Climatic Change

, Volume 115, Issue 3, pp 759–776

Assessing temporary carbon sequestration and storage projects through land use, land-use change and forestry: comparison of dynamic life cycle assessment with ton-year approaches

Authors

    • CIRAIG, Department of Chemical Engineering, École Polytechnique de Montréal
  • Pascal Lesage
    • CIRAIG, Department of Chemical Engineering, École Polytechnique de Montréal
  • Manuele Margni
    • CIRAIG, Department of Mathematical and Industrial Engineering, École Polytechnique de Montréal
  • Miguel Brandão
    • Sustainability Assessment Unit, Institute for Environment and Sustainability, Joint Research Centre, European Commission
    • International Life Cycle Academy
  • Réjean Samson
    • CIRAIG, Department of Chemical Engineering, École Polytechnique de Montréal
Article

DOI: 10.1007/s10584-012-0473-x

Cite this article as:
Levasseur, A., Lesage, P., Margni, M. et al. Climatic Change (2012) 115: 759. doi:10.1007/s10584-012-0473-x

Abstract

In order to properly assess the climate impact of temporary carbon sequestration and storage projects through land-use, land-use change and forestry (LULUCF), it is important to consider their temporal aspect. Dynamic life cycle assessment (dynamic LCA) was developed to account for time while assessing the potential impact of life cycle greenhouse gases (GHG) emissions. In this paper, the dynamic LCA approach is applied to a temporary carbon sequestration project through afforestation, and the results are compared with those of the two principal ton-year approaches: the Moura-Costa and the Lashof methods. The dynamic LCA covers different scenarios, which are distinguished by the assumptions regarding what happens at the end of the sequestration period. In order to ascertain the degree of compensation of an emission through a LULUCF project, the ratio of the cumulative impact of the project to the cumulative impact of a baseline GHG emission is calculated over time. This ratio tends to 1 when assuming that, after the end of the sequestration project period, the forest is maintained indefinitely. Conversely, the ratio tends to much lower values in scenarios where part of the carbon is released back to the atmosphere due to e.g. fire or forest exploitation. The comparison of dynamic LCA with the ton-year approaches shows that it is a more flexible approach as it allows the consideration of every life cycle stage of the project and it gives decision makers the opportunity to test the sensitivity of the results to the choice of different time horizons.

1 Introduction

Growing concerns about climate change have led to an increasing number of global warming mitigation projects by temporary carbon sequestration through land-use, land-use change and forestry (LULUCF) (Taiyab 2006). Indeed, Annex I Parties to the Kyoto Protocol have in place systems for monitoring and estimating LULUCF carbon stock changes in order to calculate total GHG balances and, hence, their proximity to their specific Kyoto targets. Despite the provision of guidance by the IPCC for this purpose, countries have the liberty of adopting different methods. In practice, methods range from repeated field measurements (e.g. national forest inventories) to the compilation of historical datasets and databases, thus resulting in various degrees of completeness, accuracy and uncertainty. It is worth noting that emphasis is given to the net balances and not to the timing issues that this paper captures.

1.1 Assessment of LULUCF projects

Biotic carbon sequestration projects cannot be assessed the same way as emission reductions or permanent sequestration projects because they are potentially reversible by disturbances such as wildfires, or by changes in project management that would return the sequestered carbon back to the atmosphere (IPCC 2000; Moura-Costa 2002). This potential reversibility is the reason why LULUCF projects are often controversial. If an offsetting project based on LULUCF that is used to reduce the national greenhouse gas (GHG) inventory releases back a part of the sequestered carbon to the atmosphere, it will ultimately result in higher GHG emissions than a project aiming at actually reducing fossil GHG emissions, and thereby, worsen climate change impacts. For instance, if a one-ton fossil CO2 emission is compensated by the sequestration of one-ton CO2 in trees, and that these trees release their carbon back to the atmosphere in a wildfire, the net CO2 balance will be one-ton CO2, which has more impact on climate than if the initial one-ton emission was simply avoided. Moreover, some studies have shown that sequestering some carbon and releasing it back to the atmosphere after storing it over a given number of years can lead, at that time, to a higher CO2 atmospheric concentration and temperature than if nothing was done (Lashof and Hare 1999; Korhonen et al. 2002; Kirschbaum 2003, 2006). However, temporary sequestration and storage is still worthwhile for mitigating climate change: the amount of CO2 that would have been present in the atmosphere during the sequestration period would have had a cumulative effect on global warming (IPCC 2000). Postponing a GHG emission is thus identical to postponing some radiative forcing effect, which is favourable in the short term as it also allows buying time while technology develops in the field of GHG emission reduction and mitigation (Chomitz 2000; Marland et al. 2001; Noble and Scholes 2001).

To properly assess these types of projects, it is important to consider their temporal aspects (Feng 2005). There is currently no consensus in the scientific community on the approach to use for estimating credits associated with LULUCF projects. The ton-year accounting system (also called Mg-year) is the most widely discussed approach and has been given particular attention by the Intergovernmental Panel on Climate Change (IPCC) in its special report on LULUCF (IPCC 2000), despite the fact that the Kyoto Protocol has apparently precluded any equivalency factor approach. A ton-year system bases the determination of the credits to be given to a temporary sequestration project for carbon mitigation on the amount of carbon stored for each year that the carbon stock is maintained. Different methods, based on different data and assumptions, have been proposed to generate equivalency factors that can be used to compare temporary sequestration scenarios (Chomitz 1998; Dobes et al. 1998; Tipper and de Jong 1998).

Among them, the Moura-Costa and the Lashof approaches are distinguished by their use of Absolute Global Warming Potential (AGWP), a concept widely accepted in the scientific community and used under the Kyoto Protocol for climate change impact assessment (IPCC 2000), to generate equivalency factors. AGWP is the cumulative radiative forcing over a given time horizon of a unit mass of GHG released to the atmosphere at one time, as shown in Eq. 1 (Forster et al. 2007):
$$ AGWP = a \times \int\limits_0^{{TH}} {C(t)dt} $$
(1)
where TH is the time horizon [years], a is the instantaneous radiative forcing per unit mass increase in the atmosphere [Wm-2 kg-1] and C(t) is the time-dependent atmospheric load of the released GHG [kg], given by the Bern carbon cycle-climate model for CO2 (Joos et al. 2001), and by a first-order decay equation for other GHGs (Forster et al. 2007). The well-known Global Warming Potential (GWP) index developed by the IPCC is defined by the AGWP value of the released GHG for a given time horizon divided by the value of AGWP of CO2 for the same time horizon.
For both the Moura-Costa and the Lashof methods, which are illustrated in Fig. 1, the baseline is the impact on radiative forcing over a 100-year time period, which is the reference time frame set for the Kyoto Protocol (UNFCCC 2008), caused by an emission at time zero of 1 t of CO2. Over this period, the atmospheric load curve integral is approximately 48 ton-years (Fig. 1a), using the IPCC’s Fourth Assessment Report data (Forster et al. 2007 p.212).
https://static-content.springer.com/image/art%3A10.1007%2Fs10584-012-0473-x/MediaObjects/10584_2012_473_Fig1_HTML.gif
Fig. 1

Illustration of the two principal ton-year approaches for a 100-year time horizon: the Moura-Costa method (b) and the Lashof method (c), both of which being based on the cumulative radiative forcing of 1 t CO2 (a)

The Moura-Costa method (Moura-Costa and Wilson 2000) uses this value to generate an equivalency factor of 48 years, which means that removing 1 t of CO2 from the atmosphere and storing it for 48 years is equivalent, in a 100-year integration and in terms of avoided radiative forcing, to avoiding a 1 t pulse emission of CO2 (Fig. 1b). The credits are then distributed evenly, meaning that storing 1 t of CO2 for 1 year compensates for a pulse emission of 0.02 ton (1 ton-year divided by 48 ton-years/ton CO2).

Alternatively, the Lashof method (Fearnside et al. 2000) considers that temporary sequestration is equivalent to delaying an emission until the end of the sequestration period, and that the benefit of this sequestration is represented by the difference in the integrals of the two curves within the time horizon of 100 years. The benefit is thus the part of the area under the curve that is pushed beyond the 100-year limit (Fig. 1c). Using the IPCC’s Fourth Assessment Report data (Forster et al. 2007), the sequestration of 1 t of CO2 during 48 years would have a benefit of approximately 19 ton-years. This results from the difference between i) the atmospheric load curve integral, from 0 to 100 years, for an emission occurring at time zero (baseline curve in Fig. 1c) and ii) the atmospheric load curve integral, from 48 to 100 years, of the delayed emission (delayed emission curve in Fig. 1c). Thus, the credit given for the sequestration of 1 t of CO2 during 48 years would be 0.4 t CO2-eq (19 ton-years divided by 48 ton-years/ton CO2-eq), contrary to the 1 t calculated with the Moura-Costa method. Note that under the Lashof method, the credits are not evenly distributed in time and depend on the time by which the emission is delayed, which also means that the delay time must be known prior to its realization in order to quantify the credits. The Moura-Costa method assigns a higher value to temporary storage, as it states that the storage of carbon for 48 years compensates for 100 % of the emission, with the consequence that storing 1 t of carbon over 100 years would compensate for 208 % of that emission, compared to 100 % in the Lashof method.

1.2 LULUCF projects and life cycle assessment

Life cycle assessment (LCA) is a method used to determine the potential environmental impacts of a product or project over its entire life cycle, from raw material extraction to end-of-life (ISO 14040 2006). LCA is increasingly used to compare GHG mitigation scenarios such as fossil fuel replacement with biofuels, carbon capture and storage, etc. (Menichetti and Otto 2009; Reijnders and Huijbregts 2009). This methodology could also be very useful for comparing different LULUCF projects because it includes the contribution of every life cycle stage, such as planting, fertilisation, weeding, pest control, harvesting, infrastructures, transportation, etc. It also typically considers the potential impacts of projects on a comprehensive range of environmental problems, not only climate change (Hauschild 2005).

However, LCA does not currently meet the need for temporal resolution necessary for the assessment of temporary carbon sequestration projects because it typically excludes temporal information from the computation (ISO 14040 2006). First, during the life cycle inventory phase (LCI), the emissions of a given pollutant stemming from all the unit processes dispersed through space and time are simply summed to give one single aggregated emission (Heijungs and Suh 2002), whereby all temporal information is lost. For example, the amount of biogenic carbon that is uptaken by biomass (recorded as a negative emission in LCA), is added to the (positive) amount of biogenic carbon released back to the atmosphere at end-of-life, resulting in a net zero emission. Second, during the life cycle impact assessment phase (LCIA), the potential impacts of the aggregated emissions calculated in the inventory phase are assessed by multiplying them with characterization factors that are based on a fixed time horizon (Udo de Haes et al. 2002). While for most impact categories this time horizon is infinity, for climate change it is usually set to 100 years. What is problematic is not the chosen time horizon per se, but rather that it is fixed, irrespective of the time the GHG emission occurred. This implies that impact scores calculated from emissions inventories occurring over different time scales do not cover the same LCA temporal boundaries (the period covered by the assessed impact), but depend on the product system assessed (Levasseur et al. 2010).

The inclusion in LCA of temporary carbon storage and delayed emissions of GHGs (based on the Moura-Costa and Lashof approaches) has previously been proposed by Nebel and Cowell (2003), Clift and Brandão (2008), the British Standards Institute/Carbon Trust PAS 2050 (BSI 2008), the European Commission’s ILCD Handbook (2010), Müller-Wenk and Brandão (2010), and Courchesne et al. (2010). More recently, Levasseur et al. (2010) propose a dynamic LCA approach to account for time in LCA and first applied it to data from the US EPA lifecycle GHG analysis on renewable fuels (USEPA 2009) to show how dynamic LCA leads to more consistent results for global warming impact assessment when temporal aspects are prevalent, as is the case when considering land-use change emissions in biofuel studies. In a dynamic LCA, emissions are not aggregated in the inventory, and their temporal profiles are considered and specifically assessed with characterization factors, which have flexible time horizons to account for the timing of the emissions in respect to the chosen time horizon of the LCA. The result of this approach is a measure of the potential impact of the studied product or project on radiative forcing at any given time.

The objective of this paper is to apply the dynamic LCA approach to the assessment of a temporary carbon sequestration project by afforestation and to compare the results to those obtained with two alternative approaches: the Moura-Costa and the Lashof methods, which are based on the ton-year approach.

2 Methodology

2.1 Application of dynamic LCA to an afforestation project

A case study is presented in this paper, using data from a paper on afforestation in open woodlands coming from the irreversible conversion of closed-crown black spruce forest following a wildfire. In these open woodlands, very poor regeneration occurs at natural conditions (Gaboury et al. 2009). The aim of that study was i) to estimate the theoretical net carbon balance of a black spruce afforestation project in Québec’s boreal forest and ii) to determine the GHG emissions related to the execution of this project (harvesting, plantation activities, etc.) using life cycle assessment methodology. The CO2FIX model (Groen et al. 2006) was used to predict the carbon dynamics of the natural regeneration and afforestation scenarios. Figure 2 presents the results of the net carbon balance used in this paper, which is the difference in carbon stocks between the afforestation scenarios and natural regeneration. During the first 20 years, the carbon balance is decreasing because the loss from the decomposition of organic matter is higher than the gain from the growth of trees (Gaboury et al. 2009). This is due to the removal of existing stems, prior to plantation, and to the preparation of the soil, which increases the decomposition rate of its organic matter.
https://static-content.springer.com/image/art%3A10.1007%2Fs10584-012-0473-x/MediaObjects/10584_2012_473_Fig2_HTML.gif
Fig. 2

Net carbon balance of an open woodland afforestation project in Québec’s boreal forest over 70 years (afforestation minus natural regeneration) from Gaboury et al. (2009)

The purpose of this case study is to apply dynamic LCA on a hypothetical afforestation project which compensates for a baseline pulse-emission of 1,000 kg CO2, which corresponds, for example, to the emission caused by approximately 5,000 km of continental air travel for one person (ecoinvent database 2010). The surface needed (34.6 m2) for the afforestation project to sequester the total amount of carbon was determined linearly using the final net carbon balance after 70 years (see Eq. 2).
$$ Surface\,\left[ {\,{m^2}} \right] = \frac{{Emission\,to\,compensate\,\left[ {kg\,C{O_2}} \right] \times 0.2727\frac{{kg\,C}}{{kg\,C{O_2}}}}}{{Final\,net\,carbon\,balance\,\left[ {\frac{{kg\,C\,stored}}{{ha}}} \right] \times \frac{{1\;ha}}{{10,000\;{m^2}}}}} $$
(2)
Dynamic inventories were built for the natural regeneration and the different sequestration scenarios presented in Table 1. To do so, the time scale was divided in one-year time steps and the carbon fluxes for each time step were calculated using the carbon balance data for the afforested surface determined previously. A net carbon sequestration is defined as a negative emission, because it reduces the amount of atmospheric CO2, leading to a negative radiative forcing following a symmetric profile compared to a positive emission (Korhonen et al. 2002). The life cycle GHG emissions coming from the afforestation operations (seed production, harvesting operations, plantation, transportation, etc.) have not been added to the dynamic inventory for simplification purposes because they represent only 0.4 % of the radiative forcing of the total amount of carbon sequestered for this study (Gaboury et al. 2009) and they do not change the results significantly.
Table 1

Description of the five scenarios assessed with dynamic LCA

SCENARIO

DESCRIPTION

At time t = 0

At time 0 < t < 70

At time t > =70

Baseline emission

• 1,000 kg CO2 released

• No sequestration activities

• No sequestration activities

Neutral

• Sequestration of 1,000 kg CO2 over 70 years by afforestation of 34.6 m2 of black spruce in Quebec’s boreal forest

• No sequestration nor emission beyond t = 70 years

Fire

• At t = 70 years, all carbon sequestered in biomass is released as CO2 after a wildfire

Exploitation

• At t = 70 years, trees are cut to produce building materials and non-stem materials (branches, leaves, etc.) are burned (carbon released as CO2)

• At t = 145 years, building materials are burned (carbon released as CO2)

Fire multi-gas

• Same as fire scenario except carbon is released as CO2 (99.7 %) and methane (0.3 %)

Landfill

• Same as exploitation scenario except carbon is released as CO2 (99.7 %) and methane (0.3 %) at t = 70 years and building materials are landfilled at t = 145 years

The time-dependent potential impact on global warming GWIinst(t) [Wm-2] was computed for the natural regeneration and the different afforestation scenarios (Table 1) by combining the dynamic inventory g(t) [kg] with the instantaneous dynamic characterization factors for global warming DCF (t) [Wm-2 kg-1] developed by Levasseur et al. (2010), as shown in Eqs. 3 and 4.
$$ DCF(t) = AGWP(t) = \int\limits_{{t - 1}}^t {a \times C(t)dt} $$
(3)
$$ GW{I_{{inst}}}(t) = \sum\limits_{{i = 0}}^t {g(i) \times DCF(t - i)} $$
(4)

The dynamic characterization factors DCF(t) (see Eq. 3) are first calculated for each GHG as the absolute global warming potential at any one-year time step following an emission. They express the radiative forcing occurring between time t-1 and t caused by a pulse-emission at time zero. The total radiative forcing occurring at time t is then calculated by multiplying each GHG emission by its respective dynamic characterization factor DCF(t-i) calculated for the time elapsed between the emission (time i) and time t (see Eq. 4). For example, to determine the impact on radiative forcing at year 100, one would characterize an emission occurring at year 99 with a DCF calculated over a period of 1 year DCF(1), at year 98 with DCF(2), etc. Ultimately, the emission occurring at year 0 would be characterized with DCF(100). Equation 4 applies to all the GHG life cycle emissions. GWIinst(t) is thus the sum of the radiative forcing occurring at time t for all these emissions. Finally, the net impact was calculated by subtracting the impact GWIinst(t) of the natural regeneration scenario from the impact of each afforestation scenario.

Six different scenarios were developed, summarized in Table 1. The “baseline emission scenario” consists in the release of 1,000 kg CO2 into the atmosphere at time zero without any sequestration. The other scenarios consist in the same baseline emission of 1,000 kg CO2, followed by the sequestration of 1,000 kg CO2 over 70 years by afforestation, an attempt to compensate for it. What differentiates the latter scenarios is the fate of the forest at the end of the project. In the “neutral scenario”, nothing happens with the forest and we suppose that no net sequestration, nor emission, will happen after the 70-years target. In the “fire scenario”, we assume that a wildfire destroys the forest the first year following the end of the project. In this case, only the amount of carbon sequestered in biomass (not in the soil) is released back to the atmosphere, which represents 74 % of the total amount of carbon sequestered (Gaboury et al. 2009). In the “exploitation scenario”, all the trees are cut down at the end of the project so that the commercial part of the stem (39 % of total sequestered carbon) is used in building materials with a 75-year lifetime and carbon coming from other wooden parts (branches, roots, sawmill residues, etc.) is released in the first year following the end of the project through burning with no energy recovery. We emphasize that only the carbon sequestered in the soil is assumed to be permanently sequestered and stored. This assumption is justified only if the land remains as forest since any land-use change might release the soil carbon back to the atmosphere (IPCC 2000). For every scenario discussed up to this point in the text, every emission, whether positive or negative, is assumed to be in the form of CO2.

For the “fire multi gas scenario”, the assumptions of the “fire scenario” were adopted, i.e. 74 % of the sequestered carbon is released the year following the end of the project. However, the carbon emissions are now assumed to be CO2 and methane (CH4) in order to investigate the bias induced in the results by assuming all carbon emissions are CO2. Amiro et al. (2009) have studied the GHG emissions coming from Canadian boreal forest fires. They used emission factors from the database developed by Andreae and Merlet (2001) that express the mass of different gases emitted per unit mass of dry fuel. The details of the calculations can be found in the supporting material (Online Resource). Finally, the “landfill scenario” is the same as the “exploitation scenario”, except that the carbon released by the burning of the wooden residues at t = 70 years is divided between CO2 and CH4, as in the “fire multi gas scenario”, and the building materials are landfilled at t = 145 years. Wood in landfills is poorly degraded because of the presence of lignin and around 97 % of the carbon stays sequestered indefinitely (Micales and Skog 1997). The details of this scenario can also be found in the supporting material (Online Resource).

The cumulative impact on global warming GWIcum(t) at a given time is obtained for each scenario by adding the instantaneous impact of the previous time steps (see Eq. 5). By comparing the cumulative impact GWIcum(t) of two different scenarios at a given time, it is possible to determine which one causes the less total radiative forcing over that time period.
$$ GW{I_{{cum}}}(t) = \sum\limits_{{i = 0}}^t {GW{I_{{inst}}}(i)} $$
(5)
Finally, the ratio of the cumulative impact of each sequestration scenario over the cumulative impact of the baseline 1,000 kg CO2 emission to compensate ratio_seq(t) was computed (see Eq. 6).
$$ ratio\_ seq(t) = \frac{{{{\left[ {GW{I_{{cum}}}(t)} \right]}_{{baseline}}} - {{\left[ {GW{I_{{cum}}}(t)} \right]}_{{scenari{o}}}}}}{{{{\left[ {GW{I_{{cum}}}(t)} \right]}_{{baseline}}}}} $$
(6)

This ratio indicates whether the benefits of the sequestration project on radiative forcing compensate for the impact of the baseline emission. For a given time horizon, a ratio of 1 indicates that the sequestration scenario has saved, over this time horizon, a total amount of radiative forcing equal to the total radiative forcing caused by the baseline 1,000 kg CO2 emission. A ratio lower than 1 indicates that the sequestration scenario is not enough to compensate for the radiative forcing caused by the baseline emission over that time horizon, and inversely, a ratio higher than 1 means that the sequestration scenario more than fully compensates for it.

2.2 Comparison of the Moura-Costa and the Lashof methods with traditional and dynamic LCA

The Moura-Costa and the Lashof methods are used to assess the temporary sequestration project for the same afforested surface (34.6 m2), and for a time horizon of 100 years. These approaches are applied in two different ways: i) the first one, called “static”, does not consider the time-dependent carbon balance and assumes that all the carbon is sequestered at time t = 0, and then stored for 70 years, which is the time frame of the project, and ii) the second way, called “dynamic”, considers that the amount of carbon sequestered each year (as shown in Fig. 2) is stored from the year of the sequestration until the end of the 70-year project.

The Moura-Costa (MC) method gives a constant credit of 0.02 t CO2-eq for every year a ton of CO2 is stored. The static (s) result in kg CO2-eq is obtained by multiplying the total net amount of CO2 sequestered by this equivalency factor and by the duration of the sequestration, which is 70 years (see Eq. 7).
$$ MC\;(s) = 1,000\;kg\;C{O_2} \times 0.02\;ton\;C{O_2}eq/ton\;C{O_2} \cdot year \times 70years $$
(7)
To obtain the dynamic (d) result of the Moura-Costa approach, the amount of CO2 sequestered at a given year is multiplied by the credit of 0.02 t CO2-eq and by the number of years from the moment it is sequestered to the end of the project (70 years). This means that, for instance, the amount of carbon sequestered on year 50 will be multiplied by the equivalency factor 0.02 t CO2-eq/yr and by 20 years, the latter being the time it will be stored (from year 50–70). For the years when the carbon balance shows a net emission (from 0 to 20 years), the value of the emission is subtracted from the total credit (see Eq. 8). These emissions, caused by the soil preparation prior to plantation, must be taken into account, and subtracted from the credit, since this part of the credit is used to cancel them, and is no more available to compensate for the baseline emission.
$$ MC\;(d) = \sum\limits_{{t = 21}}^{{70}} {\left[ {C{O_{{2\,sequestered}}}(t) \times (70 - t) \times 0.02} \right]} - \sum\limits_{{t = 1}}^{{20}} {C{O_{{2\,released}}}(t)} $$
(8)
In the Lashof (L) approach, the credit given for temporary carbon sequestration and storage is not constant but rather depends on the number of years carbon will be stored. The static result, which assumes that the total net amount of carbon is sequester at time t = 0, is obtained by multiplying the total amount of CO2 sequestered by the part of the area under the load curve which is pushed beyond the 100-year time horizon. Since the length of the storage period is 70 years, the benefits are given by the area under the baseline curve (see Fig. 1a) going from year 30–100 (AREA30 to 100). This value is then divided by 48 ton-years/ton CO2-eq (AREA0 to 100), which is the value in ton-years of a pulse emission at time t = 0, to get the value of the credit. Equation 9 shows this calculation.
$$ L\;(s) = \frac{{1,000\,kg\,C{O_2} \times ARE{A_{{30\,to\,100}}}\,ton \cdot year/ton\,C{O_2}}}{{ARE{A_{{0\,to\,100}}}\;ton \cdot year/ton\,C{O_2}eq}} $$
(9)
Equation 10 gives the formula used to get the dynamic result for the Lashof method. The benefit in ton-years is determined for every amount of CO2 sequestered according to the number of years of storage, and the total benefit is then divided by 48 ton-years/ton CO2-eq to get the value of the credit. For instance, the amount of carbon sequestered on year 50 is stored from year 50 to year 70, i.e. over a time period of 20 years. Since storing carbon for 20 years is equivalent to delaying an emission for 20 years, the benefit is given by the area under the baseline curve from year 80–100. Just as for the Moura-Costa method, for the years when the carbon balance shows a net positive emission, the value of the emission is subtracted from the total credit.
$$ L\;(d) = \sum\limits_{{t = 21}}^{{70}} {\frac{{\left[ {ARE{A_{{(t + 30)\,to\,100}}} \times C{O_{{2\,sequestered}}}(t)} \right]}}{{ARE{A_{{0\,to\,100}}}}} - \sum\limits_{{t = 1}}^{{20}} {C{O_{{2\,released}}}(t)} } $$
(10)
Finally, for comparison purposes, a traditional static LCA and a dynamic LCA are performed on the sequestration project. The application of the Moura-Costa and the Lashof approaches implies that the carbon is released back to the atmosphere at the end of the storage period. To be consistent with this assumption, the static and dynamic LCA results are calculated assuming that the total amount of CO2 sequestered (1,000 kg) is released back to the atmosphere after 70 years. The traditional static LCA result LCA(s) is simply given by the sum of every CO2 flux, multiplied by the characterization factor, which is the GWP value for a 100-year time horizon (see Eq. 11).
$$ LCA\,(s) = \sum\limits_{{t = 1}}^{{70}} {\left( {C{O_{{2\;released}}}(t) - C{O_{{2\;sequestered}}}(t)} \right)} \times GW{P_{{100}}} $$
(11)
The dynamic LCA result LCA(d) is obtained by dividing the cumulative impact on global warming GWIcum for a 100-year time horizon (obtained with Eq. 5) by the cumulative radiative forcing of a 1 kg CO2 pulse emission occurring at t = 0 (obtained with Eq. 1) to get a value in CO2-eq (see Eq. 12).
$$ LCA\,(d) = \frac{{GW{I_{{cum}}}\,(100)}}{{AGWP\,(100)}} $$
(12)

3 Results and discussion

3.1 Application of dynamic LCA to an afforestation project

Figure 3 presents the results obtained with dynamic LCA for the assessment of the afforestation project i.e. the instantaneous (a) and cumulative (b) impact on global warming as computed with Eqs. 4 and 5, respectively, and the ratio (c) of the cumulative impact of the sequestration project over the cumulative impact of the baseline emission as computed with Eq. 6.
https://static-content.springer.com/image/art%3A10.1007%2Fs10584-012-0473-x/MediaObjects/10584_2012_473_Fig3_HTML.gif
Fig. 3

(a) Instantaneous and (b) cumulative impact on global warming obtained with dynamic LCA for a baseline emission of 1,000 kg CO2 and five different sequestration scenarios of afforestation in a boreal forest and (c) ratio of the cumulative impact of the baseline emission of 1,000 kg CO2 compensated by the afforestation project

The curves for the five sequestration scenarios are overlaying each other for the first 70 years because the difference between the scenarios resides in what is happening with the project at the end of the sequestration period. Instantaneous dynamic LCA results (Fig. 3a) show the time-dependent potential impact on global warming, for a baseline emission scenario and five different mitigation scenarios (see Table 1) starting from t = 0, time when the emission of 1,000 kg CO2 occurs. The net radiative forcing for every sequestration scenario is positive for the first 50 years, and is even greater than the radiative forcing of the baseline emission itself for the first 25 years, because the net carbon balance of the afforestation project is negative in this period (Fig. 2), meaning that a net CO2 emission is added to the baseline 1,000 kg CO2, increasing the total impact on global warming during that period of time. This shows that the assumptions regarding carbon exchanges between the forest and the atmosphere have a significant impact on the results.

The curve of the cumulative impact (Fig. 3b) for the neutral scenario tends to zero with time as the total amount of carbon stored in the biomass, which is the same amount of carbon released at time zero, is assumed to stay sequestered indefinitely. This is also observed in Fig. 3c, where the ratio of the impact of the baseline emission compensated by the afforestation project of the neutral scenario tends actually to 1 for an infinite time horizon. This means that a traditional LCA of the sequestration project with the assumptions of the neutral scenario gives the same result as a dynamic LCA integrated to infinity. In this case, if one were to assume an infinite time horizon, the sequestration project would fully compensate the total effect of the baseline GHG emission. But, as shown in Fig. 3c, the shorter the time horizon, the lesser is the compensation. A shorter time horizon increases the relative importance of what is happening earlier, which is a positive radiative forcing for the first 50 years. The benefits of the sequestration come later and, for short time horizons, do not have the time to substantially compensate for the baseline emission. Figure 3 shows the sensitivity of the results to the choice of a time horizon. For instance, the fire scenario reaches a compensation ratio of 0.23 for a 100-year time horizon, and of 0.04 for a 500-year time horizon. Under very long (a few centuries) or infinite time frames, there may be insignificant or no benefits at all to temporary carbon storage.

The plausibility of the assumptions made for the neutral scenario is questionable. In reality, black spruces in boreal forests reach maturity around 70 years following their plantation (MRNFP 2003), but they still sequester or release CO2 after that time. This means that the carbon balance following the end of the project is probably not neutral. More research would be needed to quantify this additional amount of carbon sequestered or released in order to modify the neutral scenario, which intends to reflect the most optimistic conditions. On the other hand, wildfires are very common in boreal forests and their probabilities must be considered while determining the more realistic mean age of a typical mature tree (Bergeron et al. 2006).

The fire and exploitation scenarios draw the attention on an important and controversial aspect of sequestration through forestry: the management of the project after its end. The assumption made for the neutral scenario, in which the total amount of carbon sequestered during the project stays indefinitely in the biomass, is not realistic. Practical projects have a finite duration and the responsibility of the manager for the carbon pool has an end (Korhonen et al. 2002). Forests are also vulnerable to natural disturbances such as wildfires and pests that would cause at least a partial release of the sequestered carbon back to the atmosphere (García-Oliva and Masera 2004). Of course, the results presented in Fig. 3 are very much affected by the assumption that no forest is replanted after the fire or the exploitation, since no carbon flux is considered following these events. Therefore, the general conclusion that the ability of an afforestation project to compensate for a fossil-based GHG emission is very sensitive to what happens at the end of the project still stands.

The effect of the carbon released back to the atmosphere at different times after the end of the project, different across scenarios, can be seen in the instantaneous radiative forcing results shown in Fig. 3a. The results for cumulative radiative forcing shown in Fig. 3b do not tend to zero for these scenarios as there is more carbon emitted to the atmosphere than sequestered as a final balance. This effect can also be observed from the ratios of the cumulative impacts over time, shown in Fig. 3c. Increasing the time horizon beyond the sequestration time decreases the ratio of the impact of the baseline emission recovered by the afforestation project.

We emphasise that the assumptions made for the fire and the exploitation scenarios could be refined. Growing trees sequester carbon by absorbing carbon dioxide present in the atmosphere during photosynthesis. However, the carbon released back to the atmosphere during a fire, although mostly in the form of CO2, is also partly released as methane (Amiro et al. 2009), which has a GWP100 25 times higher than that of CO2 (Forster et al. 2007). The fire multi gas scenario considers both gases. Even if the amount of CO2 released by the fire is 334 times higher than the amount of methane (on a mass basis), the fraction of the impact of the baseline emission that is compensated by the fire multi-gas scenario is significantly different than that by the simplified fire scenario. For a time horizon of 100 years, for example, the compensation for the fire multi-gas scenario is 16 % of the impact of the baseline emission, while the compensation is 23 % for the CO2-only fire scenario. The dynamic LCA approach can determine the impact on radiative forcing of any given GHG if inventory data are available, which could be applied to any scenario. The landfill scenario also raises the important question of permanent sequestration. As most of the wood contained in the building materials landfilled at their end-of-life will not be degraded, a part of the carbon can be considered as permanently stored. The dynamic LCA allows assessing it as shown in Fig. 3c, where it is shown that the ratio of the baseline emission compensated by the landfill scenario is much higher than that of the scenarios which are releasing the total amount of carbon back to the atmosphere.

3.2 Comparison of the Moura-Costa and the Lashof methods with traditional and dynamic LCA

Table 2 shows the results obtained by applying the Moura-Costa (Eqs. 7 and 8), the Lashof (Eqs. 9 and 10) and the LCA (Eqs. 11 and 12) methods, in both the so-called “static” and “dynamic” interpretations, with a time horizon of 100 years. The calculated credit (in kg CO2-eq) is presented on the first line and expresses the ability of the project to compensate for a baseline GHG emission. On the second line, the credit is transformed into a ratio of the baseline emission by dividing it by 1,000 kg CO2-eq. A ratio below 1 indicates that the afforestation of 34.6 m2, as assessed with the respective methods, is not enough to compensate for the baseline emission, and a ratio over 1 indicates the opposite. On the third line, the surface needed to obtain a ratio of 1, according to each approach, is also computed. For example, if the initial afforestation project, which covers 34.6 m2, compensates for 0.5 of the baseline emission, the surface needed to obtain the full compensation is twice the initial surface, i.e. 69.2 m2. Finally, the last line of the table gives the fraction of the initial surface that this new surface represents. Using the last example, the fraction of the initial surface would be 2.0.
Table 2

Static and dynamic results of the Moura-Costa, the Lashof and the LCA methods applied to an afforestation project in boreal forest for a time horizon of 100 years

 

Moura-Costa

Lashof

LCA

 

Static

Dynamic

Static

Dynamic

Static

Dynamic

Calculated credit (kg CO2-eq)

1,464

500

605

182

0

226

Ratio of the baseline emission compensated by the project

1.46

0.50

0.61

0.18

0

0.23

Afforested surface needed to compensate the baseline emission (m2)

24

69

57

190

N/A

153

Fraction of the initial 34.6 m2 surface

0.68

2.00

1.65

5.48

N/A

4.4

The high variability of the results presented in Table 2 shows that the ability of a temporary sequestration and storage project to compensate for a certain GHG emission highly depends on the chosen assessment method, and on the related assumptions. Because they assume that all the carbon is sequestered at time x = 0, which is an overestimation of the real storage time, the “static” interpretations of the Moura-Costa and the Lashof methods obtain higher values for the calculated credit than their dynamic equivalent. The credit calculated for the static LCA method (0 kg CO2-eq) shows that current LCA does not allow giving a value to temporary carbon storage. Since the total amount of carbon sequestered is released back to the atmosphere at the end of the project, the net emission is zero, corresponding to a zero impact. As explained in the introduction, the Moura-Costa method gives a much higher value than the Lashof and LCA approaches, since it considers that the storage of carbon for 48 years compensates for 100 % of the equivalent emission, compared to 100 years for the Lashof approach. The static Moura-Costa result implies that 1,000 kg CO2 is stored for 70 years, which is longer than this 48 years. This is why the calculated credit is higher than the baseline emission.

The same scientific basis is used for the Moura-Costa, the Lashof, and the dynamic LCA approaches, which is the cumulative radiative forcing concept (AGWP) developed by the IPCC, but the way it is used to give value to temporary carbon storage differs. The dynamic LCA approach goes a step further by giving the value of the impact on radiative forcing caused by the project at any given time, for any temporal profile and for every GHG. Both Moura-Costa and Lashof methods are developed for CO2 only; they do not allow for the consideration of other GHGs, nor of other activities associated with the project such as the different scenarios presented in this paper. Dynamic LCA is a more flexible and comprehensive approach as it allows considering all the GHG fluxes emitted or sequestered during the different steps of the project life cycle, which gives an insight of the impact of the whole project on global warming. The dynamic LCA approach has the advantage to be consistently applicable to other impact categories by developing further sets of dynamic characterization factors (Levasseur et al. 2010), in the case where the objective would be to expand the scope of the analysis of the LULUCF project to environmental problems other than global warming.

Decision makers often need to make value judgements on the relative importance of short- and long-term impacts (Hellweg et al. 2003). Giving permanent credits for temporary sequestration activities is made possible only by the choice of a time horizon beyond which impacts are not considered (100 years in the present case); a larger importance is implicitly given to what happens in the short term, and the long term is overlooked. The choice of time horizons rely more on political decisions than on scientific approaches (Fearnside 2002; Shine 2009). Both Moura-Costa and Lashof methods use an implicitly fixed time horizon of 100 years to determine the carbon credits of specific projects. This is unfortunately not always clear for decision makers. Dynamic LCA allows the use of flexible time horizons, which can be determined at anytime in the process, since the results are detailed in time. This flexibility supports the performance of sensitivity analyses, helping decision makers to understand how the arbitrary choice of time horizon affects the results of the analysis. Dynamic LCA also offers a consistent way to look at the impact of different events occurring after the end of the afforestation project such as wildfires or different uses of the wood, which cannot be done with ton-year approaches.

The dynamic characterization factors developed for global warming have the same principal limitation as any method based on IPCC’s data for radiative forcing, which is the use of an instantaneous radiative forcing value and an atmospheric GHG load expression that do not vary with CO2 atmospheric concentration. In the Fourth Assessment Report (Forster et al. 2007), these variables are calculated for an atmospheric concentration of 378 ppm CO2. Korhonen et al. (2002) have shown that the use of a more complex model for carbon dynamics, where the variation of CO2 atmospheric concentration is considered, can change the conclusions regarding the benefits of a sequestration project. The integration of a dynamic carbon model into the dynamic LCA method could therefore be an interesting approach. However, it should be noted that modeling with a variable CO2 concentration implies the use of future emission scenarios (such the ones developed by the IPCC) which are highly uncertain.

Another limitation of dynamic LCA and ton-year approaches is the lack of consideration for the impacts of afforestation or deforestation activities on albedo effect. Indeed, some researchers have shown that the plantation of conifers in regions covered with snow or the deforestation activities in some particular geographical zones can have the opposite impact on global warming than the objective of the project by a change in surface albedo (Betts 2000; Bala et al. 2007). Schwaiger and Bird (2010) propose an approach to incorporate albedo changes in carbon accounting and suggest that LCA studies done on land use change projects include these non-GHG effects on climate in the future. Munoz et al. (2010) present a methodology to include the impact of land surface albedo changes as CO2-eq in LCA studies. In this paper, we made the choice to restrain the scope of our research to the impact pathway linked to GHG emissions with dynamic LCA. However, combining this with approaches for assessing the albedo effects would certainly improve the relevance of assessing LULUCF projects and therefore provide more robust recommendations

4 Conclusions

Giving permanent credits for temporary carbon sequestration and storage projects is still a hotly-debated issue. This paper compared the dynamic LCA approach (Levasseur et al. 2010) with two well-known methods proposed to determine the credits due to LULUCF projects: the Moura-Costa and the Lashof approaches. We showed that the dynamic LCA approach has the following key advantages: (1) it is a flexible approach as it consistently evaluates the impact on radiative forcing occurring at any given time caused by every GHG flux; (2) it allows the consideration of every life cycle step of the project, such as harvesting, plantation, transport and also different end-of-life scenarios; and (3) it gives decision makers the opportunity to test the sensitivity of the choice of different time horizons.

This analysis shows that the results from the assessment of LULUCF projects vary significantly across the alternative approaches. Therefore, it is important that the method used be as transparent as possible so that the sensitivity to the different assumptions and temporal choices can be determined and made explicit. Dynamic LCA is capable of fulfilling this need, but further research is necessary to improve the method so that it better represents the impact of temporary sequestration projects. Also, as discussed in this paper, practitioners using dynamic LCA for the assessment of temporary sequestration projects must be careful while making assumptions regarding the carbon flows between the forest and the atmosphere while trees are growing and what is happening following the end of the sequestration project.

Acknowledgements

The authors acknowledge Jean-François Boucher, Simon Gaboury, Achille-Benjamin Laurent, Claude Villeneuve and Jean-Robert Wells form Université du Québec à Chicoutimi for their advice and the industrial partners of the International Chair in Life Cycle Assessment (a research unit of CIRAIG) for their financial support: Alcan, Arcelor-Mittal, Bell Canada, Cascades, Eco-Entreprises-Québec/Recyc-Québec, Groupe EDF/GDF-SUEZ, Hydro-Québec, Johnson&Johnson, Mouvement des caisses Desjardins, RONA, Total and Veolia Environnement.

Supplementary material

10584_2012_473_MOESM1_ESM.pdf (34 kb)
ESM 1(PDF 34 kb)

Copyright information

© Springer Science+Business Media B.V. 2012