You are on page 1of 10

Journal of Cleaner Production 76 (2014) 64e73

Contents lists available at ScienceDirect

Journal of Cleaner Production

journal homepage:

Environmental impacts of roundwood supply chain options in

Michigan: life-cycle assessment of harvest and transport stages
Robert M. Handler a, *, David R. Shonnard b, Pasi Lautala c, Dalia Abbas d,1, Ajit Srivastava d
Sustainable Futures Institute, Michigan Technological University, 1400 Townsend Dr., Houghton, MI 49931, USA
Department of Chemical Engineering, Michigan Technological University, 1400 Townsend Dr., Houghton, MI 49931, USA
Department of Civil and Environmental Engineering, Michigan Technological University, 1400 Townsend Dr., Houghton, MI 49931, USA
Department of Biosystems & Agricultural Engineering, Michigan State University, 524 S. Shaw Lane, East Lansing, MI 48824, USA

a r t i c l e i n f o a b s t r a c t

Article history: Here, we analyze greenhouse gas emissions and fossil energy demand for roundwood supply chain
Received 6 November 2013 activities (harvesting and transport) within the state of Michigan. A life-cycle assessment was completed,
Received in revised form relying on a combination of peer-reviewed literature, national databases, and primary data collected
7 April 2014
from Michigan loggers and truckers. Several equipment configurations and operating scenarios for
Accepted 18 April 2014
Available online 6 May 2014
roundwood harvesting have been considered. Results indicated that a full processor/forwarder is the best
combination of harvesting equipment, with greenhouse gas emissions of 9.9e14.7 kg CO2eq/green tonne,
due to relatively low inputs and high reported productivity, although environmental impacts of har-
Life-cycle assessment
vesting depend strongly on the intensity of harvest being conducted. Bimodal truck þ rail transport has
Wood supply chain environmental burdens roughly one third to half that of typical log truck transport at longer transport
Michigan distances, directly related to the increased fuel efficiency of rail transport. Aggregated results for
Forest products industry roundwood supply within Michigan are comparable to similar studies in other regions, although the
Wood harvesting mechanization of the harvesting industry and large size of Michigan log trucks are factors in the smaller
Wood transport environmental burden. A sensitivity analysis indicated that a variety of factors related to truck transport
(distance, fuel economy, load factor, truck capacity) are the most influential for overall environmental
impacts of the forest biomass supply chain. Environmental impacts associated with roundwood supply
are quite low in comparison to the carbon content and embodied energy of delivered wood, implying
that roundwood supply activities do not preclude beneficial use of this feedstock in biofuels or bioenergy
Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction (Intergovernmental Panel on Climate Change, 2007a,b). Biofuels

made from renewable feedstocks are among the largest expected
Emissions of greenhouse gases from transportation are a major contributors to the transportation industry’s planned emission re-
contributor to human-caused climate impacts on a global scale. ductions over the foreseeable future (Hoekman, 2009). Addition-
Recent studies have predicted serious consequences from a “busi- ally, in recent years the U.S. has roughly half of its oil needs from
ness as usual” approach to energy production and use, including foreign sources (U.S. Energy Information Administration, 2012).
increasing global temperatures, sea level rise, displacements of Such a high dependence increases U.S. strategic vulnerability, and a
human populations from submerged lands, changing weather domestic biofuels industry is increasingly seen as a way to combat
patterns, and increase in incidence of certain diseases this trend while increasing employment in rural areas of the
country (Perez-Verdin et al., 2008).
Biomass from forest resources will be an important feedstock
* Corresponding author. Tel.: þ1 906 487 1092; fax: þ1 906 487 2943. source in forested regions of the U.S., and are a significant
E-mail addresses:, (R.M. Handler), contributor to the available biomass inventory nationwide (D.R. Shonnard), (P. Lautala),
(Department of Energy, 2011). According to statewide U.S. Forest
(D. Abbas), (A. Srivastava).
Present address: Department of Agricultural and Environmental Science, Ten- Service data, the state of Michigan has the seventh largest
nessee State University, 3500 John A Merritt Blvd, Nashville, TN 37209, USA. timberland area in the United States, and current annual growth
0959-6526/Ó 2014 Elsevier Ltd. All rights reserved.
R.M. Handler et al. / Journal of Cleaner Production 76 (2014) 64e73 65

exceeds removals and mortality in most forest types (Van Deusen stages in the forest-based biomass supply chain that occur prior to
and Roesch, 2008). Decline of traditional roundwood-utilizing in- biomass conversion into biofuels, bioproducts, or bioenergy at a
dustries in the region over recent years (pulp and lumber) have processing facility (Fig. 1). Material and energy inputs used directly
opened up a portion of forest biomass to alternate uses (Leefers during feedstock supply chain activities, i.e. wood harvesting and
et al., 2010), while alternative markets for this timber would sta- transport, will be considered. Of these inputs, fuel is the most
bilize prices and maintain forest products industry infrastructure in important, but other inputs are also included, including major
the region [e.g., (Dwivedi and Alavalapati, 2009). Additionally, equipment used to harvest and transport wood (harvesters, for-
several thousand acres of abandoned or unproductive agricultural warders, log trucks, etc.) and estimates of lubricants and inputs
lands in Michigan could be converted to short-rotation forest (SRF) used to maintain equipment.
crops of fast-growing willow or poplar, suitable to the growing We also do not include inputs from any activities that would
conditions in Michigan [e.g., (Froese and Abbott, 2012). occur ‘upstream’ of the wood harvesting process, such as forest re-
To be a sustainable and environmentally beneficial industry, planting or carbon stock changes on the landscape resulting from
alternative fuels should be produced in a responsible manner that direct or indirect land-use change. The vast majority of wood
achieves measurable and significant gains in environmental per- currently harvested in Michigan is from forest stands which were
formance over ‘business as usual’ use of fossil fuels. For biofuels, initiated by natural processes, with no external inputs. These stands
this typically translates to reductions in greenhouse gas emissions have been cultured by foresters under the guidance of forest
and fossil energy demand over the cradle-to-grave life cycle of the management plans that dictate the intensity of harvest operations
alternative fuel, compared to conventional gasoline or diesel. Many and intended outcomes for future forest regeneration and growth.
industry sectors are addressing sustainability issues by reducing Michigan’s forested land base has been increasing in size and vol-
the emission of greenhouse gases across the entire production ume of wood fiber for many decades, and we do not expect new
chain. Previous life cycle assessments (LCAs) of biofuels produced uses of forest biomass to contribute to losses of forestland within
from a variety of feedstocks have highlighted the importance of the state. A growing body of research on impacts of different har-
feedstock production and supply in the overall life cycle of a given vesting regimes on soil carbon storage reveals a range of potential
fuel product (Fan et al., 2011; Handler et al., 2012; Shonnard et al., outcomes that are likely dependent on specific site characteristics,
2010). Commercial biofuel operations will rely on inputs of feed- although the average result seems to be little or no change in soil
stock grown over a large area, with potentially variable supply over carbon stocks over time [e.g., (Johnson and Curtis, 2001)]. Potential
the course of a year. Assessing supply chain options for this type of environmental impacts of increased harvest activity on landscape
emerging industry will be critical for continued success. carbon stocks, through shorter rotation ages on actively managed
The supply chain environmental assessment presented here lands or an increased willingness of landowners to actively manage
focuses on forest biomass grown within the state of Michigan. Our their forestlands for timber production, is an important issue that
objective was to develop environmental metrics for greenhouse gas should be addressed in further research.
emissions and fossil energy demand for forest-based biomass
harvesting and transport within Michigan. Improved estimates of 2.2. Functional unit
woody biomass sustainability require a more detailed description
of the full supply chain for these materials, to fully assess costs and The functional unit for this study was one green (50% moisture
environmental impacts (Department of Energy, 2011), which were content) (Argonne National Laboratory, 2013) metric tonne of forest
achieved in this study through collection of data from a large biomass, defined here as roundwood pulpwood that can be har-
sample of workers in Michigan’s forest products industry. vested and collected with forestry equipment commonly used in
Furthermore, the intent is that the study results could be utilized Michigan. Forest products industry workers in Michigan were most
both to understand the impacts associated with a particular forest comfortable with English units as our data collection efforts were
biomass supply scenario and to be generally applicable to the cur- devised and tested on sample respondents, but they have been
rent forest product industry as it exists in aggregate. To this end, we converted here. Harvesting activities are normalized to this func-
have developed a limited-scope life-cycle assessment (LCA) pro- tional unit, while transportation activities have been normalized on
cedure for several forest biomass harvesting and transportation the basis of a tonne-mile, due to the dependence of transport
scenarios, using a well-detailed process assumptions and inventory burdens on the particular distance moved. No specific origin-
data. The intended audience of results from this study is other LCA destination pairs of feedstock location and processing facilities
professionals, forest industry companies, environmental managers, were utilized in this study, so the transportation data can be uti-
government regulators, and biofuel developers. Results and lized by parties interested in specific case studies by multiplying
methods from this study may be later used at different levels of environmental burdens per tonne-mile by the mileage of the spe-
data aggregation when considering specific biofuels or bioenergy cific transport step, as is done in a few examples presented below.
projects within the state of Michigan, and may be applicable to For the purposes of comparison to other studies, results are also
forest-based biomass supply within the broader Great Lakes region. presented in English units (lb CO2eq per U.S. short ton) and on a dry
weight basis.
2. Research methods
2.3. Life cycle input data
2.1. Goal and scope
The data and assumptions required for this study came from a
The goal of our LCA is to determine greenhouse gas (GHG) variety of sources. An important component of our life cycle in-
emissions and fossil energy demand associated with harvesting ventory was the use of primary input data from loggers within the
and transport of forest-based biomass within the state of Michigan. state of Michigan. A comprehensive logging and transportation
For the purposes of this study, harvesting includes cutting trees survey was mailed to loggers within the State of Michigan from
from the stump, processing into typical pulpwood length (100 2009 to 2010. Information was collected on current equipment and
inches), and moving the logs to a forest landing. Transport refers to operations using in harvesting and transportation supply stages.
movements of wood from the forest landing to a processing facility. The combined results of over 220 unique survey respondents
The scope is limited in the sense that our focus is only on those represent the most current and accurate picture of forest products
66 R.M. Handler et al. / Journal of Cleaner Production 76 (2014) 64e73

Fig. 1. Diagram of system boundary for limited-scope life-cycle assessment of the forest biomass supply chain in Michigan.

harvesting operations over the entire state of Michigan. A detailed essential to make several key assumptions regarding the treatment
explanation of survey results and analysis is explained in Abbas of this data. It was assumed that loggers had an average productive
et al. (2014). In the following sections, we detail how life cycle work day of 8 h, and that average harvest rates were maintained
input data was developed for harvesting and transport of forest continuously throughout the workday. The lifetime productivity of
biomass. a major piece of harvesting equipment (harvester, forwarder,
Estimates of harvesting (cutting wood) and forwarding (moving skidder, etc.) was assumed to be 144,000 tonnes (10 years,
harvested wood to a truck-accessible collection point) activity were 40 weeks/year, 8 loads/day, 45 tonnes/load), based on guidance
taken primarily from the state of Michigan logger survey, with from industrial partners and selected loggers in the region. This
supplementary information from other sources (Table 1). Three estimate can be easily changed to suit local conditions, but overall
main harvesting/forwarding equipment configurations were used LCA burdens for equipment fabrication and repair are likely to be
to characterize the logging industry in Michigan: small, as shown in results.

a)cut-to-length full processor/forwarder

b)feller-buncher/skidder/slasher 2.3.2. Harvest productivity
c)chainsaws/skidder In order to transform life cycle inventory data for harvesting and
forwarding, reported on the basis of hourly usage rates, into inputs
2.3.1. Harvest configurations and assumptions normalized on the basis of one green tonne of forest biomass, it is
From the logger survey results, we were able to obtain the necessary to know the productivity of each equipment configura-
average fuel use of various key pieces of forestry equipment. This tion in tonnes of wood per hour. In the state of Michigan logger
data was reported in gallons/hour, which was the most common surveys, respondents were asked to list their average harvest pro-
and reliable metric of fuel use for loggers in the sample population. ductivity in short tons or cords of green timber per hr e most re-
As seen in Table 1, the amount of variability in fuel use estimates is spondents preferred cords. Information was collected for three
large. Sensitivity analyses in the following sections will illustrate theoretical harvest types e clearcutting all merchantable timber, a
the importance of obtaining good estimates of fuel use for the ac- 70% (shelterwood) removal treatment, and a 30% (selective cut)
curacy of overall forest biomass supply chain burdens. Estimates of removal treatment. In each case, respondents were also asked to list
lubricants and grease for each equipment configuration came from which of the harvest equipment configurations they would likely
operating data collected from experienced loggers by foresters use in each treatment. For each of these treatment scenarios, there
working for J.M. Longyear (JML), LLC, a forest management com- were also separate entry sections for entering productivity esti-
pany operating in the Lakes States region with extensive experience mates for each of four potential forest types e natural hardwood
contracting with logging crews in Michigan. In an effort to convert stands, natural softwood stands, mixed hardwood/softwood
data to a consistent format and make valid comparisons, it was stands, and softwood plantations (Abbas et al., 2014).

Table 1
Summary of inputs for harvesting configurations.

Item Unit Total Data source/comments

Configuration A: full processor/forwarder

Fuel use L/hr 30.7  11.3 Logger survey data, full processor ¼ 18.5  8.7 (na ¼ 142), forwarder ¼ 12.1  7.2 (n ¼ 159)
Lubricants L/d 25.3 JMLb, no variability given
Grease kg/d 0.9 JML, no variability given
Equipment 2 Major pieces of equipment
Configuration B: feller-buncher/grapple skidder/slasher
Fuel use L/hr 55.4  14.1 Logger survey data, feller buncher ¼ 6.3  2.6 (n ¼ 37), slasher ¼ 3.9  1.8 (n ¼ 18),
weighted average of grapple (5.1  2.3, n ¼ 33) and cable skidders (2.4  1.0, n ¼ 11)
Lubricants L/d 9.4 JML, no variability given
Grease kg/d 0.45 JML, no variability given
Saw gas L/d 3.79 JML, no variability given
Equipment 3 Major pieces of equipment
Configuration C: chainsaw/cable skidder
Fuel use L/hr 27.2  7.8 Logger survey data, chainsaws ¼ 1.1  0.6 (n ¼ 35), 2.5 average chainsaws used per logging
crew, weighted average of grapple (5.1  2.3, n ¼ 33) and cable skidders (2.4  1.0, n ¼ 11)
Lubricants L/d 1.5 JML, no variability given
Equipment 1 Major piece of equipment
n-values listed in Comments refer to the number of survey responses included in the reported average.
JML data was obtained through personal communication with forest industry experts and is used as representative of logging crews in the state.
R.M. Handler et al. / Journal of Cleaner Production 76 (2014) 64e73 67

Estimates of harvest productivity were wide-ranging, and this Table 2

analysis required some standardization to ensure that accurate Combined state of MI productivity estimates for different logging equipment
comparisons were being made. In previous survey sections, re-
spondents were asked which equipment configurations (processor/ Treatment Forest type Productivity per harvester
forwarder, feller-buncher/skidder/slasher, etc.) were used in their (tonnes/hr)

operations, along with the number of pieces of equipment owned. Na Average Std. dev
For the full processor and feller-buncher configurations, only re- A: full processor/forwarder
spondents indicating that one or two pieces of harvesting equipment 30% Cut Natural hardwoods 54 7.12 2.95
(one or two processors, one or two feller-bunchers) were included in (Selective) Mixed hardwood/softwood 48 8.17 3.15
Natural softwoods 47 8.41 4.59
the productivity analysis. In situations where respondents indicated
Softwood plantations 37 9.74 4.49
three or more processors or feller bunchers, it was more likely that 70% Cut Natural hardwoods 43 8.72 3.83
these pieces of equipment were working on different sites, or not all (Shelterwood) Mixed hardwood/softwood 41 9.62 3.86
working at the same time, and therefore would not yield produc- Natural softwoods 38 9.93 4.57
tivity data that was reflective of the capability of each machine. No Softwood plantations 29 10.58 4.54
Clearcutting Natural hardwoods 43 11.75 5.84
such distinction was made for the equipment configurations
Mixed hardwood/softwood 47 12.09 5.32
involving chainsaws as the main harvesting equipment. However, an Natural softwoods 40 12.95 5.94
average ownership of 2.6 chainsaws was indicated in the survey Softwood plantations 35 14.86 8.57
responses for loggers who used chainsaws as a tool to cut more than Treatment Forest type Productivity per harvester
50% of their total production in 2009e2010 (Abbas et al., 2014). (tonnes/hr)
Weighted averages for each category were calculated as follows: N Average Std. dev

B: feller-buncher/skidder/slasher
Average Productivity ðcords=hrÞ ¼ ðN1*P1 þ N2*P2Þ= 30% Cut Natural hardwoods 15 7.93 3.24
ðN1 þ N2*2Þ (Selective) Mixed hardwood/softwood 15 7.79 2.79
Natural softwoods 13 7.19 2.80
Softwood plantations 8 8.55 1.98
where N1 and N2 are the number of 1-harvester and 2-harvester 70%Cut Natural hardwoods 14 10.11 3.05
respondents, respectively, and P1 and P2 represent average pro- (Shelterwood) Mixed hardwood/softwood 15 9.86 3.02
Natural softwoods 16 10.71 3.41
ductivity values for 1-harvester and 2-harvester respondents (in Softwood plantations 9 11.49 3.68
cords/hr). Clearcutting Natural hardwoods 13 14.53 5.71
Below in Table 2 is a summary of productivity estimates (average Mixed hardwood/softwood 13 14.04 6.36
tonnes green timber/hr) for survey respondents that indicated a Natural softwoods 11 13.69 6.04
Softwood plantations 9 15.13 8.92
particular equipment configuration would be used in each cutting
prescription and forest type. To convert these values from cords Treatment Forest type Productivity (tonnes/hr)
into green tonnes/hour, an average conversion factor of 2.35 U.S. N Average Std. dev
short tons per cord has been applied (Blinn and Hendricks, 1997).
C: chainsaws/skidder
This value can vary between regions and tree species, and more 30% Cut Natural hardwoods 32 4.31 2.83
specific data may be substituted if values are known for target (Selective) Mixed hardwood/softwood 19 4.15 3.11
species in a certain area. As expected, average productivity for Natural softwoods 17 3.93 3.32
Softwood plantations 13 3.75 1.84
chainsaws is lower than the more mechanized systems, roughly 4
70% Cut Natural hardwoods 20 4.68 3.44
tonnes/hour across most harvest types and forest types (Table 2). In (Shelterwood) Mixed hardwood/softwood 18 4.15 2.99
both fully-mechanized systems (A and B), productivity increased as Natural softwoods 14 4.01 3.16
harvest treatment intensity rose from 30% to 70% to 100%, with Softwood plantations 12 3.70 2.25
feller-bunchers slightly more productive than full processors in 70% Clearcutting Natural hardwoods 12 4.26 2.39
Mixed hardwood/softwood 14 4.07 1.97
and clearcutting operations.
Natural softwoods 13 3.03 1.27
Softwood plantations 9 3.79 2.34
2.3.3. Data aggregation a
n-values listed refer to the number of survey responses included in the reported
In order to simplify the analysis for this study, the following average.
data aggregation steps have been made. Productivities for all
natural stands were averaged for each harvest configuration
(Table 2: A, B and C) in each harvest type (Table 2: Treatments), and used hereafter, but these estimations can be altered based on
resulting in 9 total productivity estimates. Plantations were left the planned operations for a specific facility and its supply area.
out of the analysis for now because they are still relatively un- Sensitivity analyses that follow indicate the potential changes that
common in the state of Michigan, but this may change in the would result if the proportions of each harvest scenario or
future. In order to estimate productivity for each harvest scenario, equipment configuration are changed.
the data for different equipment configurations was combined
with input from an external review committee (with members 2.3.4. Road and rail transportation
from academia, local government, and private industry) to yield a For forest biomass transportation from a forest landing to a
weighted average for each harvest scenario in the following conversion facility, the two modes of transportation considered
manner (Table 3). In this way, we now have one estimate of pro- here are road and rail. Over-the-road transport can occur in log
ductivity for each of the three harvest scenarios. If one single trucks (roundwood logs) or chip vans (processed biomass). In
metric to encompass all potential harvest activity is desired, the Michigan, log trucks are allowed to attain a gross vehicle weight of
data could be further aggregated by taking a weighted average of 74.4 tonnes (164,000 lb), which is considerably larger than other
the three harvest scenarios to represent their relative importance northern states such as MN or WI (36.3 tonnes, 80,000 lb) (Stewart
in terms of the overall harvest of forest biomass within the state of et al., 2010). Large log trucks are the primary method of roundwood
Michigan. One example of this aggregation is presented in Table 3 transport in the state. We developed LCA profiles of transport based
68 R.M. Handler et al. / Journal of Cleaner Production 76 (2014) 64e73

Table 3
Proportion of harvesting done in each scenario by each equipment configuration (%).

on an average log truck reported within the state of Michigan, but derived from national or regional databases and peer-reviewed
include estimates of fuel use for larger MI-only trucks (10 or 11 literature sources. Emissions of different greenhouse gases (CO2,
axles) and chip vans if LCA burdens for these modes of transport are N2O, CH4, etc.) are normalized on the basis of global warming po-
desired in future work. Fuel usage data was collected from truck tential (CO2 e equivalents, CO2eq) using either the IPCC GWP 100-
operators on the basis of an average fuel economy for a tank of year average (for Ecoinvent input data) or other life-cycle sources
gasoline or a given operation time, which included both loaded and (see Table 5) and aggregated to estimate the overall impact of a
unloaded segments. We assume here that the proportion of product or process.
unloaded miles traveled is 50%, according to guidance from in-
dustrial partners, and use an average fuel economy for all segments 3. Results and discussion
of transportation. Rail transport of forest biomass is typically per-
formed by 72.6 tonne log cars for roundwood logs. Rail is Combining the life cycle inputs (Tables 1, 2 and 4) with the
commonly perceived as being more fuel efficient than truck environmental impacts listed above (Table 5) and normalizing the
transport by a factor of 4 or more (Federal Railroad Administration, data to the basis of one green tonne, we arrive at greenhouse gas
2009). Our estimates of fuel use for rail cars operating in MI come emissions and fossil energy demand per green tonne of forest
from national averages of a major rail company operating in the biomass for harvesting and transportation stages within the state of
Upper Peninsula, and are in agreement with general estimates of Michigan (Table 6). Due to the different units commonly employed
rail fuel use (Table 4). We also consider the fuel use required for the in the areas of life-cycle assessment and forest products (English vs.
loading activity required to transfer roundwood from the forest metric units, green freshly cut wood vs. dry biomass), the main
landing to a truck. Most log trucks in MI are equipped with hy- results are presented using a variety of unit configurations. For
draulic loaders to accomplish this transfer, and one loading and conversions between green recently harvested timber and dry
unloading cycle of a self-loading log truck was incorporated into biomass, a moisture content of 50% was assumed in all cases. Re-
the estimates of supply chain fuel use (Table 4). sults appear to scale up or down by a factor of 2 due to unit
2.4. Environmental impact assessment Harvesting activity is the most complicated part of the analysis
(Table 6) due to the many possible levels of data aggregation.
Environmental impact factors of the production of material and Chainsaw harvesting does not rate as the option with the lowest
energy inputs, in addition to their direct use in this supply chain, environmental footprint despite the low relative machinery and
were included as part of this assessment through use of the input requirements, due to the low efficiency of production
Ecoinvent 2.1 database (Frischknecht et al., 2004), peer-reviewed compared to other harvesting scenarios. Within the full processor
literature, expert opinion, or other sources. We combined the life and feller-buncher harvesting scenarios, overall environmental
cycle input data detailed above with estimates of greenhouse gas impacts fall drastically as harvest intensity is increased from 30% to
emissions and fossil energy demand resulting from production and clearcutting, due to the increase in productivity in cords per hour. In
use of each of the inputs listed. Environmental impact factors and our analysis, we combine one average fuel use value (L/hr) for the
their sources are detailed in Table 5. A majority of the factors are equipment configuration with different productivity estimates

Table 4
Key input data and assumptions regarding transport of forest biomass in Michigan.

Item Data Comment

Fuel use required per tonne of green timber 17.0 L/hr Average of one full-day trial conducted with 2007 MI log truck
1 h to load or unload equipped with self-loader
36.3 green tonne average load
0.77 L/tonne
Truck transportation
Log truck fuel economy per tonne-km 1.90  0.76 km/L Logger survey, average of loaded and unloaded transport
36.3 green tonne average load Fuel economy for average of all forest biomass hauling trucks
50% load factor reported in survey (10-11 axle trucks ¼ 1.57  0.31 km/L,
0.0290 L/tonne-km chip vans ¼ 1.78  0.42 km/L
Lifetime tonne-km of log truck 15 yr productive life Logger survey data, estimates from industry experts Same load
88,500 km/yr factor and load weight used as above
Rail transportation
Rail fuel economy per tonne-km 0.0066 L/tonne-km CN Railroad (2009), comparable to Ecoinvent estimates
(Frischknecht et al., 2004)
Lifetime tonne-km of rail equipment 32,180,000 lifetime km Assumed values
1800 tonnes per load
R.M. Handler et al. / Journal of Cleaner Production 76 (2014) 64e73 69

Table 5
Environmental impact factors and major assumptions.

Item Data Comment

GHG emissions factor for fuels 3.28 kg CO2eq/L diesel, 2.97 kg (Skone and Gerdes, 2008), combining data on emissions
CO2eq/L gasoline per MJ of fuel, energy content of fuels, density of fuels
Energy demand of fuels 40.6 MJ/L, used for diesel and (Klvac et al., 2003), roughly 10% due to production of fuels
for gasoline
GHG Emissions factor for oils, lubricants 4.22 kg CO2eq/kg material 1.05 kg GHG emissions from production (Ecoinvent)
(Frischknecht et al., 2004) plus 3.17 kg GHG emissions for
mineralization to CO2
Energy demand of oils, lubricants 57.9 MJ/L (Klvac et al., 2003) values for synthetic oil
Emissions factor of grease 0 Assumed to be fairly recalcitrant, not combusted or biodegraded
Energy demand of Grease 76.7 MJ/kg Ecoinvent factor for lubricating oil production
(Frischknecht et al., 2004)
Emissions factors of harvesting/forwarding 0.43 kg CO2eq/green tonne, Calculations based on Swedish forwarder (Athanassiadis et al., 2002),
machine fabrication and repair for each large machine 41,873 kg CO2eq per original machine, plus 50% extra for lifetime of
involved repairs and maintenance, then normalized to lifetime estimated
wood production
Energy demand for machine fabrication 8.4 MJ/green tonne, for full Calculations based on Swedish forwarder (Athanassiadis et al., 2002),
and repair processor/feller/bunchers 66 MJ/kg for original machine, assumed 15,000 kg for harvesters and
6.7 MJ/green tonne for 12,000 kg for forwarders/skidders, plus 50% extra for lifetime of
forwarders/skidders repairs and maintenance, then normalized to estimated lifetime
Emissions for log truck production, 55,400 kg CO2eq Ecoinvent for 40-t lorry production, maintenance
maintenance (Frischknecht et al., 2004)
Energy demand for log truck production, 1,308,350 MJ Ecoinvent for 40-t lorry production, maintenance
maintenance (Frischknecht et al., 2004)
Emissions for rail equipment production, 2,537,000 kg CO2eq Ecoinvent for long-distance train production, maintenance
maintenance (Frischknecht et al., 2004)
Energy demand for rail equipment 54,368,890 MJ Ecoinvent for long-distance train production, maintenance
production, maintenance (Frischknecht et al., 2004)

(tonnes/hr) for different harvest types to arrive at different fuel chain (Sonne, 2006), and this stage of the supply chain deserves as
intensity factors (L/tonne of wood). While it may have been more much scrutiny in regards to potential optimization as other life
accurate to ask loggers to report an average fuel use for each cycle stages. The small environmental metrics displayed here
equipment type as it was being used in a variety of harvest types, (Table 7) are normalized on the basis of a tonne-km as opposed to a
initial testing of the survey instrument deemed this line of ques- tonne, so multiplication of these values by an actual transport
tioning too difficult for the majority of respondents to answer distance will yield an environmental burden with the same func-
clearly. Environmental impacts from fabrication and maintenance tional unit as the harvesting life cycle stage. For instance, if biomass
of equipment represents between 2 and 10% of overall greenhouse is to be transported 100 km by truck, greenhouse gas emissions
gas emissions and 3e15% of fossil energy demand, a small but non- for one-way transport become 0.097 kg CO2eq/tonne-
trivial component of the environmental burdens for this life cycle km  100 km ¼ 9.7 kg CO2eq/tonne feedstock, which is comparable
stage. to estimates of the harvesting stage. If no backhauls are possible
Transportation of biofuels or bioenergy feedstocks is potentially from the end-use facility, which is often the case in roundwood
the largest source of environmental impacts in the entire supply truck transport, impacts of the truck return trip must also be

Table 6
Environmental impacts of harvesting/forwarding at different levels of data aggregation.

Greenhouse gas emissions Fossil energy demand

lb CO2 eq kg CO2 eq kg CO2 eq MJ MJ MJ

green tona green tonneb dry tonne green ton green tonne dry tonne

A: full processor/forwarder
30% Cut (selective) 29.4 14.7 29.4 178.9 197.2 394.4
70% Cut (shelterwood) 24.6 12.3 24.6 150.5 165.9 331.7
Clearcutting 19.8 9.9 19.8 122.7 135.2 270.4
B: feller-buncher/skidder/slasher
30% Cut (selective) 52.6 26.3 52.6 305.7 337.0 674.0
70% Cut (shelterwood) 38.3 19.1 38.3 225.1 248.1 496.3
Clearcutting 27.2 13.6 27.2 162.4 179.0 358.0
C: chainsaws/skidder
30% Cut (selective) 48.6 24.3 48.6 276.0 304.2 608.5
70% Cut (shelterwood) 46.6 23.3 46.6 264.8 291.9 583.7
Clearcutting 44.0 22.0 44.0 250.0 275.5 551.1
All 30% selective cut harvesting 41.8 20.9 41.8 245.7 270.8 541.6
All 70% shelterwood cut harvesting 32.5 16.3 32.5 193.5 213.3 426.6
All clearcut harvesting 20.6 10.3 20.6 126.6 139.6 279.2
All harvesting activityc 35.7 17.8 35.7 211.4 233.1 466.1
‘ton’ refers to U.S. short ton.
‘tonne’ refers to metric tonne.
Baseline result that will be used in the rest of the study.
70 R.M. Handler et al. / Journal of Cleaner Production 76 (2014) 64e73

Table 7
Environmental impacts of forest biomass transport.

Item Greenhouse gas emissions Fossil energy demand

lb CO2eq kg CO2eq kg CO2eq dry MJ MJ MJ dry

ton e mile tonne e km tonne e km ton e mile tonne e km tonne e km

Log truck operations and equipment 0.313 0.097 0.194 1.79 1.23 2.46
Percentage due to equipment 2.3% 4.4%
Rail operations and equipment 0.069 0.022 0.043 0.39 0.27 0.53
Percentage due to equipment 0.2% 0.3%

allocated to the feedstock, effectively doubling the impact of the had a large impact on overall results, while emissions factors for
transport stage. A range of possible environmental burdens for large machinery fabrication and maintenance were small, as
sample truck trips is presented below in Table 8, using a range of would be expected. Changing the productivity of harvest systems
assumed transport distances, indicating that transportation could A, B, or C (full processor, feller-buncher, or chainsaws) had a larger
potentially be the most significant stage of the biomass supply impact than increasing or decreasing the proportion of each har-
chain if backhaul opportunities are limited and transport distance vest system in our data aggregation strategy outlined in Table 3.
is increased. Bimodal transportation that combines a short truck For variables involved in aggregation of harvest data, an increase
movement with a longer rail transport segment has the ability to of 10% in one harvest system was paired with an equivalent
move forest biomass a greater distance while maintaining an decrease split equally between the other two harvest systems (e.g.,
environmental burden similar to truck movements from a smaller increasing the use of feller-bunchers from 50 to 55% and reducing
supply radius, as the example in Table 8 shows. As expected, systems A and C by 2.5% each). The variable that induced the most
environmental burdens from the equipment fabrication and change in GHG emissions was the GHG emissions intensity of fuels
maintenance considered in both truck and rail cases represents a being used, however this variable is a nation-wide estimate and
small component of the environmental impacts considered here. local operators have little control over changes in this variable. Key
Other transportation infrastructure could be considered, such as harvesting variables which are under some control of forest op-
roads or rail lines, but normalization of this specific use among the erators include system productivity for all harvest systems A, B
lifetime of potential use experienced by that transportation infra- and C, and use of harvest system A (increasing values for these
structure would inevitably make the impacts small enough to be variables results in decreased overall GHG emissions). Decreasing
disregarded in this type of analysis. the use of 30% selective cuts would also decrease GHG emissions,
Not explicitly shown in Tables 7 or 8 below is the environmental but selective cuts serve other important ecosystem functions as
impact of the loading/unloading steps in the forest feedstock sup- part of a sustainable harvesting plan, and their use is not likely to
ply chain, which amount to 3.1 kg CO2eq/green tonne and 17.3 MJ/ be reduced within Michigan. Because so many harvest system
green tonne for greenhouse gas emissions and fossil energy de- variables are aggregated to derive a single value for harvesting
mand, respectively. These figures are based on fuel use estimates GHG emissions, the individual impact of harvest system variables
for self-loaders commonly supplied with Michigan log trucks on the overall supply chain emissions was reduced. In the trans-
(Table 4), but could be altered to reflect specific operating condi- portation system, however, most input data was used without
tions. Bimodal transportation with rail will require at least one aggregation or weighting to directly calculate environmental im-
extra loading cycle to complete the truck-rail interchange. In order pacts in the baseline assumption of round-trip truck use. An in-
to minimize the need for additional loading/unloading re- crease of 10% in truck capacity, fuel economy, or load factor
quirements in bimodal transportation, the final destination should resulted in decreases of roughly 5% in total cradle-to-gate emis-
have rail access and logs should ideally be loaded directly from log sions. Notably, an increase in use of bimodal rail transportation
truck to waiting rail cars at the rail siding. (90% truck/10% bimodal) compared to our base case (100% truck)
In order to explore the relative importance of inputs to our LCA using transport distances outlined in Table 9 would reduce GHG
model of the forest biomass supply chain, a sensitivity analysis emissions by 2.25%, even though the average overall transport
was performed on several variables in the harvesting and trans- distance would increase by 7%, increasing the potential supply
portation unit operations. Input variables were increased or radius while reducing environmental impacts.
decreased 10%, and overall supply chain GHG emissions were Environmental impacts for supplying forest-based biomass
recalculated and compared to the baseline (Fig. 2). Emissions within the state of Michigan calculated in this work can be
factors for petroleum fuels used in harvest or transport activities compared to similar estimates made in the literature, although the
comparisons depend upon different assumptions and scenario
boundaries among studies. Table 9 provides a comparison of our
Table 8 work with other published results. Our results for the state of
Potential environmental burdens associated with different trucking distances. Michigan are in reasonable agreement with studies from Europe
One-way trip distancea 50 km 100 kmb 150 km 200 km 20 km Truck þ and different regions in the United States that consider harvest and
(baseline) 150 km rail transport of forest-based biomass grown in natural stands. It was
(bimodal) unclear from Sonne (2006) what transport distance was used to
GHG emissions (kg CO2 12.8 22.5 32.3 42.3 16.5 calculate emissions for this stage of the supply chain, but their GHG
eq/green tonne)c emissions value is roughly twice as large as the baseline assump-
Fossil energy demand 140.1 263.0 385.8 508.7 163.7
tion here considering a 200 km roundtrip distance. Johnson et al.
(MJ/green tonne)
(2005) consider a transport distance in line with our own
Environmental burdens calculated on the basis of round-trip impacts, assuming assumption, but harvesting systems for the US Southeast and Pa-
no backhauls.
100 km is assumed baseline scenario for additional comparisons.
cific Northwest appear to be more energy intensive than our
Includes emissions due to one loading/unloading cycle for truck trips and two Michigan logger survey data and default assumptions indicate e
cycles for bimodal trips. although if we were to assume a combination of feller-buncher and
R.M. Handler et al. / Journal of Cleaner Production 76 (2014) 64e73 71

Fig. 2. Sensitivity analysis for key inputs to harvesting and transportation unit operations. Input variables were increased or decreased by 10% to observe resulting changes in overall
GHG emissions for harvesting or transport unit operations.

chainsaw harvesting, Michigan harvesting impacts could be much associated with incorporating a renewable carbon feedstock into a
higher. biofuel or bioenergy system. Assuming the roundwood contains
The supply chain environmental impacts presented in Table 9 50% carbon (Argonne National Laboratory, 2013), a green tonne of
are significant, but they should not be viewed as a barrier to forest biomass would correspond to 916.7 kg CO2eq available for
sustainable development of forest-based biofuels or bioenergy. To conversion into a useful product. If the wood is used to create
place the results in context, a cumulative GHG emissions value of ethanol (LHV 26.8 MJ/kg) via fermentation with a conservative
40.4 kg CO2eq/green tonne is much smaller than the carbon credit yield of 334 L/dry tonne (Argonne National Laboratory, 2013), the

Table 9
Comparison of forest biomass supply life cycle environmental impacts.

Source GHG emissions kg CO2 Fossil energy demand Comments

eq/tonnea MJ/tonne

(Sonne, 2006) 17.4 Harvesting e U.S. Pacific NW, 2.9 Mg CO2 eq/300 m3 timber and 5.5 Mg
38.2 Transport CO2 eq/300 m3 timber for mechanized harvest and transport,
55.6 Total respectively, Douglass fir density 0.48 g/cm3 (Seely, 2011) used
for all density assumptions needed in subsequent comparisons
(Johnson et al., 2005) w50e58 Total w615e715 Total Table 4b, harvesting and hauling, fuel use and lubricant data
for U.S. SE and Pacific NW, 90e120 km one-way transport,
CORRIM group
(Gonzalez-Garcia et al., 2009) e 283e340 harvesting Data for Spain (eucalyptus plantation, low value) and Sweden
226e100 transport (softwoods, high value), 90 km transport, 40% moisture assumed
509e440 total
(Slade et al., 2009) 23.8 Harvesting e Softwood logs, UK and Swedish data, 107 km transport, assume
9.2 Transport 50% moisture to convert data from dry tonnes to green tonnes
33 Total
(Klvac et al., 2003) e 214e250 Harvesting Only Estimates from Sweden and Ireland
(Keoleian and Volk, 2005) 5.9 Harvesting Only 157.1 Harvesting Only Willow plantation, high intensity growth with periodic coppice
harvest every 3 years, very different system
(Valente et al., 2011) 15.2 Harvesting 204 Harvesting Norwegian forestry in mountainous regions, 64 km transport
10.2 Transport 155 Transport
25.4 Total 359 Total
This Study 17.8 Harvest 233 Harvest Assuming aggregated harvest data and baseline transport
3.1 Loading 17.3 Loading scenario as discussed above
19.4 Transport 246 Transport
40.4 Total 495.3 Total
All values in table listed on the basis of green tonnes.
72 R.M. Handler et al. / Journal of Cleaner Production 76 (2014) 64e73

supply chain GHG emissions would translate to roughly 11.5 g Disclaimer

CO2eq/MJ final ethanol product. These emissions are only for the
feedstock supply, but this level of emission prior to the fuel con- This report was prepared as an account of work sponsored by an
version stage should still allow for production of fuel that has a agency of the United States Government. Neither the United States
cradle-to-grave emissions profile well below the comparable Government nor any agency thereof, nor any of their employees,
standard for petroleum gasoline (91.3 g CO2eq/MJ) (Skone and makes any warranty, express or implied, or assumes any legal lia-
Gerdes, 2008). This is especially true in this case, because ligno- bility or responsibility for the accuracy, completeness, or usefulness
cellulosic ethanol production often assumes a significant envi- of any information, apparatus, product, or process disclosed, or
ronmental credit from lignin that is used to create an electricity represents that its use would not infringe privately owned rights.
co-product (Argonne National Laboratory, 2013). Similarly, fossil Reference herein to any specific commercial product, process, or
energy demands of wood supply amount to 495.3 MJ/green tonne, service by trade name, trademark, manufacturer, or otherwise does
but this energy use is corresponds to roughly 3% of the energy that not necessarily constitute or imply its endorsement, recommen-
is embodied in the forest biomass (LHV 16.5 MJ/kg) (Argonne dation, or favoring by the United States Government or any agency
National Laboratory, 2013). If the same ethanol yield of 334 L/ thereof. The views and opinions of authors expressed herein do not
dry tonne is applied, the baseline fossil energy demand for necessarily state or reflect those of the United States Government
roundwood feedstock would correspond to 14% of the embodied or any agency thereof.
energy within the ethanol product.
4. Conclusions
Abbas, D., Handler, R., Hartsough, B., Dykstra, D., Srivastava, A., Hembroff, L., et al.,
2014. Forest logging and transportation Logistics in Michigan. Croat. J. For. Eng.
Using a combination of data from a survey of the Michigan forest (under review).
products industry, expert opinion, literature, and database sources, Argonne National Laboratory, 2013. The Greenhouse Gases, Regulated Emissions,
and Energy Use in Transportation (GREET) Model. Available at:
a limited-scope life-cycle assessment of the forest biomass supply
chain in Michigan was constructed. Greenhouse gas emissions and Athanassiadis, D., Lidestav, G., Nordfjell, T., 2002. Energy use and emissions due to
fossil energy demand for wood harvesting were calculated for a the manufacture of a forwarder. Resour. Conserv. Recycl. 34, 149.
Blinn, C.R., Hendricks, L.T., 1997. Marketing Timber from the Private Woodland.
range of equipment configurations and harvest scenarios, revealing
University of Minnesota Extension Services, St. Paul, MN.
important differences between the harvesting practices commonly CN Railroad, 2009. Canadian National Railroad Greenhouse Gas Calculator Emission
observed within the state. Transport of forest biomass by truck Factors.
Department of Energy, 2011. U.S. Billion-Ton Update: Biomass Supply for a Bio-
carries a higher environmental burden per ton-mile than an
energy and Bioproducts Industry. Oak Ridge National Laboratory, Oak Ridge,
equivalent distance of rail transport, usually by a factor of 4e5. One TN, U.S.
loading and unloading cycle of wood accounted for roughly 8% of Dwivedi, P., Alavalapati, J.R.R., 2009. Stakeholders perceptions on forest
overall GHG emissions and 3.5% of fossil energy demand, while biomass-based bioenergy development in the southern US. Energy Policy 37,
remaining environmental burdens were divided relatively equally Fan, J., Kalnes, T.N., Alward, M., Klinger, J., Sadehvandi, A., Shonnard, D.R., 2011. Life
between harvest and transport steps. Calculated environmental cycle assessment of electricity generation using fast pyrolysis bio-oil. Renew.
burdens for a defined harvest and transport supply chain base case Energy 36, 632.
Federal Railroad Administration, 2009. Comparative Evaluation of Rail and Truck
in MI were within the range of values reported for similar opera- Fuel Efficiency on Competitive Corridors. Prepared for the U.S. Department of
tions in the literature. Baseline scenario results indicate that wood Transportation.
harvest and transport can be equally important components of an Frischknecht, R., Althaus, H.-J., Doka, G., Dones, R., Heck, T., Hellweg, S., et al., 2004.
Overview and Methodology, Final Report. Ecoinvent v2.0 No. 1. Swiss Centre for
overall wood product, biofuels, or bioenergy supply chain, but the Life Cycle Inventories, Duebendorf, CH.
level of impacts observed here do not prevent forest biomass from Froese, R., Abbott, Z., 2012. Energy Crop Opportunities in the Western Upper
serving as an environmentally friendly feedstock source for biofuels Peninsula of Michigan. Michigan Technological University, Houghton, MI.
Gonzalez-Garcia, S., Berg, S., Feijoo, G., Moreira, M.T., 2009. Environmental impacts
or bioenergy. For instance, fossil energy demand in our baseline
of forest production and supply of pulpwood: Spanish and Swedish case
scenario is 495.3 MJ/green tonne of wood, which is only 3% of the studies. Int. J. Life Cycle Assess. 14, 340.
energy content embodied in the delivered wood. We hope that a Handler, R.M., Canter, C.E., Kalnes, T.N., Lupton, F.S., Kholiqov, O., Shonnard, D.R.,
et al., 2012. Evaluation of environmental impacts from microalgae cultivation in
detailed summary of our approach to arriving at these values will
open-air raceway ponds: analysis of the prior literature and investigation of
allow policy makers, business developers, and other stakeholders wide variance in predicted impacts. Algal Res. 1, 83.
in the forest biomass industry to utilize these values with some Hoekman, S.K., 2009. Biofuels in the US: challenges and opportunities. Renew.
degree of confidence when considering the environmental burdens Energy 34, 14.
Intergovernmental Panel on Climate Change, 2007a. Climate Change 2007 e the
of the forest biomass supply chain for planned biofuels and bio- Physical Science Basis: Working Group I Contribution to the Fourth Assessment
energy facilities, or highlight areas where more location-specific Report of the IPCC. Cambridge University Press, New York, NY.
data would improve the accuracy of a particular assessment. Intergovernmental Panel on Climate Change, 2007b. Climate Change 2007: Miti-
gation Contribution of Working Group III to the Fourth Assessment Report of
the IPCC. Cambridge University Press, New York, NY.
Johnson, D.W., Curtis, P.S., 2001. Effects of forest management on soil C and N
Acknowledgements storage: meta analysis. For. Ecol. Manag. 140, 227.
Johnson, L.R., Lippke, B., Marshall, J.D., Comnick, J., 2005. Life-cycle impacts of forest
This research was supported through an agreement with the resource activities in the Pacific Northwest and Southeast United States. Wood
Fiber Sci. 37, 30.
Michigan Economic Development Corporation with funding from Keoleian, G.A., Volk, T.A., 2005. Renewable energy from willow biomass crops: life
the U.S. Department of Energy award DE-EE-0000280 and from the cycle energy, environmental and economic performance. Crit. Rev. Plant Sci.
Michigan Economic Development Corporation Center of Energy 24, 385.
Klvac, R., Ward, S., Owende, P.M.O., Lyons, J., 2003. Energy audit of wood harvesting
Excellence Program, a program funded through the State’s 21st
systems. Scand. J. For. Res. 18, 176.
Century Jobs Fund. The authors wish to acknowledge the generous Leefers, L.A., Vasievich, J.M., 2010. Timber Resources and Factors Affecting Timber
input from several MI-based industries and logging firms and op- Availability and Sustainability for Kinross, Michigan. Michigan State University,
erators, from Michigan Technological University: Rui Shi for assis- Department of Forestry, East Lansing, MI.
Perez-Verdin, G., Grebner, D.L., Munn, I.A., Sun, C., Grado, S.C., 2008. Economic
tance in report preparation, and from Michigan State University: impacts of woody biomass utilization for bioenergy in Mississippi. For. Prod. J.
Mr. Michael D. Schira (MSU Extension). p. 75þ.
R.M. Handler et al. / Journal of Cleaner Production 76 (2014) 64e73 73

Seely, Oliver, 2011. Physical Properties of Common Woods. California State Uni- Stewart, R.D., Cao, M., Lautala, P., Paterson, K., Graman, G., Hicks, J., et al., 2010.
versity, Dominguez Hills. Available at: Study of Greenhouse Gas Savings Associated with Congestion Reduction Using
woods.htm (Web Accessed 2011). Multi-modal Optimization of Timber Shipments in the North Central United
Shonnard, D.R., Williams, L., Kalnes, T.N., 2010. Camelina-derived jet fuel and diesel: States. Transportation Logistics and Research Center, University of Wisconsin-
sustainable advanced biofuels. Environ. Prog. Sustain Energy 29, 382. Superior.
Skone, T., Gerdes, K., 2008. Development of Baseline Data and Analysis of Life Cycle U.S. Energy Information Administration, 2012. Annual Energy Outlook 2012, with
Greenhouse Gas Emissions of Petroleum-Based Fuels. US Department of Energy, Projections to 2035. U.S. Department of Energy.
National Energy Technology Laboratory, Office of Systems, Analysis and Valente, C., Hillring, B.G., Solberg, B., 2011. Bioenergy from mountain forest: a life
Planning. cycle assessment of the Norwegian woody biomass supply chain. Scand. J. For.
Slade, R., Bauen, A., Shah, N., 2009. The greenhouse gas emissions performance of Res. 26, 429.
cellulosic ethanol supply chains in Europe. Biotechnol. Biofuels, 2. Van Deusen, P.C., Roesch, F.A., 2008. Alternative definitions of growth and removals
Sonne, E., 2006. Greenhouse gas emissions from forestry operations: a life cycle and implications for forest sustainability. Forestry 81, 177.
assessment. J. Environ. Qual. 35, 1439.