fuel consumption, emissions estimation, and emissions cost estimates using global positioning data

Fuel Consumption, Emissions Estimation, and EmissionsCost Estimates Using Global Positioning Data

Betsy J. Agar and Brian W. BaetzMcMaster University, Hamilton, Ontario, Canada

Bruce G. WilsonUniversity of New Brunswick, Fredericton, New Brunswick, Canada

ABSTRACTThe methodology laid out in this paper shows that typicaloperational data from vehicle fleets monitored by a globalpositioning system (GPS) can be used to estimate heavy-duty diesel vehicle (HDDV) emissions, thereby enablingwaste managers and governing bodies to internalize theresponsibility for socioenvironmental costs traditionallyabsorbed by external parties. Although municipal solidwaste (MSW) collection vehicles are the subjects of thisparticular study, the methodology presented here can beapplied to any fleet of vehicles monitored by GPS. Thisstudy indicates that MSW collection trucks may be con-siderably less fuel efficient in the field than publishedvalues for HDDV fuel efficiency suggest. The average fuelefficiency of one MSW collection truck was estimated as0.90 � 0.44 km/L (2.12 � 1.03 mi/gal). This same truckwould generate �42 metric tons of CO2 equivalents/yr,which is comparable to the greenhouse gas emissions of alarge sport utility vehicle driving six times the distance, intown, for a year. In terms of the impacts such emissionshave, projections for the monetary cost of emissions areavailable but highly variable. They suggest that the exter-nal monetary costs of emissions range between 6 and 39%of the annual fuel costs for the studied MSW collectiontruck. The results of this study indicate a need for furtherresearch into valuation of the hidden, external costs ofemissions, borne by local and global socioecological com-munities. The possible implications of this result includepoorly advised fleet procurement decisions and underes-timation of MSW collection fleet emissions.

INTRODUCTIONOver the past four decades, municipal solid waste (MSW)managers have steadily reduced the greenhouse gases(GHGs) and criteria air pollutants (CAPs) associated withMSW management practices. Changes in MSW manage-ment practices (including treatment and transportation)in the United States have prevented 3 million metric tonsCO2 equivalent (MTCO2E) emissions between 1974 and1997, despite waste production nearly doubling.1 How-ever, over this same period, the proportion of GHG emis-sions contributed by collection practices has increasedfrom 1.4% to 12.5% of MSW management activities. Al-though this increase is likely a reflection of the stridesmade in other aspects of MSW management practices,collection vehicle emissions are clearly a growing share ofthe total MSW management practice emissions. Theseemissions could be internalized rather than disregarded asunfortunate, incidental byproducts of MSW collectionmanagement.

The primary objectives of this paper are as follows: (1)to describe a straightforward methodology for estimatingthe driving and idling fuel consumption rates of vehiclefleets, using standard global positioning system (GPS)data and recorded fuel usage; (2) to offer a comparison ofthose estimates with published in-use fuel consumptionrates for heavy-duty diesel vehicle (HDDV) engines; (3) toprovide an estimate of the potential GHGs and CAPsemitted from an MSW collection vehicle; and (4) to illus-trate the potential external costs associated with suchemissions.

These objectives are addressed in four primary sec-tions. First, a brief description of the equipment and datais presented, followed by the procedure for estimating fuelconsumption while idling and driving an MSW collectionvehicle. Next, the procedure for estimating emissionamounts, based on published GHG and CAP emissionrates, is outlined. Finally, the potential range of monetaryimpacts that these emissions impose is discussed. Afterthe methodology section, a case study from Hamilton,Ontario, Canada, is presented.

Fuel Consumption, Emissions, and CostsAccording to Environment Canada, MSW collection ve-hicles are classified as HDDVs,2 which is the same classi-fication assigned to long-haul tractors. U.S. Environmen-tal Protection Agency’s (EPA’s) emission estimation

IMPLICATIONSThis paper shows (1) that published fuel consumption ratesfor HDDV may not be representative for MSW collectionvehicles because of the stop-and-go nature of the activity;(2) there is little information available for MSW fleet man-agers to help them estimate emissions or the external costof those emissions, but we have estimated that externalcost could be as much as 39% of the cost of fuel; and (3)MSW fleet managers may want to consider alternative fuelsbecause of rising fuel costs and emissions from diesel-fuelled trucks.

TECHNICAL PAPER ISSN 1047-3289 J. Air & Waste Manage. Assoc. 57:348–354

Copyright 2007 Air & Waste Management Association

348 Journal of the Air & Waste Management Association Volume 57 March 2007

model, MOBILE6, provides an approximate basic fuelconsumption rate for older HDDVs of 6 mi/gal (unfac-tored), which continues to improve with newer engines.3

Because waste collection trucks mostly drive in stop-and-go conditions, lower fuel efficiency would be ex-pected, yet Wang et al4 propose a fuel efficiency of 9mi/gal (4 km/L) for the MSW collection trucks in theirstudy. That rate was used by Sonesson5 when he con-cluded that “transport does not have a major impact onthe results” of environmental analyses of MSW manage-ment practices.

Lim6 estimates the fuel consumption rate for long-haul tractor engines to be 3.1 L/hr (0.82 gal/hr), whileidling in a laboratory. Although the hauling demands onlong-haul tractor and MSW collection vehicle engines arevery different, the behavior of an engine idling in a lab-oratory setting should not be affected by the type of truckbody that encases it. To overcome the limited informa-tion available for HDDV engines in stop-and-go drivingconditions, Lim’s6 long-haul HDDV idling fuel consump-tion rate is applied in the present study and is the basis forestimating the fuel used by the MSW collection vehiclewhile driving.

Currently there is no practical method to estimate thevolume of GHGs or CAPs emitted by HDDVs in the spe-cific context of MSW collection, a unique application inthat collection vehicles start and stop hundreds of timesevery day. Existing emission estimation models, such asEPA’s MOBILE6, require characteristics such as the vehicleclassification (HDDV), the mix of vehicle miles traveled(the proportion of miles spent on the highway or intown), and the average vehicle speed of the facility orroadway type.7 Although MOBILE6 is a sophisticated anduseful tool, the required inputs for vehicle operation in-sufficiently describe the operational demands on MSWcollection vehicles.

In terms of emission rates, research into HDDVs tendsto focus on long-haul transportation. Reul-Chen et al8

studied the relative performance of passive diesel partic-ulate filters in the MSW context but did not publishaverage emission rates. For the purposes of this study,rates for the most widely studied emissions, including theGHGs, CO2, CH4, and N2O, and CAPs, nitrogen oxides(NOx), CO, and particulate matter (PM; particulate matter�10 �g in diameter, PM10), are adopted from studiesabout long-haul HDDVs2,6,9–11 and listed in Table 1.

GHGs are reported in terms of MTCO2E, an interna-tionally accepted measure that is used in the NationalGHG Inventory issued annually by Environment Canadato the United Nations Framework Convention on ClimateChange (UNFCCC).2

Vehicle operational and maintenance costs tradition-ally include refueling and tune-ups, whereas the effects ofemissions are typically disregarded as external costs, orexternalities, and are unaccounted for by fleet managers.Externalities are costs paid by people who do not directlyexhaust the emissions in question. Health risks, habitatloss, and global warming are just a few of the externalitiescaused by vehicle exhaust emissions for which local andglobal societies “pay.” Unlike many environmental exter-nalities, those generated by public service sectors tend toimpose a cost on the people that they are designed to

serve in the forms of social programs, such as healthcareor clean air campaigns, for example.

In 2004, the average cost of fuel in Hamilton wasUS$1.85/gal ($0.49/L) of diesel. In addition to the at-the-pump costs of fuel, Stodolsky et al10 estimate that idlingincurs the need for additional maintenance, which isestimated as 7% of the fuel cost for oil changes and 7% forengine overhauls. EPA New England estimates that idlingincurs twice the wear-and-tear on engines as comparedwith driving.12 It follows that maintenance because ofdriving costs is 7%, and maintenance because of idlingcosts is 14% of the at-the-pump cost of fuel. Less obviousare the external costs of consuming fuel, including theimpacts of emitting pollutants, the costs of which arevariable and uncertain.

The growing interest in estimating environmentaland social externalities is evident by the rapid appearanceof new online carbon and air pollutant markets in manyof the world’s industrialized countries (including Canada,United States, European Union, United Kingdom, andJapan) and by the recent jump in prices. Chicago ClimateExchange (CCX) credits are issued based on CH4 destruc-tion, agricultural or forestry practices, renewable energy,and other projects. Emissions markets can be tracked onthe Internet in much the same way as long-establishedfinancial markets. On December 12, 2003, the CCX wentonline.13 In January 2004, GHG credits were selling forprices between US$0.90 and US$0.98 per MTCO2E (vin-tage 2004 and 2005, where “vintage year” is described asthe first year a credit can be used for compliance). A yearlater, credits were selling for US$1.74 to US$1.96 percredit (vintage 2005 and 2006). As of May 11, 2006, theprice per MTCO2E ranged between US$3.05 and US$3.15per MTCO2E (vintage 2005 and 2006), an increase of asmuch as 238% in �2 yr. The European Carbon Exchangemarket prices are an impressive three to six times higherthan the North American range of carbon prices on theCCX (US$3.00–5.00/t at close in April 2006). The recentdrop in carbon prices trading in the European carbonmarket, from a high of US$29.00 (€34.85) in April 2006 toa low of US$11.06 (€13.30) in the same month,13 is a

Table 1. Summary of the emission rates used in this study.

Activity Type Emission RateStandardDeviation Source

Idling GHGs N2O 0.08 g/L Environment Canada2

CH4 0.12 g/L Environment Canada2

0.13 g/L0.15 g/L

CO2 2730 g/L Environment Canada2

CAPs NOx 144 g/L 72 g/L Lim6

CO 94.6 g/L EPA9; Stodolsky et al.10

PM10 2.57 g/L EPA9; Stodolsky et al.10

Driving GHGs N2O 0.08 g/L Environment Canada2

CH4 0.12 g/L Environment Canada2

0.13 g/L0.15 g/L

CO2 2730 g/L Environment Canada2

CAPs NOx 6.68 g/km Lindhjem and Jackson11

CO 26.6 g/km Lindhjem and Jackson11

PM10 0.17 g/km Lindhjem and Jackson11

Agar, Baetz, and Wilson

Volume 57 March 2007 Journal of the Air & Waste Management Association 349

result of European countries surpassing emissions reduc-tions, creating a surplus of units. The market value of NOX

credits can also be tracked online.14

Open market trading is just one way of costing theenvironmental externalities of exhaust emissions. Emis-sions costs have also been estimated based on humanmorbidity and mortality because of respiratory damage,the cost and effectiveness of various carbon sequestrationschemes, and willingness-to-pay surveys, to name a few.15

Carbon sequestered (taken up through growth) by forestscan be costed using such methods as avoided damage,contingent valuation, alternative cost, marginal social op-portunity cost, replacement cost, and substitute cost.15 A“conservative” range of median carbon sequestrationcosts lies between US$4.56 and US$4.91 per MTCO2E. Atthe extremes, the value of carbon sequestration has beenestimated within the range of US$0.65 and US$140 perMTCO2E.15

As illustrated by Kulshreshtha et al.,15 the costs ofexposure to CAPs can be estimated by a broad range ofmethods, including the cost of healthcare or the cost ofreducing CAPs. The city of Boulder, CO, has inventoriedthe urban forest based on the ability of the trees to re-moving GHGs and CAPs. The value for removing CAPswas derived from the expected cost of removing air pol-lutants from industrial sources by technical means, as wellas the long-term healthcare costs incurred by people whosuffer from CAP-related diseases. They estimate that theability of urban trees in their dry climate to remove NO2,CO, and PM10 is worth approximately US$7600/t,US$1070/t, and US$5080/t in 2006 dollars, respectively.16

Ideally, costs derived from reducing the CAP emissionsfrom Hamilton’s local industries and treating health is-sues suffered by Hamiltonians would be applied to localemissions measurements.

Externalities incurred by GHGs are “paid for” differ-ently than the externalities incurred by CAPs. GHGs con-tribute to the now widely accepted global issue of climatechange, whereas CAPs affect the health of humans, ani-mals, and plants local to the polluter. As the term “exter-nality” suggests, people do not directly pay for the costsassociated with climate change. International agree-ments, such as the UNFCCC, call for signatories to reduceGHGs through regulatory means. The city of Hamilton isaddressing climate change via programs such as an anti-idling campaign and a tree-planting program. Hamiltoni-ans are subject to global release of GHGs, just as the GHGssequestered or avoided in Hamilton will benefit the globalpopulation. A summary of estimates available for the costof emissions is provided in Table 2.

Emission cost estimates are complicated, and theirrange is consequently broad. The list of estimates in Table2 is not exhaustive; however, it provides a sense of theuncertainty of present estimates, the need for furtherstudy, and, most importantly, the potential long-termimpacts of ignoring such externalities.

EXPERIMENTAL WORKBackground

The city of Hamilton is located at the western tip of LakeOntario and has a population of �490,000. In November2003, Hamilton equipped 5 of their 40 MSW collection

trucks with GPS remote sensor units. The logistics of in-stalling and collecting and analyzing data using GPS areprovided by Agar et al.17 and are not repeated here. Thefleet management system was provided and installed byELM TECHnologies, Ltd., who also collect and store theGPS data. The Telvisant CrossCheck GPRS 1900 was in-stalled to track the truck’s position (longitude–latitude),speed (miles per hour) and heading (degrees, north being0° and 90° being east) every minute from the time that theignition is turned on to the time the ignition is turned off.The GPS also logs the mileage and the corresponding timeof the reading. The data are communicated via satellitethrough a Telvisant Echo LDX. The data analyzed to de-velop the methodology in this work refer to one 2003,25-yd3, rear packer MSW collection truck by Interna-tional, grossing an unloaded weight of 29 t (64,000 lb)and powered by engine model DT 530.

DataFor this study, the data from all of the routes serviced bythe mixed solid waste collection truck were analyzed overa period of 13 months, December 2003 through Decem-ber 2004. The primary GPS data required for this study arethe date, time, speed, times the engine was turned on oroff, and mileage. The data were sampled at 1-min inter-vals. It would be impractical to record an MSW vehicle’severy action, because collection vehicles stop and start sofrequently and briefly. The volume of data would be pro-hibitively large, expensive to store, and such detail havebeen shown to be unnecessary.18

The truck route was plotted using the minute-by-minute position records to illustrate the general activitiesperformed by the crew: morning preparation, travel toand from the route, collection on the route, end-of-daycleanup, and truck maintenance. Each GPS logs 200–300readings per route, per day, a significant and statisticallysound sample size to adequately represent the truck’sdaily activities. At times the GPS was offline or the truckwas undergoing maintenance and data were not recorded,

Table 2. Summary of emission cost estimates (in 2006 US$) used in thisstudy.

Emission Methodology High Low

CO2 equivalent Carbon $405/t $2/tSequestration15

Chicago open market (April 2006)13 $5/t $3/tEuropean Union open market

(April 2006)13

$29/t $11/t

NOx Open market (average forApril 2006)14

$2474/t

NO2 Urban forest16 $7600/tCO Urban forest16 $1070/tPM10 Urban forest16 $5080/t

Notes: The values in the above table were calculated using the appropriatecombination of the online Bank of Canada historical currency converter orinflation calculator, the European Central Bank foreign exchange rates, or theonline Federal Reserve Bank of Minneapolis inflation calculator depending onthe original currency in which the estimate was reported, available at http://www.bankofcanada.ca/en/exchangeconvert.htm; http://www.bankofcanada.ca/en/inflation_calc.htm; http://www.ecb.int/stats/exchange/eurofxref/html/index.en.html; and http://minneapolisfed.org/research/data/us/calc/, respectively.



however, the abundance of records and the randomnessof sampling should assuage concerns about missing data.In addition to the GPS information, the city of Hamiltonprovided the fuel records corresponding with the studyperiod. Fuel records include the time, date, volume, andodometer reading. The time and date were matched to theGPS data based on the time and date of refueling.

Fuel ConsumptionThe GPS unit is programmed to signal when the enginehas been turned on or off. The sum of the time differencesbetween initial “ignition on” and the sequential “ignitionoff” events is the total time that the engine is running.The total time is further broken down based on whetherthe truck is “moving” or “not moving” during the mo-ment between present and previous time stamps. Thetruck activity is based on the instantaneous speed re-corded by the GPS unit at the end of each moment. Theamount of time that the truck is idle (i.e., registering aspeed of 0 mi/hr) is tallied for the period of study, and thedifference between the total and idle times is assumed tobe the time spent driving. It is also assumed that the fuelis consumed completely while idling and driving. That isto say, fuel lost to evaporation, leaks, or spills is rolled intothe total fuel used while idling or driving.

Typically, fuel consumption rates while driving arebased on distance driven. In this case, the rate of fuelconsumed while driving is derived from Lim’s6 fuel con-sumption rate while idling, which is based on time; theamount of fuel consumed while idling is subtracted fromthe total amount of fuel consumed to obtain the amountof fuel consumed while driving during the study period.The fuel consumption rate while driving is subsequentlythe quotient between the total distance driven and theamount of fuel not consumed while idling. Using thismethod, the average fuel efficiency for the truck over the13-month study period was estimated to be 0.90 � 0.44km/L (2.12 � 1.03 mi/gal).

EmissionsOnce established, the fuel consumption rates are thencombined with average GHG and CAP emission rates. Theemission outputs are estimated as the product of mea-sured fuel consumption rates and published average emis-sion rates. The emission rates used in this study are drawnfrom several sources, although preference is given to themost recent studies and to Canadian studies, because thesubject vehicle operates in Canadian climate conditionsand under Canadian emission control regulations. It isimportant to note that emissions and engine performancedirectly depend on the climate in which the vehicle op-erates. The rates that we apply refer to engine perfor-mance in different regions. Local emissions monitoring isnecessary to improve the accuracy of the emissions out-put estimates.

The different CH4 emission rates suggested by Envi-ronment Canada reflect the CH4 emission controls avail-able for HDDVs. The level of control in place on the studytruck is unknown. Therefore, the average of the threerates (0.13 g/L) was used.

The hourly rate of fuel used while idling was multi-plied by each emission rate to estimate the hourly emis-sions output. For example, 2730 g/L * 3.1 L/hr equals 8463g/hr of CO2. At these rates, the truck would emit 8MTCO2E, idling 940 hr, and 42 MTCO2E, driving 14,000km/yr (based on 52 weeks, 5 days/week, 7 hr/day). Therelative magnitude of these emissions can be understoodby way of comparison with a large family automobile.According to www.fueleconomy.gov, an interactive web-site developed by U.S. Department of Energy and EPA, a2003, gasoline powered, four-wheel drive Lincoln Aviatorwould have to drive 81,000 km in town to emit the sameamount of CO2 equivalents (42 MTCO2E). In this frame ofreference, the MSW collection truck emits almost sixtimes as much GHG as a large family sport utility vehicleper kilometer driven.

Direct Economic ImpactsBetween December of 2003 and December 2004, the cityof Hamilton paid an average of $1.85/gal ($0.49/L) of fuel.The subject truck in the present study uses an estimatedaverage of $1855 worth of fuel while idling and $9820worth while driving, annually. At 14% of the fuel cost, theadditional cost of wear-and-tear caused by idling is�$260/yr. At 7% of the fuel cost, the cost of enginemaintenance associated with driving is �$690/yr.

RESULTS AND DISCUSSIONFuel Consumption

In contrast to published rates, the fuel consumption rateestimated in the present study is �0.9 km/L (2.1 mi/gal),which is consistent with the rate historically used in MO-BILE6,3 when the stop-and-go driving patterns of MSWcollection vehicles are taken into account, but consider-ably less than the rate presented by Wang et al.4 andapplied by Sonesson.5 The significant dispersion indicatedby the variance of the data is potentially caused by load orseasonal variations. Table 3 summarizes the publishedfuel efficiency rates for HDDVs.

EmissionsTable 4 lists the GHG and CAP emissions in order ofoutput amount. The time spent idling is approximatelyequal to the time spent driving for the study vehicle. Notsurprisingly, more fuel is consumed while driving thanwhile idling. More GHGs (CO2, N2O, and CH4) are emit-ted while driving, which is a result of applying emissionrates reported in terms of volume of fuel consumed, re-gardless of the operational details of the engine. Thispoint can be illustrated by comparing the ratios betweenthe GHGs emitted and the fuel consumed, while drivingand while idling. For example, the ratio while idling is

Table 3. Comparison of published fuel efficiency rates for HDDVs.

Source Rate

Wang et al.4 9 mi/galSonesson5

EPA MOBILE63 6 mi/galCurrent study 2.1 mi/gal



42,201 MTCO2E/84%, which is equal to the ratio whiledriving 7978 MTCO2E/16%, or 0.5 MTCO2E/%.

The rates used to estimate the CAP (NOx, CO, andPM10) volumes are based on time or distance. In thesecases, the ratio between the volume of CAP emitted andthe proportion of fuel consumed while idling does notnecessarily equal the emission-to-fuel consumption ratiowhile driving (e.g., 136 kg/16% � 96 kg/84%). In fact, theestimates for NOx and PM10 suggest that these CAPs areemitted at higher rates while idling than while driving.The CO emission-to-fuel consumption ratios are almostequal, suggesting that CO is emitted in direct proportionto the amount of fuel consumed regardless of the vehicle’sactivity.

External Costs of EmissionsThe average costs of fuel and vehicle maintenance, in-cluding oil changes and engine overhauls, have been es-timated and summarized in Table 5. The average cost offuel is based on the purchase records provided by the city.The final column of Table 5 highlights the dramatic in-crease in fuel costs over the last 2 yr (from $0.49/L in 1994to $0.74/L in April 2006, according to Statistics Cana-da19). With rising fuel costs and the high fuel consump-tion rate determined in the present results, MSW collec-tion managers have cause to reevaluate decisions aboutengine choice for their fleets, because alternative fuelsmay prove more cost-effective in the present economicclimate.

Assigning costs to environmental externalities is, atbest, a challenge. The costs presented in Table 6 are in-tended to raise awareness that emissions have conse-quences and to introduce the possibility of internalizing

the costs. Several researchers have attempted to evaluatethose consequences, and each study reports costs that arehighly variable from the others. For this reason, upperand lower bounds for each cost estimate are presented inTable 6. The external emission costs of consuming fuelduring MSW collection potentially amount to as much as39% and are, at a minimum, at 6% of the fuel cost.

Assumptions and LimitationsLim’s6 emission rates are based on steady state conditions,which were reached after 3 hr of idling. MSW collectionvehicles do not typically idle for long periods.

By virtue of loading waste into the compactor alongthe collection route, the load on the MSW collectiontruck engine continually increases. In addition, for themajority of time, MSW collection vehicles travel localroads deliberately designed with curves, stops, and slopes.The fuel consumption and vehicle emission rates used inthis study are based on average driving activity and vehi-cle operation. As a result, they may not directly reflecttypical conditions for MSW collection vehicles.

Under UNFCCCs regulations, Environment Canada isnot required to inventory CAPs, because they have locallyrather than globally detrimental effects on health and thenatural landscape. That said, Environment Canada indi-cates that PM emissions are inadvertently accounted for,to some extent, in their GHG emissions factors.2 In addi-tion, CO emissions are considered part of the CO2 emis-sion rates, because CO is assumed to completely oxidizewithin 20 weeks of release and becomes CO2.2 Therefore,the factors included for CO and PM9–11 potentially over-lap with Environment Canada’s estimates, resulting inthe potential for some overestimation of these emissions.

Similarly, the costs presented in this research do notaccount for potential overlapping costs or benefits that mayresult from mutual effects of reducing CAPs and GHGs. Ithas been argued that the interdependence between CAPsand GHGs should be considered in policy decisions to avoidunnecessarily high costs of regulations.20

This study was conducted in Southern Ontario wherethe climate varies significantly between winter and sum-mer. The time frame of this work does not permit inves-tigation into the differences in fuel consumption betweenseasons, because each season has only been sampled once(13 consecutive months of data). Ideally, the fuel con-sumption and emissions rates should come from local

Table 4. Summary of the annual emissions estimated for the MSW collection vehicle operating 52 weeks/yr, 5days/week, and 7 hr/day.

Activity Idling Driving

Time spent per activity (percentage of total) 52% 48%Proportion of fuel consumed 16% 84%Average rate of fuel consumption 3.1 l/hr 0.9 km/LCO2 (per liter rate) 7978�10�3MTCO2E 42,201�10�3MTCO2EN2O (per liter rate) 72�10�3MTCO2E 383�10�3MTCO2ECH4 (per liter rate) 8�10�3MTCO2E 43�10�3MTCO2ENOX 135.6 kg 95.7 kgCO 89.1 kg 370.8 kgPM10 2.4 kg 2.4 kg

Table 5. Summary of the average idling and driving fuel consumptionand engine maintenance cost estimates for the study period.

VariableFuel

(2004 US$)Maintenance(2004 US$)

Fuel(at 2006 at-the-pump rates)

Idling $1855 $260 $3195Driving $9820 $690 $16,910Total $11,675 $950 $20,100

Notes: The sudden rise in the oil market is inconsistent with annual Canadianinflation rates. As such, 2006 projections for the maintenance costs have notbeen projected here.



measurements and studies. However, as noted previously,such studies are not presently available.

CONCLUSIONSIn this study, emissions are estimated by multiplying therate of fuel consumption during driving and idling bypublished emissions factors. This work asserts that esti-mating emissions from MSW collection vehicles usingGPS data is feasible.

This paper presents a method of estimating the emis-sions and their associated costs of one MSW collectionvehicle. The results of this study do not conclude that allMSW collection trucks consume fuel at a rate of 0.9 km/L(2.1 mi/gal). However, the present results do give cause forMSW collection researchers and managers to reconsiderthe apparent efficiency of MSW collection trucks. The fuelefficiency rate while driving that was estimated in thisstudy is considerably lower (less efficient) than publishedrates.3–5 Further investigation into the efficiency of MSWcollection trucks is needed.

Compounding the underestimation of fuel consump-tion is the lack of emissions research specific to MSWcollection vehicle operational context. Specifically, thisstudy illustrates that the rate of emission of CAPs dependson the operating conditions (idling or driving in a stop-and-go pattern). The present methodology provides anopportunity for fleet managers to contribute to a bot-tom-up emissions inventory, using “borrowed” emissionsestimates, estimates that can be easily substituted as re-search specific to MSW collection vehicles emerges.

Frey et al.21 demonstrate the feasibility of directlymeasuring a selection of GHGs and CAPs using “portableonboard tailpipe emissions measurement systems” in-stalled on gasoline-powered highway vehicles. Their ap-proach requires an additional computer program tomatch time stages, a step that could be eliminated byintegrating the tailpipe data into existing geographic in-formation systems.

The environmental challenges that waste managersare accustomed to tackling are traditionally rooted intreating the waste itself, but mounting environmental

concerns are forcing new approaches to not only thetreatment of waste but also the handling of waste, such asMSW collection. Without supporting information, man-agers are ill-equipped to make holistic, environmentallysound decisions about programs that potentially remainin place for decades.

As the current upward trend of fuel costs continues,MSW managers making fuel or engine choices should giveconsiderable attention to the fuel efficiency specific toMSW collection in the present study. The subject vehiclewas new at the time that this research was conducted, yetits fuel efficiency while driving was shown to be consid-erably less than published values. Coupled with the ex-ternalities not accounted for, MSW collection managersmay be forming decisions about engine types without afull understanding of their in-service performance withinthe MSW collection context.

This study shows that the costs of emitting CAPs andGHGs are not trivial relative to the magnitude of thedirect operational costs. Inevitably, local governments in-ternalize the impacts of MSW collection truck emissionsthrough other municipal programs. MSW collection man-agers could potentially use the present methodology tocompare MSW collection choices: tipping the scale infavor of the option expected to incur the lowest externalcost. For example, queuing at the treatment facility mightbe addressed via prescheduled unloading times or build-ing a new facility. Prescheduled loading times can resultin more trips between the facility and the collection. Anew facility could result in shorter distances to somecollection areas. Each of these options likely increases thetime spent driving but also decreases the time spentidling. The present methodology offers a means to com-pare such options.

In addition, MSW collection managers could be con-sulted during the planning stages of roadways, housingdensity, treatment facility location, or community loca-tion. Based on comparisons of emissions expected to re-sult from the relevant options, MSW collection managerscould argue for reducing deadheading, decreasing the fre-quency of stops, and minimizing the distance traveled to

Table 6. Summary of the annual externality cost range estimates for emissions (2006 US$).

Emission

Idling Driving Total

High Low High Low High Low

N2O (per g/L) $8.01 $0.04 $42.38 $0.19 $42.39 $0.23CH4 (per g/L) $0.90 $0.00 $4.78 $0.02 $5.68 $0.02CO2 (per g/L) $881.92 $4.06 $4664.85 $21.46 $5546.77 $25.52NOx idling (per g/L) and driving (per g/km) $1030.14 $335.42 $727.02 $236.72 $1757.15 $572.14CO $95.41 $397.18 $492.59Idling (per g/L)Driving (per g/km)PM10 $12.30 $12.16 $24.46Idling (per g/L)Driving (per g/km)Total $2030 $445 $5850 $670 $7880 $1,115Percent of projected 2006 fuel costs 39% 6%

Notes: The cost estimates presented here are derived from the highest and lowest published rates summarized inTable 5.



the treatment facility. In general, factoring in the esti-mated external emissions costs of municipal pollutionsources could change decisions about collection fre-quency, fuel options, and urban planning and reduce theburden on healthcare, for example.

ACKNOWLEDGMENTSThis research was supported by the Natural Sciences andEngineering Research Council of Canada, and the coop-eration and participation of the City of Hamilton hasbeen invaluable.

REFERENCES1. Weitz, K.A.; Thorneloe, S.A.; Nishtala, S.R.; Yarkosky, S.; Zannes, M.

The Impact of Municipal Solid Waste Management on GreenhouseGas Emissions in the United States; J. Air & Waste Manage. Assoc. 2002,52, 1000-1011.

2. Environment Canada. Canada’s National Inventory Report (NIR); Green-house Gas Division Environment Canada; available at http://www.ec.gc.ca/pdb/ghg/inventory_e.cfm (accessed 2006).

3. U.S. Environmental Protection Agency. Update Heavy-Duty EngineEmission Conversion Factors for MOBILE6: Analysis of BSFCs and Calcu-lation of Heavy-Duty Engine Emission Conversion Factors; EPA420-P98-015; Government Printing Office: Washington, DC, 1998; available athttp://www.epa.gov/otaq/models/mobile6/r02005.pdf (accessed 2006).

4. Wang, F.S.; Richardson, A.J.; Roddick, F.A. SWIM Ð A Computer Modelfor Solid Waste Integrated Management; Comput. Environ. Urban Sys-tems 1996, 20, 233-246.

5. Sonesson, U. Modelling of Waste Collection: a General Approach toCalculate Fuel Consumption and Time; Waste Manage. Res. 2000, 18,115-123.

6. Lim, H. Study of Exhaust Emissions from Idling Heavy-Duty Diesel Trucksand Commercially Available Idle-Reducing Devices; EPA420-R-02-025;U.S. Environmental Protection Agency: Washington, DC, 2002; avail-able at http://epa.gov/smartway/documents/epaidlingtesting.pdf#search%22EPA420-R-02-025%22 (accessed 2006).

7. U.S. Environmental Protection Agency. User’s Guide to MOBILE6.1 andMOBILE 6.2 (Mobile Source Emission Factor Model); EPA420-R-03-010;Government Printing Office: Washington, DC, 2003; available athttp://www.epa.gov/otaq/models/mobile6/420r03010.pdf (accessed 2006).

8. Reul-Chen, C.K.; Ross, C.; Steele, N.L.C., Winer, A.M. Exhaust Tem-perature Profiles for Application of Passive Diesel Particulate Filters toSolid Waste Collection Vehicles in California; J. Air & Waste Manage.Assoc. 2005, 55, 241-247.

9. U.S. Environmental Protection Agency. Emission Facts; Idling VehicleEmissions; EPA420-F-98-014; Government Printing Office: Washing-ton, DC, 1998; available at http://www.epa.gov/OMS/consumer/f98014.pdf#search%22EPA420-F-98-014%22 (accessed 2006).

10. Stodolsky, F.; Gaines, L.; Vyas, A. Analysis of Technology Options toReduce the Fuel Consumption of Idling Trucks; ANL/ESD-43; Center forTransportation Research, Energy Systems Division, Argonne NationalLibrary: Argonne, IL, 2000; available at http://www.transportation.anl.gov/pdfs/TA/15.pdf (accessed 2006).

11. Lindhjem, C.; Jackson, T.; Beardsley, M. Update of Heavy-Duty EmissionLevels (Model Years 1988–2004) for Use in MOBILE6; EPA420-R-02-018;U.S. Environmental Protection Agency: Washington, DC, 2002.

12. What You Should Know About Truck Engine Idling; Save Money and ReducePollution [Factsheet]. U.S. Environmental Protection Agency New En-gland: Boston, MA, 2003; available at http://www.epa.gov/region01/eco/diesel/assets/pdfs/Diesel_Factsheet_Truck_Idling.pdf (accessed 2003).

13. Chicago Climate Exchange. Online Carbon Trading; available at http://www.chicagoclimatex.com/ (accessed 2005).

14. Evolution Markets LLC. Online NOX Trading; available at http://www.evomarkets.com/mdata/index.php (accessed 2005).

15. Kulshreshtha, S.N.; Lac, S.; Johnston, M.; Kinar, C. Carbon Sequestra-tion in Protected Areas of Canada: an Economic Valuation. ResearchReport for the Federal-Provincial Parks Council of Canada, Depart-ment of Agricultural Economics, University of Saskatchewan: Saska-toon, Saskatchewan, 2000; p 142; available at http://www.cd.gov.ab.ca/preserving/parks/fppc/carbonsequestration.pdf (accessed 2006).

16. City of Boulder Water Conservation Office. Calculating the Value ofBoulder’s Urban Forest; City of Boulder Water Conservation Office:Boulder, CO, 2002; available at http://bcn.boulder.co.us/basin/boulder/urbanforest/ (accessed 2006).

17. Agar, B.; Wilson, B.G.; Winning, A.; Baetz, B.W. Measuring the Per-formance of Solid Waste Collection Vehicles Using Global PositioningSystems. In 9th Environmental Engineering Specialty Conference; Cana-dian Society for Civil Engineering: Toronto, Ontario, Canada, 2005;Paper EV-110, p 10.

18. Wilson, B.G., Baetz, B.W. Modeling Municipal Solid Waste CollectionSystems Using Derived Probability Distributions. I: Model Develop-ment. J. Environ. Eng. ASCE 2001, 127, 1031–1038.

19. Statistics Canada. Gasoline and Fuel Oil, Average Retail Prices by UrbanCentre (monthly) (Toronto); available at http://www40.statcan.ca/l01/cst01/econ152h.htm (accessed 2006).

20. Burtraw, D.; Krupnick, A.; Palmer, K.; Paul, A.; Toman, M.; Bloyd, C.Ancillary Benefits of Reduced Air Pollution in the US from ModerateGreenhouse Gas Mitigation Policies in the Electricity; J. Environ. Econ.Manage. 2003, 45, 652-673.

21. Frey, H.C.; Unal, A.; Rouphall, N.M.; Colyar, J.D. On-Road Measure-ment of Vehicle Tailpipe Emissions Using a Portable Instrument; J. Air& Waste Manage. Assoc. 2003, 53, 992-1002.

About the AuthorsBetsy Agar is a research engineer in civil engineering atMcMaster University. Bruce Wilson is an associate profes-sor in civil engineering at the University of New Brunswick.Brian Baetz is a professor of civil engineering and thedirector of the Engineering and Society Programme at Mc-Master University. Address correspondence to Betsy Agar,McMaster University, 1280 Main Street West, Hamilton,Ontario, Canada L8S 4K1; phone: 1-905-308-9952; fax:1-905-529-9688; e-mail: [email protected].



fuel consumption, emissions estimation, and emissions cost estimates using global positioning data

Documents