Waste Management xxx (2017) xxx–xxx

Contents lists available at ScienceDirect

Waste Management
journal homepage: www.elsevier.com/locate/wasman

A mass balance model to estimate the rate of composting, methane

oxidation and anaerobic digestion in soil covers and shallow waste
layers
Reza Rafiee, Lizanne Obersky, Sihuang Xie, William P. Clarke ⇑
Centre for Solid Waste Bioprocessing, Schools of Civil and Chemical Engineering, The University of Queensland, Brisbane 4072, Australia

a r t i c l e i n f o a b s t r a c t

Article history: Although CH4 oxidation in landfill soil covers is widely studied, the extent of composting and CH4 oxida-
Received 1 June 2016 tion in underlying waste layers has been speculated but not measured. The objective of this study was to
Revised 2 December 2016 develop and validate a mass balance model to estimate the simultaneous rates of anaerobic digestion
Accepted 17 December 2016
(rAD), CH4 oxidation (rOX) and composting (rCOM) in environments where O2 penetration is variable and
Available online xxxx
zones of aerobic and anaerobic activity are intermingled. The modelled domain could include, as an
example, a soil cover and the underlying shallow waste to a nominated depth. The proposed model
Keywords:
was demonstrated on a blend of biogas from three separate known sources of gas representing the three
Landfill
Aerobic processes
reaction processes: (i) a bottle of laboratory grade 50:50% CH4:CO2 gas representing anaerobic digestion
Composting biogas; (ii) an aerated 250 mL bottle containing food waste that represented composting activity; and (iii)
Mass balance an aerated 250 mL bottle containing non-degradable graphite granules inoculated with methanotrophs
Carbon isotopes and incubated with CH4 and O2 to represent methanotrophic activity. CO2, CH4, O2 and the stable isotope
13
C-CO2 were chosen as the components for the mass balance model. The three reaction rates, r (=rAD, rOX,
rCOM) were calculated as fitting parameters to the overdetermined set of 4 mass balance equations with
the net flux of these components from the bottles q (=qCH4 , qCO2 , qO2 and qCO2
d13 C-CO2 ) as inputs to the
model. The coefficient of determination (r2) for observed versus modelled values of r were 1.00, 0.97, 0.98
when the stoichiometry of each reaction was based on gas yields measured in the individual bottles and q
was calculated by summing yields from the three bottles. r2 deteriorated to 0.95, 0.96, 0.87 when using an
average stoichiometry from 11 incubations of each of the composting and methane oxidation processes.
The significant deterioration in the estimation of rCOM showed that this output is highly sensitive to the
evaluated stoichiometry coefficients for the reactions. r2 deteriorated further to 0.86, 0.77, 0.74 when
using the average stoichiometry and experimental measurement of the composition and volume of the
blended biogas to determine q. This was primarily attributed to average errors of 8%, 7%, 11% and 14%
in the measurement of qCH4 , qCO2 , qO2 and qCO2
d13 C-CO2 relative to the measurement of the same quan-
tities from the individual bottles.
Ó 2016 Elsevier Ltd. All rights reserved.

1. Introduction
The oxidation of CH4 in landfill cover systems is widely studied.
However, few studies have explored aerobic processes in waste
beneath soil covers, even though a number of authors have
observed O2 penetration into landfilled waste (Kallel et al., 2006;
Lefebvre et al., 2000; Yoshida et al., 2012). Peneteration of oxygen
into the waste layer can be variable and channelled, leading to
intermingled zones of aerobic and anaerobic activity. It is proposed
that high concentrations of CO2 relative to CH4 in gaseous emis-
⇑ Corresponding author.
E-mail address: william.clarke@uq.edu.au (W.P. Clarke).
sions from the surface of soil covers in active landfill cells could
be partially attributed to aerobic activity within shallow waste
(Abushammala et al., 2013; Bogner et al., 2011).
Three main techniques are used to measure the fraction of CH4
that is oxidized in a landfill cover: (1) a mass balance approach
(Bogner and Spokas, 1993; Christophersen et al., 2001); (2) a stable
carbon isotope method (Bergmaschi and Harris, 1995; Liptay et al.,
1998b) and (3) an approach based on the ratio of CH4:CO2 (Gebert
et al., 2011a). These approaches assume CO2 production from com-
posting activity in the soil layer is insignificant compared to CO2
production from CH4 oxidation. Scheutz et al. (2011) and Röwer
et al. (2011) accounted for composting activity in their measure-

http://dx.doi.org/10.1016/j.wasman.2016.12.025
0956-053X/Ó 2016 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Rafiee, R., et al. A mass balance model to estimate the rate of composting, methane oxidation and anaerobic digestion in
soil covers and shallow waste layers. Waste Management (2017), http://dx.doi.org/10.1016/j.wasman.2016.12.025
2 R. Rafiee et al. / Waste Management xxx (2017) xxx–xxx
ments of CH4 oxidation activity in organic rich soil by performing
parallel aerated incubations of the soil without CH4.
The aim of this paper was to develop and validate a mass bal-
ance model to estimate the rate of anaerobic digestion, CH4 oxida-
tion and composting from gas blended from three sources, each
source representing biogas generated by one reaction process. Val-
idation of the model was performed by comparing reaction rates
calculated from measurements of the blended gas, as would be
made when measuring gas fluxes in any real system such as a land-
fill, with the rates based on gas production from the individual
sources.
2. Materials and methods
A key assumption in the mass balance model is that the stoi-
chiometry of each of the three reaction processes is stable.
Although composition of biogas from a landfill cell will change
with time, these changes are gradual and the assumption of stable
yields for each degradation process over the duration of flux mea-
surements (e.g., with a static chamber on the surface of a landfill) is
adopted.
In order to validate the model, stable sources of gas representa-
tive of each reaction process was required. This is readily achieved
with composting where the product gas will be CO2 provided con-
ditions are fully aerobic, regardless of the stage of decomposition
of the waste. A stable CH4 oxidation process can also be maintained
by feeding CH4 to a methanotrophic culture supported on an inert
substratum (Xie et al., 2016). Maintaining stable biogas production
from the anaerobic digestion of organic waste is more difficult and
not achievable in batch mode because faster fermentation pro-
cesses would dominate before establishment of methanogenic
activity (Nopharatana et al., 2003). A stable biogas from anaerobic
digestion could be achieved with continuous operation; however
this would still require careful maintenance of reactor conditions
and feedstocks. A further complication is that anaerobic conditions
are difficult to maintain in the experimental procedure described
below, where the headspaces of the anaerobic digestion, compost-
ing and CH4 oxidation reactors are connected to produce a blended
gas. This would expose the anaerobic reactor to O2. Therefore, a
bottle of laboratory grade 50:50% CH4:CO2 gas was used as a source
of biogas from anaerobic digestion.
2.1. Experimental plan
In order to validate the model, the model was applied to 11 gas
blends generated by 11 sets of composting and CH4 oxidation incu-
bations blended with the 50:50% CH4:CO2 gas. Apart from this, an
additional 22 composting incubations and 29 CH4 oxidation incu-
bations were performed. The average yields from all 33 compost-
ing incubations and 40 CH4 oxidation incubations were used to
determine global molecular stoichiometric coefficients for the
composting and CH4 oxidation reactions.
2.2. Batch incubation experiments
The composting incubations were conducted in 250 mL serum
bottles amended with 3.0 g of blended synthetic waste (Table 1),
flushed with compressed air and sealed with butyl rubber septum
(20 mm thick) and aluminium crimps. All incubations were per-
formed in a clean bottle with fresh waste. The butyl septum
allowed gas samples to be withdrawn or injected using 22G hypo-
dermic needles and gas-tight syringes. An initial sample of 10 mL
was collected to characterise O2 and background CO2 levels. These
incubations were performed over 24 h at 25 °C.
Please cite this article in press as: Rafiee, R., et al. A mass balance model to estimate the rate of composting, methane oxidation and anaerobic digestion in
soil covers and shallow waste layers. Waste Management (2017), http://dx.doi.org/10.1016/j.wasman.2016.12.025
Table 1
Composition of synthetic organic waste.
Composition Type Weight (w/w, wet basis %)
Vegetable Lettuce 20
Cauliflower
Cabbage
Potato and onion peel
Fruit Banana 30
Golden apple
Red apple
Orange
Food scraps Lamb meat 50
Rice
Macaroni
Bread
The CH4 oxidation test was performed using 3.0 g of graphite
granules colonised with methanotrophic bacteria (Xie et al.,
2016). 5 mL of a nutrient solution with the following composition
was added to the granules: KH2PO4 (0.26 g/l), Na2HPO47(H2O)
(0.62 g/l), (NH4)2SO4 (0.25 g/l), FeSO4 (5 mg/l) and CuCl22H2O
(0.07 mg/l).
The headspace of the 250 mL bottle was flushed with com-
pressed air and sealed. 100 mL of laboratory grade 50:50% CH4:
CO2 gas (BOC) and 200 mL of additional air was then injected to
the bottle. Temperature, pressure and concentration of CH4, CO2
and O2 were recorded initially (t = 0 h) and after 24 h to calculate
gas generation and consumption.
2.3. Blending of source gases
To blend the source gases (i.e., the 50:50% CH4:CO2 gas mixture
and the product and residual reactant gases in the composting and
CH4 oxidation incubations), the headspace of the 250 mL bottles
were connected by Tygon Master Flex tubing (ID 6 mm) with each
bottle headspace readily isolated by a two-way valve. The volume
and composition of each headspace was characterised before the
headspaces were mixed.
To ensure that the gas from the three bottles was thoroughly
mixed, all two-way valves were opened and a 200 mL syringe
was used to drive gas movement into each bottle headspace by
drawing and pushing the plunger several times. A one hour time
period was used to allow the combined headspace to reach an
equilibrated state. The pressure, temperature and composition of
the mixed gas were measured. The entrained O2 in the connection
tubes was accounted for in the calculations.
2.4. Analytical methods
For molecular gas composition monitoring, samples were anal-
ysed using a Shimadzu GC-8A FID with 100/120 mesh ShinCarbon
73 ST micropacked column for CH4 and CO2 analysis and a Shi-
madzu GC-8A TCD with a 80/100 mesh Molecular Sieve 5A column
for O2 and N2 analysis. Calibrations for both GCs were performed
using external gas standards (BOC).
Headspace temperatures were measured by an infrared ther-
mometer (Testo-model 830-T1) and the pressure was monitored
using an Omega gauge (Omega, model DPG 1001B-15G). Pressure
and temperature were measured at the start and the end of the
24 h incubation period.
2.4.1. Isotope analysis
Like the molecular components, the concentration of the iso-
tope d13C-CO2 in the blended gas depends on the rate of each reac-
 R. Rafiee et al. / Waste Management xxx (2017) xxx–xxx 3

tion process. The concentration of d13C-CO2 is conventionally The mass balances for the molecular components CO2, CH4 and
expressed as a relative abundance defined as (Liptay et al., 1998a): O2 are shown in Eqs. (5)–(7). In a practical application such as the

 top layer of a landfill, fluxes would be integrated over the boundary
Rsam
d13 C-CO2 ¼  1
1000 ð‰Þ ð1Þ of the control volume to arrive at the net production rates, qCH4 ,
Rstd
qCO2 and qO2 , qCH4 , qCO2 and qO2 .
where Rsam is the 13C/12C ratio of the sample and Rstd the 13C/12C
ratio of a reference standard VPDB (Vienna Peedee Belemnite: rAD þ r OX þ r COM ¼ qCO2 ð5Þ
0.0112372).
In all cases, analyses were performed on 100 mL gas samples Y CH4 ;AD r AD  Y CH4 ;ox r ox ¼ qCH4 ð6Þ
injected into 1 L SKC FlexFoil gas bags. All isotope samples were
analysed in the Stable Isotope Geochemistry Laboratory, University Y O2 ;ox rOX þ Y O2 ;COM r COM ¼ qO2 ð7Þ
of Queensland using an Isoprime/Agilent Gas Chromatograph-com
bustion-isotope ratio mass spectrometer (GC-c-IRMS). The where rAD, rOX and rCOM (mole CO2/d) represent the rate of anaerobic
required sampling for the three sources of d13C-CO2 and the digestion, CH4 oxidation, and composting. qCO2 (mole CO2/d), qCH4
blended gas were as follows: (mole CH4/d) and qO2 (mole O2/d) represent the net production or
consumption of CO2, CH4 and O2.
Anaerobic digestion. The d13C-CO2 signature for the 50:50% CH4: The mass balance for the d13C-CO2 isotope is expressed in terms
CO2 bottled gas was analysed on one 100 mL sample from the of relative abundance d13C-CO2 using a linear mixing model:
one gas cylinder used throughout the experiments.
 
ðd13 C-CO2 Þout
qCO2 ¼ ðd13 C-CO2 ÞAD
rAD
Composting. Composting does not exhibit isotopic fractionation.  
Consequently the d13C-CO2 signature of the product gas was the þ ðd13 C  CO2 Þox
rOX
same as the relative abundance of 13C in the solid. d13C-CO2 of  
þ ðd13 C  CO2 ÞCOM
r COM ð8Þ
the product gas from composting was measured in three incuba-
tions that were not in the set of 11 incubations used to valid the  
where d13 C-CO2 out ; ðd13 C-CO2 ÞAD , (d13 C-CO2)OX and (d13 C-CO2)COM
mass balance model. Analyses for d13C-CO2 were performed by col-
lecting 100 mL of headspace gas from each of the three bottles after are the relative abundances of 13C-CO2 in the blended gas, the
24 h incubation. The average of these three measurements was 50:50% CH4:CO2 bottled gas, the headspace of the CH4 oxidation
assumed to be the d13C-CO2 signature from the composting incubation bottle and the headspace of the composting incubation
process. bottle.
The mass balance equations for CH4, CO2, O2 and 13C-CO2 (Eq.
CH4 oxidation. d13C-CO2 was measured at the end of the 24 h incu- (9)) represent an over-determined system for calculating r = (rAD,
bation period in three incubation experiments that were not in the rox, rcom):
set of 11 incubations used to validate the mass balance model.
100 mL of headspace gas was collected from each of the three bot- Ar ¼ B
2 3 2 3
tles at the end of the 24 h incubation period. 1:0 1:0 1:0 2 3 qCO2
6 7 rAD 6 7
6 Y CH4 ;AD Y CH4 ;OX 0 7 6 qCH4 7
Blended gas. Isotopic analysis of the blended biogas was performed 6 76 7 6 7 ð9Þ
6 76
4 r OX
7¼6
5 6
7
7
on a 100 mL sample withdrawn from the headspace of bottles after 6 0 Y O2 ;OX 7
Y O2 ;COM 5 6 qO2 7
4 4 5
mixing the gases. Three 5 mL samples were collected from the r COM
100 mL sample and each measured for d13C-CO2. The 100 mL vol- dAD dOX dCOM qCO2
dout
ume set aside for analyses was taken into account when calculat-
ing q. where the abbreviated notation d is used in place of d13C-CO2.
Matrix A defines the stoichiometric coefficients of all reactions
2.5. The mass balance model and matrix B defines the net production or consumption of the com-
ponents CO2, CH4, O2 and 13C-CO2.
The mass balance model can be applied to any control volume
where the net flux of gas components in and out of the control vol-
2.6. Solving for r
ume can be measured. Such an environment could be the upper
layers of a landfill (Fig. 1) where fluxes to the base of the control
The rate values were solved using Excel SolverÒ with the Gener-
volume can be estimated as a function of surface fluxes and the
alized Reduced Gradient (GRG) nonlinear algorithm. It was tar-
concentration of gas components at the base of the control volume
geted to minimize the Root Mean Square Error (RMSE) between
(Christophersen et al., 2001; Gebert et al., 2011).
the observed and model estimation of net fluxes, q, by varying
CO2 was selected as a basis to express all rates [mole CO2/d]
the r vector subject to one constraint that the r vector consists of
because CO2 is common to all of the reactions. Only the rates of
non-negative values.
production or consumption of the other components with respect
to the rate or production of CO2 need be specified to complete
the mass balances. The stoichiometric coefficient for the organic 3. Results and discussion
matter is not required. The partial stoichiometry for the 3 reaction
processes is: 3.1. Stoichiometry coefficients
Anaerobic digestion : Organic matter ! CO2 þ Y CH4 ;AD CH4 ð2Þ
Table 2 presents stoichiometric coefficients determined in 33
composting and 40 CH4 oxidation incubations. The small standard
CH4 oxidation : Y CH4 ;ox CH4 þ Y O2 ;ox O2 ! CO2 ð3Þ
deviations in the coefficients demonstrate repeatable degradation
conditions in the incubations. The stoichiometric coefficients are
Composting : Organic matter þ Y O2 ;COM O2 ! CO2 ð4Þ
comparable with previous studies. The respiratory quotient (RQ),

Please cite this article in press as: Rafiee, R., et al. A mass balance model to estimate the rate of composting, methane oxidation and anaerobic digestion in
soil covers and shallow waste layers. Waste Management (2017), http://dx.doi.org/10.1016/j.wasman.2016.12.025
4 R. Rafiee et al. / Waste Management xxx (2017) xxx–xxx

Fig. 1. Possible fluxes and reactions in the upper layers of a landfill.

Table 2
Average reaction stoichiometries and d13C-CO2 values (SEs are included where applicable).

Process Reaction/equation Measured values No. of incubations

Composting r com Y O2 ;com = 1.0 ± 0.1 33
OM þ Y O2 ;com O2 ! CO2 þ H2 O
CH4 oxidation r OX Y CH4 ;ox = 1.7 ± 0.1 40
Y CH4 ;ox CH4 þ Y O2 ;ox O2 ! CO2 þ H2 O
Y O2 ;ox = 2.7 ± 0.2
Isotope signatures CH4 oxidation composting d13C-CO2,ox = 62.0 ± 5.0‰ 3
d13C-CO2,COM = 20.4 ± 4.0‰ 3
50:50% CH4:CO2 cylinder gas representative of anaerobic digestion d13C-CO2,AD = 31.5‰ 3

equal to the ratio of CO2 evolution and O2 uptake, was 1.04 for the Therefore the mass balance is as follows:
composting incubations. The RQ is equal to 1 for carbohydrates
Ar ¼ B
under fully aerobic conditions (Atkinson et al., 1997). Values of 2 3 2 3
1:0 1:0 1:0 2 3 qCO2
RQ measured for municipal solid waste vary (0.95 to 1.02, Smårs 6 1:0 r
0 7 6 7
AD
et al., 2001; 1.24, Gea et al., 2004; and 0.86, Adani et al., 2004). 6 1:7 76 7 6 qCH4 7 ð10Þ
6 74 rox 5 ¼ 6 7
For the case of CH4 oxidation activity, the estimated stoichio- 4 0 2:7 1 5 4 qO2 5
r COM 13
metric coefficients were significantly different to the theoretical 31:5 62:0 20:4 qCO2
d C-CO2
values defined by basic chemical stoichiometry where 2 mol of
O2 are consumed per 1 mol of CH4 to produce 1 mol of CO2. In real-
ity, some fraction of carbon is assimilated to build biomass 3.2. Incubations
(Hanson and Hanson, 1996).
In terms of the isotope signatures, the observed values in Table 2 The results of the 11 sets of incubations are presented in Tables
were within the expected ranges for the activity types. For com- 3 and 4. The tables includes (i) the measured production or con-
posting, d13C-CO2 = 20.4 ± 4.0‰. For the CH4 oxidation batch sumption of CH4, CO2 and O2 in the CH4 oxidation and composting
incubations, d13C-CO2 = 62.0 ± 5.0‰ which is within the expected bottles and the supplied amount of CH4 and CO2 from the 50:50%
shift response for methanotrophic behaviour (Whiticar, 1999). The CH4:CO2 bottled gas; (ii) qCH4 , qCO2 , qO2 and d13C-CO2 based on sum-
CO2 arising from CH4 oxidation is typically highly depleted in 13C mations of the three gas sources; (iii) r based on the direct mea-
(Börjesson et al., 2001; Cabral et al., 2010; Chanton et al., 2011; surements from the bottles; (iv) qCH4 , qCO2 , qO2 and d13C-CO2
Liptay et al., 1998a; Widory et al., 2012). based on measurement of the blended gas; and (v) the relative

Please cite this article in press as: Rafiee, R., et al. A mass balance model to estimate the rate of composting, methane oxidation and anaerobic digestion in
soil covers and shallow waste layers. Waste Management (2017), http://dx.doi.org/10.1016/j.wasman.2016.12.025
 R. Rafiee et al. / Waste Management xxx (2017) xxx–xxx 5

Table 3
The gas measurements in 11 incubations.

Incubation set Gas generation/consumption

Supplied CH4:CO2 gas CH4 Oxidation bottles (mmol) Composting bottles
(mmol) (mmol)
CH4 CO2 O2 CO2 CH4 O2 CO2
1 5.83 5.88 1.23 0.44 0.82 0.15 0.16
2 7.57 7.49 0.50 0.17 0.29 1.72 1.61
3 5.74 5.74 0.49 0.24 0.29 3.01 2.38
4 6.52 6.42 2.24 0.75 1.39 1.50 1.53
5 7.03 6.96 0.49 0.20 0.27 0.81 0.77
6 7.56 7.46 1.16 0.42 0.72 0.91 0.91
7 6.41 6.34 2.16 0.95 1.45 0.46 0.51
8 8.33 8.15 0.98 0.30 0.59 1.63 2.48
9 8.98 8.84 0.94 0.31 0.54 0.20 0.23
10 6.27 6.26 1.17 0.62 0.61 1.30 1.05
11 7.13 7.14 0.58 0.22 0.32 0.85 0.91

Table 4
Component gas fluxes based on i) measurement of blended gas; and ii) measurement of individual gas sources prior to blending.

Set no. Flux of gas: summation of three gas sources Flux of gas: measurement of the blended gas Relative error (%)
(mmol/d) (mmol/d)
qCH4 qCO2 qO2 qCO2
d13 C-CO2 qCH4 qCO2 qO2 qCO2
d13 C-CO2 qCH4 qCO2 qO2 qCO2
d13 C-CO2

1 6.48 5.01 1.37 215.7 6.42 5.01 1.58 206.7 0.87 0.06 13.4 4.36
2 9.27 7.28 2.22 279.3 9.02 6.31 2.13 255.2 2.78 15.3 4.33 9.40
3 8.36 5.45 3.50 244.4 8.43 5.46 3.70 239.5 0.81 0.24 5.62 2.03
4 8.70 5.13 3.74 279.9 9.38 4.94 4.07 267.0 7.25 3.85 8.11 4.85
5 7.93 6.76 1.30 247.4 6.99 5.98 1.60 183.2 13.5 13.0 18.8 35.0
6 8.79 6.84 2.07 279.6 7.90 6.75 1.79 231.8 11.3 1.33 15.6 20.6
7 7.80 4.96 2.62 269.0 7.27 4.80 2.41 236.7 7.29 3.33 8.71 13.6
8 10.9 7.74 2.61 325.9 9.86 7.12 2.86 285.5 10.9 8.71 8.74 14.2
9 9.38 8.44 1.14 302.4 8.55 7.71 1.42 260.3 9.71 9.47 19.7 16.2
10 7.93 5.66 2.47 257.1 7.21 5.03 2.54 218.3 9.99 12.5 2.76 17.8
11 8.27 6.81 1.43 257.1 7.45 6.56 1.27 215.6 11.0 3.81 12.6 19.3

error in the measurement of qCH4 , qCO2 and qO2 . The amount of
50:50% CH4:CO2 gas that was supplied in parallel with the com-
posting and CH4 oxidation incubation typically comprised more
than 80% of the total gas blend volume, in order to represent a sys-
tem dominated by anaerobic digestion activity. The reaction rate in
the composting incubations, rCOM, was always greater than the rate
of CH4 oxidation, rOX. Despite identical treatments being applied in
the 11 sets of incubations, rOX and rCOM varied significantly from
one set of incubations to the next. The highest error in the mea-
surement of component net fluxes, based on a comparison with
the summation of consumption or generation of the component
in the individual bottles, was observed in qCO2
d13C-CO2 and qO2 .
3.3. Application of mass balance model
A stepwise approach was used to determine the sensitivity of r
to the stoichiometry of the mass balance equations, matrix A, and
values of the net production term, q. Mathematically, this sensitiv-
ity would be readily identified if the number of components
matched the number of unknown reaction rates and A could be
inverted. A practical approach to illustrate the sensitivity of calcu-
lated values of r to the inputs q and A is to apply the model with
increasing uncertainty in the input values. The series of trials with
the model were as follows:
 Approach 1: qCH4 , qCO2 and qO2 were taken as the sum of the pro-
duction or consumption of each component in the CH4 oxida-
tion bottle, the composting bottle and the supplied 50:50%
CH4:CO2 bottle gas (Column 1–4, Table 4). This approach elim-
inated error in the measurement of the blended gas volume and
composition. The d13C-CO2 value for the blended gas was taken
as the value determined by a linear mixing model applied to the
three gas sources (Eq. (8)).
For matrix A, the stoichiometric coefficients for the CH4 oxida-
tion and composting reactions were based on the gas yields in
the incubation being modelled. Global d13C-CO2 values were
adopted for all three gas sources (Table 2). Given that qCH4 ,
qCO2 and qO2 and the stoichiometric coefficients for CH4, CO2
and O2 in Matrix A were based on the same data, and that
d13C-CO2 in the blended gas was proportional to the sources
of d13C-CO2 according to Eq. (8), it is expected that the model
should exactly predict the experimentally observed values of
r. Some error is expected however because global values of
d13C-CO2 were used rather than values specific to the incuba-
tion being modelled.
 Approach 2: q was calculated as in Approach 1, but the indepen-
dently determined stoichiometries listed in Table 2 were used
for Matrix A.
 Approach 3: Matrix A was calculated as in Approach 2, but the
measured volume and measured composition of the blended
gas were used to calculate q (=qCH4 , qCO2 , qO2 ; and qCO2
d13C-
CO2), including a measurement of d13C-CO2 on each of the
blends from the 11 sets of incubations.
Fig. 2 shows plots of modelled versus measured values for the
reaction rates r for the three approaches. Approach 1 which used
the same data to calculate q and A resulted in good correlations
between the modelled and directly measured values of r (Table 4).
Good agreement between the measured and calculated values of
rAD and rOX were achieved when using Approach 2, where indepen-
dently determined global stoichiometric coefficients for the CH4
oxidation and composting reactions were used in A (Table 5).

Please cite this article in press as: Rafiee, R., et al. A mass balance model to estimate the rate of composting, methane oxidation and anaerobic digestion in
soil covers and shallow waste layers. Waste Management (2017), http://dx.doi.org/10.1016/j.wasman.2016.12.025
6 R. Rafiee et al. / Waste Management xxx (2017) xxx–xxx

Fig. 2. Scatter plot of modelled vs observed values for rAD, rOX and rCOM. Approach 1:4; Approach 2: ; and Approach 3:d.

Table 5
Values of rAD, rOX and rCOM calculated from the 11 sets of incubations. Standard deviations in brackets.

Approach 1 Approach 2 Approach 3

r2 RMSE (mmol CO2/d) NRMSE (%) r2 RMSE (mmol CO2/d) NRMSE (%) r2 RMSE (mmol CO2/d) NRMSE (%)
rAD 1 0.04 0.6 0.95 0.24 3.4 0.86 0.89 12.8
rOX 0.97 0.04 9.1 0.96 0.07 15.9 0.77 0.27 62
rCOM 0.98 0.11 9.3 0.87 0.29 24.5 0.74 0.73 61.9

However, significant deterioration occurred in the fit of rCOM using
Approach 2. The deterioration in evaluation of rCOM shows that,
when compared to other outputs, rCOM is more sensitive to the
evaluated parameters used in matrix A. Using the sum of RMSE
as a basis for fitting r will tend to produce an absolute error of sim-
ilar magnitude for all unknowns (Hossain et al., 2013). In the case
of Approach 2, the RMSE for rAD and rCOM was similar (Table 5), but
the relative error in the fit of rAD was smaller because rAD was
almost an order of magnitude larger than rCOM.
The lack of fit of the model using global stoichiometries for
composting and CH4 oxidation and experimental measurements
of q (Approach 3) are almost entirely attributable to errors in the
measurement of q, as illustrated by model results using Approach
2 and 3. There is considerable error between q values based on
summations of the individual gas sources (Column 1–4, Table 4)
and q based on measurements on the blended gas (Column 5–8,
Table 4).
4. Conclusions
A mass balance approach was proposed to estimate the rates of
anaerobic digestion, CH4 oxidation and composting occurring
simultaneously in environments with a mixture of aerobic chan-
nels and anaerobic pockets. This is relevant to packed beds subject
to an upper atmospheric boundary condition, the configuration
that exists at a landfill. The model was validated by analysing a
blend of gas arising from known flows of gas from composting
and CH4 oxidation incubations, mixed with a known quantity of
50:50% CH4:CO2 gas that represented anaerobic digestion activity.
As expected, the rates of anaerobic digestion, CH4 oxidation and
composting could be accurately calculated if the exact generation
or consumption rates of the component gases from each sources
was used in the model. Error was introduced if measurements of
the blended biogas were used instead to estimate the component
gas fluxes, which then propagated to error in the estimate of the
reaction rates.
The model parameters and mass balance equations can be
updated based on the context of model application. For example,
a mass balance on O2 consumption was included in the current
model. O2 consumption can be measured with acceptable accuracy
in the laboratory but there is no established method for measuring
passive O2 uptake in a landfill. Instead, the O2 mass balance could
be replaced by a mass balance on another isotope such as d13C-CH4.
This model can be used to understand waste degradation pro-
cesses in active landfill cells and in other static bed biodegradation
processes such as windrow composting. Also, the model can be
used to evaluate the impact of design parameters such as depth
of soil cover on the initial phase of waste degradation in a landfill.
Acknowledgments
The research was funded by the Queensland Government
through the Smart Futures Research Partnerships Program, Remon-
dis Pty Ltd and the University of Queensland.

Please cite this article in press as: Rafiee, R., et al. A mass balance model to estimate the rate of composting, methane oxidation and anaerobic digestion in
soil covers and shallow waste layers. Waste Management (2017), http://dx.doi.org/10.1016/j.wasman.2016.12.025
 R. Rafiee et al. / Waste Management xxx (2017) xxx–xxx 7

References Hossain, M.A., Ngo, H.H., Guo, W., 2013. Introductory of microsoft EXCEL SOLVER
function-spreadsheet method for isotherm and kinetics modelling of metals
biosorption in water and wastewater. J. Water Sustainability 3, 223–237.
Abushammala, M.F., Basri, N.E.A., Basri, H., Kadhum, A.A.H., El-Shafie, A.H., 2013.
Kallel, A., Tanaka, N., Tojo, Y., Matsuto, T., Hanada, S., 2006. Oxygen intrusion into
Empirical gas emission and oxidation measurement at cover soil of dumping
waste in old landfills of low organic content. Waste Manage. Res. 24, 242–249.
site: example from Malaysia. Environ. Monit. Assess. 185, 4919–4932.
Lefebvre, X., Lanini, S., Houi, D., 2000. The role of aerobic activity on refuse
Adani, F., Confalonieri, R., Tambone, F., 2004. Dynamic respiration index as a
temperature rise, I. Landfill experimental study. Waste Manage. Res. 18, 444–
descriptor of the biological stability of organic wastes. J. Environ. Qual. 33,
452.
1866–1876.
Liptay, K., Chanton, J., Czepiel, P., Mosher, B., 1998a. Use of stable isotopes to
Atkinson, C.F., Jones, D.D., Gauthier, J.J., 1997. Microbial activities during
determine methane oxidation in landfill cover soils. J. Geophys. Res.: Atmos.
composting of pulp and paper-mill primary solids. World J. Microbiol.
(1984–2012) 103 (D7), 8243–8250.
Biotechnol. 13, 519–525.
Liptay, K., Chanton, J., Czepiel, P., Mosher, B., 1998b. Use of stable isotopes to
Bergmaschi, P., Harris, G.W., 1995. Measurements of stable isotope ratios
determine methane oxidation in landfill cover soils. J. Geophys. Res.: Atmos.
(13CH4/12CH4; 12CH3D/12CH4) in landfill methane using a tuneable diode laser
103 (D7), 8243–8250.
absorption spectrometer. Global Biogeochem. Cycles 9, 439–447.
Nopharatana, A., Pullammanappallil, P.C., Clarke, W.P., 2003. A dynamic
Bogner, J., Spokas, K., 1993. Landfill CH4: rates, fates, and role in global carbon cycle.
mathematical model for sequential leach bed anaerobic digestion of organic
Chemosphere 26, 369–386.
fraction of municipal solid waste. Biochem. Eng. J. 13, 21–33.
Bogner, J.E., Spokas, K.A., Chanton, J.P., 2011. Seasonal greenhouse gas emissions
Röwer, I.U., Geck, C., Gebert, J., Pfeiffer, E.-M., 2011. Spatial variability of soil gas
(methane, carbon dioxide, nitrous oxide) from engineered landfills: daily,
concentration and methane oxidation capacity in landfill covers. Waste
intermediate, and final California cover soils. J. Environ. Qual. 40, 1010–1020.
Manage. 31, 926–934.
Börjesson, G., Chanton, J., Svensson, B.H., 2001. Methane oxidation in two Swedish
Scheutz, C., Pedicone, A., Pedersen, G.B., Kjeldsen, P., 2011. Evaluation of respiration
landfill covers measured with carbon-13 to carbon-12 isotope ratios. J. Environ.
in compost landfill biocovers intended for methane oxidation. Waste Manage.
Qual. 30, 369–376.
31, 895–902.
Cabral, A., Capanema, M., Gebert, J., Moreira, J., Jugnia, L., 2010. Quantifying
Smårs, S., Beck-Friis, B., Jönsson, H., Kirchmann, H., 2001. SE—structures and
microbial methane oxidation efficiencies in two experimental landfill biocovers
environment: an advanced experimental composting reactor for systematic
using stable isotopes. Water Air Soil Pollut. 209, 157–172.
simulation studies. J. Agric. Eng. Res. 78, 415–422.
Chanton, J., Abichou, T., Langford, C., Hater, G., Green, R., Goldsmith, D., Swan, N.,
Whiticar, M.J., 1999. Carbon and hydrogen isotope systematic of bacterial formation
2011. Landfill methane oxidation across climate types in the US. Environ. Sci.
and oxidation of methane. Chem. Geol. 161, 291–314.
Technol. 45, 313–319.
Widory, D., Proust, E., Bellenfant, G., Bour, O., 2012. Assessing methane oxidation
Christophersen, M., Kjeldsen, P., Holst, H., Chanton, J., 2001. Lateral gas transport in
under landfill covers and its contribution to the above atmospheric CO2 levels:
soil adjacent to an old landfill: factors governing emissions and methane
The added value of the isotope (d13C and d18O CO2; d13C and dD CH4) approach.
oxidation. Waste Manage. Res. 19, 595–612.
Waste Manage. 32, 1685–1692.
Gea, T., Barrena, R., Artola, A., Sánchez, A., 2004. Monitoring the biological activity of
Xie, S., O’Dwyer, T., Freguia, S., Pikaar, I., Clarke, W.P., 2016. Effect of biomass
the composting process: oxygen uptake rate (OUR), respirometric index (RI),
concentration on methane oxidation activity using mature compost and
and respiratory quotient (RQ). Biotechnol. Bioeng. 88, 520–527.
graphite granules as substrata. Waste Manage. 56, 290–297.
Gebert, J., Röwer, I.U., Scharff, H., Roncato, C.D., Cabral, A.R., 2011. Can soil gas
Yoshida, H., Gable, J.J., Park, J.K., 2012. Evaluation of organic waste diversion
profiles be used to assess microbial CH4 oxidation in landfill covers? Waste
alternatives for greenhouse gas reduction. Resour. Conserv. Recycl. 60, 1–9.
Manage. 31, 987–994.
Hanson, R.S., Hanson, T.E., 1996. Methanotrophic bacteria. Microbiol. Rev. 60, 439–
471.

Please cite this article in press as: Rafiee, R., et al. A mass balance model to estimate the rate of composting, methane oxidation and anaerobic digestion in
soil covers and shallow waste layers. Waste Management (2017), http://dx.doi.org/10.1016/j.wasman.2016.12.025