Empirical gas emission and oxidation measurementat cover soil of dumping site: example from Malaysia

Mohammed F. M. Abushammala & Noor Ezlin Ahmad Basri &Hassan Basri & Abdul Amir H. Kadhum &

Ahmed Hussein El-Shafie

Received: 15 May 2012 /Accepted: 24 September 2012 /Published online: 2 October 2012# Springer Science+Business Media Dordrecht 2012

Abstract Methane (CH4) is one of the most relevantgreenhouse gases and it has a global warming poten-tial 25 times greater than that of carbon dioxide (CO2),risking human health and the environment. MicrobialCH4 oxidation in landfill cover soils may constitute ameans of controlling CH4 emissions. The study wasintended to quantify CH4 and CO2 emissions rates atthe Sungai Sedu open dumping landfill during the dryseason, characterize their spatial and temporal varia-tions, and measure the CH4 oxidation associated withthe landfill cover soil using a homemade static fluxchamber. Concentrations of the gases were analyzedby a Micro-GC CP-4900. Two methods, kriging val-ues and inverse distance weighting (IDW), were foundalmost identical. The findings of the proposed methodshow that the ratio of CH4 to CO2 emissions was25.4 %, indicating higher CO2 emissions than CH4

emissions. Also, the average CH4 oxidation in the

landfill cover soil was 52.5 %. The CH4 and CO2

emissions did not show fixed-pattern temporal varia-tion based on daytime measurements. Statistically, anegative relationship was found between CH4 emis-sions and oxidation (R200.46). It can be concludedthat the variation in the CH4 oxidation was mainlyattributed to the properties of the landfill cover soil.

Keywords Sungai Sedu . Open dumping landfill .

Methane emission .Methane oxidation . Kriging . IDW

Introduction

Decomposition of municipal solid waste (MSW) inlandfills under anaerobic conditions produces gas con-taining approximately 50–60 %methane (CH4) and 30–40 % carbon dioxide (CO2) by volume. The processtakes place in three stages: hydrolyses and fermentation,acetogenesis, and methanogenesis (Themelis and Ulloa2007). The Intergovernmental Panel on Climate Change(IPCC) estimated that CH4 emissions from landfillscontribute approximately less than 5 % of global CH4

emissions (IPCC 2007). In Malaysia, landfills are themain source of CH4 emissions (53%), followed by palmoil mill effluent (38 %), swine manure (6 %), andindustrial effluent (3 %) (Ministry of Energy, Water,and Communications et al. 2004; Kamarudin 2008).The total CH4 emission from landfills in Malaysia wasestimated to be 318.8 Gg in 2009 (Abushammala et al.2011). CH4 is one of the most relevant greenhouse

Environ Monit Assess (2013) 185:4919–4932DOI 10.1007/s10661-012-2913-5

M. F. M. Abushammala (*) :N. E. A. Basri :H. Basri :A. H. El-ShafieDepartment of Civil and Structural Engineering,Faculty of Engineering and Built Environment,Universiti Kebangsaan Malaysia,43600 Bangi, Selangor, Malaysiae-mail: eng_abushammala@yahoo.com

A. A. H. KadhumDepartment of Chemical and Process Engineering,Faculty of Engineering and Built Environment,Universiti Kebangsaan Malaysia,43600 Bangi, Selangor, Malaysia

gases; it has a global warming potential 25 times that ofCO2, creating health and environmental risks. Further-more, venting of these gases into the atmospherewithout pretreatment or collection may be affectingecosystem balance (Abushammala et al. 2010).

There is no precise method of measuring of landfillgas (LFG) emissions. The static flux chamber methodis the simplest and most frequently used method ofmeasuring LFG emissions (Maurice and Lagerkvist2003; Ishigaki et al. 2005; Chen et al. 2008; Zhanget al. 2008). LFG emission values were highly variedwithin and between landfills because of the multiplefactors that govern the emission process (Zhang et al.2008; Abushammala et al. 2009; Cheng et al. 2010).Field measurements indicated that CH4 emission lev-els range between 0.0004 and more than 4000 gm−2

day−1 (Bogner et al. 1997).There are many methods used in the literature to

estimate mean surface emissions from landfill. Theseinclude calculation of the arithmetic, geometric, geo-spatial, and tributary area means (Spokas et al. 2003;Fourie and Morris 2004; Ishigaki et al. 2005; Abichouet al. 2006a). The geospatial mean is the most precisefigure and has been widely used to estimate total LFGemissions (Spokas et al. 2003; Fourie and Morris2004; Ishigaki et al. 2005; Abichou et al. 2006b; Tecleet al. 2008). There are two common interpolationmethods available in geostatistics: kriging and inversedistance weighting (IDW) methods (Spokas et al.2003). The kriging method is a technique that utilizesthe spatial structures of semivariograms to predict thevalues of a property at an unknown location basedon the relationship found in the sampled locations(Spokas et al. 2003). Semivariogram modeling is aquantitative technique in which the variance betweensampled points decreases as separation distancedecreases. Kriging is used to generate interpolationcontour maps. The IDW method is used to calculatethe interpolation contours by weighing neighboringdata using the inverse of the separation distance to apower (Abichou et al. 2006b). Abichou et al. (2006a)reported approximately identical means of CH4 emis-sions estimated using the kriging and IDW methods.However, Spokas et al. (2003) performed a compari-son between the kriging and IDW methods and foundthat the IDW method is the most precise.

Spatial and temporal variations in LFG emissionsfrom landfill surface is attributed to biological, chem-ical, and physical processes occurring within landfill

cover soils. Microbial CH4 oxidation in landfill coversoils may provide a means of controlling CH4 emis-sions. Several previous studies have reported that theCH4 oxidation process in landfill cover soils can effi-ciently reduce CH4 emissions (Huber-Humer 2004;Stern et al. 2007; Huber-Humer et al. 2008). A valueof 0 to 10 % oxidation is recommended in the IPCCguidelines for national greenhouse gas inventories.However, field studies show that CH4 oxidation ca-pacity ranges from 0 to 64 % (Abichou et al. 2006b).The objectives of this research were to quantify theCH4 and CO2 emissions from the Sungai Sedu opendumping landfill in Malaysia during the dry seasonand to characterize their spatial and temporal varia-tions. Furthermore, the study was intended to mea-sure the CH4 oxidation associated with the landfillcover soil.

Methodology

Site description

The Sungai Sedu open dumping landfill is located inKuala Langat, Selangor, Malaysia. The landfill is sit-uated adjacent to the Sungai Sedu River, lying exactlybetween 2°50′39″ N and 101°31′01″ E (Fig. 1). Thelandfill has a total area of 5.5 ha and received approx-imately 340 t/day of domestic and nonhazardous in-dustrial wastes. The total amount of waste disposed ofin the landfill was estimated to be 1,613,300 t. Thelandfill was operated and managed by Alam FloraSdn. Bhd. The landfill was filled and has not been inuse since 2010.

In the present study, measurements were performedat the inactive part of the landfill, cell number five,where a flat surface was chosen (Fig. 1). Cell numberfive consists of waste that was between 2 and 12 yearsold; the cell started receiving waste in 1998 and closedin 2008. The waste depth of the cell is approximately34 m. The cell is covered by approximately 10 to30 cm of poorly graded sand, which functions as aninterim soil layer. The vegetation was not well estab-lished at some parts of the cell at the time of thestudy, whereas other parts were covered with grassesand herbs. The landfill did not have a gas collectionsystem at that time, and thus, waste gases directlyescaped through the cover soil and were released intothe atmosphere.

4920 Environ Monit Assess (2013) 185:4919–4932

Experimental design

To provide the data necessary to quantify the spatialvariability of the CH4 and CO2 emissions from thestudy surface area using geostatistical analysis, asquare portion of the cell surface of approximately1,000 m2 was overlaid with a 3.5×3.5-m grid to iden-tify the measuring points. The centers of the gridswere marked with wooden sticks to specify the sam-pling point locations (Börjesson et al. 1999; Mulla andMcBratney 2002). The total number of measuringpoints was 80, providing data pairs for modeling thespatial distribution of CH4 and CO2. The number ofsampling points satisfied the minimum criteria (n>20–30; n is number of samples) for unbiased statistics onspatial data as recommended by Livingston andHutchinson (1995). This number was also consistentwith the number of samples used by Abichou et al.(2006b).

A square static flux chamber was constructed tomeasure the CH4 and CO2 emissions (Fig. 2). Thechamber volume was 80 L, whereas the area was0.4 m2 (Stern et al. 2007). The chamber contained adigital temperature module for measuring headspacegas temperature and a small propeller for attainingsufficient gas mixing inside the chamber headspace

(Cheng et al. 2010). Four sequential gas samples wereextracted from the chamber headspace into a 50-mLgastight syringe at the predetermined intervals (5 min).

Based on 20 years of rainfall data (1990–2009) col-lected from the Department of Irrigation and Drainage(Jabatan Pengairan dan Saliran; JPS) in Ampang,Malaysia, February, March, and May were selected asthe driest months at the Sungai Sedu open dumpinglandfill location. Thus, CH4 and CO2 emission measure-ments were performed during February, March, andMay 2011 to analyze the amount of emissions duringthe dry season. Emission measurements were undertak-en from 8:00 am to 11:00 am to minimize the effect oftime. The atmospheric pressure and air temperaturewere monitored during the emission measurements.

Four monitoring stations were chosen randomly foruse in investigating the CH4 oxidation capacity (Fig. 1).The CH4 and CO2 emissions on the surface and the soilgas concentrations profiles (10, 20, and 30 cm) at thosestations were simultaneously measured between 9:00and 11:00 am twice per month from September 2010to June 2011. Soil gas was trapped in preinstalled stain-less steel tubes constructed according to Kiese andButterbach-Bahl (2002) and collected using 10-mL gas-tight syringes for direct analysis. Three main soil gasconcentrations were investigated: the concentrations of

Fig. 1 Aerial view ofthe Sungai Sedu landfillshowing location of thestudy area and monitoringstations

Environ Monit Assess (2013) 185:4919–4932 4921

CH4, CO2, and O2. To monitor the temporal variation inCH4 and CO2 emissions, two of the four monitoringstations were selected randomly for sampling once permonth in the morning (8:00–10:00), at noon (12:00–14:00), and during the afternoon (15:00–17:00) fromSeptember 2010 to June 2011. The air temperature andsoil temperature at 10 cm depth were simultaneouslymeasured with the emission and soil gas profile meas-urements at each monitoring station. To investigate theproperties of the soil at each monitoring station, dis-turbed and undisturbed soil samples were collected froman area near the soil gas collection tubes to minimize theeffect of soil spatial variability, whichmay be significantbeyond 1m (Kennedy et al. 2003). The results of the soilproperties at each monitoring station are shown inTable 1.

Gas concentrations analysis and emission calculation

AVarian Micro-GC (CP-4900) was used to analyze theCH4 and CO2 concentrations from the flux chambersamples and the two gases concentrations besides theO2 concentration from the soil gas samples. Helium gas(99.999 %) was used as the carrier gas for the thermalconductivity detector (TCD) at a pressure of 80 psi.Further details of the Micro-GC are shown in Table 2.

Each gas sample was analyzed at least twice, andthe average was determined (Eklund 1992). CH4 andCO2 emission flux was calculated following Abichouet al. (2006b). Nonzero flux was reported only whenthe regression coefficient (R2) for the linear regression

of the four sequential concentrations over time waslarger than 0.85 (Zhang et al. 2008); otherwise, zeroflux was reported (Abichou et al. 2006b).

Estimation of CH4 oxidation

Surface CH4 and CO2 emissions and soil gas concen-tration profiles at the monitoring stations were testedto estimate percent CH4 oxidation within cover soilsfollowing Christophersen et al. (2001). The CH4 flowrate at the bottom layer of the landfill cover (FinCH4, gm−2day−1) was estimated as given in Eq. 1:

FinCH4 ¼ FoutCH4 þ FoutCO2ð Þ � CinCH4

CinCH4 þ CinCO2

� �ð1Þ

where FoutCH4 and FoutCO2 are the outflow of CH4 and

Fig. 2 In situ static fluxchamber

Table 1 Physicochemical characteristics of cover soils at eachmonitoring station

Parameters StationA

StationB

StationC

StationD

Cover soil type (USCSa) Poorly graded sand

Particle density (mg/m3) 2.61 2.63 2.62 2.57

Moisture content (%dry wt.) 14 9 17 20

pH 8.5 3.7 7.4 8.5

Dry bulk density (mg/m3) 1.79 1.75 1.65 1.8

Porosity (%) 31.4 33.5 37 30

Soil organic matter (%) 0.4 0.2 0.8 0.4

a Unified Soil Classification System

4922 Environ Monit Assess (2013) 185:4919–4932

CO2, respectively, from the landfill surface (gram persquare meter per day). The CinCH4 and CinCO2 are CH4

and CO2 concentrations (percent v/v), respectively, inthe bottom layer of landfill cover. The difference be-tween CH4 flow rate from the landfill surface and thatat the bottom layer of the landfill cover was the CH4

oxidation rate (Rox, gram per square meter per day)(Eq. 2). The percent CH4 oxidation (%OX) was esti-mated as given in Eq. 3:

Rox ¼ FinCH4 � FoutCH4 ð2Þ

%OX ¼ Rox

FinCH4

� �� 100: ð3Þ

Geospatial analysis

In the current study, 80 sample points were used toquantify the CH4 and CO2 emissions. A total of 67points were analyzed in the Surfer 8 software to gen-erate experimental variograms for the CH4 and CO2

emissions. The remaining 13 emission measurementswere excluded from the geostatistics analysis andanalyzed separately because they exhibited highlyskewed distributions, especially in the areas with dif-ferent cover soil thickness and in the vegetated areas.Linear models were fitted to the experimental vario-grams and tested in different directions. The experi-mental variograms for both the CH4 and the CO2

emissions were shown to have different slopes indifferent directions (anisotropy). Therefore, the anisot-ropy angle and ratio were estimated for each model.The Auto Fit button was used to fine-tune the modelusing the least-squares method. Thus, the y-axis

intercept was used as the nugget value (the errorvariance and the micro variance) for each model.The key variables for a linear variogram model, in-cluding the nugget, slope, and anisotropy angle andratio, were used with the point kriging method todevelop contour maps for both the CH4 and CO2

emissions. When the IDW method was employed, apower of two was used, and the anisotropy angle andratio were set to be equal to those resulting fromvariograms modeling for both the CH4 and the CO2

emissions (Abichou et al. 2006b). The net volumes ofthe total CH4 and CO2 emissions from the entire areawere estimated by subtracting the volume of the neg-ative contour emissions from that of the positive con-tour emissions. The geospatial means for the CH4 andCO2 emissions were then calculated by dividing thenet emission volume by the total area. To provide addi-tional estimates of the total emission rates, geometricand arithmetic means were estimated and multiplied bythe total surface area.

Analytical methods

Atmospheric pressure and air temperature were mea-sured using Skymaster, SM-28, and Speedtech instru-ments. The soil temperature at a depth of 10 cm wasmeasured using a Thermometer Dual Channel, TypeK, Monarch 306 connected with two K wire thermo-couple probes (Monarch Instrument, Columbia, USA).Three soil temperature measurements were performed,and the average was determined. Soil sieve analysis,particle density, moisture content, bulk density, pH,and organic matter were performed following Head(1992). The correlation analysis using Pearson's cor-relation coefficients was implemented to measure thestrength of the relationship between air and soil tem-peratures with the CH4 and CO2 emissions during thetemporal investigation. The statistical analysis wasperformed using SPSS 16.0.

Results

CH4 and CO2 flux measurements

The dates of the CH4 and CO2 emission measure-ments, the ambient temperature readings, and themean atmospheric pressure readings are recorded inTable 3. Nine measuring points were sampled during

Table 2 Representative parameters for CH4, CO2, and O2

measurements

Gas chromatography Varian Micro-GC CP-4900

Column MS5Å,10 m

PPQ,10 m

Detector TCDa TCD

Oven temperature (°C) 80 40

Column head pressure (kPa) 150 150

Carrier and reference gas Helium Helium

Detection of O2 CH4 and CO2

a Thermal conductivity detector

Environ Monit Assess (2013) 185:4919–4932 4923

each sampling date. The summary and descriptive sta-tistics for the CH4 and CO2 flux measurements are listedin Table 4. The numbers of flux measurements in thefirst column of Table 4 exclude the hot spot fluxes; thesecond column includes the hot spot measurements.

Spatial variability of CH4 and CO2 emissions

Spatial structures were evaluated by fitting linear modelsto directional experimental variograms generated fromthe emission measurements (Fig. 3). The spatial struc-tures of theCH4 andCO2 emissions are shown in Table 5.The nugget to total semivariance ratio for the CH4 andCO2 emissions were 0.42 and 0.47, respectively, whichindicated moderate spatial dependence according toSpokas et al. (2003). The nonzero nugget value inTable 5 indicates that spatial variability at smaller distan-ces than our sampling distance results from micro anderror variances. The contour maps generated for bothCH4 and CO2 emission data using the point kriging andIDW methods are shown in Figs. 4 and 5, respectively.

The geospatial means, cross-validation residual meansquares, and model residual mean squares derived usingthe kriging and IDW methods for CH4 and CO2 emis-sions are shown in Table 5. The cross-validation residualwas performed by removing an observation from thedata set and using the remaining data set and the modelto predict the removed value. The difference betweenthe observed and predicted values is the cross-validationresidual (Surfer 2002). The difference between themeasurements and the model values is termed the modelresidual. A comparison of the cross-validation residualmean squares obtained using the two methods showsthat the IDW method yielded a smaller degree of error

than the kriging method. Therefore, CH4 and CO2 emis-sions were estimated using the geospatial mean obtainedvia the IDW method.

Table 6 shows the results for the total CH4 and CO2

emissions estimated using the total surface area (squaremeter) with the arithmetic, geometric, and geospatialmeans (gram per square meter per day) taken from 67measuring points, including summation of the hot spotemission values.

Soil gas concentration profiles

The composition of the LFG samples collected fromthe preinstalled tubes at four monitoring stations pro-vides a snapshot of the gas concentrations with thecover soil profile. Figure 6 shows the mean gas con-centration profiles obtained in this study at the fourmonitoring stations.

All monitored stations had depleted O2 concentra-tions from the top of the cover soil to the bottom,whereas the CH4 and CO2 were depleted from thebottom to the top. The O2 concentrations at a depthof 30 cm for A, B, C, and D stations were 70.6, 76.3,72.1, and 45.5 % less, respectively, than the concen-trations at a 10 cm depth. The CH4 concentrations at

Table 3 Date and weather condition during emission sampling

Sampling date Ambienttemperature (°C)

Mean barometricpressure (mbar)

03 February 2011 28–35 1,011

08 February 2011 27–37 1,008

17 February 2011 25–35 1,010

25 February 2011 27–37 1,008

08 March 2011 26–33 1,009

18 March 2011 28–37 1,008

09 May 2011 28–36 1,008

10 May 2011 27–34 1,010

24 May 2011 28–32 1,008

Table 4 Descriptive statistics of CH4 and CO2 emission data(gm−2day−1)

Properties na067 nb080

Methane flux

Minimum −1.48 −1.48Median 1.32 1.749

Maximum 74.92 1,645.94

Mean 6.93 88.50

Standard deviation (s) 14.95 288.33

Geometric mean (meang) 2.64 7.58

Geometric standard deviation (sg) 2.57 6.07

Carbon dioxide flux

Minimum 0 0

Median 28.17 32.96

Maximum 245.85 2,578.01

Mean 40.89 170.26

Standard deviation (s) 40.74 457.56

Geometric mean (meang) 28.21 46.13

Geometric standard deviation (sg) 2.45 4.08

a Number of flux measurements exclude the hot spot fluxesb Number of flux measurements include the hot spot fluxes

4924 Environ Monit Assess (2013) 185:4919–4932

all monitoring stations (A to D) decreased by 89.5,56.2, 99.5, and 64.1 %, respectively, from bottom totop. The CO2 concentrations decreased from bottom totop by 77.7, 54.2, 63.8, and 32.9 %, respectively, fromstations A to D.

CH4 oxidation capacity

CH4 oxidation capacity was estimated at each moni-toring station in the area under study. The CH4

oxidation ranged from 0.5 to 97.9 % with a mean of52.5 % and 30.6 standard deviations at all monitoringstations (Table 7). The mean CH4 oxidation values forstations A, B, C, and D are shown in Table 7. The twomeans for stations A and C were almost identical,whereas those for stations B and D showed a lowercapacity for CH4 oxidation. The relationship betweenthe CH4 emissions and the oxidation at all monitoringstations is shown in Fig. 7, with a negative relationship(R200.46) at all monitoring stations.

Temporal variation of CH4 and CO2 emissions

The temporal variations in CH4 and CO2 emissions atstations B and D and the air and soil temperatures areshown in Fig. 8. There was no fixed pattern in emis-sions during the daytime at either station. The statisti-cal analysis using Pearson's correlation showed that

(a)

(b)

0 2 4 6 8 10 12

Lag Distance (m)

Lag Distance (m)

0

50

100

150

200

250

300

Var

iogr

amV

ario

gram

Direction: 120.0 Tolerance: 30.0

Direction: 120.0 Tolerance: 30.0

0 2 4 6 8 10 120

500

1000

1500

2000

2500

3000

Fig. 3 Linear models fitting to directional experimental vario-grams from emission data of a CH4 and b CO2

Table 5 Kriging and IDW parameters for interpolating spatialCH4 and CO2 emission data

Parameters Kriging IDW

Methane flux

Model Linear model withnugget effect

Inverse distancesquared

Nugget 82.6 N/A

Slope 4.19 N/A

Anisotropy angle (°) 133.7 133.7

Anisotropy ratio 2.0 2.0

Geospatial mean(gm−2day−1)

7.35 6.95

Cross-validation residualmean square

216.09 196.75

Model residual meansquare

97.60 0.21

Carbon dioxide flux

Model Linear model withnugget effect

Inverse distancesquared

Nugget 1,120 N/A

Slope 25.4 N/A

Anisotropy angle (°) 136.0 136.0

Anisotropy ratio 2.0 2.0

Geospatial mean(gm−2day−1)

42.88 41.77

Cross-validation residualmean square

2,030.22 1,617.19

Model residual meansquare

1,087.06 2.62

N/A not available

Environ Monit Assess (2013) 185:4919–4932 4925

there was no significant correlation between the CH4

emissions at station B with air temperature (r00.173,p00.378) and soil temperature (r00.005, p00.979)and also no significant correlation between the CO2

emissions with air temperature (r00.234, p00.231)and soil temperature (r00.093, p00.639). There wasno significant correlation between the CH4 emissionsat station D with air temperature (r00.290, p00.215)and soil temperature (r00.310, p00.183) and alsono significant correlation between the CO2 emissions

with air temperature (r00.325, p00.163) and soil tem-perature (r00.413, p00.07).

Discussion

The CH4 emissions ranged from −1.48 to 1,645.94 gm−2

day−1. The negative CH4 fluxes indicate that soil coverconsumes atmospheric CH4. The CO2 emissions rangedfrom zero to 2,578.01 gm−2day−1. The maximum CH4

(a)

-20246810121416182022242628303234

(b)

-5

0

5

10

15

20

25

30

35

40

45

50

55

60

65

70

75

Fig. 4 CH4 emission (gm−2day−1) contours obtained using a kriging and b IDW

4926 Environ Monit Assess (2013) 185:4919–4932

emission value found in this study was almost similar tothat found by Abichou et al. (2006b). The variation in theCH4 and CO2 emissions through the landfill surface wasattributed to variation in the CH4 oxidation activity, soilcover thickness, underlying waste composition, and hotspot emission measurements. The hot spot measure-ments exhibited higher emission values for both CH4

and CO2 emissions. These may be a result of the

variations in the thickness of the cover soil, as someareas may have been poorly covered, with garbagespread out on the surfaces. However, the hot spot emis-sions might also be a function of the cracks in the coversoil that resulted from the underlying waste settlements.Nevertheless, vegetated soils can improve air capacity byabsorbing soil water through their roots, thus enhancingCH4 oxidation capacity. According to Bohn and Jager

(a)

20

24

28

32

36

40

44

48

52

56

60

64

68

72

76

(b)

0102030405060708090100110120130140150160170180190200210220230240250

Fig. 5 CO2 emission (gm−2d−1) contours obtained using a kriging and b IDW

Environ Monit Assess (2013) 185:4919–4932 4927

(2009), the CH4 oxidation rate can be at least 50 %higher in vegetated soils. Therefore, the vegetated mea-suring points showed very low CH4 emissions and zeroCH4 emissions in most cases, whereas the CO2 emis-sions were much higher.

The nugget values in Table 5 indicated the moderatespatial dependence of both the CH4 and the CO2 emis-sions. The anisotropy angles for CH4 and CO2 emis-sions were 133.7° and 136°, respectively (Table 5). Thisanisotropy indicates that CH4 emissions were moresimilar in the 133.7° direction (W43.7°S). The CO2

emissions exhibited the same trend as the CH4 emis-sions; the direction of the CO2 emissions was W46°S.This is shown in Figs. 4 and 5, in which it is clear thatthe W43.7°S direction showed relatively similar CH4

observations (Fig. 4), while W46°S direction showedrelatively similar CO2 emissions (Fig. 5).

The IDW method did provide better cross-validationresiduals for both the CH4 and the CO2 emission datathan the kriging method (Table 5). Consequently, itprovides a better model for the emission data set. Theseresults are consistent with those of Spokas et al. (2003)and Abichou et al. (2006b). The geospatial means of theCH4 and CO2 emissions generated using the point

Table 6 Estimation of total CH4 and CO2 emissions usinggeospatial, arithmetic, and geometric means

Method Total emission (kgday−1)

Methane flux

Arithmetic mean 13.31

Geometric mean 9.05

Geospatial mean IDW 13.3

Carbon dioxide flux

Arithmetic mean 51.45

Geometric mean 38.9

Geospatial mean IDW 52.3

Station 'A'

Depth (cm)

0 10 20 30 40

Con

cent

ratio

n (%

v/v)

0

5

10

15

20

25

30

35

40

45

50

55

CH4

CO2

O2

Station 'B'

Depth (cm)

0 10 20 30 40

Con

cent

ratio

n (v

/v)

0

5

10

15

20

25

30

35

40

45

50

55

CH4

CO2

O2

Station 'C'

Depth (cm)

0 10 20 30 40

Con

cent

ratio

n (%

v/v)

0

5

10

15

20

25

30

35

40

45

50

CH4CO2

O2

Station 'D'

Depth (cm)

0 10 20 30 40

Con

cent

ratio

n (%

v/v)

0

5

10

15

20

25

30

35

40

45

50

55

60

65

CH4

CO2

Fig. 6 Mean concentrations of CH4, CO2, and O2 with error bars representing positive standard deviation; n (number of observa-tions)019, 48, 19, and 48 for stations A, B, C, and D, respectively

4928 Environ Monit Assess (2013) 185:4919–4932

kriging and IDW methods were almost identical, andtheir arithmetic means were approximately equal. Thisis consistent with the results of Abichou et al. (2006b)which indicated that the kriging, IDW, and arithmeticmeans were almost identical. The geospatial means ofthe CH4 emissions in this research were approximately2.6 times higher than the geometric means, whereas thegeospatial mean for CO2 was only 1.5 times higher thanthe geometric mean. Abichou et al. (2006b) found thatthe geospatial means derived using the point kriging andIDW methods were 3.4 times higher than the geometricmean. However, Spokas et al. (2003) reported that theemissions estimated using the IDW method were twicethe arithmetic mean and 45 times higher than the geo-metric mean. The total CH4 emissions estimated usingthe IDW method for the entire study area were 13.3 kgday−1, whereas the total CO2 emissions were 52.3 kgday−1 (Table 6). The CH4 to CO2 ratio was 25.4 %, and

the higher CO2 emissions (as compared with the CH4

emission figures) is attributed to the CH4 oxidationactivity in the cover soil, plants, and CO2 content ofthe LFG.

The shapes of the soil gas concentration profiles atall monitoring stations were similar to those shown inother researchers' results (Abichou et al. 2006b; Zhanget al. 2008). The CH4 concentrations at all monitoringstations from A to D decreased by 89.5, 56.2, 99.5,and 64.1 %, respectively, from bottom (30 cm) to top(10 cm). This finding is consistent with the mean CH4

oxidation results at all monitoring stations (Table 7),where station C exhibited the highest CH4 oxidation,followed by stations A, D, and B. The steepest declinein the CH4 concentration indicated the optimum CH4

oxidation zone; however, the overlap between gradientof the CH4 and O2 concentration profiles in the coversoil occurred at the maximum level of CH4 oxidation,and the depth of the overlap denotes the optimumdepth for maximizing CH4 oxidation. In this study,the optimum depths for CH4 oxidation at stations Aand C were 13.5 and 22 cm, respectively. The depthrange from 20 to 30 cm was the optimum oxidationdepth at station B, whereas the depth range from 10 to20 cm was the optimum oxidation depth for station D.

The mean CH4 oxidation found in this study at thefour monitoring stations was 52.5 %. Station Cexhibited the highest CH4 oxidation at 71.8 %, fol-lowed by stations A, D, and B with CH4 oxidationmeans of 70.7 %, 49.8 %, and 17.6 %, respectively.The variations between the CH4 oxidation levels at the

Table 7 CH4 oxidation results (%) from all monitoring stations

Station StationA

StationB

StationC

StationD

Allstations

na 16.0 39.0 9.0 36.0 100

Maximum 97.9 91.2 95.7 92.5 97.9

Minimum 34.1 0.5 9.5 22.5 0.5

Mean 70.7 17.6 71.8 49.8 52.5

sb 17.2 23.7 28.8 19.9 30.6

a Number of observationsb Standard deviation

R2 = 0.46

CH4emission (g m-2 d-1)

0 10 20 30 40 50 60 70 80 90 100 110 120

CH

4oxid

atio

n (%

)

-30

-20

-10

0

10

20

30

40

50

60

70

80

90

100

110

Fraction oxidized (%)

Fig. 7 CH4 emissionsversus CH4 oxidation at allmonitoring stations

Environ Monit Assess (2013) 185:4919–4932 4929

different monitoring stations might be attributed to thesoil properties reported in Table 1. The cover soilswere acidic at station B (3.7), basic at station A andD (8.5), and neutral at station C (7.4). The soil poros-ity at station C was higher than those at other stations.The organic matter at station C was two times higherthan the organic matter at stations A and D and four

times higher than the organic matter at stationB. Accord-ing to Kightley et al. (1995), soils rich in organic matterproduced more CH4 oxidation than those with lowerorganic matter. Organic matter plays a central role byproviding good soil structure and supplying nutrients(nitrogen, phosphate, and sulfur), through its water hold-ing capacity, and by improving ion exchange capacity in

Station 'B'

Time10

:3013

:3016

:3010

:2513

:2516

:2509

:3512

:3316

:3509

:3012

:3015

:3009

:3212

:3415

:3709

:0012

:0015

:0009

:0512

:3515

:3308

:3012

:3815

:4008

:2512

:3115

:3908

:2312

:39

Em

issi

on (

mol

/m2 /h

)

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

Air

and

soi

l tem

pera

ture

(o C

)

0

10

20

30

40

50

Date

01-Sep-10 01-Nov-10 01-Jan-11 01-Mar-11 01-May-11 01-Jul-11

CH4

CO2

Air temp.Soil temp.

Station 'D'

Time09

:3012

:3215

:3309

:2012

:2715

:3109

:0512

:3316

:0009

:0812

:0015

:0509

:1012

:1515

:2008

:3012

:0815

:1808

:3312

:1315

:1908

:2612

:0415

:2108

:1012

:0715

:2208

:2312

:11

Em

issi

on (

mol

/m2 /h

)

0.0000

0.0025

0.0050

0.0075

0.0100

0.0125

0.0150

0.0175

0.0200

0.0225

0.0250

0.0275

0.0300

Air

and

soi

l tem

pera

ture

(o C

)

10

20

30

40

50

Date01-Sep-10 01-Nov-10 01-Jan-11 01-Mar-11 01-May-11 01-Jul-11

CH4

CO2

Air temp.Soil temp.

Fig. 8 Temporal variabilityof CH4 and CO2 emissionsat stations B and D

4930 Environ Monit Assess (2013) 185:4919–4932

tropical soils (Ross 1993). The soil pH, porosity, andorganic matter results are consistent with the fact that themean CH4 oxidation at station C was higher than at anyother station, whereas station B exhibited the lowestmean CH4 oxidation. The negative relationship foundbetween the CH4 emission and oxidation (Fig. 7) indi-cates that CH4 oxidation controls the CH4 emission fromthe landfill. This was in agreement with other research-ers' results (Abichou et al. 2006b; Stern et al. 2007;Zhang et al. 2012). The weak of relationship is probablydue to the fact that the data were collected from differentmonitoring stations.

Conclusions

This study quantified the CH4 and CO2 emission ratesand CH4 oxidation capacity of the Sungai Sedu opendumping landfill in Malaysia. Averages CH4 and CO2

emissions measured in this research were consistentwith the other results found in the literature withhigher CO2 emissions than CH4. Negative CH4 emis-sions indicated that the cover soil consumed atmo-spheric CH4. Spatial variations in CH4 and CO2

emissions were evident at the study area. These spatialvariations may be attributed to variation in LFG gen-eration or in the thickness and other properties of thecover soil which mainly affect microbial CH4 oxida-tion process. The IDW method was considered toprovide a more precise mode than the kriging method;the cross-validation residual mean square was lower inthe IDW than that obtained using the kriging method.It is recommended that future studies consider moremeasurements at small separation distances than thosefound in the present study to more reliably evaluatethe variogram model. A negative relationship wasfound between CH4 emissions and oxidation (R20

0.46), indicating the importance of CH4 oxidationprocess in controlling CH4 emissions. Throughoutthe temporal measurements, it was observed thatthe CH4 and CO2 emissions did not show a fixed-pattern temporal variation or a correlation with air andsoil temperatures.

Acknowledgments This work was financed by UniversitiKebangsaan Malaysia (UKM) under research grant UKM-GUP-ASPL-08-06-208. The authors gratefully acknowledgethe Alam Flora Company staff for their assistance with thisresearch and Tuwati B. Tuwati for his assistance with fieldsampling.

References

Abichou, T., Chanton, J., Powelson, D., Fleiger, J., Escoriaza,S., Lei, Y., & Stern, J. (2006). Methane flux and oxidationat two types of intermediate landfill covers. Waste Man-agement, 26, 1305–1312.

Abichou, T., Powelson, D., Chanton, J., & Escoriaza, S. (2006).Characterization of methane oxidation at a solid wastelandfill. Journal of Environmental Engineering, 132,220–229.

Abushammala, M. F. M., Basri, N. E. A., Basri, H., & Kadhum,A. A. H. (2009). Review on landfill gas emission to theatmosphere. European Journal of Scientific Research, 30,427–436.

Abushammala, M. F. M., Basri, N. E. A., Basri, H., Kadhum, A.A. H., & El-Shafie, A. H. (2010). Estimation of methaneemission from landfills in Malaysia using the IPCC 2006FOD model. Journal of Applied Sciences, 10, 1603–1609.

Abushammala, M. F. M., Basri, N. E. A., Basri, H., El-Shafie,A. H., & Kadhum, A. A. H. (2011). Regional landfillsmethane emission inventory inMalaysia.WasteManagement& Research, 29, 863–873.

Bogner, J., Meadows, M., & Czepiel, P. (1997). Fluxes of CH4

between landfills and the atmosphere: Natural engineeredcontrol. Soil Use and Management, 3, 268–277.

Bohn, S., & Jager, J. (2009). Microbial methane oxidation inlandfill top covers-process study on an MBT landfill. In:Proceedings from Sardinia 2009, Twelfth InternationalWaste Management and Landfill Symposium S. Margheritadi Pula, Cagliari, Italy; 5–9 October 2009.

Börjesson, G., Danielsson, Å., & Svensson, B. H. (1999).Methanefluxes from Swedish landfills. Final report by Department ofWater and Environmental Studies, University of Linköping(in Linköping) and IVL Swedish Environmental ResearchInstitute (in Stockholm), Printed by Swedish EnvironmentalProtection Agency, Stockholm, Sweden.

Chen, I. C., Hegde, U., Chang, C. H., & Yang, S. S. (2008).Methane and carbon dioxide emissions from closed landfillin Taiwan. Chemosphere, 70, 1484–1491.

Cheng, X., Luo, Y., Xu, Q., Lin, G., Zhan, Q., Chen, J., & Li, B.(2010). Seasonal variation in CH4 emission and its 13C-isotopic signature from Spartina alterniflora and Scirpusmariqueter soils in an estuarine wetland. Plant and Soil,327, 85–94.

Christophersen, M., Kjeldsen, P., Holst, H., & Chanton, J.(2001). Lateral gas transport in soil adjacent to an oldlandfill: Factors governing emissions and methane oxida-tion. Waste Management & Research, 19, 595–612.

Eklund, B. (1992). Practical guidance for flux chamber meas-urements of fugitive volatile organic emission rates. Air &Waste Management Association, 42, 1583–1591.

Fourie, A. B., & Morris, J. W. F. (2004). Measured gas emissionsfrom four landfills in South Africa and some implications forlandfill design and methane recovery in semi-arid climates.Waste Management & Research, 22, 440–453.

Head, K. H. (1992). Manual of soil laboratory testing (2nd ed.,Vol. 1). London: Pentech.

Huber-Humer, M. (2004). International research into landfill gasemissions and mitigation strategies-IWWG working group“CLEAR”. Waste Management, 24, 425–427.

Environ Monit Assess (2013) 185:4919–4932 4931

Huber-Humer, M., Gebert, J., & Hilger, H. (2008). Biotic systemsto mitigate landfill methane emissions.WasteManagement &Research, 26, 33–46.

IPCC (Intergovernmental Panel on Climate Change). (2007).Climate change 2007: working group III: mitigation ofclimate change. Cambridge: Cambridge University Press.

Ishigaki, T., Yamada, M., Nagamori, M., Ono, Y., & Inoue, Y.(2005). Estimation of methane emission from whole wastelandfill site using correlation between flux and groundtemperature. Environmental Geology, 48, 845–853.

Kamarudin, W. N. B. (2008). The CDM/sustainable energymarket in Malaysia. Malaysian Energy Center (PTM),Kuala Lumpur, Malaysia, Copenhagen, Denmark.

Kennedy, J., Keefer, T., Paige, G., & Barnes, E. (2003). Evalu-ation of dielectric constant-based soil moisture sensors in asemiarid rangeland. In: Proceedings of the First InteragencyConference on Research in the Watersheds (ICRW), 27–30October 2003, Benson, AZ.

Kiese, R., & Butterbach-Bahl, K. (2002). N2O and CO2 emis-sions from three different tropic forest sites in the wettropics of Queensland, Australia. Soil Biology and Bio-chemistry, 34, 975–987.

Kightley, D., Nedwell, D. B., & Cooper, M. (1995). Capacity formethane oxidation in landfill cover soils measured inlaboratory-scale soil microcosms. Applied and Environ-mental Microbiology, 61, 592–601.

Livingston, G. P., & Hutchinson, G. L. (1995). Enclosure-basedmeasurement of trace gas exchange: applications and sourcesof error. In P. A. Matson & R. C. Harriss (Eds.), Biogenictrace gases: measuring emissions from soil and water(pp. 14–51). Cambridge: Blackwell.

Maurice, C., & Lagerkvist, A. (2003). LFG emission measure-ments in cold climatic conditions: Seasonal variations andmethane emissions mitigation. Cold Regions Science andTechnology, 36, 37–46.

Ministry of energy, water and communications, Malaysia energycenter, Danida (2004). Study on clean development mecha-nism potential in the waste sectors in Malaysia. RenewableEnergy & Energy Efficiency Component (Sub-ComponentIII: CDM Action Plan). Final report activity 1.3 (a).

Mulla, D. J., & McBratney, A. B. (2002). Soil spatial variability.In A. W. Warrick (Ed.), Soil physics companion (pp. 343–373). Boca Raton: CRC Press.

Ross, S. M. (1993). Organic matter in tropical soils: Currentconditions, concerns and prospects for conservation. Progressin Physical Geography, 17, 265–305.

Spokas, K., Graff, C., Morcet, M., &Aran, C. (2003). Implicationsof the spatial variability of landfill emission rates on geo-spatial analyses. Waste Management, 23, 599–607.

Stern, J. C., Chanton, J., Abichou, T., Powelson, D., Yuan, L.,Escoriza, S., & Bogner, J. (2007). Use of biologicallyactive cover to reduce landfill methane emissions andenhance methane oxidation. Waste Management, 27,1248–1258.

Surfer. (2002). Surfer 8 user's guide. Golden: Golden Software, Inc.Tecle, D., Lee, J., & Hasan, S. (2008). Quantitative analysis of

physical and geotechnical factors affecting methane emissionin municipal solid waste landfill.Environmental Geology, 56,1135–1143.

Themelis, N. J., & Ulloa, P. A. (2007). Methane generation inlandfills. Renewable Energy, 32, 1243–1257.

Zhang, H., He, P., & Shao, L. (2008). Methane emissions fromMSW landfill with sandy soil covers under leachate recir-culation and subsurface irrigation.Atmospheric Environment,42, 5579–5588.

Zhang, G., Ji, Y., Ma, J., Xu, H., Cai, Z., & Yagi, K. (2012).Intermittent irrigation changes production, oxidation, andemission of CH4 in paddy fields determined with stablecarbon isotope technique. Soil Biology and Biochemistry,52, 108–116.

4932 Environ Monit Assess (2013) 185:4919–4932