer et al., 2008, Kazimoglu et al., 2005, Korfiatis et al., 1984, Orta de Velasquez et al., 2003, Powrie and Beaven, 1999, Staub et al., 2009, Stoltz et al., 2012, Tinet et al., 2011, Zornberg et al., 1999). In these studies, water content measurements were taken: (i) using gravimetric methods or (ii) introducing moisture sensors into waste samples (such as neutron probes, time domain reflectometry [TDR] or time domain transmissivity [TDT]). The waste hydrodynamic parameters of these samples were assessed to improve the hydrodynamic models used to predict leachate flow. However, hydrodynamic parameters determined at a small scale in the laboratory may not be appropriate to characterize full waste deposit cells, because of the heterogeneous nature of this medium (Fellner et al., 2009). Waste materials at the laboratory scale are generally shredded or sieved to a smaller grain size and waste confinement in cells may undergo edge effects. Moreover, compaction of waste into layers during landfilling leads to anisotropy within the landfill (Beaven et al., 2008, Fellner and Brunner, 2010). Consequently, hydraulic conductivity in the horizontal direction is potentially at least one order of magnitude greater than in the vertical direction (Landva et al., 1998, Powrie and Beaven, 1999, Stoltz et al., 2010). This anisotropy is not always taken into account at the laboratory scale in the assessment of hydrodynamic parameters.

Moreover, as sensor measurements provide only local information, it is difficult and costly to obtain good representativeness of the water content’s spatial distribution, because of the high heterogeneity of waste and the large number of probes, that would be required to instrument the landfill. Poor contact between probes and waste is also a problem currently encountered in these measurements (Grellier et al., 2006).

For all these reasons, other techniques are required to assess hydrodynamic parameters of waste at the field scale.

Many studies have shown that electrical resistivity tomography (ERT) is a suitable method to provide spatial information on leachate flow at the field scale (Clément et al., 2011a, Clément et al., 2010, Grellier et al., 2008, Guérin et al., 2004, Mondelli et al., 2007, Moreau et al., 2003, Morris et al., 2003, Olofsson et al., 2006, Rosqvist et al., 2003). Time-lapse ERT can be used to monitor changes in electrical resistivity related to leachate content variation. Indeed, the leachate injection process implies an increase in water content and consequently a corresponding decrease in electrical resistivity. ERT time-lapse monitoring consists in conducting the same ERT surveys several times at the same place (Daily et al., 1992), namely, before, during and after the leachate injection period. However, this geophysical method does not allow ERT users to directly measure water content and assess hydrodynamic parameters.

In hydrology, several studies have combined ERT measurements and subsurface flow modelling to constrain hydrodynamic models and assess the hydrodynamic parameters of a sandstone medium (Binley et al., 2002, Cassiani and Binley, 2005). The two papers used an inversion approach to derive hydrodynamic parameters from ERT measurements (combined with radar). This approach is based on two steps. First, the interpreted resistivity data resulting from the inversion process were converted to water content using Archie’s petrophysical relationship (Archie, 1942). Second, water content data from ERT and water content simulated by hydrodynamic models were compared to constrain hydrodynamic models and assess hydrodynamic parameters. More recently, Beaujean et al. (2014) also proposed a coupled inversion approach to assess hydrodynamic parameters from ERT measurements and showed the importance of taking into account the variable resolution of ERT images.

Many approaches are based on the use of a calibrated petrophysical relationship (i.e. with one parameter set), which seems to be appropriate for a relatively homogeneous medium such as sandstone. However, the studied medium is often highly heterogeneous in particular, waste.

Grellier et al. (2005) and Dumont et al. (2016) attempted to calibrate Archie’s law for waste in assessing the related in-laboratory parameters. The calibrated law was then used on field data sets to assess waste water content from resistivity (Grellier et al., 2008, Dumont et al., 2016, Ling et al., 2012). As mentioned above, the use of one calibrated Archie’s law for the entire landfill is probably not appropriate. Moreover, it can be assumed that Archie’s law calibrated at the laboratory scale is ill suited to assessing water content at the field scale for the above-mentioned reasons (i.e. different sizes, densities, compaction and anisotropy). To avoid the dependance on petrophysics, Hermans et al. (2015) used a probabilistic approach to define the relationship between resistivity and facies. However, this requires a large amount of borehole data to validate the relationships. In addition, temperature may strongly affect the results obtained by resistivity (Hermans et al., 2014), in particular in waste deposits where it may vary spatially.

In their study, Audebert et al. (2014) developed the MICS (“multiple inversions and clustering strategy”) methodology, which is based on a razor-sharp delimitation of the infiltration area on the ERT results. It also allowed ERT users to avoid the use of petrophysical relationships.

Other methods could also be applied, such as blocky inversion, also known as L1 model norm (Loke et al., 2003), or minimum support gradient (Nguyen et al., 2016) to obtain sharp boundaries.

As seen in the first paper (Audebert et al., 2016), the infiltration shape obtained with MICS allows ERT users to extract hydrodynamic information from their time-lapse surveys (i.e. wetting rate, infiltration shape anisotropy and the pore volume used for leachate flow).

The aim of the first paper was to improve the understanding of leachate flow through an analysis of hydrodynamic information derived from MICS on two different waste deposit cells. In spite of the heterogeneous nature of waste, this paper identified similarities between leachate flows into several waste deposit cells. It is probably due to: (i) the composition of the waste mass, which is approximately the same between the different waste deposit cells and (ii) the similar installation of waste into the landfills implying horizontal anisotropy due to the compaction and preferential flows.

Following the results of the first paper, we supposed that one hydrodynamic model with one parameter set could predict leachate flows for the Champs-Jouault landfill site. Indeed, due to the similarities mentioned above, the ERT method is probably not enough sensitive to detect differences of the leachate flows between the waste deposit cells of the same landfill site.

The aim of the present paper is to propose a new constraint methodology for the assessment of hydrodynamic parameters to provide a hydrodynamic model predicting leachate flow at the landfill scale.

ERT field data are used to estimate infiltration shapes and volume affected by infiltration using MICS (Audebert et al., 2016). Then the subsurface flow model is run using a wide range of hydrodynamic parameters. Finally, the range of hydrodynamic parameters is constrained to those for which the subsurface flow model and MICS results are in good agreement.

This methodology is applied to both single and dual continuum models to compare their ability to reproduce hydrodynamic information obtained with MICS at the field scale and consequently to improve the understanding of leachate flow into waste porous medium.

Section snippets

Field description

The Champs-Jouault MSW landfill site, located in western France (Cuves, Manche – Fig. 1a), is divided into 20 waste deposit cells. A leachate injection experiment was carried out on waste deposit cell 4 (Fig. 1b). It covers more than 5000 m2, and is approximately 100 m long, 50 m wide and 15 m at its maximum height (Fig. 1c). At the top, a horizontal trench (25 m long and 1.5 m wide) is dedicated to both biogas extraction and leachate injection through a 24-m-long perforated pipe. Finally, waste

Single continuum approach

The constraint methodology was applied to the Champs-Jouault ERT field data set to assess the three hydrodynamic parameters, θi, Ks and SF, of the single continuum approach.

The surface responses of the two comparison criteria, namely CISC and ΔV, are displayed in Fig. 4 – 1.a and – 1.b, respectively. Each figure has five subplots corresponding to the five time steps of the injection. The results are presented in 3D, each dimension corresponding to one hydrodynamic parameter investigated in the

Comparison between single and dual continuum approaches

Fig. 5 presents the RSF, with the hydrodynamic parameter sets calibrated by the constraint methodology (Table 2) for the five time steps of the infiltration monitoring (2 h, 3 h 30, 5 h, 6 h 30 and 9 h). Fig. 5a corresponds to the single continuum approach with a value of Ks and SF equal to 6 · 10−6 m · s−1 and 10, respectively, and θi was set at 0.64, corresponding to a state very close to saturation (θs equal to 0.69). These results (Fig. 5a) show a saturated infiltration area growing with time and

Conclusion

This study was based on two papers that have attempted to improve the understanding of leachate flow in MSW landfill by time-lapse ERT and subsurface flow modelling. Assessment of hydrodynamic parameters for leachate flow simulation is not an easy task because of the heterogeneous nature of waste. Therefore, a shape-based constraint methodology was introduced to assess hydrodynamic parameters for hydrodynamic models predicting leachate flow at the landfill scale. The constraint methodology

Acknowledgements

This study was conducted by IRSTEA (National Research Institute of Science and Technology for Environment and Agriculture) with ERT field data sets recorded on the SAS Les Champs-Jouault landfill site. Experimental work was supported by the company Les Champs-Jouault, Feder and the Région Basse Normandie.

References (69)

  • R. Guérin et al.

    Leachate recirculation: moisture content assessment by means of a geophysical technique

    Waste Manage.

    (2004)
  • B. Han et al.

    Measuring fluid flow properties of waste and assessing alternative conceptual models of pore structure

    Waste Manage.

    (2011)
  • Y.S. Jang et al.

    Hydraulic properties and leachate level analysis of Kimpo metropolitan landfill, Korea

    Waste Manage.

    (2002)
  • M. Khire et al.

    Leachate injection using vertical wells in bioreactor landfills

    Waste Manage.

    (2007)
  • P.T. McCreanor et al.

    Mathematical modeling of leachate routing in a leachate recirculating landfill

    Water Res.

    (2000)
  • J.W.F. Morris et al.

    Findings from long-term monitoring studies at MSW landfill facilities with leachate recirculation

    Waste Manage.

    (2003)
  • H. Rosqvist et al.

    Solute transport through preferential pathways in municipal solid waste

    J. Contam. Hydrol.

    (2000)
  • J. Šimůnek et al.

    Review and comparison of models for describing nonequilibrium and preferential flow and transport in the vadose zone

    J. Hydrol.

    (2003)
  • G. Stoltz et al.

    Liquid and gas permeabilities of unsaturated municipal solid waste under compression

    J. Contam. Hydrol.

    (2010)
  • A.J. Tinet et al.

    Experimental and theoretical assessment of the multi-domain flow behaviour in a waste body during leachate infiltration

    Waste Manage.

    (2011)
  • G.E. Archie

    The electrical resistivity log as an aid in determining some reservoir characteristics

    Trans. Am. Inst. Min. Metall. Eng.

    (1942)
  • J. Beaujean et al.

    Calibration of seawater intrusion models: inverse parameter estimation using surface electrical resistivity tomography and borehole data

    Water Resour. Res.

    (2014)
  • R.P. Beaven et al.

    Description of a tracer test through waste and application of a double porosity model

  • R.P. Beaven et al.

    Hydraulic properties of MSW

  • D. Bendz et al.

    Kinematic wave model for water movement in municipal solid waste

    Water Resour. Res.

    (1998)
  • C. Benson et al.

    Soil water characteristic curves for solid waste

    Environ. Geotech. Rep.

    (1998)
  • R.J. Breitmeyer et al.

    Field-scale lysimeter experiment to study hydrologic and mechanical properties of municipal solid waste

  • R. Clément et al.

    COMSOL multiphysics modelling for measurement device of electrical resistivity in laboratory test cell

  • Code-Permanent-Environnement-et-Nuisances, 1997. Arrêté Du 9 Septembre 1997. Décharges D’ordures Ménagères – Décharges...
  • W. Daily et al.

    Electrical resistivity tomography of vadose water movement

    Water Resour. Res.

    (1992)
  • A.C. Demetracopoulos et al.

    Modeling leachate production from municipal landfills

    J. Environ. Eng.

    (1986)
  • G. Dumont et al.

    Gravimetric water distribution assessment from geoelectrical methods (ERT and EMI) in municipal solid waste landfill

    Waste Manage.

    (2016)
  • M. El-Fadel et al.

    Temperature effects in modelling solid waste biodegradation

    Environ. Technol.

    (1996)
  • H.H. Gerke et al.

    Dual-porosity model for simulating the preferential movement of water and solutes in structured porous media

    Water Resour. Res.

    (1993)
  • Cited by (22)

    • Joint interpretation of geophysical data: Applying machine learning to the modeling of an evaporitic sequence in Villar de Cañas (Spain)

      2021, Engineering Geology
      Citation Excerpt :

      The combination of geophysical methodologies is broadly used in projects involving exploration of natural resources, geological storage, environmental monitoring, civil engineering practice, as well as natural and anthropic hazard assessment (Yordkayhun et al., 2009; Carbonell et al., 2010; Font-Capó et al., 2011; Alcalde et al., 2014; Seillé et al., 2015). Among other techniques, seismic and electrical methods, combined with ground truth provided by boreholes, have proven to be extremely useful for shallow subsurface exploration and characterization (Ogaya et al., 2016; Festa et al., 2016; Audebert et al., 2016). For example, controlled seismic source data provides information on the structural architecture and the mechanical properties of the subsurface.

    • Comprehensive overview of numerical modeling of coupled landfill processes

      2020, Waste Management
      Citation Excerpt :

      Several other significant issues related to long-term behaviors of MSW and landfill performance remain. For example, some researchers have shown that it is more reasonable to simulate leachate seepage within the MSW skeleton using a dual-porosity model (Audebert et al., 2016; Feng et al., 2018b). In addition, many behaviors remain poorly described, such as those related to the effects of heterogeneity and anisotropy, the effect of temperature on MSW hydraulic and mechanical properties, and the interaction between MSW and infrastructures such as liner systems, collection systems, and drainage systems.

    View all citing articles on Scopus
    View full text