Ap

TE

a

AA

KAFIMST

1

dtpTocnamH

nf

U

0d

Ecological Modelling 221 (2010) 2215–2225

Contents lists available at ScienceDirect

Ecological Modelling

journa l homepage: www.e lsev ier .com/ locate /eco lmodel

ssigning results of the Tool for Sustainability Impact Assessment (ToSIA) toroducts of a forest-wood-chain

aru Palosuo ∗, Tommi Suominen, Wendelin Werhahn-Mees, Jordi Garcia-Gonzalo1, Marcus Lindneruropean Forest Institute (EFI), Torikatu 34, FIN-80100 Joensuu, Finland

r t i c l e i n f o

rticle history:vailable online 21 April 2010

eywords:llocationorestrympact assessment

ulti-output processesustainability impactsoSIA

a b s t r a c t

A Tool for Sustainability Impact Assessment (ToSIA) has been developed for assessing sustainabilityimpacts of forest-wood-chains (FWCs). Sustainability is determined by analysing environmental, eco-nomic, and social sustainability indicators for all the production processes along the FWC. Results of thetool can be analysed at an aggregated level for complete FWCs, but for some applications it is useful toassign the indicator results to products of the chain.

This paper presents a procedure in ToSIA to assign sustainability impacts to multiple output products ofFWC. The procedure was tested and demonstrated with an example FWC from Scandinavia that includedfurniture and bio-energy production. Two different allocation criteria, carbon-based and economic value-based, were applied with different options for assigning the impacts on the sub-products of the chain.Three indicators representing the three pillars of the sustainability were chosen to demonstrate theprocedure: production costs (economic), employment (social) and transport intensity (environmental).

The results indicated that the allocation criteria greatly affect the indicator results assigned to the

different products of FWCs. The selection of the allocation criterion depends on the question approachedand on the availability of the needed process related data. The data availability is assured for the carbon-based allocation within ToSIA, as following the carbon flows within the chain is mandatory for any ToSIAapplication. Economic values of products, on the other hand, are more closely linked to the aims of theproduction processes of the value chains and are thereby meaningful allocation criteria in many cases.The allocation procedure of ToSIA was proved to be flexible allowing different criteria and still consistent

s sus

in allocation of the variou

. Introduction

The European Union (EU) has established ambitious goals toevelop a more sustainable society. It adopted a strategy for sus-ainable development (EC, 2001) in which it declared that all EUolicies must have sustainable development as their core concern.herefore, all major policy proposals should include an assessmentf the potential economic, environmental and social benefits andosts of action or lack of action. This has consequently created a

eed for reliable and transparent ex-ante assessment of sustain-bility impacts of planned policies, and new integrated assessmentethods and tools are being developed for different sectors (e.g.elming et al., 2008; Ewert et al., 2009).

∗ Corresponding author. Current address: MTT Agrifood Research Finland, Luut-antintie 13, FIN-00410 Helsinki, Finland. Tel.: +358 40 186 9114;

ax: +358 9 563 1164.E-mail address: taru.palosuo@mtt.fi (T. Palosuo).

1 Current address: Centro de Estudos Florestais, Instituto Superior de Agronomia,niversidade Técnica de Lisboa, Tapada da Ajuda, 1349-017 Lisboa, Portugal.

304-3800/$ – see front matter © 2010 Elsevier B.V. All rights reserved.oi:10.1016/j.ecolmodel.2010.03.020

tainability impacts of the FWCs.© 2010 Elsevier B.V. All rights reserved.

The forest-based sector makes a significant contribution to theEuropean economy. The sector employs around 3.9 million peopleand the annual sales are on average 400 billion EUR (Blombäck etal., 2003). The challenge of measuring the sustainability impactsof the forest-based sector was tackled in the European-wideEFORWOOD project (www.eforwood.org), in which the Tool forSustainability Impact Assessment (ToSIA) for complete forest-wood-chains (FWCs) was developed. The tool can be used tohighlight changes in sustainability due to deliberate actions (e.g.in policies or business activities) or due to external forces (e.g.climate change, global markets). Impacts are calculated for vari-ous economic, social and environmental sustainability indicatorsthat are linked to processes of the FWCs in a similar way asin Life Cycle Assessment (LCA) (ISO 14044, 2006). The tool wasdeveloped to be able to do wide-scale impact assessment analysiscovering forest sector processes and products up to a continen-

tal scale. ToSIA integrates several different methods, and allowsan overall sustainability impact assessment to be performed byfurther processing results with the incorporated multi-criteriaanalysis (MCA) (Mendoza and Martins, 2006) or cost–benefit anal-ysis (CBA) (Nas, 1996) functionalities. A detailed description of the

2216 T. Palosuo et al. / Ecological Modelling 221 (2010) 2215–2225

Fig. 1. Simplified illustration of the Tool for Sustainability Impact Assessment (ToSIA), input–output modelling (IOM) and environmental life cycle assessment (LCA). Big boxesrepresent geographical system boundaries (e.g. country) within which all studied impacts are to be covered. Small boxes represent production sectors. LCA follows singularp linkd onoma

Ti

ctiwtscf2tcataoiaswoatdtfmeepf

roduction chains as far as possible, also over the country and sector borders. IOMone by ToSIA concentrates on the forest-based sector and covers also social and ecnalysis (MCA) or policy analysis (PA).

oSIA approach is given in Lindner et al. (2010) in this specialssue.

As ToSIA was developed to be a decision support tool, the mainriteria for the ToSIA development were reliability, flexibility andransparency, the latter of which requires a certain level of simplic-ty from the methods used. Because the integrative nature of ToSIA

as an important consideration, the tool was developed to be ableo handle a large number of varying sustainability indicators at theame time and to be able to cover systems from single productionhains up to the whole EU forest-based sector. These aspects dif-erentiate ToSIA from the environmental product LCA (ISO 14044,006), in which the number of studied indicators at one time isypically narrower, but where the impacts of the production orhanges in production are studied at a more detailed level includinglso the avoided impacts of alternative production. Whereas LCA orhe related carbon or ecological footprint assessment (Wiedmannnd Minx, 2007) focus on improving the environmental impactsf a product, ex-ante sustainability impact assessment as appliedn ToSIA aims to evaluate the impacts of alternative policies orlternative production processes on sustainable development, con-idering the different dimensions of sustainability in a balanceday. FWCs studied in ToSIA do not currently cover processes from

ther industrial sectors whereas in LCA the production is followeds far as possible, also beyond sectoral borders (Fig. 1). This limita-ion of ToSIA is not, however, a conceptual decision, but is ratherue to limited resources that required priority to be placed onhe development of methods and collection of sustainability dataor the forest-based sector in Europe. On the other hand, ToSIA is

ore detailed than environmentally extended input-output mod-ls (Turner et al., 2007; Wiedmann et al., 2007), in which thenvironmental impacts of production sectors are linked to com-iled trade statistics of inter-sectoral or national level. The primaryunction of environmental input–output analysis is in a static ex-

s the environmental impacts to trade statistics. Sustainability impact assessmentic impacts that can then be evaluated by cost–benefit analysis (CBA), multi-criteria

post accounting mode to quantify the inter-dependence of differentactivities and environmental loads within the economy (Fig. 1).

Production chains in the forest-based sector, as well as many ofthe processes of FWCs are multi-functional, i.e. they yield more thanone functional product. For example, material flows from a harvestof one forest stand can lead to several products from high qual-ity carpentry products to soft paper and from house constructionmaterials to energy production. The issue of multi-functionalityleads to the question of how the sustainability impacts of produc-tion chains should be divided among the multiple output productsof chains. This problem is commonly called an allocation problemand has been widely discussed in the context of Life Cycle Inven-tories (LCI) (e.g. ISO 14044, 2006; Jungmeier et al., 2002a; Ekvalland Finnveden, 2001). It has been noticed that the choice of theallocation procedure has a notable impact on the results of thesestudies (see e.g. Guinee and Heiijungs, 2007; Jungmeier et al., 2003).The International Organization for Standardization (ISO) has pre-sented a standard for LCI (ISO 14044, 2006), that defines possibleprocedures for allocation.

The focus of ToSIA lies on sustainability impact analysis ofchanges to the FWC, rather than on assessing sustainability of indi-vidual FWCs (Lindner et al., 2010). The tool is therefore meant foreffect-oriented sustainability assessments and it approaches thistask in a descriptive way by facilitating the consistent data col-lection and analysis of FWCs to be compared. For example, thetool could be applied to study the impacts of policy actions takento enhance the bio-energy production on the sustainability of theforestry sector at country level by comparing two alternative FWCs.

Here, allocation serves the purpose of enabling comparisons of sus-tainability impacts between the alternative scenarios in relation toa common functional unit quantifying the output of the productionsystems in question (e.g. Schmidt et al., 2009). Allocation in thesecomparisons is applied separately for compared chains in order

Model

ttabaT

T. Palosuo et al. / Ecological

o see how different products of alternative chains contribute to

otal sustainability impacts of the chains. Therefore the allocationpproaches applied in ToSIA can be closer to those applied in attri-utional LCI (Heijungs, 1997) even though the basic aims of ToSIAre similar to consequential LCI (Ekvall and Weidema, 2004). InoSIA, the allocation is not only implemented at the process level,

Fig. 2. Flow diagram of the Scandi

ling 221 (2010) 2215–2225 2217

but the sustainability impacts along the chain are also assigned to

the main products of an entire chain. Therefore, the readily avail-able allocation methods in LCI are not directly applicable in ToSIA.

The objective of this study was to develop a procedure to assignsustainability impacts to FWC products within ToSIA. The use of theprocedure was demonstrated with an example FWC from Scandi-

navian pine chain example.

2 Model

nvaod

2

2

icmciawaaeitibor

tIatFovciTaahoeP2dcpEwp

itcFompmoofac

218 T. Palosuo et al. / Ecological

avia. Two different allocation criteria (carbon-based and economicalue-based) were applied for the FWC with different options forssigning the impacts to the sub-products of the chain. Finally, rec-mmendations for the use of the allocation procedure in ToSIA wereiscussed.

. Material and methods

.1. ToSIA description

ToSIA is a software tool that analyses sustainability impactsn the forest-based sector. In ToSIA, the FWC is specified as ahain of production processes covering segments of forest resourceanagement (FWCS1), forest to industry interaction (FWCS2), pro-

essing and manufacturing (FWCS3), and industry to consumernteractions (FWCS4) (Fig. 2). The start and end points of the FWCre usually forest regeneration and the end-of-life of the consumedood product. The production processes are connected by input

nd output products, i.e. output products of a production processre input to the following production process (see Fig. 3 in Lindnert al., 2010). The material flows are calculated based on the topolog-cal information of the FWC, which contains the information neededo divide and merge the material flows at different processes. Thenternal reference unit for material flows in ToSIA is tons of car-on from the harvested trees to the consumption and end-of-lifef the wood products (FWCS2–FWCS4), whereas within the forestesource management (FWCS1) the unit is hectares.

The basic principle in the sustainability calculation of ToSIA iso link the sustainability impacts to the processes of the FWCs.mpacts are represented by different sustainability indicators thatre externally collected for each analysis at the process level. Sus-ainability indicator values are provided for processes of the studiedWC per unit of material flowing through the process (e.g. unitf production costs indicator is given in [D m−3]). The calculatedolume of material flowing through the processes determines thealculated process-level indicator values (i.e. calculated processndicator value = indicator value per material flow × material flow).oSIA can calculate a wide range of indicators defined by the user,s long as the indicators are such that they can be linked to themounts of the forest to wood products (FWCS2–FWCS4) or toectares in the forest (FWCS1). The indicators selected and devel-ped in the EFORWOOD project were mainly based on alreadyxisting indicator sets (e.g. MCPFE (Ministerial Conference on therotection of Forests in Europe), 2003; EC (European Commission),005; EUROSTAT, 2005; UN (United Nations), 2007). The flexibleesign of ToSIA allows for applications with large number of indi-ators and long or wide production chains with large numbers ofrocesses. For example, one of the applications of ToSIA in theFORWOOD project is to analyse the sustainability impacts of thehole EU-FWC (Werhahn-Mees et al., 2010) covering hundreds ofrocesses.

The purpose of ToSIA is to analyse and assess FWC sustainabilitympacts of changes in the FWCs. Therefore, there must be at leastwo variations of the same FWC study case to be able to make aomparison. The comparisons can be, for example, of two differentWC technologies under the same circumstances, policy changes,r consumer changes. Comparisons for current conditions may beeaningful, but most scenario analyses should also include a tem-

oral dimension, as new policies need time to be implemented andarket changes will evolve gradually. The results can be compared

n a quantitative basis, for example by analysing the relative changef different indicator values for a whole FWC or its segments. To giveurther meaning to the changes detected in comparisons, ToSIAlso provides components to perform multi-criteria analysis andost–benefit analysis.

ling 221 (2010) 2215–2225

2.2. Allocation within ToSIA

Elementary level ToSIA results are the sustainability indicatorvalues calculated for processes, such as forest management actions,harvesting, transporting and sawmilling. Impact analyses aggre-gating results for complete FWCs or FWC segments are the moststraightforward application of the tool. For some applications, how-ever, it is useful to assign the results to a certain product or productgroup, e.g. when assessing the effects of bio-energy policies on sus-tainability impacts of pulp and paper production. Another need toassign sustainability impacts to products arises when applying thetool to assess sustainability impacts in a FWC that is producingone main product. The material amounts handled in the FWCs usu-ally diminish from the early to later phases of the FWC because ofthe residues and sub-products that are produced besides the mainproduct (see e.g. Fig. 3). Therefore, the sustainability impacts of theearly phases of the chain are not only linked to the main productand it could be argued that the sub-products should also carry someof the sustainability impact. Thus, to be able to assign the sustain-ability impacts to final products produced in the FWC, an allocationprocedure for the tool is needed.

How sustainability impacts calculated along the FWCs are con-nected to output products of the FWC depends both on theprocess-level division of impacts between the output products ofmulti-functional processes and on the position of the source pro-cesses within the chain topology. In the following sub-sections, wedescribe the allocation procedure developed for ToSIA.

2.2.1. Identifying the productsAs FWCs are multi-functional, they are always comprised of

more than one product (see Fig. 3 as an example of a multiple prod-ucts produced in a FWC). For the allocation, a selection needs to bemade of the products (FWC products) among which the sustainabil-ity impacts of a chain will be allocated. These products are a subsetof products linking the processes within the FWC topology. Gener-ally, the FWC products should include all products in the use-stageprocesses (i.e. the processes where the products are used by con-sumers, see e.g. processes 22 Using the chair, and 26 Consumption,energy and heat production in Fig. 2) and at the system boundaries ofthe chain (i.e. products that are leaving the chain—product arrowsin Fig. 3). The selection of the FWC products is, however, not alwaysunambiguous, because FWC structures can be complicated withseveral, and possibly even hundreds of products. Also, FWC prod-ucts are usually not located at the end of the FWC. This is becausethe chain typically includes also the processes describing possibledisposal or recycling of products after use (e.g. recovery processesafter the use-stage in our example chain in Fig. 2).

The FWC products are classified as main products and sub-products. The main products define the principal functional purposeof the chain, while sub-products are those that have only negligi-ble economic value and a minor material share. In principle, none ofthe output products of processes of the FWC are called waste. Thisis because only very seldom does any wood-based material withinthe production chains have a negative market value. The weight-ing factor (˛) allows additional control on the allocation of impactsto those sub-products. There are three different options on howto treat the sub-products: (1) the sustainability impacts assignedto sub-products are the same as for the main products accord-ing to selected criterion (˛ = 1); (2) allocation to sub-products isdecreased with an ˛ in the range]0,1[; or (3) sustainability impactsare not allocated to sub-products at all (˛ = 0).

2.2.2. Allocation criterionThe selected allocation criterion is consistently used through-

out the chain, as this promotes transparency in ToSIA, where FWCsconsisting of hundreds of processes may be defined. The allocation

T. Palosuo et al. / Ecological Modelling 221 (2010) 2215–2225 2219

F he Scas rvest r

cbttctuAi

gabwippomu

= ∑

= ∑

ig. 3. Material flow (calculated based on the carbon content of the products) of tystem boundaries relative to the total amount of harvested wood products and ha

riterion that determines the basis for the partitioning needs toe such that it can be linked to all products of the chain. Further,he criterion should be relevant for the final goals of the alloca-ion and studied question, and it should be feasible throughout thehain. Allocation based on the carbon content of products is easiesto implement within ToSIA, as wood-based carbon is the internalnit for the material flow calculations of the tool (FWCS2–FWCS4).vailability of the carbon amounts of all products along the chain

s therefore assured. ToSIA defines that 100% of the carbon amount

oing into a process is accounted for in the output products ors a documented loss and the carbon amount is not impactedy changes in moisture content or addition of, for example, non-oody materials such as metals. Product amounts (for allocation)

n other units can be attained by using product specific multi-

⎧⎪⎪⎨⎪⎪⎩

amainproduct

asub-product

liers (i.e. conversion factors), which can be used to convert theroduct’s carbon amount to another defined unit of product. Onef those units suitable as an economic allocation criterion is thearket price of products. Price information is needed for the prod-

cts along the FWC. Price information is used in the cost–benefit

ndinavian chain. The figure shows the carbon content of the products leaving theesidues.

analysis and is needed in calculating certain indicators relevantonly at the chain level, such as total production or the gross valueadded.

2.2.3. Allocation calculationsAfter the selection of the FWC products, their relative shares

among all FWC products are calculated based on the chosen allo-cation criteria. The FWC product shares (aproduct) are calculatedwith

mmainproductnmainproduct=1mmainproduct +

∑psub-product=1˛msub-product

˛msub-productnmainproduct=1mmainproduct +

∑psub-product=1˛msub-product

(1)

where m is the property of the FWC products selected as the alloca-tion criterion, n is the number of main products, and p the numberof sub-products. FWC product shares are calculated using productinformation from the use-stage processes or from the processeswhere the product is leaving the FWC.

Not all the impacts of the FWC are allocated to FWC productsdirectly on the basis of the FWC product shares, because this couldresult in biased allocation. This is due to the fact that the processesare not equal in relation to the indicator values, and also not thesame sets of processes are needed to process different products.

2 Model

Faitr

rpntcpt6napwictwcts(mcbpp

mibtFe(

c

wbois

aHohasiFtbtaw

pp

220 T. Palosuo et al. / Ecological

or example, a process can play a large role in the total value ofn indicator over the whole FWC, but this process may be neededn the production of only one of the FWC products. This brings outhe need to look at the impacts at process level and how they areelated to different FWC products.

To be able to relate the process-level impacts to FWC productsequires information from the chain topology; i.e. whether a certainrocess is a part of the production chain of a certain FWC product orot, and if there are other processes or production chain segmentshat are parallel to this process and are involved in the productionhain of the same product. For example, saw logs for the sawmillingrocess may come both from final felling and from thinning opera-ions through parallel chain branches (e.g. the branch of processesand 8 vs. the branch of processes 7 and 9 in Fig. 2) meaning thatot all sawn material come from both of these processes. Within thellocation procedure of ToSIA, this effect is taken into account usingrocess shares (bij). The bij reflects the share of a FWC product i thatas treated in process j. This is calculated in carbon amounts, which

s the internal calculation unit of ToSIA. Material flow in terms ofarbon is followed from the use-stage process of the FWC producto each process of the chain and for each process it is calculatedhat share of the carbon in the FWC product in the use-stage pro-

ess was treated in them. To illustrate the process shares we can usehe flow diagram in Fig. 2, where all the material coming to the use-tage process 26 (i.e. consumption) comes through the processes 25transport) and 19 (pellet production). So, in these processes, as all

aterial was treated, the process share for the pellets burnt in pro-ess 26 is 100%. On the other hand, since there are two branchesringing material to process 19 (processes 16 and 17), for theserevious processes the process shares are determined by the inputroduct shares of process 19.

The process shares are calculated using the topological infor-ation of the graph that is a representation of the FWC and the

nformation on the shares of products at processes: we defineij = cuv, where u refers to an arbitrary process j with a path to v,he final process of product i (i.e. the process used to calculate theWC product share of product i). cuv is calculated recursively forach process in a chain assuming a path P in the directed graph Gthe topology) connecting the processes u and v. For each n ∈ Pu,v:

n,v = cncn−1,v (2)

here cn is the share of input product for process n determinedased on the carbon amounts. When n = 1, cn−1,v is the output sharef queried product i in process v. When n = u, cn = 1. In case theres more than one path between the processes u and v, the cu,v areummed over the alternative paths.

Thinning operations in forests are done to enhance forest growthnd to increase wood quality of the trees that are left in the forest.owever, the living trees left in the forest are not output productsf the FWCS2 thinning processes, because the output products ofarvesting processes include only the harvested wood assortmentsnd harvest residues. As the allocation procedure described abovetrictly follows the material flows, the impacts of processes affect-ng a product in an indirect manner would not be allocated to theWC products without an exception to the procedure. This meanshat, for example, saw logs from final felling would not be affectedy impacts of thinning operations. The users can therefore definehe process shares of these special processes in FWCS2 to allow for

llocation also to those FWC products which are not directly linkedith a process.

To get the final allocation share (sij) for each process (j) and FWCroduct (i), the FWC product shares (ai) calculated for the FWCroducts (i) based on the selected allocation criterion are weighted

ling 221 (2010) 2215–2225

with the process shares (bij):

sij = aibij∑ni=1aibij

(3)

FWCs may also include internal loops, e.g. in the case of recyclingthere is a material flow from a process further in the chain enteringa process earlier in the chain. For the chains including these loops,the procedure described above is also feasible. With the procedure,the impacts of recycling processes are allocated between the oldand the new products according to their shares at chain and processlevels, exactly as for the allocation of other processes.

2.2.4. Allocation of the sustainability indicator valuesAllocation of the sustainability indicator values Iij of process (j)

related to FWC product (i) is calculated by multiplying the processindicator values Ij with the allocation shares (sij):

Iij = sijIj (4)

2.2.5. Calculation of the chain indicator values with varyingsystem boundaries

In principle, the system boundaries of the FWCs are set withthe chain structure. It determines which processes are taken intoaccount and how far each product is followed along its produc-tion, use and recycling chain. Applying the allocation proceduredescribed above, it is possible to redefine the system boundarieswithin the chains. This allows for the assessment of sustainabil-ity impacts of some particular products, and to exclude the effectsthat are due to the handling of material leading to sub-products.This is particularly meaningful when comparing the sustainabilityimpacts of the processes along the chain. For example, one maylike to exclude the sustainability effects of growing and handlingof pulp wood in the processes of early phases of a FWC (e.g. in har-vesting and forwarding processes), when the main aim is to studythe sustainability impacts of some sawnwood-products.

2.3. Scandinavian pine chain example

As an example, the allocation procedure is applied to a furni-ture and bio-energy production FWC in Scandinavia. This FWC isbuilt from the forest resource perspective; the forest resources arein northern Sweden and industrial processing and consumption ofthe products are followed also outside of that region. The chainis based on the wood material flows emanated from a Scots pine(Pinus sylvestris L.) management system. The FWC has three mainproducts: (i) chairs representing wooden furniture, (ii) pellets asa bio-energy product, and (iii) pulp wood, for which the systemboundary was set at the mill gate, after measuring (process 11in Fig. 2). For simplicity, the pulp and paper production processeswere not included. Typical production processes for a wooden fur-niture value chain have been specified and the wood residues thataccrue as by-products of sawmilling and furniture manufacturingare utilised to produce pellets for use in private households. Thechain consists of 26 processes (Fig. 2).

The yield level of final cutting in the boundary of the segments offorest resource management (FWCS1) and forest to industry inter-action (FWCS2) are fixed to the equivalent of 1000 tons of carbonin order to initialise the flow amounts of the FWC. The carbon flowfrom other cuttings and thinning was then calculated based on thefixed product shares of the chain. The harvest amounts were then

converted to hectares by assuming an estate of normal forest (i.e.a fully regulated forest) consisting of a myriad of stands, with ageclasses covering the whole range from 0 years to clear-felling age(Davis and Johnson, 1987). This resulted in a forest area of 1803 haas the basis for indicator calculations in FWCS1.

Model

uttm(f

2

tfiTsfsnwotwatTpp

sacttpbipa‘pip

T. Palosuo et al. / Ecological

The data used in this study cover the input and output prod-ct shares within the processes, the required unit conversions forhe products, and the sustainability impacts with relative indica-or values per material flow. The calculations in this study were

ade based on the data collected in the EFORWOOD projectwww.eforwood.org), complemented with some additional valuesrom the literature.

.4. Applying different allocation criteria

The allocation procedure described above was applied for thehree main products of the Scandinavian pine chain example andve sub-products leaving the system boundaries along the FWC.he three main products were chairs, pellets and pulp wood. Theub-products were harvest residues of the harvesting operations,uelwood from sorting at the sawmill, as well as residues from theawmill and the chair components and chair production (that wereot used for pellet production). Two alternative allocation criteriaere tested: (1) carbon-based allocation and (2) allocation based

n economic value, i.e. market prices. Also, three options for how toreat the sub-products were tested: (1) the sustainability impactsere assigned to sub-products as for the main products (˛ = 1), (2)

llocation to sub-products was decreased (˛ = 0.25), and (3) sus-ainability impacts were not allocated to sub-products at all (˛ = 0).he process shares of the pre-commercial thinning and thinningrocesses (felling with large harvester) were set to 100% for all FWCroducts.

Three indicators were selected, representing the three pillars ofustainability: production costs (economic), employment (social)nd transport intensity (environmental). These indicators werehosen, because extensive data along the chain were available andhe indicators represent different patterns in their sources alonghe chain. It must be emphasised that with this simplistic exam-le we do not aim to make a thorough sustainability assessment,ut only demonstrate the allocation procedure and its results. The

ndicator ‘production costs’ includes all costs occurring in the singlerocesses, for example, raw material costs, labour and energy costs,

dministrative costs, sales expenditures and taxes. The indicator

employment’ represents the employment effect (i.e. number oferson years) that occurs in the processes along the FWC. ‘Transport

ntensity’ indicates the distance of the transported material multi-lied by the mass of the transported material in tons. Its unit is ton

Fig. 4. Indicator results for processes along the chain: (a) total produ

ling 221 (2010) 2215–2225 2221

km. Transport intensity is important since it determines many otherindicators, like energy use or emissions of the transport processes.

3. Results

In the studied Scandinavian pine chain, the pulp wood covered39% and harvest residues 22% of the total removals. Only 10% and17% of the total removals ended up to chairs and pellets, respec-tively (Fig. 3).

Sources of different sustainability impacts, i.e. processes withhigh indicator values, along the FWCs varied between different sus-tainability indicators. The production costs were mainly due to thesawmill process and the chair and chair component production pro-cesses (Fig. 4a). These same processes were also the main sourcesof employment (Fig. 4b). However, the employment effects weremore spread along the processes of the chain. The transport inten-sity indicator was naturally meaningful only for those processesthat involve transportation (Fig. 4c). The largest transport intensitywas for the process of transport from forest to sawmill after finalfelling.

FWC product shares (Eq. (1)) determined for the FWC productsdiffered considerably when using either economic value (marketprice) or carbon content as the allocation criterion (Fig. 5). Inclusionof the sub-products affected the carbon-based FWC shares, whereasthe value-based shares were hardly changed at all as the value of thesub-products was very small. Carbon-based FWC product shares ofpulp wood increased from 39% to even 60% when sub-productswere excluded. The FWC product shares of pellets increased from17% to 26% when sub-products were excluded and FWC productshares of chairs increased from 10% to 14%. Chairs covered 88% ofthe total economic value of the FWC products when sub-productswere included and 89% if they were excluded. Pellets and pulp wood(at the mill gate) covered only 7% and 4%, respectively, regardlessof the inclusion of the sub-products. The carbon-based FWC prod-uct share of the sub-product harvest residues was 22%, and thecarbon-based share of residues from chair component and chairproduction were both 4%. The value-based FWC product shares of

all sub-products were less than 1%.

Similarly with the FWC product shares, the final allocationshares (Eq. (3)) of the studied indicators calculated on a value basiswere not sensitive to inclusion of sub-products (Fig. 6a–c). Whensub-products were included and allocation was made on a carbon

ction costs, (b) employment, and (c) road transport intensity.

2222 T. Palosuo et al. / Ecological Modelling 221 (2010) 2215–2225

Fig. 5. FWC product shares at the chain-level assessed on carbon and value basis and with sub-products (˛ = 1) and without sub-products (˛ = 0).

Fig. 6. Indicators total values at the FWC level: (a) total production costs, (b) employment, and (c) transport intensity, allocated to main and sub-products of the Scandinavianpine chain on a value basis and on a carbon basis and with different weights for sub-products: with sub-products (˛ = 1), without sub-products (˛ = 0), and with weightedsub-products (˛ = 0.25).

T. Palosuo et al. / Ecological Modelling 221 (2010) 2215–2225 2223

F each Fw

boip

pmpcttp

aspomvpt

4

adtitrtaasno2

atfaotm

ig. 7. Production cost indicator allocated among the FWC products separately inith sub-products (˛ = 1).

asis, the sub-products covered altogether a bit less than one-thirdf the production costs, employment, and the transport intensityndicators. When the allocation was made on a value basis, the sub-roducts together were allocated less than 1% of all the impacts.

Production costs and employment impacts were allocated toulp wood only by minor shares (Fig. 6a and b) as they wereainly originating from the chair manufacturing processes. Trans-

ort intensity, on the other hand, was allocated to pellets withonsiderable shares (Fig. 6c), since the material handled in tworansport processes was fully assigned to the pellets (transport ofhe residues from the sawmill to pellet production and transport ofellets to users).

Allocation of the same indicator along the FWC is not same forll processes. It varies for processes and thereby also for the FWCegments. In the earlier phases of the chain there are typically moreroducts that are affected by the processes than in the later phasesf the chain. For example, allocation of the production cost indicatorade both on carbon (Fig. 7a) and value (Fig. 7b) basis showed clear

ariation along the chain. Allocation of production costs to mainroducts, chairs and pellets, increased towards the later phases ofhe chain.

. Discussion

The allocation procedure described in this paper enables thessignment of the sustainability impacts calculated with ToSIA toifferent products of FWCs. It is of interest to know where and howhe different sustainability impacts arise in the FWC, and how thesempacts are affected by different actions or different external fac-ors. It is, however, also of interest to know how these impactselate to the products of the FWC, as these products and their valueo consumers, are essentially the goal that motivates the processesnd activities creating the sustainability impact. Furthermore, thellocation serves also the purpose of enabling comparisons ofustainability impacts between, e.g. alternative policies or tech-ologies, in relation to a common functional unit quantifying theutput of the production systems in question (e.g. Schmidt et al.,009).

In FWCs, the material amounts in the early phases of the chainre large and the economic values of the products relatively low. Inhe later phases of the chains, on the other hand, the values of the

urther processed products are higher, but the material amountsre much smaller. These opposite patterns of volume and valuef the products along the chain raise the question which one ofhese properties related to products would be a more meaningful

easure to use as the basis of the sustainability impact allocation in

WC segments: (a) on a carbon basis and (b) on a value basis. Allocation was done

FWCs. This depends on the case studied. For example, if the targetof the decision makers would be to reduce the demand for productswhose production is unsustainable, a value-based allocation seemsappropriate as the demand is reflected in the values of the finaland intermediate goods. On the other hand, if the target would beto minimise the impacts of traffic related to alternative biomassresource use scenarios, a mass-based allocation (or carbon-basedas a close approximate) could be preferred as the product mass ismore relevant for transport movements.

In this study we demonstrated the use and results of anallocation procedure with three indicators (production costs,employment and transport intensity) calculated for the Scandina-vian pine chain including wooden furniture (chairs), a bio-energyproduct (pellets), and pulp wood as the main products. Theresults indicated that different criteria used, here carbon-basedvs. value-based, greatly affect the indicator results assigned to thedifferent FWC products (Fig. 6a–c). For example, the allocationshare (Eq. (3)) of the total production costs, employment andtransport intensity indicators for chairs were more than doubledwith value-based allocation compared to carbon-based allocation.The importance of the allocation criterion has been demonstratedby several allocation studies in LCA for different sectors (e.g.Jungmeier et al., 2002b; Schau and Fet, 2008).

Economic value is preferred by the ISO standard (ISO 14044,2006), as it reflects the idea of using causal relation between theburden and functions. Value-based allocation typically gives onlyminor weight to the sub-products (see e.g. Dove and Boustead,1998) as can be seen in Fig. 6. On the other hand, economic val-ues are dependent on the prevailing conditions at markets andare therefore highly variable. The uncertainty of the product priceestimates needed for scenario analysis can be high, which is prob-lematic for the scenario analysis typically applied within ToSIA.In the forest-based sector, also the relative prices among differentproduct groups can vary quite drastically as recently documentedin the fluctuations of the energy wood prices vs. pulp wood prices(see e.g. Hawkings Wright, 2009). For the carbon-based allocation,on the other hand, the data availability is assured also for scenarioprojections as it applies the carbon amounts from flow calculationsof ToSIA that are based on data that are mandatory for any ToSIAapplications. Carbon as an allocation criterion in ToSIA is thereby aconsistent and transparent option.

The sustainability impact assessment approach of ToSIA has dif-ferent characteristics than LCA and input–output modelling (seeFig. 1). Each of the methods has specific strengths and limitations.Input–output models provide detailed static ex-post accountingtools to link environmental impacts to the production of com-

2 Mode

maspamacaowsLio

rrfiguTttd

iaucaiitlc

atshvTcdmric

cTsclisdstietfot

224 T. Palosuo et al. / Ecological

odities (Wiedmann et al., 2007). But as they lack informationbout changes in producer and consumer demand, they are not welluited to describe change in a predictive (ex-ante) way. ToSIA wasarticularly developed to support ex-ante sustainability impactssessment based on ‘what if’ scenario analysis. The current imple-entation of the method limits its scope for consequential scenario

nalysis of production systems as practiced in recent LCA appli-ations (see also Lindner et al., 2010; Schmidt et al., 2009, Ekvallnd Weidema, 2004). This limitation is mainly because the scopef the sustainability impact assessment in ToSIA is much broaderith respect to many indicators of sustainability, covering also the

ocial and economic dimensions which are very rarely covered inCA studies. Although this paper used only three indicators forllustration of the method, in EFORWOOD a much larger numberf indicators were collected.

Particularly for the scenario projections, the allocation crite-ion may need to be selected between a criterion that is moreelevant for the question handled and one that is supported by suf-cient, consistent and transparent data. Carbon-based allocationives weight also to the sub-products that may be economicallynimportant. To reduce that effect, the allocation procedure ofoSIA involves the weighting factor ˛ that can be used to reducehe allocation of the impacts to sub-products. Carbon-based alloca-ion without sub-products or with weighted sub-products in Fig. 6emonstrates this option.

In the development of an allocation procedure of ToSIA, the flex-bility of the allocation criterion was seen important. As ToSIA isimed to be used to study different aspects of sustainability, theser should be able to apply the criterion most appropriate for eachase and to assess the impacts of different allocation criteria. Inddition, it was important to generate an allocation procedure thats easy to implement and does not require too much additionalnformation from the user. For the sake of simplicity and consis-ency, the allocation criterion can only be selected at the chainevel, i.e. the same criterion is used consistently throughout thehain.

The setting of the system boundaries of the studied chain alsoffects the allocation. For example, if the pulp and paper produc-ion processes would have been included in the test chain of thistudy, the value of pulp and paper products would have beenigher (or the carbon content of these products lower) than thealue (or carbon content) of the pulp wood at road-side used here.herefore, the allocation of the impacts to pulp wood of those pro-esses that are linked to pulp production would have increased (orecreased). Mainly this concerns the processes in forest resourceanagement segments of the chains. It is therefore important to

emember that the allocation shares of processes are case-specific,.e. dependent on the system boundaries of the studied productionhains.

One important aspect to take into account when building thehain topologies is the subdivision of processes of FWCs withinoSIA. In this context, the level of detail of the chain structurehould allow for sufficient analysis of the chain. Larger units as pro-esses decrease the effort of data collection. On the other hand,arger units may oversimplify the allocation of the sustainabil-ty impacts to the products of the processes. For example, inawmilling, which covers several sub-processes like sawing andrying, the drying process is responsible for most of the energy con-umption. Thus the energy-related impacts should not be allocatedo residues that are not dried. On the other hand, drying is not work-ntensive and therefore there is a need to allocate employment

ffects differently from energy impacts. This is not possible withhe current implementation of the allocation procedure, where dif-erent indicators are treated similarly at the process level. The onlyption would be to split the sawmilling into separate sub-processeso follow the principle of the causal allocation.

lling 221 (2010) 2215–2225

The allocation procedure of ToSIA was proved to be flexibleallowing different allocation criteria and still consistent in allo-cation of the various sustainability impacts of the FWCs. Theallocation procedure of ToSIA was created to be as transparent aspossible, but one should notice that allocation will always involvesubjective decisions by the persons making the allocation (Wernerand Scholz, 2002). This should be remembered when interpretingand applying the allocated indicator results.

Acknowledgements

This work was funded by the European Commission (FP6)through the EFORWOOD project (Project no. 518128). We wouldlike to thank Jörg Schweinle, Bo Jellesmark-Thorsen and threereviewers for their constructive comments on the manuscript.Data on sustainability indicators and other FWC parameters forthe illustration case were contributed by David Gil and variousproject partners. Furthermore, we thank Tim Green for checkingthe English language of the manuscript.

References

Blombäck, P., Poschen, P., Lövgren, M., 2003. Employment Trends and Prospectsin the European Forest Sector. Geneva timber and forest discussion papersECE/TIM/DP/29, Geneva, Switzerland.

Davis, L.S., Johnson, K.N., 1987. Forest Management. McGraw-Hill, New York, 790pp.

Dove, W.T., Boustead, I., 1998. The effect of sulphur on primary zinc ecoprofile calcu-lations. In: Proceedings of the 3rd International Conference on Ecobalance, pp.413–416.

EC (European Commission), 2001. A Sustainable Europe for a Better World: A Euro-pean Union Strategy for Sustainable Development.

EC (European Commission), 2005. Impact Assessment Guidelines, SEC 791.Ekvall, T., Finnveden, G., 2001. Allocation in ISO 14041—a critical review. J. Clean.

Prod. 9, 197–208.Ekvall, T., Weidema, B.P., 2004. System boundaries and input data in consequential

life cycle inventory analysis. Int. J. Life Cycle Assess. 9, 161–171.EUROSTAT, 2005. Measuring Progress towards a more Sustainable Europe. Sustain-

able Development Indicators for the European Union, Luxembourg.Ewert, F., van Ittersum, M.K., Bezlepkina, I., Therond, O., Andersen, E., Belhouchette,

H., Bockstaller, C., Brouwer, F., Heckelei, T., Janssen, S., Knapen, R., Kuiper, M.,Louhichi, K., Olsson, J.A., Turpin, N., Wery, J., Wien, J.E., Wolf, J., 2009. A method-ology for enhanced flexibility of integrated assessment in agriculture. Environ.Sci. Policy 12, 543–630.

Guinee, J.B., Heiijungs, R., 2007. Calculating the influence of alternative allocationscenarios in fossil fuel chains. Int. J. Life Cycle Assess. 12, 173–180.

Hawkings Wright, 2009. Forest Energy Monitor, vol. 1, pp. 1–23.Heijungs, R., 1997. Economic Drama and the Environmental Stage; Formal derivation

of algorithmic tools for environmental analysis and decision-support from aunified epistemological principle. Ph.D. Thesis, Rijksuniversiteit Leiden, pp. 203.

Helming, K., Pérez-Soba, M., Tabbush, P., 2008. Sustainability Impact Assessment ofLand Use Changes. Springer Verlag, Berlin Heidelberg, 502 pp.

ISO 14044, 2006. Environmental Management—Life Cycle Assessment—Requirements and Guidelines.

Jungmeier, G., McDarby, F., Evald, A., Hohenthal, C., Petersen, A.K., Schwaiger, H.P.,Zimmer, B., 2003. Energy aspects in LCA of forest products: guidelines from CostAction E9. Int. J. Life Cycle Assess. 8, 99–105.

Jungmeier, G., Werner, F., Jarnehammar, A., Hohenthal, C., Richter, K., 2002a. Allo-cation in LCA of wood-based products—experiences of Cost Action E9. Part I.Methodology. Int. J. Life Cycle Assess. 7, 290–294.

Jungmeier, G., Werner, F., Jarnehammar, A., Hohenthal, C., Richter, K., 2002b. Allo-cation in LCA of wood-based products—experiences of Cost Action E9. Part II.Examples. Int. J. Life Cycle Assess. 7, 369–375.

Lindner, M., Suominen, T., Palosuo, T., Garcia-Gonzalo, J., Verweij, P., Zudin, S.,Päivinen, R., 2010. ToSIA-A tool for sustainability impact assessment of forest-wood-chains. Ecol Model. 221, 2197–2205.

MCPFE (Ministerial Conference on the Protection of Forests in Europe), 2003.Improved Pan-European Indicators for Sustainable Forest Management asadopted by the MCPFE Expert Level Meeting, Vienna, Austria, 7–8 October2002.

Mendoza, G.A., Martins, H., 2006. Multi-criteria decision analysis in natural resource

management: a critical review of methods and new modelling paradigms. For.Ecol. Manage. 230, 1–22.

Nas, T.F., 1996. Cost-Benefit Analysis: Theory and Application. Sage Publications, 220pp.

Schau, E.M., Fet, A.M., 2008. LCA studies of food products as background for envi-ronmental product declarations. Int. J. Life Cycle Assess. 13, 255–264.

Model

S

T

U

W

T. Palosuo et al. / Ecological

chmidt, J.H., Christensen, P., Christensen, T.S., 2009. Assessing the land use impli-cations of biodiesel use from an LCA perspective. J. Land Use Sci. 4, 35–52.

urner, K., Lenzen, M., Wiedmann, T., Barrett, J., 2007. Examining the global environ-

mental impact of regional consumption activities—Part 1. A technical note oncombining input–output and ecological footprint analysis. Ecol. Econ. 62, 37–44.

N (United Nations), 2007. Indicators of Sustainable Development: Guidelines andMethodologies, third edition, New York.

erner, F., Scholz, R.W., 2002. Ambiguities in decision-oriented life cycle invento-ries: the role of mental models. Int. J. Life Cycle Assess. 7, 330–338.

ling 221 (2010) 2215–2225 2225

Werhahn-Mees, W., Vötter, D., Suominen, T., Lindner, M., 2010. Documentation ofToSIA developments up to month 52. EFORWOOD Deliverable Report D1.4.6(update), April 2010, European Forest Institute, Joensuu, Finland.

Wiedmann, T., Lenzen, M., Turner, K., Barrett, J., 2007. Examining the globalenvironmental impact of regional consumption activities—Part 2. Review ofinput–output models for the assessment of environmental impacts embodiedin trade. Ecol. Econ. 61, 15–26.

Wiedmann, T., Minx, J., 2007. A Definition of ‘Carbon Footprint’. ISA-UK ResearchReport 07-01, Durham, UK.