United States
                     Environmental Protection
                     Agency
Air and Energy Engineering
Research Laboratory
Research Triangle Park NC 27711
                     Research and Development
EPA/600/S2-91/055 Dec. 1991
&EPA       Project Summary
                     Analysis  of  Factors Affecting
                     Methane  Gas  Recovery from  Six
                     Landfills
                     Darcy Campbell, David Epperson, Lee Davis,
                     Rebecca Peer, and Walter Gray
                        In 1990, EPA's Air and Energy Engi-
                     neering Research Laboratory (AEERL)
                     began a research program with the goal
                     of improving global landfill  methane
                     (CH4)  emissions estimates. Part of
                     AEERL's program Includes developing
                     a field testing program to gather data to
                     identify key variables that affect  CH4
                     generation and to develop an empirical
                     model of CH4 generation based on those
                     variables. The first step in developing
                     the field testing program was a pilot
                     study of six U.S. landfills that have CH,
                     gas recovery systems. In order to evalu-
                     ate the effects of climate on  CH4 pro-
                     duction and recovery, the sites were
                     chosen to represent a variety of mois-
                     ture and temperature patterns (I.e., hot
                     and wet, cool and wet, hot and dry).
                     Landfill gas was tested at each of the
                     six landfills In  order to evaluate the
                     quality of the gas recovery data avail-
                     able at each site. The testing program
                     included assessing the adequacy of on-
                     site instrumentation and scanning the
                     landfill surfaces for organic vapors that
                     would Indicate emissions of CH4. In ad-
                     dition, information on waste composi-
                     tion and  landfill  characteristics was
                     sought for each landfill. Except for flow
                     measurements, the test procedures se-
                     lected for this project were well suited
                     to the types of gas recovery installa-
                     tions at the landfills visited. Based on
                     comparisons between EPA Reference
                     Method 3C and  instrument analyses of
                     the landfill gas composition, all on-site
                     analysis instruments appeared to be
                     operating with  reasonable accuracy.
                     Reviews of calibration procedures and
                     records indicate that long-term instru-
                     ment accuracy  should be comparable
                     to the accuracies noted during on-site
 testing. A negative correlation between
 refuse age and CH4 recovery per ton
 was found; a weak positive correlation
 was found for normal annual precipita-
 tion and CH4 recovery per ton. The re-
 sults of this pilot study are sufficiently
 encouraging to  warrant further data
 gathering and analyses.
    This  Project Summary  was  devel-
 oped by EPA's Air and Energy Engi-
 neering Research Laboratory, Research
 Triangle Park, NC, to announce key find-
 Ings of the research project that Is fully
 documented In a separate report of the
 same title (see Project Report ordering
 Information at back).

 Introduction
    In response to concerns about global
 warming, the U.S. Environmental Protec-
 tion Agency's (EPA's) Office of Research
 and Development (ORD) has initiated  a
 program  to characterize the  causes and
 effects of global climate change,  and to
 identify and quantify emission sources of
 greenhouse gases. To assist in this under-
 taking, EPA's Air and Energy Engineering
 Research Laboratory (AEERL) has begun
 research to improve emissions inventories
 of greenhouse gases in  the U.S.  and
 throughout the world.
    One greenhouse gas of particular con-
 cern is methane (CH4). Methane's radia-
 tive-forcing potential is thought to be much
 greater than that of carbon dioxide (C02).
 In 1990, AEERL began a research  pro-
 gram with the goal of  improving global
 landfill CH4 emissions estimates.
    To determine  the factors that affect
 CH4 generation in landfills on a global ba-
 sis, a model  is needed that is responsive
 to a wide range of climates and types of
 waste. Part of AEERL's program to create

-------
 a CH4 landfill emissions database, there-
 fore, includes  developing  a field testing
 program to gather data to: (I) identify key
 variables that affect CH4 generation; and
 (2)  develop  an empirical  model  of  CH4
 generation based on those variables.
    Landfills  with gas recovery systems
 were chosen for this study. The landfill gas
 is being collected and measured by the
 gas recovery operators; if those data can
 be verified to be reasonably accurate, and
 if sufficient data are available on the land-
 fill  itself,  the landfill  gas measurements
 collected over several years may be used
 to estimate total CH4 generation.
    The first  step in developing the field
 testing  program was a pilot study of six
 U.S. landfills that have landfill gas recov-
 ery  systems. The objectives of the pilot
 study were to:
  (1)  Determine the type and quality of
      landfill  data on landfill gas recovery
      rates, gas composition, and refuse
      characteristics available at landfills
      with gas recovery systems;
  (2)  Use these data to determine trends
      in the effects of climate, refuse age,
      and landfill characteristics on landfill
      gas recovery;  and
  (3)  Use the results of the emissions test-
      ing and data analysis to assess the
      relationship between gas  recovery
      and gas generation, and the feasibil-
      ity of expanding the study to include
      other sites.
   To meet these objectives, a pilot study
 of six sites, chosen to represent a range of
 climates, was undertaken. The general pro-
 cedures and methodologies planned were:
  (1)   Identify potential sites;
  (2)   Visit  the landfills  to  collect  data
      records from the facility;
  (3)   Independently measure landfill  gas
      flow;
  (4)  Assess accuracy and adequacy of
      the data; and
  (5)   Develop statistical methods for analy-
      sis of the data.
   Although the CH4 content of the landfill
 gas  is of most  importance at this time,
 other constituents were also measured (car-
 bon   dioxide,  oxygen,  nitrogen,  and
 nonmethane organic compounds).

 Site Selection and Description
   The  pilot  study  included visits to six
 landfills  in the U.S. to gather data on CH4
 recovery rates and factors thought to influ-
 ence these rates. The primary criterion in
 selecting a landfill for study was that it
 have a gas recovery system in place. The
 recovery system needed to be well-con-
trolled (i.e.,  operating  under good engi-
 neering  practices to minimize leaks and
 maximize CH4 recovery) so that the CH4
 recovery data would be useful in estimat-
 ing  total  CH4  production at the site. In
 addition, well-maintained records on rou-
 tine monitoring were needed for possible
 gas migration at the perimeter and surface
 of the landfill.
    To evaluate the effects of climate on
 CH4 production and recovery, sites were
 sought in geographic regions representing
 a variety of moisture and temperature pat-
 terns (i.e., hot and wet, cool and wet, hot
 and dry). Initial recommendations provided
 by landfill gas recovery experts in the U.S.
 were used to identify potential sites. Final
 site selection was influenced largely by the
 assurance that  long-term gas production
 and refuse composition data were  avail-
 able at the site, the suitability of the site for
 sample  acquisition,  and the  landfill
 operator's willingness to cooperate in the
 study.
    A landfill  survey form was sent to the
 operators of selected sites prior to visiting
 them so that they could begin gathering
 the  records.  Sites were visited  between
 Augusts and 24, 1990.

 Landfill Gas Test Procedures
 and Results
    Landfill gas was tested at each of the
 six  landfills in the pilot study in order to
 evaluate the  quality of the gas recovery
 data  available at each site.  The testing
 program included assessing the adequacy
 of on-site instrumentation  and scanning
 the landfill surfaces for organic vapors that
 would indicate emissions of CH4.
   The following test procedures  were
 used:

 (1)  EPA Reference Method (RM) 3C was
      used to  determine CH4, CO2,  nitro-
      gen (N2), and oxygen (O2) levels;
 (2)  EPA RM 25C was used to test for
      nonmethane  organic compounds
      (NMOC); and
 (3)  EPA RM 4 was used to test for the
     volume  percentage of moisture.
 Initially, the volumetric flow rates of landfill
gas were to be tested using EPA RM 2.
However, mechanical difficulties were en-
countered at the landfills, making this test
impossible. In lieu of this test, calibration
records of on-site flow measurement in-
struments were obtained for three of the
six landfills. Tests for the presence of or-
ganic vapors near the landfill surface were
conducted using  an organic  vapor  ana-
lyzer (OVA).
   The on-site gas analyzer measurements
were compared to the results of EPA RM
3C  testing  and  found  to  be reasonably
accurate; three of the CH4  analyzers ex-
 hibited relative accuracies better than 10%,
 and five were within 12%. In general, all of
 the instruments were observed to be oper-
 ating  in a manner consistent with good
 operating practices.

 Statistical  Methods
 Development and Results
   The ultimate objective of this research
 program is to determine which variables
 relating to refuse characteristics,  landfill
 characteristics, or climate  are significant
 determinants  of gas production. This pilot
 study  addressed a small number of sites,
 and the results were not intended to  be
 representative of all landfills.  Rather, the
 study was intended to provide the basis for
 development of statistical methods for use
 in a larger study,  to  identify data  quality
 issues, and to look for trends.
   The data  obtained from each  landfill
 consisted of  computer printouts or hand-
 written data sheets listing  total  gas flow,
 percent CH4  composition of the total gas
 flow, and other information applicable to
 the individual  landfills. The data were usu-
 ally in the form of daily averages of hourly
 flow rates, and were reported for each on-
 line gas recovery unit.
   The descriptive data for each  landfill
 are summarized ir Table 1. The average
 CH4 flow in standard cubic feet per  minute
 (cfm)* was calculated from  daily averages
 supplied by site operators. Although the
 between-landfill variation is large, ranging
 from 590 to 3477 cfm (16.71 to 98.47 m3/
 min), the day-to-day variability is relatively
 small, as shown by the coefficients of varia-
 tion, which were generally below 10% ex-
 cept for Landfill 6 (12.4%).
   One objective of this study was to de-
 termine if sufficient data were available for
 a  time series analysis  of  CH4  emission
 rates from individual landfills. Methane re-
 covery  is  a relatively new process, and
 none of these  landfills had records for CH4
 emissions  of sufficient  length (several
 years) and completeness for  time series
 analysis. It is highly probable that  emis-
 sions are autocorrelated so that any at-
 tempt  to find correlations  between CH4
 recovery and  weather data on a daily  or
 monthly basis  is likely to be confounded by
 autocorrelations in the data. Since the
 strength of autocorrelation decreases with
 averaging period,  only  annual averages
 were used in the statistical analysis of the
 relationship between long-term CH4 emis-
 sions and weather data between landfills.
The annual CH4 averages were correlated
to  annual averages of  temperature and
precipitation obtained from 30  years  of
                                                                                    * Flow rate at 25°C and 1 atmosphere

-------
 T»bl0 1.
Summary Statistics for Each Landfill Calculated from Daily CH4 and Weather Data
                                                                         Landfill
Parameter
Analysis period
Number of wells
Average well depth (m)
Number of hectares
Refuse mass (10* Mg)
Average landfill depth (m)
1990 average age (years)
Total Methane Flow
Number of days
Mean (trf/min) (cfm)
Standard deviation
Coefficient of variation (%)
Temperature
Mean (°C) during analysis period
30-year normal
Precipitation
Total (in.) (cm) during
analysis period
30-year normal
1
5/89 to 4/90
45
14
35
6.3
67
8

194
55.36 (1995)
2.12
3.80

7.34
7.51

31.7(80.5)

28.8 (73.2)
2
10/89 to 7/90
65(44 VA)*
14
55
6.1
26
10

302
18.04 (637)
1.19
6.60

7.67
9.28

34.0 (86.4)

35.6(90.4)
3
8/89 to 7/90
31
23
51
7.3
66
10

314
40.07(1415)
2.32
5.80

10.51
12.23

43.9(111.5)

42.4(107.7)
4
7/89 to 6/90
111
21
57
13.8
56
9.50

85
98.47(3477)
1.33
1.40

24.96
23.96

40.0 (101.6)

61.4(156.0)
5
1/90 to 8/90
102
34
32
10.9
46
15

209
24.86 (878)
1.70
6.80

16.12
17.12

9.0 (22.9)

17.0(43.2)
6
5/89 to 4/90
68
10
40
2.6
10
7

37
16.71 (590)
2.07
12.40

16.57
16.18

16.6 (42.2)

17.9(45.5)
 *VA = very active; other wells were primarily for odor control.
data, as well as to other landfill param-
eters.
   Table 2 shows the Pearson correlation
coefficients between annual CH4 flow rates
and  CH4 flow rates  per unit mass with the
annual and long-term (normal) weather data
and  other landfill parameters for the six
landfills. Only two  correlations with  CH4
flow rate were found to be significant at the
95% confidence level, and no correlation
coefficients were significant with the CH4
flow rate per unit mass. The low number of
significant correlations can be attributed,
at least in part, to the low number of obser-
vations. The normal annual  precipitation
correlated fairly well with the 1 -year annual
mean CH4 flow rate and,  even though it
was not significant for the CH4 flow rate per
unit mass, it had the largest positive corre-
lation coefficient. The correlation coeffi-
cient for refuse mass with CH4 flow rate
was  just under the  cutoff point for signifi-
cance at the 95% confidence level, but its
value of 0.71 suggests that perhaps  with
more data it would be significant.

Conclusions
   Except for flow measurements, the test
procedures selected for this project were
well  suited to the types of gas recovery
installations  encountered at  the landfills
visited. Alternative flow measurement meth-
ods that are  more appropriate to the site
conditions must be  identified if flow mea-
                               Tabl* 2.      Correlation Coefficients of CH4 Recovery Variables with Landfill Parameters and
                                           Summarized Weather Data (n=6)

                                                                                     Dependent Variables

Independent Variables
Annual temperature (1989-1990)
Normal annual temperature
Annual precipitation (1989-1990)
Normal annual precipitation
1990 mean age of landfill
Number of wells
Tons of refuse
Mean depth of landfill
Area of landfill
Volume of landfill
Mean well depth
Annual Methane
Recovery Rate

0.56
0.51
0.55
o.er
-0.15
0.37
0.71
0.62
0.37
0.74'
0.10
Annual Methane
Recovery
Rate per Unit Mass

0.12
0.01
0.33
0.25
-0.80"
-0.15
-0.18
0.26
-0.04
0.24
-0.58
                                 Correlation coefficient significant at 95% confidence level.
                                 Correlation coefficient significant at 90% confidence level.
                              surements are desired in the future. Since
                              all sites record flow data, however, a qual-
                              ity assurance program could  be used to
                              determine  the acceptability  of the on-site
                              data.
                                 Based  on comparisons  between the
                              RM  3C and instrument analyses of the
landfill gas composition, all on-site analy-
sis instruments appeared to be operating
within  reasonable accuracy  ranges.  Re-
views of calibration procedures and records
indicate that long-term instrument accu-
racy should be comparable to the accura-
cies  noted during on-site testing.
                                                                       •^ U.S. GOVERNMENT PRINTING OFFICE: W»2 - 648-080/401M

-------
   Although the results of this pilot study
are sufficiently encouraging to warrant fur-
ther data gathering and  analyses, some
limitations need to be recognized. The main
problem was that the collection efficien-
cies of the CH4 recovery systems were not
known. Where emission control was one
(or the  only)  reason for the collection
system's existence, efficiency appeared to
be high. However, this is  a qualitative as-
sessment  based on  visual inspection of
the landfills and an assessment of operat-
ing practices at the landfills.
   One key piece of information is missing
from most landfills: the average composi-
tion of the refuse. Waste composition un-
doubtedly contributes to data variability but,
unfortunately, it is not possible to get com-
position information for most landfills in the
U.S. Also, it is impossible to fully account
for differences in the structure and operat-
ing characteristics of landfills. All of these
unknowns contribute to the variability of
the CH4 flow rate data. Although it should
be possible to explain some of the variabil-
ity,  a certain amount will always remain.
    It is likely that the functional  relation-
ship between CH4 per ton of refuse and
age and climate are nonlinear, or that in-
teractions between these variables  pro-
duce nonlinearities. With a larger sample,
it  may be  possible  to  identify  these
nonlinearities, and fit the data to the appro-
priate model.
 D. Campbell, D. Epperson, L Davis, P. Peer, and W. Gray are with Radian Corp.,
   Research Triangle Park, NC 27709.
 Susan A. Thorn eloe is the EPA Project Officer, (see  below).
 The complete report, entitled "Analysis of Factors Affecting Methane Gas Recovery from
   Six Landfills," (Order No. PB92-101351/AS; Cost: $35.00, subject to change) will be
   available only from:
         National Technical Information Seivice
         5285 Port Royal Road
         Springfield, VA22161
          Telephone: 703-487-4650
 The EPA Project Officer can be contacted at:
         Air and Energy Engineering Research Laboratory
         U.S. Environmental Protection Agency
         Research Triangle Park, NC 27711
United States
Environmental Protection
Agency
       Center for Environmental Research
       Information
       Cincinnati, OH 45268
                BULK RATE
          POSTAGE & FEES PAID
           EPA PERMIT NO. G-35
Official Business
Penalty for Private Use $300
EPA/600/S2-91/055

-------