p a p e r 05 2005 characterization of s o l i d w a s t e disposed at columbia sanitary ...

http://wmr.sagepub.com

Waste Management & Research

DOI: 10.1177/0734242X05050995 2005; 23; 62 Waste Management Research

Yinghui Zeng, Kathleen M. Trauth, Robert L. Peyton and Shankha K. Banerji Characterization of solid waste disposed at Columbia Sanitary Landfill in Missouri

http://wmr.sagepub.com/cgi/content/abstract/23/1/62 The online version of this article can be found at:

Published by:

http://www.sagepublications.com

On behalf of:

International Solid Waste Association

can be found at:Waste Management & Research Additional services and information for

http://wmr.sagepub.com/cgi/alerts Email Alerts:

http://wmr.sagepub.com/subscriptions Subscriptions:

http://www.sagepub.com/journalsReprints.navReprints:

http://www.sagepub.com/journalsPermissions.navPermissions:

© 2005 International Solid Waste Association. All rights reserved. Not for commercial use or unauthorized distribution. by IGNACIO GARCIA MARTINEZ on November 22, 2007 http://wmr.sagepub.comDownloaded from

http://www.iswa.org//web/guest/home

http://wmr.sagepub.com/cgi/alerts

http://wmr.sagepub.com/subscriptions

http://www.sagepub.com/journalsReprints.nav

http://www.sagepub.com/journalsPermissions.nav


62 Waste Management & Research

Waste Manage Res 2005: 23: 62–71Printed in UK – all right reserved

Copyright © ISWA 2005Waste Management & Research

ISSN 0734–242X

Characterization of solid waste disposed at Columbia Sanitary Landfill in Missouri

Waste sorts were conducted during each of the four quarters(or seasons) of 1996 at the City of Columbia Sanitary Landfill.A detailed physical sampling protocol was outlined. Weightfractions of 32 waste components were quantified from all geo-graphic areas that contribute to the Columbia Sanitary Land-fill using a two-way stratification method, which accountedfor variations in geographical regions and seasons. Compari-sons of solid waste generated between locations and seasonswere conducted at the 80% confidence level. The composi-tion of the entire waste stream was 41% paper, 21% organic,16% plastic, 6% metal, 3% glass and 13% other waste. Paperwas the largest composition and glass was the smallest com-position for all geographical regions. The result of this studywas also compared with a 1987 Columbia, Missouri studyconducted by EIERA (1987), with studies conducted in otherstates such as Minnesota, Wisconsin, Oregon and withnational study conducted by the USEPA (USEPA 530-R-96-001, PB96-152 160. US Environmental Protection Agency,Office of Solid Waste, Washington, DC). The results of studiesfrom other states are different from this study due to differentlocal conditions, different methodologies and a differentscope. There was a small (5%) increase in per capita weightfrom 1987 to 1996. The total per capita weight in thepresent study was 60% greater than the national per capitaweight reported by the USEPA (1996) due to that theUSEPA report excluded industrial, construction and certaincommercial waste. The total per capita weight agrees with thenational per capita weight for municipal waste reported byTchobanoglous (1993), which included industrial, construc-tion and commercial sources. The geographical and seasonaleffects on the waste composition are evaluated and discussed.Statistical analysis indicates that waste characteristics are dif-ferent among geographical regions and seasons. The potentialfor waste recovery and reduction is also discussed.

Yinghui ZengOffice of Social Economic Data Analysis, University of Missouri,Columbia, MO, USA

Kathleen M. TrauthRobert L. PeytonShankha K. BanerjiDepartment of Civil & Environmental Engineering, Universityof Missouri, Columbia, MO, USA

Keywords: Solid waste characterization, sample survey, inte-grated solid waste management planning, solid waste composi-tion, wmr 633–8

Corresponding author: Y. Zeng, Office of Social EconomicData Analysis, University of Missouri, Columbia, MO 65211,USATel: +1 (573) 884-9137; fax: +1 (573) 884-4635; e-mail: [email protected]

DOI: 10.1177/0734242X05050995

Received 7 January 2003; accepted in revised form 17 November 2004



Characteristics of solid waste disposed at a landfill in Missouri

Waste Management & Research 63

Introduction

The percentage and weight of waste components in a solidwaste stream are important data for decision-makers. Thisinformation is necessary in order to identify waste compo-nents to target for source reduction and recycling pro-grammes, and to allow technical professionals to design mate-rial recovery facilities (MRF) and waste-to-energy (WTE)projects. For a MRF, the waste component weight will affectthe sizing of the equipment, tipping floor area, recoveredproduct storage areas and possible economic benefits fromthe sale of recovered products. The waste component per-centage will affect the separation and processing configura-tions. For a WTE facility, the weight and composition willaffect the sizing of the facility and the quantity of energy to beproduced. National average values of solid waste compositionmay not accurately reflect conditions in local communities.The magnitudes of the variations of waste components areunknown, but they are needed by solid waste managementplanners, local officials and MRF designers.

In conducting a study of local conditions, a variety ofwaste characterization methods can be used. Computer mod-els can use national averages for waste generation rates andother community features to calculate waste quantities. Thisis a quick method, but it does not account for local wastecharacteristics that can vary significantly from national orregional averages. An alternative is to use materials-flow sur-veys based on ‘production data for the materials and productsin the waste stream, with adjustments for imports, exports andproduct lifetimes’ (USEPA 1996). Based on production data,an estimate is made for the total weight of waste generated.Then an estimate is made of the portion of this generated wastethat is recycled or composted. The remaining waste is definedas discards. This approach never collects physical samples andis difficult to apply when evaluating waste characteristics at afacility such as a landfill or a treatment facility. It is more suit-able for conducting national studies where collecting physi-cal samples from a rather broad area is difficult. The approach

chosen for this study was to determine a statistically signifi-cant sampling size, then use physical sampling, separationand direct measurement to quantify waste characteristics. Thisapproach accounts for the unique characteristics of each sourcelocation and the landfill as a whole.

The objectives of this study were to: (1) quantify 32 com-ponents of the waste entering the City of Columbia SanitaryLandfill by source location, sector and season; (2) present aphysical sampling protocol to collect the desired number ofsolid waste samples determined with the two-way stratifica-tion method for minimizing sampling errors; (3) contributedata to a state-wide waste characterization study and developa database that can be used for economic analysis of a mate-rial recovery facility and for assessment of existing wastereduction programmes within the landfill service area; and(4) evaluate the potential for waste recovery and reduction.

Methodology

The data were collected at the City of Columbia SanitaryLandfill in Columbia, Missouri during 1996. Thirty-two tar-geted materials or sorting categories were selected. The compo-sition was then categorized into six categories: paper, organic,plastic, metal, glass and other (Table 1).

Because seasonal variation and geographical variation canhave a significant impact on waste characteristics, samplingwas designed to be two-way stratified. The first level of strat-ification is seasonal stratification. The second level is geo-graphical stratification. Sampling was designed to take placeduring each of the four quarters of the year. Quarter 1 wasfrom 22 February to 29 March 1996, quarter 2 was from 1 Mayto 29 May 1996, quarter 3 was from 5 August to 11 September1996 and quarter 4 was from 4 November to 23 December 1996.The service area of the landfill was subdivided into six ‘regions’consisting of the cities of Centralia, Columbia and Mexicoand the unincorporated areas of Audrain, Boone and Calla-

Table 1: Waste composition category.

Waste composition category Waste components

Paper Corrugated board, box board, newsprint, magazines, office paper, mixed paper (all paper that does not fit into other category)

Organic Food, wood, textiles, manure, other

Plastic PET(#1), HDPE(#2), PVC(#3), LDPE(#4), PP(#5), PS(#6), Other

Metal Ferrous, non-ferrous, bi-metal, aluminium cans, other aluminium

Glass Clear, brown, green, other

Other Nappies/sanitary products, banned items*, fines (pass through 63-mm-opening sieve), medical waste, miscellaneous (demolition waste and any other waste)

* Missouri State Law bans the acceptance of waste such as yard waste, tyres, batteries and large appliances, motor oil, etc.



Y. Zeng, K.M. Trauth, R.L. Peyton, S.K. Banerji


way counties. The City of Columbia was further subdividedinto three sectors: commercial/industrial, residential and uni-versity (Table 2). For the first quarter of the year, no localdata were available for estimating the number of samplesneeded. Twelve waste components in the ASTM International(1992) national data set were used to estimate the requirednumber of samples. For an error, e = 0.05, 0.10 and 0.20, respec-tively, the number of samples needed were 522, 131 and 34 toachieve an 80% confidence level. One hundred and fifty-onesamples corresponding to an error between e = 0.10 and e = 0.05were selected for the present study. During the subsequentsampling periods, due to cost and resource constraints, thenumber of samples collected was adjusted to 128, 130 and127, respectively. A two-way stratification method was devel-oped to account for variations between the geographical regionsand seasons in calculating waste compositions.

Sampling protocol

Sample weight

The term sample size is sometimes used to refer to two differ-ent parameters in solid waste characterization studies. Oneis the number of sample units to be sorted. The other is theweight of each unit. In this paper, sample size means thenumber of sample units, and sample weight, means the weightof each unit. The sample weight affects the variability of esti-mation. Obviously, if a sample weight is too small, it wouldgive inaccurate results, because, for example, a large piece ofwood could not physically be included in a small sample.However, separating and sorting raw solid waste is expensive.A typical vehicle load of commercial solid waste weighsbetween 4500 and 9000 kg. It is not practical to separate theentire vehicle load or a load from an even larger vehicle.

Klee (1980) indicated that the smaller the sample weight,the greater the variance of the waste sample composition.He stated that as sample weights decreased from approxi-mately 91 kg, the sample variance increased rapidly, but thatabove approximately 140 kg, the variance increased muchmore slowly. He thus recommended a sample weight between91 and 140 kg. This recommendation was also adopted by theASTM (1992) standard. Therefore, the target sample weightfor this study was set at 140 kg.

Field protocol

Trucks from each geographical area were numbered and ran-domly selected from the geographical area during a samplingperiod. The previously identified truck needed to unload thewaste at a working area. A landfill worker then used a front-end loader to mix the waste and collect a 140 kg (within 10%error) waste sample. The waste sample was identified by a scalehouse ticket with the record of the source of waste, totalweight measured at the scale house, date and time of arrival atthe scale house. The identified waste sample was then trans-ported to the sorting shed and was deposited onto a high-density polyethylene (HDPE) 60-mil liner for sorting.

Sorters removed the large items such as large pieces ofcorrugated board and put them in an identified container.Trash bags were torn open. Portions of the waste were placedonto a sorting table and were then sorted by hand and placedinto identified containers. An estimate of wetness of thesample was made and recorded on the data sheet. After thesorting was completed, each container was weighed with anaccuracy of ± 0.23 kg. Standard personnel safety procedureswere followed during the sorting process such as wearinggloves, apron/coverall, safety glasses and boots, sorting thenearest item first, etc.

Table 2: Component weight of waste entering Columbia Landfill during 1996.

Geographical source*Weight of waste composition category (tonne) Per capita

(kg year–1)Paper Plastic Metal Glass Organic Other Total

Audrain Countya 900 300 100 37 600 200 2100 300

Boone Countyb 16 800 4800 2900 1300 7700 5700 39 200 900

City of Columbiac 25 600 11 200 3100 1500 13 200 7700 62 300 800

Commercial/industrial 15 300 8100 1900 700 7500 4700 38 200

Residential 8000 2400 1000 700 4800 2300 19 200

University of Missouri 2300 700 200 100 900 700 4900

Callaway Countyd 1300 400 300 200 1000 400 3600 600

Centraliae 1400 800 200 100 1000 500 4000 1200

Mexicof 1200 500 100 46 500 300 2600 4600

All waste stream 47 200 18 000 6700 3200 24 000 14 800 113 800 800

*Waste came from a50% of population in Audrain County outside of City of Mexico and 5% of Montgomery County; bpopulation from Boone County excluding City of Columbia, City of Centralia and City of Ashland; cpopulation from City of Columbia and 5% of Osage County; dpopu-lation in Callaway County outside of City of Fulton and 5% of Osage County and 5% of Montgomery County; epopulation from City of Centralia; and f5% of population in City of Mexico.





Results and discussions

Waste component weight and compositionAs shown in Table 2, a total of 113 800 t of waste entered theCity of Columbia Sanitary Landfill during 1996 (City ofColumbia’s Solid Waste Utility, 1996). The City of Columbiacontributed the most (55%, 62 300 t) to the total wastestream entering the landfill in 1996. The City of Columbia andthe remainder of Boone County contributed approximately89% (101 500 t) of the waste stream. They contributed themost to each of the waste composition categories. 61% ofColumbia waste came from the commercial/industrial wastesector, 31% came from the residential sector and 8% from theUniversity of Missouri. The per capita weight was also calcu-lated using estimated service population of each geographicalsource. It was estimated that waste came from 50% of thepopulation in Audrain County outside the City of Mexicoand 5% of the population in Montgomery County; the popu-lation in Boone County excluding that in the City of Colum-bia, the City of Centralia and the City of Ashland; the popu-lation in the City of Columbia and 5% of the population inOsage County; the population in Callaway County outside ofthe City of Fulton and 5% of the population in both OsageCounty and Montgomery County; the population in the Cityof Centralia and 5% of the population in the City of Mexico;and the remaining 95% of the population of the city of Mex-ico. The percentages of the population were estimated by talk-ing to waste haulers and City of Columbia Solid Waste Utilitypersonnel. As some portion of the waste of some counties orcities goes to different landfills, it is difficult to estimate anaccurate service population. However, because all waste inthe City of Columbia goes to the City of Columbia Landfill,the per capita weight for the City of Columbia has a higherdegree of confidence (820 kg/year, 2.3 kg/day).

The waste composition for the entire waste stream andgeographical sources in 1996 is shown in Table 3. The per-

centage composition of waste combined from all locationswas 41% paper, 21% organic, 16% plastic, 6% metal, 3%glass and 13% other waste. Paper was the largest compositionand glass was the smallest composition for all locations.

Table 4 presents the percentage error for the mean weightfractions for four quarters with an 80% confidence level, whichmeans that the true mean will lie within the range of the esti-mated mean ± error associated with the 80% confidence.Because only a limited number of samples of the waste streamwere collected for measurement of waste component weightfraction, there is a degree of uncertainty for each mean weightfraction value. This uncertainty is represented by the per-centage errors and confidence associated with the errors. Thepercentage error depends on the number of samples collected:the more samples collected, the lower the error. The percent-age error also depends on the variability of the waste stream.The less variable (or more uniform) the waste stream, thelower the errors. The percentage errors tend to be larger forwaste components that make up the smaller fractions of thewaste stream (e.g., other organics, medical waste, etc.)because with the smaller fraction comes the larger variability.The seasonal and geographical variations in the waste streamand the sample weight of each sample also contribute to theerror. The percentage errors tend to be larger for those geo-graphical locations where the fewest samples were collected.

Table 3: Composition of waste entering Columbia Landfill during 1996.

Geographical sourceWaste composition(%)

Paper Plastic Metal Glass Organic Other Total

Audrain County 43 14 5 2 29 10 100

Boone County 43 12 7 3 20 15 100

City of Columbia 41 18 5 2 21 12 100

Commercial/industrial 40 21 5 2 20 12 100

Residential 42 13 5 4 25 12 100

University of Missouri 47 14 4 2 18 14 100

Callaway County 36 11 8 6 28 11 100

Centralia 35 20 5 3 25 13 100

Mexico 46 19 4 2 19 12 100

All waste stream 41 16 6 3 21 13 100

Table 4: Percentage error of mean weight fraction of four quarters with 80% confidence level.

Waste component Q1 Q2 Q3 Q4

Total paper 9 11 9 8

Total plastic 13 12 10 14

Total metal 9 23 12 12

Total glass 10 15 13 17

Total organic 9 9 10 7

Total other 18 24 19 16

Weighted average 19 23 19 18





The number of samples needed to achieve a certain degreeof accuracy is affected by seasonal and geographical varia-tions in the local waste stream. For the first quarter of theyear, no local data were available for estimating the numberof samples needed. One hundred and fifty-one samples, corre-sponding to an error between e = 0.10 and e = 0.05 was selectedbased on the national data set and simple sampling method(ASTM 1992).

The simple sampling method suggests that the 151 sam-ples would generate an error of approximately 10%. Afteractually collecting 151 samples from the Columbia landfill,and analysing them using the two-way stratified method tocalculate the percentage error, the actual error was approxi-mately 20% (Table 4), rather than the 10% predicted by thesimple sampling method. This indicates that the use of anational dataset and the simple sampling method did affectthe accuracy of the results. To achieve greater accuracy, moresamples would be needed.

Comparison with other studiesA comparison of this study with other studies is presented inTable 5. The waste components analysed in each study weredifferent and regrouped to match the components defined inthe national dataset (ASTM 1992).

In a Columbia, Missouri 1987 waste characterization study(EIERA 1987), waste sampling was conducted in May andAugust of 1987. There was no reporting of the number ofsamples taken, the selection criteria, etc. in 1987 EIERAdocument. Based on an estimation from the City of Colum-bia Solid Waste Utility (1996), the number of samples col-lected was less than 10. With the very small number of sam-ples collected, the percentage error associated with the result

would be expected to be larger than that generated by thepresent study.

A 1999 Minnesota study (Minnesota Solid Waste Man-agement Coordinating Board 2000) was conducted in theTwin Cities Metropolitan Area. Waste sorts were conductedfor five disposal sites through four seasons. The sorts wereconducted each season for a 1-week period at each site. A 90to 180 kg sample was taken from a sample truck. The numberof samples collected for each sort was reported as 40 to 60. Atotal of 1170 samples were sorted.

The 2001 Wisconsin study (Cascadia Consulting Group,Inc. 2003) was a state-wide waste characterization study. Sam-ples were collected from 14 landfills during two sampling days.A total of 400 waste samples of 90 to 140 kg were each sortedinto 64 categories.

The Oregon 2002 study (State of Oregon Department ofEnvironmental Quality 2002) was also a state-wide study. Atotal of 884 samples were collected for 60 waste substreamsand the results were averaged.

The USEPA annually publishes a national municipal solidwaste characterization report. The USEPA report publishedin 1996 was most comparable to this study. It presents theresults of a study conducted in 1995 which was based on 1994data. All studies except the USEPA study were based on aphysical sampling method, whereas the USEPA study wasbased on a material flow survey method.

The results of studies from other states are different fromthis study for several reasons. The sampling design and dataanalysis method for this study was a two-way stratificationmethod, which accounts for variations among seasons andregions when calculating the mean weight fraction. Themethod used for some of the other studies was a simple sam-

Table 5: Comparison with other studies.

Waste Composition (%)

Wastecomponent Columbia 1987 Columbia 1996 EPA

1994 Minnesota 1999 Wisconsin 2001 Oregon 2002

Paper 41 41 33 34 21 21

Plastic 7 16 12 11 11 11

Metal 6 6 6 5 6 8

Glass 4 3 6 3 2 2

Wood 15 7 8 8 14 9

Textiles 4 4 4 3 1 2

Food 7 9 9 12 11 16

Inorganic 7 1 2 6 2 13

Yard waste 7 1 15 2 1 7

Other 2 12 5 16 31 11





pling method in which a simple arithmetic mean was com-puted for the weight fraction. The studies for Minnesota, Wis-consin and Oregon were all state-wide studies which includedmultiple landfills whereas this study only investigated thewaste stream entering the City of Columbia landfill. The localconditions such as the type of waste accepted at landfills,climate, economic activities, life styles, recycling and wastemanagement programmes all have a great impact on the wastecomposition. There was also a lack of a standard definitionfor waste sorting categories. Thus, each state defined some oftheir waste categories differently from other states.

The results from the USEPA waste characterization studyare different from the results of the present study for threereasons. First, the USEPA methodology was different. Insteadof using waste sampling data, the USEPA used a materials-flow approach based on ‘production data for the materials andproducts in the waste stream, with adjustments for imports,exports and product lifetimes’. Second, the USEPA charac-terization study did not cover all materials that enter SubtitleD landfills, such as the City of Columbia Sanitary Landfill.The materials included by the USEPA are defined in their1996 report (USEPA 1996). This is a different waste streamfrom that entering the City of Columbia Sanitary Landfill.For instance, the Columbia landfill does accept certain indus-trial process wastes and construction and demolition debrisbut does not accept yard trimmings, large appliances, or auto-mobile tyres, whereas the USEPA study did not include wastesfrom construction and demolition debris, industrial processwastes but did include yard trimmings, large appliances andautomobile tyres. Third, the USEPA study is a nationwidestudy, whereas the study presented here covers a small region.The differences in methodology and definitions describedabove serve to confound the comparisons.

The total waste entering the landfill increased by 40%from 1987 to 1996; most of this increase was due to an increasein the service population. The per capita increase was only5%. The largest changes in percentage weight among the cat-egories in Table 5 were for plastics, which increased from 7to 16% (a 129% increase). Inorganic and yard waste bothdecreased from 7 to 1% (86% decreases). Wood decreasedfrom 15 to 7% (a 53% decrease) and glass decreased from 4 to3% (a 25% decrease).

The large decrease in yard waste composition is due toMissouri Statute RSMo 260.250 (2004), which mandatedthat after 1 January 1992, yard waste was not to be allowedto enter landfills. In addition, the decrease in glass composi-tion indicates that recycling for glass was effective. The totalrecycled glass in 1995 was 200 t (City of Columbia SolidWaste Utility 1996). The composition of waste with respectto corrugated board and newsprint for Columbia landfilldecreased from 17 to 13% and 8 to 5%, respectively, from 1987

to 1996. This agrees with the 1995 recycling data of 500 tand 400 t, respectively (City of Columbia Solid Waste Util-ity 1996). However, other un-recycled paper components suchas magazines increased. This caused the total paper composi-tion to remain the same from 1987 to 1996.

A detailed breakdown of the waste category comparisons isshown in Table 6. Among the five categories in Table 6, theweight percentages for paper, plastic and metal were higher atthe Columbia landfill (30, 21 and 2% higher, respectively)and glass and other materials were lower at the Columbialandfill (50 and 23% lower, respectively).

The per capita weights from Columbia were higher in fourof the five categories (not glass). The total per capita weightat the Columbia landfill (800 kg year–1) is 60% higher than thenational average total per capita weight reported by USEPA(1996), which excluded waste from construction and demoli-tion debris and industrial process waste. A typical per capitavalue of 1000 kg year–1 for total municipal solid waste gener-ation in US was reported by Tchobanoglous et al. (1993).This value was converted to 800 kg year–1 of discarded wastefor total municipal solid waste by using a discard-to-genera-tion ratio of 0.773 (USEPA 1996). The Columbia per capitaweight agrees with the typical national value reported byTchobanoglous et al. (1993), which included constructionand industrial waste that was excluded in the USEPA report(1996).

The result with respect to yard waste and waste tyres is ratherdramatic: ‘Finally, all individual components had higher percapita weights at the Columbia landfill except for yard trim-mings and rubber tyres, which were lower by 88%. This is aclear indication of the effectiveness of the ban on yard wasteand tyres at the Columbia landfill’ (Center for Environmen-tal Technology and Energy Systems and Resources Program1997).

Comparison between locations and sectorsTable 7 presents a listing of locations with waste characteris-tics that are different from each other at the 80% confidencelevel. It is logical to conclude that annual mean weight frac-tions that are close to each other may indicate waste charac-teristics that are similar. Statistical tests were conducted todetermine whether there were significant differences betweenthe weight fractions of the waste components at the eight dif-ferent locations. The tests not only consider the mean valuebut also consider the extent of spread, or variance, of all meas-ured values about the mean value and consider the number ofmeasured values.

Two possibilities were considered. One possibility was thatthe variances of the two populations, p, were unknown butequal. The other possibility was that the variances of the twopopulations were unknown and unequal. An F test (Milton





& Arnold 1995) was first conducted to compare the vari-ances of the two populations. If the F test indicated that the

variances were equal, a test for comparing means with equalvariances was performed; otherwise, the test for comparingmeans with unequal variances was performed. When it wasassumed that the variances were equal, a pooled sample vari-ance was computed and a pooled t test was conducted; other-wise, the individual variances were used to conduct the t test(Milton & Arnold 1995). The General Linear Model in SAS(1990) was used to conduct the tests, using a 80% confi-dence level. The results are summarized below.

The weight fractions varied from location to location.Table 7 shows the locations where waste characteristics weredifferent. The letters are the components for which the wastecharacteristics were different. Every location had at least onewaste component that was different from that of other loca-tions. The locations that had the greatest number of differentcomponents were Callaway County and the City of Colum-bia. There were no significant differences between total ‘otherwaste’ characteristics at any locations. These results support thenotion that waste stream characteristics vary with geographi-cal region and this variation should be taken into considera-tion when designing a sampling strategy for a waste charac-terization study.

Comparison between seasonsTable 8 presents a summary of the annual mean weight frac-tions of the six waste categories for each of the four quarterlysorts. A statistical test was conducted to determine whetherthere were significant differences among the waste character-istics between quarters. The method was the same as used totest for significant differences between locations explained inthe section entitled ‘Comparison between locations and sec-tors’ above. Every quarter had at least one component thatwas different in other quarters, except that quarter 2 was notdifferent from quarter 3. This lack of difference was probablybecause of the fact that quarter 2 has an average of 23% of errorwhich is larger than the 20% for other quarters (Table 4). Thedifferences in quarter 3 could be related to special summerevents that draw out-of-town visitors and the transition in stu-dent population. The differences in quarter 4 could be relatedto holiday activities. The differences in quarter 1 could prob-ably be due to the lack of special events, transitions and hol-idays that create the sharp contrasts between quarters 3 and 4.

Potential for recovery and reductionHereafter, the potential for waste recovery and reduction isdiscussed from the view of market value. Table 9 presents themarket values for recyclable materials in the Columbia wastestream. The prices were obtained from Associated Recyclersof the Midwest, 2004. The cost of waste collection is notincluded because it would be incurred whether the waste wasrecycled or landfilled.

Table 6: Waste composition comparison with national average (USEPA 1996).

Component

National 1994(%)a

Columbia 1996(%)

Per capitab 1994

(kg year–1)

Per capitac

Columbia1996

(kg year–1)

Corrugated board

8 13 40 100

Box board 5 9 30 70

Newsprint 5 5 30 40

Magazines 1 3 5 30

Office paper 2 3 10 20

Mixed paper 12 10 70 80

Total paper 33 43 200 300

PET (#1) 1 2 3 10

HDPE (#2) 2 2 10 10

PVC (#3) 1 2 5 20

LDPE (#4) 4 5 20 40

PP (#5) 2 2 8 10

PS (#6) 2 2 8 20

Other plastic 2 2 9 20

Total plastic 14 17 60 100

Aluminium 1 1 7 7

Ferrous and bi-metal

5 5 30 40

Non-ferrous 0 1 2 4

Total metal 6 7 40 50

Total glass 6 3 40 20

Food waste 9 9 50 70

Yard trimmings and rubber tyresd

17 1 90 10

Textiles 4 4 20 30

Wood 8 7 50 60

Other wastes 6 13 30 100

Total other 44 34 200 300

Total 100 100 500 800aMaterials from ‘municipal solid waste’ that were discarded after materials and compost recovery. bEqual to percentage of materials discarded divided by 100, multiplied by 159 760 000 tons total materials discarded in US in 1994, multiplied by 2000 lb/ton, multi-plied by 0.45 kg/lb, divided by the July 1994 US population of 260 372 174. cEqual to percentage of total weight that entered Columbia landfill in 1996 divided by 100, multiplied by 125 790 tons total weight that entered Columbia landfill in 1996, multiplied by 2000 lb/ton, multiplied by 0.45 kg/lb, divided by the estimated 1996 total service population of 138 341. dAssumed equivalent to banned items in present study.





The total weight of recyclables was 39 300 t (108 t/d). Ifthe recovery rate is assumed to be 80%, the recovered mate-rials are then 108 × 80% = 86 t/d. The revenue is 9 000 $/d(Table 9). The cost of recycling is the sum of the capital andoperating cost (O&M) of the material recovery centre (MRF).Although detailed costs vary by community, the configura-tion of the MRF and many other factors, one can make pre-liminary estimates from the general average cost data. Thetypical unit capital cost for a low-tech MRF is $10 000 pertonne of daily capacity (Tchobanoglous and Kreith, 2002).Thus, the capital cost for an MRF is approximately 1.7 mil-lion dollars. The typical O&M cost for a low-tech MRF is20 $/t. The waste collected from the City of Columbia is171 t/d (62 300 t/y) (Table 2). Thus, the O&M cost for anMRF would be 3420 $/d. Revenues thus exceed costs by5580 $/d or 2 031 120 $/yr (not considering the time valueof money within the year). The revenues from one year ofoperation would pay for the construction of the MRF. Thelong-term implications of recycling can be seen from consid-ering the present worth of the value of revenue minus O&Mcost. One can convert an annual value to a present worthvalue using standard economic tables if the life of a facilityand an interest rate are specified. Present worth values for

the revenues over cost are shown in Table 10 for variousfacility lives and interest rates.

The present worth of the revenues minus O&M cost farexceeds the capital cost, even considering short facility livesand high interest rates. Such present worth would still showthe value of recycling even if revenues were less than calcu-lated above or if excess costs associated with bags or bins forrecyclable collection were considered.

However, this is just a rough estimate, many factors are com-munity specific. The feasibility of an MRF should not dependpurely on economics as it did in the past. The USEPA haspromulgated regulations for municipal solid waste landfills asrequired by subtitle D of the Resource Conservation andRecovery Act of 1976 (RCRA 42 U.S.C. 6901 et seq., 2004),effective October 9, 1993. Both existing and new landfillswere affected by the statute. Available landfill volume isdecreasing because of stringent regulations (The above calcu-lations do not consider the cost of environmental protectionfeatures now required for landfills or the cost of groundwatercleanup from potential contamination.). The wastes thatused to be discarded need to be recovered and managed in asustainable way. The USEPA recommends that recycling bethe top priority option used in an integrated solid waste man-

Table 7: Listing of locations with waste characteristics that are different from the location shown at the top of column, at 80% confidence level.

Comparisons among geographical regionsa

Location name Locationnumber

Location number

1 2 3 4 5 6 7 8

Audrain 1 . M, G, O P, M, G P, L, M, G L, O G G, O L, O

Boone 2 M, G, O . P, M, G, O P, L, M, O L, M, G M, O P, M L, M, G

Callaway 3 P, M, G P, M, G, O . L, M, G P, L, M, G, O P, M, G, O P, M, G, O P, L, M, G, O

Centralia 4 P, L, M, G P, L, M, O L, M, G P, G, O P, G P, L, M, O P, G, O

Columbia-Comm./Ind.

5 L, O L, M, G P, L, M, G, O P, G, O . G, O P, L, M, G L

Columbia-Res. 6 G M, O P, M, G, O P, L, G L, G, O . P, M, G, O L, G, O

Columbia-MU 7 G, O P, M P, M, G, O P, L, M, O P, L, M, G P, M, G, O . L, M, G

Mexico 8 L, O L, M, G P, L, M, G, O P, G, O L G, O L, M, G .aP, total paper; L, total plastics; M, total metal, G, total glass; O, total organics.

Table 8: Comparison between seasons.

Mean weight fraction

Seasons Total paper Total plastic Total metal Total glass Total organic Total other

Quarter 1 0.391 0.156 0.059 0.027 0.240 0.127

Quarter 2 0.429 0.173 0.064 0.026 0.190 0.118

Quarter 3 0.383 0.150 0.063 0.030 0.210 0.164

Quarter 4 0.448 0.152 0.055 0.030 0.209 0.106

Annual mean 0.412 0.158 0.061 0.029 0.212 0.128





agement system. Economical feasibility is not the only factorthat drives the solid waste management system. A sustainablesolid waste management system also needs to be environmen-

tally sound and socially acceptable. Therefore, the revenuefrom recyclables can only be viewed as a supplemental bene-fit, not as a determining factor.

Table 9: Recyclable materials waste stream entering Columbia landfill in 1996.

Recovered material 2004 pricea

(US$ t–1)Weight

(t)Recovered weightb

(t)Market value

(US$)

Metal

Aluminium cansc 772 600 480 370560

Other aluminiumc 639 400 320 204480

Steel cans 72 2400 1920 138240

Ferrous 72 3000 2400 172800

Plasticd

HDPE natural 507 500 400 202800

HDPE mixed colour 265 500 400 106000

HDPE/PET mixed 22 1800 1440 31680

PET clear loose 11 400 320 3520

PET mixed colour 11 400 320 3520

Paper and paper board

Newspaper loose 61 5400 4320 263520

Corrugated loose 110 14500 11600 1276000

Office mixed loose 50 2900 2320 116000

Magazines loose 110 3600 2880 316800

Glass

Clear 44 1900 1520 66880

Brown 33 700 560 18480

Green 22 300 240 5280

Total 39300 31440 3296560aPrice is obtained from Associated Recyclers of the Midwest http://recyclingcoop.org/market.htm. bAssumes that 90% of the weight is recovera-ble. cAnnual weight was computed as the mean percentage entering the landfill from this source during quarters 3 and 4 times the total annual weight of all waste entering the landfill from this source, since quarters 3 and 4 were the only quarters when measurements were taken for alu-minium cans. dArbitrarily assumes the following distribution: PET (25% clear loose, 25% mixed colour, 50% mixed with HDPE); HDPE (25% natu-ral, 25% mixed colour, 50% mixed with PET).

Table 10: Present worth values for excess of revenues over O&M costs (E. Grant et al., 1982).

Facility Life (years)

Interest Rate (%)

Present Worth Factor Present Worth ($)

20 20 4.870 9 891 554

20 10 8.514 17 292 956

20 5 12.462 25 311 817

20 2 16.351 33 210 843

10 20 4.192 8 514 455

10 10 6.144 12 479 201

10 5 7.722 15 684 309

10 2 8.983 18 245 551





Conclusions

A total of 536 samples were collected at the City of ColumbiaSanitary landfill for 32 waste components in 1996. The wastestream was subdivided into two-way stratified sub-levels withthe geographical region being the first level of stratificationand the season being the second level of stratification. Randomsamples were collected for each sub-level. A detailed physicalsampling protocol was outlined for the sampling schemedetermined by the two-way stratification method. It is esti-mated that a total of 113 800 t of waste entered the City ofColumbia Sanitary Landfill during 1996. The City of Colum-bia contributed the most to the waste stream. The per capitaweight for the City of Columbia in 1996 was 2.19 kg person–1

day–1 or 800 kg person–1 year–1 (Table 2). The composition ofthe waste stream was 41% paper, 21% organic, 16% plastic, 6%metal, 3% glass and 13% other waste. Seven components rep-resented almost 60% of the total waste entering the landfill.Ranked by weight, these were corrugated board, mixed paper,box board, food, miscellaneous other waste (primarily con-struction-related waste), wood and newsprint. The total weightof waste that entered the landfill increased by 40% from 1987

to 1996. Most of this increase was due to an increase in theservice population. The per capita increase was only 5%.

It was difficult to accurately compare the results of thisstudy with other waste characterization studies because ofthe lack of consistency in methodology and waste compo-nent definitions. For instance, this study sampled all wasteentering the landfill. In the national study, the USEPA(1996) excluded certain waste streams from their analysis.The comparisons that were possible suggest variability inwaste characteristics by geography and seasons that shouldbe addressed by site-specific sampling for integrated solidwaste management.

Acknowledgement

This project was funded by a grant from the Missouri Depart-ment of Natural Resources Solid Waste Management Pro-gram. Portions of the information presented in the paper arealso documented in the 1997 final project report submittedto the funding agency. R.L.P., the principal investigator andmajor author of the 1997 report, served as dissertation advi-sor to Y.Z. during this time.

References

ASTM International (1992): ASTM D 5231–92 Standard Test Method forDetermination of the Composition of Unprocessed Municipal Solid Waste.For referenced ASTM standards, visit the ASTM website, http://www.astm.org, or contact ASTM Customer Service at [email protected] the Annual Book of ASTM Standards volume information, referto the standard's Document Summary page on the ASTM website

Associated Recyclers of the Midwest (2004): http@//recyclingcoop.org/mar-ket.htm

Cascadia Consulting Group, Inc. (2003): Wisconsin Statewide Waste Charac-terization Study. Cascadia Consulting Group, Inc, Wisconsin.

Center for Environmental Technology and Energy Systems and ResourcesProgram. University of Missouri-Columbia (1997): Waste Characteriza-tion Study for City of Columbia Sanitary Landfill. Columbia, Missouri.

City of Columbia’s Solid Waste Utility (1996): Oral conversation with Colum-bia’s solid waste utility. Contact information or other details are ontheir web site: http://www.gocolumbiamo.com/PublicWorks/Solidwaste

EIERA (1987): Statewide Resource Recovery Feasibility and Planning Study,Volume II, Solid Waste Characterization Report. EnvironmentalImprovement and Energy Resources Authority, State of MissouriDepartment of Natural Resources, Jefferson City, Missouri.

Eugene L. Grant, W. Grant Ireson, and Richard S. Leavenworth. (1982):Principles of Engineering Economy, Seventh Edition. John Wiley &Sons, Inc., New York.

Klee, A.J. (1980): Quantitative Decision Making, Design & Management forResource ® Recovery Series, Vol. 3. Ann Arbor Science, Ann Arbor,Michigan.

Milton, J.S. & Arnold. J.C. (1995): Introduction to Probability and Statistics,third edition. McGraw-Hill, New York.

Minnesota Solid Waste Management Coordinating Board (2000): StatewideMSW Composition Study Final Report. Minnesota Solid Waste Manage-ment Coordinating Board, Minnesota.

42 U.S.C. 6901 et seq. Resource Conservation and Recovery Act of 1976.http://www4.law.cornell.edu/uscode/42/ch82.html. http://web.lexisn-exis.com/congcomp.docu-ment?_m=0ed157f656565121f1cd660a49fc357&_docnum=1&wchp=dGLbVtz-zSkSA&_md5=9a4ce05df4667f0b0c5b81a1dd192fad

RSMo 260.250. (2004): Missouri Revised Statutes. Chapter 260. Environ-mental Control. Section 260.250.

SAS (1990): SAS/STAT User's Guide: Version 6. 4th edition, v.2. SAS Insti-tute Inc., Cary, NC.

SRI International (1992): Data Summary of Municipal Solid Waste Manage-ment Alternatives. NREL.

State of Oregon Department of Environmental Quality. 2002. 2002 OregonSolid Waste Characterization and Composition. Sky Valley Associates,Oregon.

Tchobanoglous, G., Theisen, H., & Vigil, S. (1993): Integrated Solid WasteManagement: Engineering Principles and Management Issues. McGraw-Hill, New York.

Tchobanoglous, G. & Kreith, F. (2002): Handbook of Solid Waste Manage-ment. 2nd edition. New York.

USEPA (1996): Characterization of Municipal Solid Waste in the United States:1995 Update. USEPA 530-R-96-001, PB96-152 160.



p a p e r 05 2005 characterization of s o l i d w a s t e disposed at columbia sanitary ...

Business