Total Fuel-Cycle Analysis of Heavy-Duty Vehicles Using Biofuels and Natural Gas-Based Alternative Fuels

ABSTRACT

Heavy-duty vehicles (HDVs) present a growing energy and environmental concern worldwide. These vehicles rely almost entirely on diesel fuel for propulsion and create problems associated with local pollution, climate change, and energy security. Given these problems and the expected global expansion of HDVs in transportation sectors, industry and governments are pursuing biofuels and natural gas as potential alternative fuels for HDVs. Using recent lifecycle datasets, this paper evaluates the energy and emissions impacts of these fuels in the HDV sector by conducting a total fuel-cycle (TFC) analysis for Class 8 HDVs for six fuel pathways: (1) petroleum to ultra low sulfur diesel; (2) petroleum and soyoil to biodiesel (methyl soy ester); (3) petroleum, ethanol, and oxygenate to e-diesel; (4) petroleum and natural gas to Fischer–Tropsch diesel; (5) natural gas to compressed natural gas; and (6) natural gas to liquefied natural gas. TFC emissions are evaluated for three greenhouse gases (GHGs) (carbon dioxide, nitrous oxide, and methane) and five other pollutants (volatile organic compounds, carbon monoxide, nitrogen oxides, particulate matter, and sulfur oxides), along with estimates of total energy and petroleum consumption associated with each of the six fuel pathways. Results show definite advantages with biodiesel and compressed natural gas for most pollutants, negligible benefits for e-diesel, and increased GHG emissions for liquefied natural gas and Fischer–Tropsch diesel (from natural gas).

IMPLICATIONS

This paper evaluates total fuel-cycle energy use and emissions of several alternative fuels used in Class 8 HDVs. The paper uses current data and includes sensitivity analysis of key variables. The results will help inform decision-makers considering programs and policies aimed at encouraging alternative fuels in the trucking sector.

INTRODUCTION

Evidence has mounted that heavy-duty vehicles (HDVs) are a significant and growing source of air pollution in the United States and globally.1–4 For example, in the United States, trucks over 8500 lb annually contribute approximately 370 million metric tons of carbon dioxide (CO₂).5This represents approximately 22% of the total transportation sector greenhouse gas (GHG) emissions and approximately 7% of the total energy-related GHG emissions for the country. In addition, in some regions, HDVs can contribute over 30% of emissions of nitrogen oxides (NO_x) and over 60% of particulate matter (PM) emissions.6 Lastly, the HDV sector is energy intensive and almost entirely reliant on petroleum, consuming nearly 600 million barrels of petroleum annually in the United States alone.7

Trucks are categorized into eight classes, depending on vehicle weight, with Class 8 accounting for vehicles greater than 33,000 lb gross vehicle weight rating.8 In 2002 there were 2.2 million Class 8 HDVs in operation in the United States, accounting for 2.5% of all trucks on U.S. roads.9 Although Class 8 trucks account for a small percentage of total trucks, they are responsible for a disproportionately large percentage of annual truck vehicle miles traveled (VMT) and fuel consumption. Whereas the average VMT for all trucks is approximately 13,000 mi/yr, Class 8 HDVs travel an average of approximately 46,000 mi/yr, accounting for 30% of total truck VMT and 21% of total truck fuel consumption.9–11 Class 8 HDVs tend to accumulate many miles per year because they are typically used for long-haul services.

The disproportionate share of fuel consumption by Class 8 HDVs is also attributable to low fuel economy relative to other transport modes. Class 8 HDVs have the lowest fuel economy of the eight vehicle classes, achieving a fleet harmonic mean of 5.7–5.9 miles per gallon (mpg).9,11,12 Whereas the fuel economy of cars, vans, pickup trucks, and sport utility vehicles increased an average of 2.1% from 1997 to 2002, the fuel economy of Class 8 HDVs actually decreased 14.8% over the same period on a mpg basis.9,13 However, despite the poor fuel economy (in mpg) of HDVs, energy intensity on a Btu per ton-mile basis is very low and has historically decreased. From 1975 to 2005, the amount of fuel required to move a given amount of freight a given distance reduced by more than half.14 Projections show that HDV energy intensity may fall 20% by 2050 from 1990 levels, but energy intensity of light-duty vehicles (LDVs) may fall more than twice as much, or 47% over the same period.15 Furthermore, projections show that domestic shipping ton-miles will increase approximately 25% from 2010 to 2035.7 These trends lead to expectations that energy consumption in the heavy-duty trucking sector will annually increase approximately 1.5% between 2010 and 2035.7

Currently in the United States, almost all large trucks use compression ignition (CI) engines that burn ultra low sulfur diesel (ULSD) fuel with a sulfur content of 15 parts per million (ppm) or less. Diesel is the fuel of choice for HDVs, rather than gasoline, because diesel engines are inherently more efficient than gasoline engines.16 Given the mounting concerns about climate change, local air pollution, and energy security, many nations are looking to alternative fuels to help power the freight trucking sector. Fuels such as biodiesel (BD), e-diesel (ED), Fischer–Tropsch diesel (FTD), compressed natural gas (CNG), and liquefied natural gas (LNG) are options being seriously considered.

Although these alternative fuels show promise, infrastructure and cost challenges must be overcome before significant market penetration will be achieved.17–19 Additionally, the overall benefits of alternative fuels must be evaluated through fuel-cycle models that account for emissions and energy use in fuel production, processing, distribution, and use (i.e., a fuel's “total fuel cycle” or TFC).20 This paper conducts a TFC analysis to assess the environmental performance of conventional and alternative fuels in the HDV sector. Although focused on Class 8 HDVs, these methods and results can be readily extended to other classes of HDVs.

ANALYTICAL APPROACH: TFC ANALYSIS

Measuring the total energy and emissions implications of HDVs requires TFC analysis (TFCA).21 TFCA considers energy use and emissions in each stage of fuel production and use, from the extraction or harvesting of feedstock (e.g., petroleum from the ground or soybeans from the field) to vehicles' end use of the processed fuel.22–25 Each stage in the fuel cycle includes activities that produce GHGs and other emissions. Figure 1 demonstrates the upstream and downstream stages of the TFC. This paper focuses on the direct energy use and emissions along a particular fuel pathway; indirect emissions are discussed in the final section of the paper.

Figure 1. Upstream and downstream stages of the TFC for conventional and alternative fuels.

Several reports and papers have discussed TFCA in a LDV context.25–38 TFCA has also been applied to different fuels in the marine sector24,39 and intermodal evaluations.40 With respect to the HDV sector, there is a limited amount of literature, only some of which includes TFCA. Some noteworthy HDV literature includes a recent TFCA of energy consumption and emissions for BD in construction vehicles41; an analysis that considers HDVs using dimethyl ether, liquid petroleum gas, CNG, BD, ethanol, methanol (M85), and hydrogen42; a 1998 analysis of HDVs using FTD43; a study on LNG and CNG in HDVs engines44; and a study on CNG, BD, and ethanol use in HDVs operating in Australia.23 This paper differs from these earlier analyses by presenting a TFCA that considers a more complete fuel cycle (i.e., greater detail of precombustion events), uses current data and recently published projections, and uses new emissions data from HDV field tests conducted in the last few years.

Because some recent work has identified the importance of local or regional assumptions, on TFC analyses,45–47 the model presented here is applied to HDV operations in New York State (NY), although the methodology could be nationally, internationally, and globally extended to other regions. The model presented here is based on the NY Greenhouse Gas, Regulated Emissions, and Energy Use in Transportation (NY-GREET) model,38 which is itself based on the LDV GREET model developed at Argonne National Laboratory30,48 but modified with NY inputs to characterize fuel and feedstock production, distribution, and use. The GREET model has been widely used and extensively peer reviewed, and the reader is referred to the literature for details about model development.25,31,49,50 NY-GREET makes use of GREET's peer-reviewed algorithms without alteration, but it differs from GREET in the following ways: NY-derived feedstock and fuel transportation and distribution distances; NY farming efficiencies, fertilizer use, and energy use; NY electricity production fuel mixes; and NY average end-use vehicular fuel mixes. This study further modifies NY-GREET to include a HDV operation module, and thus the model is referred to as NY-GREET-HDV.

NY-GREET-HDV conducts TFC analyses for six fuel pathways: (1) petroleum to ULSD; (2) petroleum and soybean oil (soyoil) to BD; (3) petroleum, ethanol, and oxygenate additive to ED; (4) petroleum and natural gas to FTD; (5) natural gas to CNG; and (6) natural gas to LNG. Figure 2 illustrates these pathways. NY-GREET-HDV calculates emissions of three GHGs (CO₂, nitrous oxide [N₂O], and methane [CH₄]) and five other pollutants (volatile organic compounds [VOCs], carbon monoxide [CO], NO_x, PM with aerodynamic diameters ≤ 10 μm [PM₁₀], and sulfur oxides [SO_x]). NY-GREET-HDV also calculates consumption of total energy and petroleum associated with each of the six fuel pathways.

Figure 2. Six fuel pathways analyzed for the HDV sector.

NY-GREET-HDV applies GREET's unmodified upstream (feedstock extraction/harvesting, fuel production, and fuel distribution) energy consumption and default emissions factors, with the exception of those modified to represent NY conditions and HDV operations (see below). The model uses HDV-specific input variables including (1) vehicle model year, freight load, and truck weight; (2) emission factors (EFs) of alternative fuel HDVs; and (3) fuel mixes for use in HDVs (i.e., percent soyoil in BD or percent ethanol in ED). In regards to freight load, it is noted that payload can vary from vehicle to vehicle and can influence engine performance, fuel economy, and EFs for any vehicle type.51,52 This work analyzes alternative fuels and, similar to other analyses of this type, a constant average duty cycle load across all fuel types is assumed to evaluate the total impact of alternative fuel usage.

The model considers these “HDV end-use” parameters along with fuel production and distribution calculations from the TFC model to calculate energy use in Btu per ton-mile (Btu/t-mi) and grams of pollutant per ton-mile (g/t-mi). Results are presented as “per ton-mile” because this is a common metric used to measure the “work done” by HDVs in moving freight.

There are recognized limitations of GREET, NY-GREET, and other TFCA models, the primary challenge being where to draw the boundary that encompasses the energy use or emissions considered in the analysis. For example, when using a fuel type that has not yet reached full market penetration (such as BD, ED, or FTD), vehicles may have to re-route to refuel until the infrastructure grows, and such phenomena are not considered by most TFCA. Furthermore, the type of fuel or energy used in upstream stages of the fuel cycle is often ambiguous and, in particular, was not considered as part of the analysis presented here. There are additional factors that lay outside of the realm of most TFCA but could have real-world impact, such as the contribution to emissions of the activities of any additional employees (displacement of behaviors) needed to develop the necessary infrastructure to deliver alternative fuels.

FUEL TYPES AND ASSUMPTIONS

Table 1 provides specific fuel characteristics for ULSD and the five alternative fuel types analyzed here. Tail-pipe EFs for each fuel type and TFC results (to be discussed later) are presented in Table 3. In NY-GREET-HDV, EFs for ULSD were compiled from GREET for CO and NO_x, the U.S. Environmental Protection Agency (EPA)53 for CH₄ and N₂O, and the Federal Highway Administration54 for VOCs and PM. CO₂ and SO_x EFs were calculated from ULSD carbon content, sulfur content, mass density, and vehicle fuel economy.55 EFs for alternative fuels were calculated by multiplying the baseline ULSD EF by an EF ratio found in studies focusing on Class 8 HDVs.1,2,12,56–59 It is noted that vehicle characteristics differed slightly within the cited literature, although these variations were negligible and present a limited degree of uncertainty.

Table 1. Fuel specifications

	ULSD	BD20	ED10	FTD100	CNG	LNG
Fuel composition	100% ULSD	80% ULSD, 20% BD	89% ULSD, 10% ethanol, 1% additive	100% natural gas-based FTD	100% CNG	100% LNG
Sulfur content ^a (ppm by weight)	15.0	12.0	13.7	0.0	6.0	0.0
Carbon content ^b	87.1%	85.2%	83.6%	85.3%	72.4%	75%
Lower heating value (Btu/gal or btu/ft3) ^b	129,488	127,500	124,038	123,670	983	74,720
Mass density (g/gal or g/ft3) ^b	3,206	3,237	3,180	3,017	22	1,621
Fuel economy ^c (mpg)	5.9	5.9	5.6	5.9	4.7	4.7
Efficiency (mi/MBtu)	45.6	45.8	42.9	45.4	36.1	36.4
Notes: ^aSulfur content for neat fuels are from the fuel specifications section of GREET, except for ULSD, which is assumed as 15 ppm to meet on-road diesel standards. Sulfur content for mixed fuels are calculated based on the ULSD baseline and composition ratio. ^bFrom GREET fuel specifications. ^cULSD fuel economy rating from the National Renewable Energy Laboratory12; rating of other fuels calculated using each fuel's energy content relative to ULSD and estimated efficiency values from the literature.2,57,59 The assumed efficiency differences are as follows: BD20 (+0.5%), ED10 (−6.0%), FTD (−0.3%), CNG (−20.7%), and LNG (−20.2%).

Table 3. Results from HDV TFC analysis for energy use, petroleum consumption, and select emissions

Item	Feedstock Stage	Fuel Processing Stage	Vehicle Use Stage (Tailpipe EFs)	Percent Vehicle Use Stage	Total	Ratio of Alternative Fuel Total to ULSD
ULSD
Total energy (Btu/t-mi)	60	120	1095	86	1280	–
Petroleum (Btu/t-mi)	10	60	1095	94	1170	–
CO2 (g/t-mi)	5	10	85	86	100	–
CH4 (mg/t-mi)	90	10	<1	<1	100	–
N2O (mg/t-mi)	<1	<1	<1	52	<1	–
GHGs (g/t-mi)	10	10	85	84	105	–
VOC (mg/t-mi)	5	5	15	65	20	–
CO (mg/t-mi)	5	5	230	95	240	–
NOx (mg/t-mi)	20	15	470	93	505	–
PM10 (mg/t-mi)	<1	5	5	39	10	–
SOx (mg/t-mi)	5	10	<1	5	15	–
BD20
Total energy (Btu/t-mi)	100	405	1090	69	1590	1.25
Petroleum (Btu/t-mi)	35	50	840	91	925	0.80
CO2 (g/t-mi)	−5	10	85	98	90	0.90
CH4 (mg/t-mi)	80	15	<1	<1	90	0.90
N2O (mg/t-mi)	<1	<1	<1	9	<1	NA
GHGs (g/t-mi)	−5	10	85	95	90	0.90
VOC (mg/t-mi)	5	20	15	35	40	2.00
CO (mg/t-mi)	10	5	215	94	230	0.95
NOx (mg/t-mi)	25	15	490	92	530	1.05
PM10 (mg/t-mi)	<1	5	5	30	10	1.00
SOx (mg/t-mi)	15	10	<1	4	25	1.65
ED10
Total energy (Btu/t-mi)	80	205	1165	80	1455	1.15
Petroleum (Btu/t-mi)	20	60	1080	93	1160	1.00
CO2 (g/t-mi)	<1	10	85	87	100	1.00
CH4 (mg/t-mi)	90	15	<1	<1	110	1.10
N2O (mg/t-mi)	5	<1	<1	8	<1	NA
GHGs (g/t-mi)	5	10	85	84	100	1.00
VOC (mg/t-mi)	5	5	15	56	25	1.25
CO (mg/t-mi)	10	5	205	93	220	0.90
NOx (mg/t-mi)	25	15	470	92	515	1.00
PM10 (mg/t-mi)	<1	5	5	29	10	1.00
SOx (mg/t-mi)	10	10	<1	4	20	1.35
FT100
Total energy (Btu/t-mi)	60	570	1100	64	1735	1.35
Petroleum (Btu/t-mi)	5	15	<1	<1	20	<0.05
CO2 (g/t-mi)	5	20	80	76	105	1.05
CH4 (mg/t-mi)	115	65	<1	<1	180	1.80
N2O (mg/t-mi)	<1	<1	<1	64	<1	NA
GHGs (g/t-mi)	10	20	80	73	110	1.05
VOC (mg/t-mi)	5	5	15	53	25	1.25
CO (mg/t-mi)	5	20	150	86	175	0.75
NOx (mg/t-mi)	15	45	410	87	475	0.95
PM10 (mg/t-mi)	<1	15	5	16	15	1.50
SOx (mg/t-mi)	10	20	<1	<1	30	2.00
CNG
Total energy (Btu/t-mi)	95	70	1385	90	1545	1.20
Petroleum (Btu/t-mi)	5	5	<1	<1	10	<0.05
CO2 (g/t-mi)	5	5	80	89	90	0.90
CH4 (mg/t-mi)	290	5	15	5	315	3.15
N2O (mg/t-mi)	<1	<1	<1	43	<1	NA
GHGs (g/t-mi)	15	5	85	82	100	0.95
VOC (mg/t-mi)	5	<1	165	96	170	8.50
CO (mg/t-mi)	10	<1	15	54	25	0.10
NOx (mg/t-mi)	25	5	345	92	375	0.75
PM10 (mg/t-mi)	<1	5	<1	6	5	0.50
SOx (mg/t-mi)	15	5	<1	2	20	1.35
LNG
Total energy (Btu/t-mi)	75	140	1375	86	1595	1.25
Petroleum (Btu/t-mi)	5	10	<1	<1	15	<0.05
CO2 (g/t-mi)	5	10	80	85	95	0.95
CH4 (mg/t-mi)	145	90	50	17	290	2.90
N2O (mg/t-mi)	<1	<1	<1	66	<1	NA
GHGs (g/t-mi)	10	10	85	80	105	1.00
VOC (mg/t-mi)	5	<1	165	95	170	8.50
CO (mg/t-mi)	10	5	15	50	25	0.10
NOx (mg/t-mi)	20	25	345	88	395	0.80
PM10 (mg/t-mi)	<1	<1	<1	14	<1	NA
SOx (mg/t-mi)	15	5	0	<1	15	1.00
Notes: Values rounded to the nearest 5. Total energy use and emissions may not equal the sum of fuel-cycle stages because of rounding. Tailpipe EFs are compiled from numerous sources and calculated assuming a truck carrying a 20-t load.1,2,12,53,54,56–59,76. NA = not applicable.

The analysis presented here is based on a HDV carrying 20 t of cargo, which is constant and is an assumption that is maintained for all fuels for comparison purposes. It is noted that, in practice, fuel consumption is a function of truck payload, just as fuel consumption is a function of driver behavior, speed, terrain, and other factors. Furthermore, the persistence of a legacy fleet or variabilities in engine design, technology, and performance can affect energy consumption and emissions. In this analysis, the relationship between these factors and fuel consumption is constrained to reduce uncertainty; the purpose of this analysis is not to capture all of these effects, but to compare emissions across a spectrum of fuels. All other inputs (e.g., carbon and sulfur content of fuels, energy and mass density, and fuel mixes) are retrieved directly from GREET, are well established in the literature, and present a negligible degree of uncertainty. Each fuel type is discussed in greater detail in the remainder of this section.

BD is a diesel alternative produced from soybean, rapeseed, mustard, sunflower, and/or palm oils. BD is a particularly promising alternative to conventional diesel because CI engines can use BD with little or no modifications. Pure BD (i.e., BD100) can be used as a fuel, but it is typically blended at different levels with petroleum diesel to create a BD blend.60 The blend rate can be restricted by vehicle manufacturers' warranty limitations. In this analysis, BD20 (20% BD and 80% ULSD) is examined using a soyoil feedstock produced from NY-farmed soybeans, the most commonly used BD blend. Vegetable oils are now also being refined using hydrogenation and present a fuel alternative that may one day compete with BD, but have not yet reached significant market penetration. BD can also be produced from animal fats, but neither vegetable oils nor BD from animal fats are considered in this analysis.

Ethanol-blended diesel, or ED, formerly referred to as “oxygenated diesel” or “oxydiesel,” is a biofuel composed of a blend of approximately 5–15% ethanol and approximately 1–3% oxygenate additive, with the remainder being petroleum diesel. Ethanol and diesel do not blend easily, so an additive is necessary to facilitate blending.4,61 This analysis uses the most common blend, known as ED10, which consists of 10% ethanol, 1% additive, and 89% ULSD. The ethanol is assumed to come from NY-based corn production using NY farming practices.

FTD is synthetic diesel fuel produced by converting gaseous hydrocarbons into liquid fuel. FTD can be substituted directly for ULSD to fuel diesel-powered HDVs without modification to the vehicle engine.43,62 Producing liquid transportation fuels from natural gas and coal using the Fischer–Tropsch process has been demonstrated on a large scale; biomass also could be used as a FTD feedstock, alone or in conjunction with fossil fuel sources.63 This analysis examines FTD produced from natural gas delivered to NY from North American sources.

Natural gas can be used as a transportation fuel—either CNG or LNG. As of 2007, there were more than 114,000 CNG vehicles and 2500 LNG vehicles in use on U.S. roadways, with many models from which to choose.9,64

Additional inputs in NY-GREET-HDV include assumptions about fuel production, transport, and delivery that are NY specific. Inputs and outputs in biofuel feedstock production (e.g., farm fertilizer use and energy use, crop yields, and electricity mix used in fuel processing) vary by region, together producing measurable differences in fuel-cycle emissions.47 Fuel that must be transported long distances creates emissions in the upstream parts of the fuel cycle that are important to consider. The vehicle modeled in this paper is representative of a NY HDV using assumptions regarding fuel transport and distribution to NY. Specifically, it is assumed that (1) NY receives its full supply of natural gas from North American sources65,66; (2) crop yields and farming energy, fertilizer, and pesticide use are representative of NY farms67; (3) upstream electricity use is characteristic of NY electric grid projections to 202068,69; and (4) transportation and distribution distances are representative of feedstock and fuel movements in NY.65–67,70,71

RESULTS

The model calculates energy consumption and emissions under assumptions discussed above for HDVs using ULSD, BD20, ED10, FT100, CNG, and LNG. Table 2 presents the HDV well-to-pump (WTP) results, reported as per million Btu (MBtu) of fuel available at fuel station pumps, as well as percent change relative to ULSD. The WTP results show particularly low emissions of CO₂ from BD20, which is not an error, rather the value is representative of net CO₂ taking into account the carbon sink created during the production of soybeans.

Table 2. HDV WTP energy consumption and emissions (reported in Btu/MBtu or g/MBtu of fuel available at fuel station pumps) with ratios of alternative fuel values to ULSD values shown in parentheses

Item	ULSD	BD20	ED10	FT100	CNG	LNG
WTP efficiency ^a	84.2%	65.7%	78.1%	60.4%	88.3%	84.7%
Energy consumption
Total energy (Btu/MBtu)	187,300	522,500 (2.79)	280,200 (1.5)	654,900 (3.5)	132,600 (0.71)	180,700 (0.96)
Petroleum (Btu/MBtu)	75,900	88,500 (1.17)	78,500 (1.03)	19,600 (0.26)	7,800 (0.10)	13,000 (0.17)
Emissions (g/MBtu)
CO2	14,500	2,000 (0.13)	12,300 (0.85)	26,100 (1.81)	8,400 (0.58)	12,000 (0.83)
CH4	104	96 (0.92)	105 (1.01)	188 (1.8)	244 (2.34)	199 (1.9)
N2O	0.2	2.7 (11.04)	2.9 (11.59)	0.2 (0.61)	0.2 (0.68)	0.3 (1.03)
GHGs ^b	17,100	5,200 (0.3)	15,800 (0.92)	30,800 (1.8)	14,500 (0.85)	17,000 (0.99)
VOC	7.6	26.5 (3.47)	11.2 (1.46)	12.5 (1.63)	6.4 (0.83)	6.6 (0.86)
CO	11.9	14.5 (1.22)	14.3 (1.2)	25.3 (2.13)	9.4 (0.79)	11.2 (0.94)
NOx	36.3	42.2 (1.16)	40.6 (1.12)	65.9 (1.81)	23.8 (0.65)	38.0 (1.04)
PM10	5.8	6.6 (1.12)	7.0 (1.19)	14.4 (2.46)	3.0 (0.51)	1.2 (0.2)
SOx	15.6	23.3 (1.49)	17.8 (1.14)	29.9 (1.91)	16.3 (1.04)	13.3 (0.85)
Notes: ^aWTP efficiency is the ratio of the energy output of the final fuel product divided by the sum of this energy output and the energy used to produce it; for example, results for ULSD (84.2%) imply that one could obtain 84.2 units of energy for every 100 units of energy involved in the production and use phases of ULSD. ^bTotal GHG emissions values are calculated in CO2 equivalents by applying a 100-yr GWP factor for each individual GHG type.

Table 3 reports well-to-wheel results, which include WTP and vehicle operation effects. Results are reported per ton-mile and are based on a HDV carrying 20 t of cargo. Table 3 also indicates the relative contributions of each stage of the fuel cycle to TFC energy use and emissions (refer to Figure 1 for components of each TFC stage). GHG emissions are reported in CO₂ equivalent emissions using a global warming potential (GWP) of 1 for CO₂, 25 for CH₄, and 298 for N₂O, as reported by the Intergovernmental Panel on Climate Change. For all fuel types, the tailpipe CO₂ emission rate is the main contributor to total GHG emissions. CO₂ emissions are primarily influenced by mpg and the fuel's energy and carbon content;these estimates correlate well to other TFCA studies. Figure 3 graphically presents these results.

Figure 3. TFC percent change vs. ULSD HDV: (a) total energy, (b) petroleum, (c) CO₂, and (d) GHGs.

The results show that a HDV using ULSD consumes the least TFC energy, followed by ED10, CNG, BD20, LNG, and FT100 from the least to the most energy-consuming pathway. The spread from ULSD to FT100 is a factor of 1.35. The upstream stages of the fuel cycle play an important role in these results; except for CNG, fuel processing contributes the second largest portion of total energy use. In the case of FT100, which consumes the greatest portion of total fuel-cycle energy consumption in the fuel processing stage, almost all of the energy is in the form of natural gas, rather than petroleum, and is due to the energy-intensive gas-to-liquids process. For BD20, which also has high energy consumption in the fuel processing stage, this energy consumption is due to the process of refining and blending soyoil with diesel. Relative to the alternatives, CNG and LNG demonstrate low energy consumption in the upstream stages because of the lack of energy-intensive petroleum refining in the fuel cycles. The TFCA results found in this analysis correlate well with those found in the literature pertaining to LDV analyses.72,73

DISCUSSION AND IMPLICATIONS

The TFCA results show that there are often conflicts between emission types of different fuels. For example, it is shown that on a TFC basis, ULSD HDV emits 100 g/mi CO₂ and 470 mg/mi NO_x, whereas a BD20 HDV emits less CO₂ (90 g/mi) and more NO_x (530 mg/mi). Policy-makers contemplating fuel-switching programs are often faced with similar situations in which tradeoffs need to be made between two or more fuel or emission types, and the impact of these tradeoffs should be considered in light of policy goals.39

To address the sensitivity of these results to uncertainty in input variables, two key variables are identified that would affect operational emissions on a per ton-mile basis; namely, truck payload (t/truck) and truck efficiency (mpg). The first (payload) is a function of logistics, cargo type, and operational characteristics of the HDVs—the greater the payload, the lower the emissions per ton-mile. The second variable (truck efficiency) is a function of vehicle technology, operating characteristics, driver behavior, route characteristics, and other factors that affect energy needed for vehicle propulsion—again, the greater the efficiency, the lower the emissions per ton-mile. The goal of this sensitivity analysis is to explore how different assumptions about these two factors affect the overall variability of TFC emissions and to compare this variability to differences observed with the alternative fuels from Table 3.

Sensitivity analysis values are shown in Table 4. Scenarios were modeled with an average value retrieved from the literature: a payload of 20 t/truck and efficiency (mpg) for each fuel as shown in Table 1. Sensitivity scenarios were conducted with low and high input values set at ±25% from the default value. The results are reported for TFC energy use, petroleum use, and GHG emissions (Table 4).

Table 4. Uncertainty results for two important operational phase variables showing the range of input variable values and the range of TFC energy use, petroleum use, and GHG emissions based on these values

Sensitivity Variable	Low and High Value	Fuel Type	Total Energy Use, Btu/t-mi (Low/High)	Total Petroleum Use, Btu/t-mi (Low/High)	Total GHG Emissions, g/t-mi (Low/High)
Truck payload (t cargo/truck)	Low = 15, high = 25	ULSD	1705/1020	1560/935	140/85
		BD20	2125/1275	1235/740	125/75
		ED10	1940/1165	1545/925	135/80
		FT100	2310/1385	25/15	145/90
		CNG	2060/1235	15/10	135/80
		LNG	2125/1275	20/15	140/85
Truck efficiency (mpg)	Low =−25%, high =+25% (from base value ^a )	ULSD	1700/1020	1560/935	140/85
		BD20	2120/1275	1230/740	125/75
		ED10	1940/1160	1545/925	140/85
		FT100	2310/1385	25/15	145/90
		CNG	2060/1235	15/10	135/80
		LNG	2125/1275	20/15	140/85
Notes: ^aSee Table 1 for base mpg values. All table values rounded to the nearest 5 units.

The results discussed earlier showed that alternative fuels (for the most part) provided TFC emissions benefits when compared with ULSD, with BD20 and CNG providing the greatest GHG emissions benefits. This sensitivity analysis demonstrates that these benefits may only apply when comparing identical trucks with identical payloads. If instead ULSD trucks happen to be more efficient than average, or if ULSD trucks happen to be hauling greater payloads than their alternative fuel counterparts, then these ULSD trucks may in fact have lower emissions on a per ton-mile basis. For example, results in Table 3 show that a BD20 HDV emits approximately 90 g/t-mi of GHGs, compared with 105 g/t-mi from a ULSD HDV. However, the sensitivity analysis demonstrates that an inefficient BD20 HDV with low mpg rating can emit more GHGs compared with an efficient ULSD HDV with a high mpg rating. (Of course, the best situation is a highly efficient BD20 HDV carrying an above-average payload).

These results have interesting policy implications that suggest that the freight sector must consider alternative fuels and efficiency improvements to reduce its petroleum usage and emissions. The implications are especially important if decision-makers are making either/or choices between improving the efficiency of conventional fuel vehicles versus deploying alternative fuels in less efficient vehicles.

More generally, the results of this sensitivity analysis imply how potential variability in input values can affect overall TFC emissions comparisons. Figure 3 shows that on a GHG emissions basis, all alternative fuels differ from ULSD by less than approximately 10%. Whether these values are statistically significant requires an analysis of input probability distributions that currently do not exist. Nevertheless, one may expect variability in the production and use phases of the TFC to create TFC distributions that overlap considerably, as shown in other work,39 implying that ULSD may perform as well as alternatives for some sets of parameter inputs beyond those evaluated here. Such statistical analysis is reserved for future work as probability distributions for key input variables are constructed over time.

The quality of certain data inputs may improve over time, and future analyses would benefit from such improvements. In particular, the collection, analysis, and presentation of EFs may improve over time. Moreover, technological improvement may yield cleaner fuels and lower EFs for traditional and alternative fuels. These improvements would lead to a more refined modeling process and potentially alter the results of future TFCA analyses.

Lastly, the authors want to explicitly call out that this analysis of biofuels does not include indirect land-use changes (ILUCs). As discussed in a growing literature, the inclusion of ILUC can have significant impacts on biofuel TFCA GHG estimates. Some researchers have attested that GHG emissions from ILUC may more than offset any GHG savings from biofuels, whereas other researchers argue that assumptions required in ILUC effects are diffuse, and assumptions are too subjective to be valuable in informing biofuel policy.74,75 The GREET/NY-GREET framework does not include ILUC emissions estimates, in part because of the high uncertainties surrounding ILUCs; it is noted that inclusion of such uncertainties in this HDV analysis may affect the GHG savings estimates for BD20 and ED10.

ACKNOWLEDGMENTS

This research was partially supported by a grant from the New York State Energy Research and Development Authority.