Journal of Environmental Management 146 (2014) 346e354

Contents lists available at ScienceDirect

Journal of Environmental Management journal homepage: www.elsevier.com/locate/jenvman

Biogenic CO2 fluxes, changes in surface albedo and biodiversity impacts from establishment of a miscanthus plantation Susanne V. Jørgensen a, b, *, Francesco Cherubini c, Ottar Michelsen c a Technical University of Denmark (DTU), Department for Management Engineering, Division for Quantitative Sustainability Assessment, 2800 Kgs. Lyngby, Denmark b Novozymes A/S, Krogshøjvej 36, 2880 Bagsværd, Denmark c Norwegian University of Science and Technology (NTNU), The Industrial Ecology Programme, Department of Energy and Process Engineering, NO-7491 Trondheim, Norway

a r t i c l e i n f o

a b s t r a c t

Article history: Received 22 January 2014 Received in revised form 19 May 2014 Accepted 3 June 2014 Available online 6 September 2014

Depletion in oil resources and environmental concern related to the use of fossil fuels has increased the interest in using second generation biomass as alternative feedstock for fuels and materials. However, the land use and land use change for producing second generation (2G) biomass impacts the environment in various ways, of which not all are usually considered in life cycle assessment. This study assesses the biogenic CO2 fluxes, surface albedo changes and biodiversity impacts for 100 years after changing land use from forest or fallow land to miscanthus plantation in Wisconsin, US. Climate change impacts are addressed in terms of effective forcing, a mid-point indicator which can be used to compare impacts from biogenic CO2 fluxes and albedo changes. Biodiversity impacts are assessed through elaboration on two different existing approaches, to express the change in biodiversity impact from one human influenced state to another. Concerning the impacts from biogenic CO2 fluxes, in the case of conversion from a forest to a miscanthus plantation (case A) there is a contribution to global warming, whereas when a fallow land is converted (case B), there is a climate cooling. When the effects from albedo changes are included, both scenarios show a net cooling impact, which is more pronounced in case B. Both cases reduce biodiversity in the area where the miscanthus plantation is established, though most in case A. The results illustrate the relevance of these issues when considering environmental impacts of land use and land use change. The apparent trade-offs in terms of environmental impacts further highlight the importance of including these aspects in LCA of land use and land use changes, in order to enable informed decision making. © 2014 Elsevier Ltd. All rights reserved.

Keywords: Land use change (LUC) impacts Climate change Effective forcing 2G biomass feedstock Life cycle assessment (LCA)

1. Introduction 1.1. Background The use of biomass as feedstock for energy and materials is one of the most promising options for replacing fossil fuels, and production of biofuels and biomaterials at various scales and with varying technological processes is increasing worldwide (Chum et al., 2011). At the moment most biofuels are first generation (1G) which has been criticized for competing with food production

* Corresponding author. Technical University of Denmark (DTU), Department for Management Engineering, Division for Quantitative Sustainability Assessment, 2800 Kgs. Lyngby, Denmark. Tel.: þ45 45254451. E-mail address: [email protected] (S.V. Jørgensen). http://dx.doi.org/10.1016/j.jenvman.2014.06.033 0301-4797/© 2014 Elsevier Ltd. All rights reserved.

and takes up large amounts of land (Cherubini et al., 2009; Melillo et al., 2009). Another option, which is currently evolving, is the use of cellulose based second generation (2G) biomass (Lange, 2007; Brown and Brown, 2013). 2G biomass feedstock, which can e.g. be agricultural residues or dedicated energy crops with high yields and thus land savings, is expected to overcome some of the challenges with the 1G biomass feedstock (Dohleman et al., 2010; Cherubini and Jungmeier, 2010). However, environmental concerns have also been raised for the 2G feedstock, especially regarding the effects of biomass harvest on soil organic carbon (SOC), either from direct or indirect land use change (LUC) (Anderson-Teixeira et al., 2009; Don et al., 2012). Life cycle assessment (LCA) is a commonly used tool for assessing environmental impacts of products and systems during their lifecycle and a number of LCAs exist on biobased fuels and products (see e.g. Cherubini et al., 2009; van der Voet et al., 2010;

S.V. Jørgensen et al. / Journal of Environmental Management 146 (2014) 346e354

Cherubini and Strømman, 2011; Weiss et al., 2012). However, many effects of land use and LUC are not usually implemented in LCAs on a routine basis, such as impacts from changes in surface albedo, biodiversity and aspects related to the skewed time distribution of biogenic carbon fluxes (e.g. Michelsen, 2008; Cherubini et al., 2012; Don et al., 2012; de Baan et al., 2013a; Koellner et al., 2013). Studies investigating the impacts of human induced land disturbances for biomass acquisition show that biogenic CO2 fluxes, including changes in SOC, are site-specific and depend on conditions like local climate, biomass type, previous land use and crop management (Cowie et al., 2006; Anderson-Teixeira et al., 2009; Don et al., 2012; Djomo et al., 2011; Cherubini et al., 2012). SOC fluxes from LUC can be both positive and negative, depending on such conditions (Cowie et al., 2006; Anderson-Teixeira et al., 2009). Beyond biogenic CO2, the so-called biogeophysical effects (i.e. changes in surface albedo, evapotranspiration, surface roughness) are increasingly included in climate impact assessments of bioenergy systems, as they are important contributors to both local and global climate, either directly or indirectly (Georgescu et al., 2009, 2011; Loarie et al., 2011; Anderson-Teixeira et al., 2012; Bright et al., 2012a,b; Cherubini et al., 2012). Albedo describes the reflectivity of a surface in terms of the fraction of the incoming radiation which is reflected. Albedo values change with the type of land cover, e.g. snow covered surfaces have a large albedo as opposed to darker surfaces such as forest cover (Bright et al., 2012a; Cherubini et al., 2012). Thus changes in albedo due to land cover change can be of major importance in relation to global warming. On the global scale, the albedo effect is found to be the dominant direct biogeophysical climate forcing, particularly in areas affected by seasonal snow cover (Claussen et al., 2001; Randerson et al., 2006; Bala et al., 2007). In some cases, the cooling contribution from albedo following (temporary) deforestation can more than offset the warming induced by the release of CO2 to the atmosphere (Randerson et al., 2006; Bala et al., 2007; Betts, 2007; Davin, 2007; Cherubini et al., 2012; O'Halloran et al., 2012). Impact on biodiversity is another recognized environmental aspect of biobased feedstock production. Changed land use is expected to be the main driver of biodiversity changes in terrestrial ecosystems due habitat loss and (local) species extinction (Sala et al., 2000). The impacts of a changed land use on biodiversity depend largely on the prior land use and may both be positive and negative (Campbell and Doswald, 2009). Those aspects makes it essential to assess the impact on biodiversity in studies including a change in land use, but despite this, there is still no consensus on how to do this in LCA (Michelsen et al., 2012; de Baan et al., 2013a; Koellner et al., 2013).

347

The energy crop miscanthus has been chosen as feedstock in this study as it is one of the most prosperous 2G biomass feedstocks due to environmental benefits, high biomass yields and low input need (Mishra et al., 2012; Anderson-Teixeira et al., 2009). Further, it is an interesting energy crop as it is a perennial crop which can be annually harvested, making it different than e.g. forests as biomass feedstock. This study addresses the hypothetic case of a miscanthus plantation established in Wisconsin US, with the prior land use being either forest or fallow land. Wisconsin has been chosen for this study as it is part of the US Midwest, also called the ‘bread basket’ of the US, where most of the national agricultural production takes place. Further, this site is representative of an area with average conditions for miscanthus cultivation and SOC properties within the US (Mishra et al., 2012). 2. Material and methods Fig. 1 shows the changes in land use analyzed in this paper. The former forest land type used in this study is deciduous forest on sandy loam soil, with a forest age of 60e80 years. For the case of former fallow land, it is assumed to have been fallow for a period of 15 years. Miscanthus is a perennial crop, and the new land use introduced with the establishment of the miscanthus plantation is continued. More details are provided for the specific methodological issues in the following sections. For the aspects addressed, we specify when proxies are used due to lack of data. Climate forcings from both biogenic CO2 fluxes and albedo changes can be measured in terms of radiative forcing (RF) (Betts, 2011; Bright et al., 2012a; Cherubini et al., 2012), which is the perturbation of the earth energy balance at the top of the atmosphere by a climate change mechanism (Knutti and Hegerl, 2008). However, the climate response, specifically the global mean temperature change, is sensitive to the characteristics of the forcing other than its magnitude measured in watts per square meter. Forcings from albedo are found to be from approximately 1.5 up to 5 times more effective than that from CO2 in altering global surface temperature (Hansen et al., 2005; Flanner et al., 2007; Bellouin and Boucher, 2010). As the RF does not take the variation in climate sensitivity to different forcings into account (Christiansen, 1999; Hansen and Nazarenko, 2004) we use the so-called effective forcing to include these. The effective forcing is the product between the instantaneous radiative forcing and the climate efficacy of the forcing. The climate efficacy of a forcing agent is defined as the global mean surface temperature change per unit of the forcing relative to that produced by a standard CO2 forcing from the same climate state (see Hansen et al., 2005 for detailed info).

1.2. Aim and scope of the study The aim of this paper is to assess the potential climate change impacts (in terms of change in effective forcing) from above and below ground biogenic carbon dynamics and changes in surface albedo, as well as biodiversity impacts, from establishing a miscanthus plantation on a former forest or a fallow land. In this way, the difference in impact when making the LUC on different former land types in the same geographic area is addressed. The impact is assessed over a time period of 100 years from establishment of the plantation and potential environmental tradeoffs between the various impacts are investigated. In terms of biodiversity impact, this paper aims at reducing uncertainty by comparing results of two of the most advanced current assessment methods, while also suggesting an elaboration to reflect the actual change in biodiversity impacts following a LUC from one human influenced land use state to another.

Fig. 1. Flow diagram of the addressed change in land use. Case A: From forest to miscanthus plantation. Case B: From fallow land to miscanthus plantation.

348

S.V. Jørgensen et al. / Journal of Environmental Management 146 (2014) 346e354

2.1. Climate impacts of biogenic CO2 fluxes In this analysis, we account for biogenic CO2 fluxes from the field to and from the atmosphere, originating from changes in stocks of above ground biomass (AGB) and belowground biomass (BGB) carbon as well as SOC. The input data and estimation of the biogenic carbon flux from the LUC in case A and B are given in the Supplementary material, along with assumptions. The resulting biogenic carbon fluxes are shown in Fig. 2. All sequestered carbon comes from atmospheric CO2, and it is assumed that all released biogenic carbon is likewise released to the atmosphere as CO2. The change in atmospheric CO2 concentration f(t) given by the biogenic carbon flux profiles e(t) for case A and case B can be computed through a mathematical convolution with the Impulse Response Function (IRF) for CO2 y(t) (see Cherubini et al., 2011 for more details).


 f t ¼

Zt

eðt 0 Þyðt  t 0 Þdt 0

(1)

0

This is the basis for the estimation of the resulting impact on global warming through the concept of RF, which is the product of f(t) and the radiative efficiency per kg CO2 emission. Since the climate efficacy of CO2 is taken as reference and is thus equal to 1 the radiative forcing in this case coincides with the effective forcing. For the aim here, we calculate the change in effective forcing on an annual basis. Rather than solving the integral analytically, we therefore use an approximate discrete solution to Equation (1) with one year time steps. 2.2. Climate impacts from changes in surface albedo Albedo values of the land cover types studied in this work are taken from MODIS satellite observations. MODIS black-sky shortwave broadband (near infrared þ visible) albedo data (Collection 5, MCD43A) are obtained from the MODIS subset data server for all sites with a spatial resolution of approximately 0.25 km2 (ORNL DAAC, 2012). Data are gathered for a forest land, an open shrub land with some extend of small vegetation cover (to represent a fallow land), and a crop land (taken as proxy for a land dedicated to

miscanthus production with annual harvest before winter). All the selected sites are within a radius of 10 km to ensure consistent climate and solar radiation. Rather than using data for one specific year, the average monthly albedo is used in order to obtain generally applicable results for the area, and avoid potentially more extreme cases associated with annual variability in phenology and climate. The monthly average is calculated for an 11 year period (February 2000eOctober 2012) for which the data were available. The history of the area is tracked using Google Earth Historical imagery (Google Earth, 2012), in order to ensure a constant land cover over the assessment period. The geographical coordinates for the different land use types used are (latitude, longitude): Forest (45.202, 90.158), fallow land (45.198, 90.088) and crop land (45.169, 90.284). The change in albedo, based on an 11-year monthly mean for those areas is shown in Fig. 3. The mean albedo clearly differs over the months for each land use type, with highest albedo in the winter due to snow cover. Further, the crop land shows the highest variation over the year. This is due to bare cropland after harvest being available for snow cover in the winter. On the contrary, both forest and fallow land has a certain level of vegetation throughout the year, which results in a snow masking effect that makes albedo decrease. The effective forcing from a change in albedo (EFalbedo) is computed following the methodology described in Cherubini et al. (2012). 2.3. Biodiversity Biodiversity is a term describing the variability between and within living species and ecosystems (UNEP, 1992) and the loss of biodiversity has been identified as a key environmental concern (e.g. Diaz and Cabido, 2001). Suggested methods on how to assess quality and quantity changes in terms of biodiversity can be separated between methods that focus on species composition and changes of this, and methods focusing on structural changes, key factors, changes in habitats etc. For all methods, lack of globally available data is a severe problem (de Baan et al., 2013a). The two only methods found sufficiently useful due to necessary spatial resolution and geographical validity, is a method presented by de Baan et al.

Fig. 2. Biogenic carbon flux profiles when converting a) forest and b) fallow land to miscanthus production for the given scenarios. The sum of AGB and BGB þ SOC change gives the total biogenic C change for each year. Positive values indicate net emissions to the atmosphere, while negative values show removal fluxes.

S.V. Jørgensen et al. / Journal of Environmental Management 146 (2014) 346e354

349

Fig. 3. Monthly mean albedo values for the different land uses over the year in the specified geographical area.

(2013a) focusing on relative changes in species diversity and a method presented by Michelsen (2008) and made globally applicable by Coelho and Michelsen (2014) focusing on rareness and structures necessary for biodiversity. Both methods are on a preliminary level and include a great amount of limitation and uncertainty. In order to reduce the uncertainty, both methods are here applied in an attempt to compare results and determine whether they will support the same conclusions. Both methods originally consider biodiversity impacts of a land use compared to the natural state of that area before human interference. The rationale for this is that any land use is regarded a postponement of the relaxation to the natural state (de Baan et al., 2013a). This entails that the use of the land before the change to a new land use is not considered. As the goal here is to assess the difference in biodiversity impact due to the shift from one human influenced land use to another, an elaboration of both methods to reflect this is therefore suggested. Further, as both methods assess biodiversity impacts of a land use compared to the natural state as reference state, they do not include the impacts during the time of transition between the states. Normally, this impact will be rather small compared to the continued land use (that is occupation) impacts, at least if including a high number of rotations (Michelsen et al., 2012) and currently methods for assessing transformation impacts are on an even more preliminary state than methods for occupation (Michelsen et al., 2014). Here, we assess occupational impacts from the time of the LUC. That is no time lag in terms of impact on biodiversity is considered for the shift from one state to another. While this is of course not the case, it is considered the best option for this study, with the current lack of options for more accurate modelling of the shift from one state to another. de Baan et al. (2013b) have also developed a method based on absolute species extinctions which could be a third alternative, but characterization factors for all but one of the relevant land use categories in the assessed areas are given as zero, partly due to limited amount of data, so this method is at present not relevant for comparison here.

2.3.1. The Michelsen method with modifications from Coelho and Michelsen and suggested elaboration The method presented by Michelsen (2008) assesses biodiversity quality as:

Q ¼ ES$EV$CMB

(2)

where Q is the ecosystem quality in terms of biodiversity, ES is ecosystem scarcity, EV is ecosystem vulnerability and CMB is conditions for maintained biodiversity.

The factors ES and EV describe the present condition of an ecosystem in terms of how rare and vulnerable it is, respectively. Thus the product of ES and EV gives the present intrinsic value of an ecosystem. ES and EV are derived on the level of ecoregions based on input data from WWF (2012) as described in Michelsen (2008) where more details on the method can be found. As two ecoregions are located very close within the area used for the assessment, both ecoregions are included in the assessment. The first ecoregion is NA0415, which is described as Nearctic; temperate broadleaf and mixed forests; Upper Midwest forest-savanna transition (WWF, 2012). It has a size of 166,017 km2 and a conservation status as ‘critical/endangered’ (WWF, 2012). The second ecoregion is NA0416, which is described as Nearctic; temperate broadleaf and mixed forests; Western Great Lakes forests, with a size of 274,019 km2 and a conservation status as ’relatively stable/intact’ (WWF, 2012). For the factor CMB, we here determine it based on the suggestion in Coelho and Michelsen (2014), who introduces a factor for ‘degree of naturalness’ for estimating the CMB, through the relationship given in Equation (3):

CMB ¼ 0 degree of naturalness0 ¼ 1  NDP

(3)

where NDP is the naturalness degradation potential, with values from Brentrup et al. (2002) as shown in Table 1. The NDP values represent a coarse choice of aggregation of different types of ecosystems into groups, defined by their degree of influence by humans, and separated with 10% human influence steps between groups. Further, far from all land use types are directly described in Brentrup et al. (2002), so the description considered the best fit is used. Both aspects introduce uncertainty. The change in biodiversity quality of a land use compared to the natural state, DQ, is calculated as the difference between the quality of the original land before human interference, Qrel, which has an inherent CMB ¼ 1, and Q1 which is the change in biodiversity quality of the land use compared to the natural state. The change is thus always positive, and the higher it is, the higher the loss of

Table 1 NDP values from Brentrup et al. (2002) considered fitting for the different land use types of this study. Land use

NDP

Description

60e80 year old forest Fallow land

0.2 0.3

Miscanthus

0.5

Small to moderate human influence, e.g. extensively managed forests Moderate human influence, e.g. permanent fallow land, fallow pasture Strong human influence, e.g. ruderal vegetation of perennials, permanent grassland (pasture or meadow) managed with medium intensity

350

S.V. Jørgensen et al. / Journal of Environmental Management 146 (2014) 346e354

biodiversity. Finally, the impact of the area and time of the land use is included simply by timing DQ with area and time. However, as the goal here is to assess the difference in biodiversity impact following a change from one human influenced state to another, not compare to the natural state, an elaboration of the method, to simply compare DQ of the unnatural land use before and after the LUC is suggested:

DQLU2 ¼ DQ1  DQ2 ¼ ðES$EV $CMB1  ES$EV$1Þ  ðES$EV $CMB2  ES$EV$1Þ ¼ ES$EV$ðCMB1  CMB2 Þ (4) where DQLU2 is the difference in biodiversity quality of a land use before and after the LUC from one unnatural state (1) to another (2), DQ1 and DQ2 are the changes in biodiversity quality of the former land use type and the new land use type, respectively, compared to the natural state, and CMB1 and CMB2 are the CMB of the former land use type and the new land use type, respectively.

2.3.2. The method presented by de Baan et al. (2013a) and suggested elaboration This method is based on values for biodiversity damage potential (BDP), using characterization factors referring to land use classes at biome level. Using the original approach from de Baan et al. (2013a) the natural biome for the area is temperate broadleaf forest. The best available categories for the 60e80 year old forest, fallow land and miscanthus plantation, respectively, are considered to be (following the same order) Secondary vegetation, Pasture/meadow and Permanent crops. For this biome, these categories have BDPs of 0.08, 0.52 and 0.02, respectively (de Baan et al., 2013a). The method of de Baan et al. (2013a) is adapted to assess the difference between the biodiversity impact of the former land use and the new one:

DBDPLU2 ¼ BDP1  BDP2

(5)

where DBDPLU2 is the difference in BDP from the former land use situation to the new, BDP1 is the BDP of the former land use type (compared to the natural state) and BDP2 is the BDP of the new land use type (compared to the natural state).

3. Results and discussion Climate impacts are shown in terms of effective forcing, first considering single contributions from biogenic CO2 fluxes and changes in albedo, and then the combined effects. Afterwards, biodiversity results according to the suggested elaboration of the two selected methods are presented. The main findings, insights, and trade-offs of these two environmental impacts are then discussed. 3.1. Effective forcing from biogenic CO2 fluxes Fig. 4 shows the effective forcing profiles for the biogenic CO2 fluxes associated with the investigated LUC, per m2 land converted, during 100 years from the time the LUC occurs. All fluxes taking place in a year are modeled as occurring at the starting point of that year, that is fluxes in year 1 are modeled at t ¼ 0, fluxes in year 2 are modeled at t ¼ 1, etc. It is clear that the former land use of the area converted to miscanthus production plays a major role in terms of biogenic CO2 flux induced climate impacts. When the plantation is established after replacement of an old, semi-natural forest, the impact is high during the first few years due to the almost simultaneous release of AGB and BGB. After that, the effective forcing gradually decreases as the emitted CO2 is removed from the atmosphere due to interactions with the global carbon cycles (as it is partially absorbed by the oceans and the terrestrial biosphere) and due to the gradual rebuilding of SOC stocks in the area. Still, effective forcing from the net biogenic CO2 fluxes is not recovered during the 100 year period from the change to miscanthus production and this LUC is a source of increased effective forcing and thus global warming from biogenic CO2 fluxes. When the miscanthus production is established on a former fallow land, the effective forcing shows a different profile. The initial loss in BGB and SOC, mainly from oxidation of soil carbon, is immediately offset by an increase in AGB. Further, SOC stock regrows and exceeds the former level over the 20 year SOC sequestration period of the miscanthus (see the Supplementary material). When fluxes cease after 20 years the effective forcing increases, as it reflects the impact of keeping CO2 out of the atmosphere, where it would similarly decrease over time due to interactions with the global carbon cycle. However, the forcing stays below the level before the LUC for the whole 100 year period from the time of the LUC. Thus this LUC is a net sink of effective forcing (negative values) and hence has a cooling effect from biogenic CO2 fluxes.

Fig. 4. Effective forcings from biogenic CO2 fluxes in the investigated LUCs: Case a (from forest to miscanthus) and case b (from fallow land to miscanthus).

S.V. Jørgensen et al. / Journal of Environmental Management 146 (2014) 346e354

3.2. Effective forcing from changes in surface albedo The albedo change impact on effective forcing following the two LUCs is illustrated in Fig. 5, covering both inter-monthly variations and the annual mean over a 22 year period. In terms of albedo change impacts, there is a decrease in effective forcing in both case A and case B, with rather similar patterns. The effective forcing profiles show that most of the cooling from the LUCs occurs in local winter months (from December to March), as the difference in albedo between vegetation covered land and cropland (as miscanthus) which is harvested before winter, is amplified by snow. In summer time, differences are reduced (although always negative). While the pattern of the albedo change impact for the two cases is similar, the increase in surface albedo, and hence decrease in effective forcing, is larger in case A than case B, due to the lower vegetation density of fallow land than forests. But the negative effective forcing (i.e. cooling effect) is still substantial also for case B, especially during winter months. 3.3. Net effective forcing Fig. 6 shows the net effective forcings of the investigated cases after combination of the contributions from biogenic CO2 fluxes and changes in surface albedo, per m2 land converted. Conversion of a forest land to miscanthus causes a net increase in effective forcing in the first few years, due to the emission of carbon from former standing biomass and SOC oxidation. Such impact is stronger than the cooling contribution from the change in surface albedo. However, as the emitted CO2 is gradually removed from the atmosphere while the effect from albedo remains constant as long as the new land use is maintained, the latter becomes the dominating contributor to global climate. The net effective forcing thus turns to negative values after less than 10 years following the LUC. When miscanthus is established on fallow land, the net effective forcing is negative from the beginning. While both the effective forcing from biogenic CO2 fluxes and albedo are negative from the beginning, the albedo impacts are dominating for the entire assessment period. The dominance of the climate forcing from changes in surface albedo in some geographical areas is in line with other findings already reported in the literature (Bala et al., 2007; Cherubini et al., 2012; Randerson et al., 2006. 3.4. Biodiversity Results for ES and EV for the two ecoregions using the Michelsen (2008) method are (ES, EV): Ecoregion NA0415 (0.964, 1), NA0416 (0.941, 0.1).

351

Results for change in biodiversity when planting miscanthus on the chosen locations using each of the two introduced methods, both in their original and elaborated forms, are shown in Fig. 7. While the two elaborated methods for assessing biodiversity are on different levels, ecoregion vs. biome, results can still be compared in terms of how much impact introducing a new land use has on the biodiversity level in the area. This is done in terms of fraction of natural state biodiversity lost/gained from the new land use compared to the previous land use (not including duration and size of land use). The impact of duration and size of the land use can be included simply by multiplying with the results; in this case the results from Fig. 7 applied to one m2 land converted to miscanthus plantation, with a land use duration of 100 years. However, as both cases assessed here are for the same area and duration, this does not impact the resulting conclusions. The comparison between and within methods shows a highly varied picture. For both methods it is obvious that the location in terms of ecoregion/biome is determining for the level of biodiversity impact. Comparing set 1 (A1, B1) and set 2 (A2, B2), results differ with more than a factor 10 depending on which of the ecoregions present in the area the miscanthus plantation is established. The large variance in these results is primarily due to the difference in the conservation status of the two neighboring ecoregions. In the case of set 3 (3A, 3B) results show that the former land use in both cases has a much higher impact than the miscanthus plantation, meaning that a substantial biodiversity degradation from the former land use is counteracted after changing to miscanthus production. This however does not seem realistic, especially in the case of shifting from former forest land use to miscanthus production, and results of set 3 are regarded as less reliable as only scarce data was available for this biome (see de Baan et al., 2013a). This is further discussed in Section 3.6. Also the mean values for this biome cover a great variance within each category and the values are highly different from those in all other biomes. Thus set 3 is supplemented by set 4, which gives the world average for the different land use types of this method. It is interesting to observe the similarities of the trends of the elaborated methods, aside from results in set 3: The impact of the LUC to the current biodiversity state is in all sets much smaller in case B than case A. Despite the difference in the results between the elaborated methods of Michelsen (2008) and de Baan et al. (2013a) and the high level of uncertainty, mainly due to lack of data, a common conclusion can thus still be drawn: Establishing a miscanthus production on former fallow land is better in terms of biodiversity impact than using former forest for the miscanthus production. But in both cases miscanthus production reduces biodiversity (excluding results from set 3).

Fig. 5. Change in mean effective forcing resulting from the change in surface albedo following conversion of one m2 from one land use type to another in the specified geographical area.

352

S.V. Jørgensen et al. / Journal of Environmental Management 146 (2014) 346e354

Fig. 6. Change in effective forcing for 100 years from the time of the LUC in the specified geographical area, in a) the case of producing miscanthus on a former forest area and b) the case of using fallow land for the production.

Compared to results based on the original approaches where the former land use is ignored, biodiversity impacts of the new land use are substantially smaller using the elaborated approaches. The suggested elaboration of the two methods seems a more realistic reflection of reality in this case, as it reflects the actual shift in biodiversity in the assessed areas. If the new land use in the assessed area would be expected to lead to indirect land use change somewhere else, the biodiversity impact of that should be assessed as well, using data for the geographic area where it would be expected to take place. No indirect land use change is however expected in the cases assessed here. 3.5. Tradeoffs between impacts The results of the cases analyzed in this paper show that there can be potential tradeoffs between the impacts on global warming and biodiversity caused by land use and land use change. Which option is perceived as the better solution thus will depend on the magnitude of the different impacts, and on which impacts are considered more

important. This is mainly a political choice, but all relevant impacts should be considered so that the choice of tradeoffs is made as an informed decision rather than due to lack of knowledge. Here, both cases of changed land use will offset global warming rather than contributing to it over the assessed period, mainly due to the decrease in effective forcing from the change in albedo. For biodiversity, on the contrary, both cases cause a decrease in biodiversity. Thus the trends in impacts are the same for the two cases, though the magnitude of the impacts is not. The relative result between the two cases is that conversion from fallow land to miscanthus plantation is better than conversion from forest in terms of both types of impacts, as it leads to the lowest decrease in biodiversity and the highest global warming offsetting effect. 3.6. Uncertainty While biogenic carbon stocks for a land use type can vary a lot, as they depend on soil type, climate conditions etc., results of this hypothetic case reflect the change from the former land use as

Fig. 7. Impacts on biodiversity following the changed land use (LU) in case A and case B, using the two methods, with two geographical alternatives for each, as described in the text. The total area pulled out of each ‘Part of natural state biodiversity level maintained’-pie gives the results of the original approaches, whereas the different patterns of the pulled out parts separate the results into impact of former and new LU following the elaboration suggested here, as indicated in the figure.

S.V. Jørgensen et al. / Journal of Environmental Management 146 (2014) 346e354

described by the literature data used. The main uncertainty in terms of biogenic carbon stock estimates in this study lies in the estimate of changes in SOC and BGB from former to new land use. Here, a loss of 30% of the SOC of the former land use has been assumed based on literature estimates as mentioned in Table S1 in the Supplementary material. As the BGB is hard to distinguish from the other biogenic carbon stocks of the land cover types used here (as it is not given separately in the literature data used) and as BGB left in the soil after a LUC is assumed to some degree to become part of the soil and litter layer, the BGB has been treated together with the SOC. For the forest data, some BGB may even be included in the estimate of the AGB as the latter is calculated as the total carbon content of the forest subtracting the total carbon content from the soil and litter layer. As all AGB is modelled as lost following the LUC, any BGB included in that estimate would mean a higher initial biogenic carbon flux from the LUC than in the case where BGB is modelled as SOC and only 30% is lost following the LUC. However, this uncertainty cannot change the conclusion of the results. In terms of albedo impact of the LUC in the two cases, a main source of uncertainty is the choice of actual land areas used as proxies for the land use types of the hypothetic case. As miscanthus production is rather new in the US, a miscanthus field of sufficient size and age could not be used and an annual crop area was used instead. While there is some differences in characteristic of an annual crop and miscanthus, the main features in terms of land cover during the year is expected to be quite similar, as miscanthus is harvested every year, thus resembling an annual crop. For the forest, an actual forest in the specified area was used for data so the resemblance is high. For the case of fallow land, an area of rather open land but with some small vegetation cover was used, as it is assumed that a land left fallow will rather fast to some degree get covered by moderate vegetation. If the fallow land in reality would be barer, the snow masking effect would be lower and the climate change benefit of the conversion to miscanthus would diminish accordingly. The simulations of this study have been undertaken under constant background climatic conditions. Future changes in climate will generally cause a decrease in the intensity and length of the snowy season, and this of course will affect the results of this analysis as a lower impact from albedo change can be foreseen. However, a warming climate will have feedbacks on the carbon cycle as well. Under the conditions of an atmosphere with higher CO2 concentration, emissions of CO2 will gradually become less effective in warming the climate (Caldeira and Kasting, 1993). A recent study has simultaneously taken these two effects in to account showing that they approximately cancel out each other (Bright et al., 2013). In terms of the uncertainty of the biodiversity impacts, those differ between the two methods, but are in both cases rather high. The results from both methods depend largely on the ‘degree of naturalness’/land use type estimated for the different land uses, and both methods are still on a rather preliminary level. In the case of the de Baan et al. (2013a) method for biome 4 (temperate broadleaf forest; set 3 in Fig. 7), results are favoring permanent crops above everything else than primary vegetation, meaning that even a rather old, recovering forest, is perceived as having less biodiversity than a permanent crop. This seems intuitively wrong, at least in most cases, and the categories should be diverted into more subcategories, based on substantial amount of data, in order to better reflect reality. Another issue is that miscanthus is not permanent, but considered to have a lifetime of ~20 years, while being cut down every year. Again, this calls for more subcategories to reflect the actual case, for results to become meaningful. In the case of the Michelsen (2008) method, with modifications from Coelho and Michelsen (2014), there is a very high difference in

353

biodiversity impact results from neighboring areas, as illustrated in Fig. 7. This is mainly due to the EV of one area being only 10% of that of the other. Such abrupt changes are not likely found in reality where a more gradual change between neighboring areas is likely. This is supported by the results from Coelho and Michelsen (2014), where use of more fine scaled local data instead of default EV values show more gradual changes. Again, this emphasizes the need for more specific data for applying reliable EV at a geographical scale. One final comment in terms of biodiversity impacts is that only impacts related to the change in land use are included, whereas biodiversity impacts as a consequence of the change in global warming following the LUC are not included. 4. Conclusion The example presented in this paper illustrates the potential importance of impacts from biogenic CO2 fluxes and changes in albedo and biodiversity when a 2G biomass plantation is established on a former fallow or forest land. Concerning the impact on climate, the assessed LUC leads to initial release of biogenic carbon from the previous land use type, while growth of the miscanthus crop leads to re-sequestration of carbon over time. The level of resequestration compared to the carbon stock levels of the previous land use type however varies substantially. When using former fallow land, there is a net effective forcing saving from sequestration of biogenic CO2 during the 100 year period following the LUC, resulting in a global cooling, which is amplified by the associated change in surface albedo. In the case of LUC from former forest assessed for a 100 year period after the LUC, the net effective forcing increases for the first years owing to emissions from oxidation of standing biomass and SOC of the previous land use. However, the associated cooling from albedo change dominates after few years, thus resulting in a net cooling effect. This finding reinforces the need to include changes in surface albedo when biomass resources are gathered from land disturbances, as they can be the dominating factor for the impacts on global climate. The suggested adapted methods for assessing biodiversity impact, following the change form one human influenced biodiversity state to another, agree that both assessed cases of establishing a miscanthus plantation reduce biodiversity in the area, and that using former fallow land is preferable over former forest. However, quantitative results are rather uncertain as the methods for estimating biodiversity impacts are still on a preliminary level and it is clear that further work needs to be done for obtaining a generally reliable method for assessing biodiversity impacts, not least in terms of sufficiently detailed geographical data. There is a tradeoff between global warming and biodiversity impacts following the LUC in both the assessed cases. A main reason for that is that the effective forcing reduction from surface albedo increases when vegetation covered areas such as forests are replaced with crops being cut down annually. On the other hand, this replacement substantially decreases biodiversity in the area. Between the two cases, the change from fallow land to miscanthus plantation is clearly preferable both in terms of highest reduction in effective forcing and lowest impact on biodiversity. Acknowledgments This paper has been written as part of an industrial PhD project which is co-funded by the Danish Agency for Science, Technology and Innovation. The work has been conducted as part of a research stay at the Norwegian University of Science and Technology (NTNU), The Industrial Ecology Programme, Department of Energy and Process Engineering, Trondheim, Norway. Francesco Cherubini

354

S.V. Jørgensen et al. / Journal of Environmental Management 146 (2014) 346e354

and Ottar Michelsen were both funded by the Norwegian Bioenergy Innovation Centre CenBio. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.jenvman.2014.06.033. References Anderson-Teixeira, K.J., Davis, S.C., Masters, M.D., Delucia, E.H., 2009. Changes in soil organic carbon under biofuel crops. GCB Bioenergy 1, 75e96. Anderson-Teixeira, K.J., Snyder, P.K., Twine, T.E., Cuadra, S.V., Costa, M.H., DeLucia, E.H., 2012. Climate-regulation services of natural and agricultural ecoregions of the Americas. Nat. Clim. Change 2, 177e181. Bala, G., Caldeira, K., Wickett, M., Phillips, T.J., Lobell, D.B., Delire, C., Mirin, A., 2007. Combined climate and carbon cycle-cycle effects of large-scale deforestation. PNAS 104, 6550e6555. Betts, R., 2007. Implications of land ecosystem-atmosphere interactions for strategies for climate change adaptation and mitigation. Tellus B 59, 602e615. Betts, R.A., 2011. Mitigation: a sweetener for biofuels. Nat. Clim. Change 1, 99e101. Bellouin, N., Boucher, O., 2010. Climate Response and Efficacy of Snow Albedo Forcing in the HadGEM2-AML Climate Model. Hadley Centre Technical Note, HCTN82. Met Office, p. 8. Brentrup, F., Küsters, J., Lammel, J., Kuhlmann, H., 2002. Life cycle impact assessment of land use based on the hemeroby concept. Int. J. Life Cycle Assess. 7 (6), 339e348.
n-Fern
Bright, R.M., Anto andez, C., Astrup, R., Cherubini, F., Kvalevåg, M., Strømman, A.H., 2013. Climate change implications of shifting forest management strategy in a boreal forest ecosystem of Norway. Glob. Change Biol. http:// dx.doi.org/10.1111/gcb.12451. Bright, R.M., Cherubini, F., Strømman, A.H., 2012a. Climate impacts of bioenergy: inclusion of carbon cycle and albedo dynamics in life cycle impact assessment. Environ. Impact Asses. Rev. 37, 2e11. Bright, R.M., Cherubini, F., Astrup, R., Bird, N., Cowie, A.L., Ducey, M.J., Marland, G., Pingoud, K., Savolainen, I., Strømman, A.H., 2012b. A comment to “large-scale bioenergy from additional harvest of forest biomass is neither sustainable nor greenhouse gas neutral”: important insights beyond greenhouse gas accounting. GCB Bioenergy 4 (6), 617e619. Brown, T.R., Brown, R.C., 2013. A review of cellulosic biofuel commercial-scale projects in the United States. Biofuels Bioprod. Bioref. 7, 235e245. Caldeira, K., Kasting, J.F., 1993. Insensitivity of global warming potentials to carbon dioxide emissions scenarios. Nature 366, 251e253. Campbell, A., Doswald, N., 2009. The Impacts of Biofuel Production on Biodiversity: a Review of the Current Literature. UNEP-WCMC, Cambridge, UK. Cherubini, F., Bird, N.D., Cowie, A., Jungmeier, G., Schlamadinger, B., WoessGallasch, S., 2009. Energy and greenhouse gas-based LCA of biofuel and bioenergy systems: key issues, ranges and recommendations. Resour. Cons. Recycl. 53, 434e447. Cherubini, F., Bright, R.M., Strømman, A.H., 2012. Site-specific global warming potentials of biogenic CO2 for bioenergy: contributions from carbon fluxes and albedo dynamics. Environ. Res. Lett. 7, 045902 (11pp). Cherubini, F., Jungmeier, G., 2010. LCA of a biorefinery concept producing bioethanol, bioenergy, and chemicals from switchgrass. Int. J. Life Cycle Assess. 15, 53e66. Cherubini, F., Peters, G.P., Berntsen, T., Strømman, A.H., Hertwich, E., 2011. CO2 emissions from biomass combustion for bioenergy: atmospheric decay and contribution to global warming. GCB Bioenergy 3, 413e426. Christiansen, B., 1999. Radiative forcing and climate sensitivity: the ozone experience. Q. J. R. Meteorol. Soc. 125, 3011e3035. Chum, H., et al., 2011. Bioenergy. In: Edenhofer, O., Pichs-Madruga, R., Sokona, Y., et al. (Eds.), IPCC Special Report on Renewable Energy Sources and Climate Change Mitigation. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA, p. 188. Claussen, M., Brovkin, V., Ganopolski, A., 2001. Biogeophysical versus biogeochemical feedbacks of large-scale land cover change. Geophys. Res. Lett. 28, 1011e1014. Coelho, C.R.V., Michelsen, O., 2014. Land use impacts on biodiversity from kiwifruit production in New Zealand assessed with global and local datasets. Int. J. Life Cycle Assess. 19, 285e296. Cowie, A.L., Smith, P., Johnson, D., 2006. Does soil carbon loss in biomass production systems negate the greenhouse benefits of bioenergy? Mitig. Adapt. Strateg. Glob. Change 11, 979e1002.
, N., Friedlingstein, P., 2007. Impact of land cover Davin, E.L., de Noblet-Ducoudre change on surface climate: relevance of the radiative forcing concept. Geophys Res. Lett. 34, L13702. de Baan, L., Alkemade, R., Koellner, T., 2013a. Land use impacts on biodiversity in LCA: a global approach. Int. J. Life Cycle Assess. http://dx.doi.org/10.1007/ s11367-012-0412-0. de Baan, L., Mutel, C.L., Curran, M., Hellweg, S., Koellner, T., 2013b. Land use in life cycle assessment: global characterization factors based on regional and global potential species extinction. Environ. Sci. Technol. 47, 9281e9290.

Diaz, S., Cabido, M., 2001. Vive la difference: plant functional diversity matters to ecosystem processes. Trends Ecol. Evol. 16, 646e655. Djomo, S.N., Kasmioui, O.E., Ceulemans, R., 2011. Energy and greenhouse gas balance of bioenergy production from poplar and willow: a review. GCB Bioenergy 3, 181e197. Dohleman, F.G., Heaton, E.A., Long, S.P., 2010. Perennial grasses as secondgeneration sustainable feedstocks without conflict with food production. In: Khanna, M., et al. (Eds.), Handbook of Bioenergy Economics and Policy, Natural Resource Management and Policy, vol. 3. Springer ScienceþBusiness Media, LLC. http://dx.doi.org/10.1007/978-1-4419-0369-3_3. Don, A., Osborne, B., Hastings, A., Skiba, U., Carter, M.S., Drewer, J., Flessa, H., €nen, N., Jones, M.B., Lanigan, G.J., Mander Ü, Monti A., Freibauer, A., Hyvo Djomo, S.N., Valentine, J., Walter, K., Zegada-Lizarazu, W., Zenone, T., 2012. Land-use change to bioenergy production in Europe: implications for the greenhouse gas balance and soil carbon. GCB Bioenergy 4, 372e391. Flanner, M.G., Zender, C.S., Randerson, J.T., Rasch, P.J., 2007. Present-day climate forcing and response from black carbon in snow. J. Geophys. Res. 112: D11202. Georgescu, M., Lobell, D.B., Field, C.B., 2009. Potential impact of U.S. biofuels on regional climate. Geophys. Res. Lett. 36. L21806 (6 pp). Georgescu, M., Lobell, D.B., Field, C.B., 2011. Direct climate effects of perennial bioenergy crops in the United States. PNAS 108 (11), 4307e4312. Google Earth Version 7.1.1. 1888. Available at: https://www.google.com/earth/ download/ge/, (accessed November 2012). Hansen, J., Nazarenko, L., 2004. Soot climate forcing via snow and ice albedos. PNAS 101, 423e428. Hansen, J., Sato, M., Ruedy, R., Nazarenko, L., Lacis, A., Schmidt, G.A., Russell, G., Aleinov, I., Bauer, M., Bauer, S., Bell, N., Cairns, B., Canuto, V., Chandler, M., Cheng, Y., Del Genio, A., Faluvegi, G., Fleming, E., Friend, A., Hall, T., Jackman, C., Kelley, M., Kiang, N., Koch, D., Lean, J., Lerner, J., Lo, K., Menon, S., Miller, R., Minnis, P., Novakov, T., Oinas, V., Ja, Perlwitz, Ju, Perlwitz, Rind, D., Romanou, A., Shindell, D., Stone, P., Sun, S., Tausnev, N., Thresher, D., Wielicki, B., Wong, T., Yao, M., Zhang, S., 2005. Efficacy of climate forcings. J. Geophys. Res. 110, D18104. http://dx.doi.org/10.1029/2005JD005776. Knutti, R., Hegerl, G.C., 2008. The equilibrium sensitivity of the Earth's temperature to radiation changes. Nat. Geosci. 1, 735e743. ~o, M., Civit, B., Margni, M., Mila  i Canals, L., Koellner, T., de Baan, L., Beck, T., Branda Saad, R., de Souza, D.M., Müller-Wenk, R., 2013. UNEP-SETAC guideline on global land use impact assessment on biodiversity and ecosystem services in LCA. Int. J. Life Cycle Assess. 18, 1188e1202. Lange, J.-P., 2007. Lignocellulose conversion: an introduction to chemistry, process and economics. Biofuels Bioprod. Bioref. 1 (1), 39e48. Loarie, S.R., Lobell, D.B., Asner, G.P., Mu, Q., Field, C.B., 2011. Direct impacts on local climate of sugar-cane expansion in Brazil. Nat. Clim. Change 1, 105e109. Melillo, J.M., et al., 2009. Indirect emissions from biofuels: how important? Science 326 (5958), 1397e1399. Michelsen, O., 2008. Assessment of land use impact on biodiversity. Proposal of a new methodology exemplified with forestry operations in Norway. Int. J. LCA 13 (1), 22e31. Michelsen, O., Cherubini, F., Strømman, A.H., 2012. Impact assessment of biodiversity and carbon pools from land use and land use changes in life cycle assessment, exemplified with forestry operations in Norway. J. Ind. Ecol. 16, 231e242. Michelsen, O., McDevitt, J.E., Coelho, C.R.V., 2014. A comparison of three methods to assess land use impacts on biodiversity in a case study of forestry plantations in New Zealand. Int. J. LCA. http://dx.doi.org/10.1007/s11367-014-0742-1. Mishra, U., Torn, M.S., Fingerman, K., 2012. Miscanthus biomass productivity within US croplands and its potential impact on soil organic carbon. GCB Bioenergy. http://dx.doi.org/10.1111/j.1757-1707.2012.01201.x. O'Halloran, T.L., Law, B.E., Goulden, M.L., Wang, Z., Barr, J.G., Schaaf, C., Brown, M., € ckede, M., Black, A., Engel, V., 2012. Radiative forcing of natural Fuentes, J.D., Go forest disturbances. Glob. Change Biol. 18 (2), 555e565. ORNL DAAC, 2012. MODIS subsetted land products, Collection 5. Available at: http:// daac.ornl.gov/MODIS/modis.html, from Oak Ridge National Laboratory Distributed Active Archive Center (ORNL DAAC), Oak Ridge, Tennessee, U.S.A. Accessed November 19, 2012. Subset obtained for MCD43A product, time period: February 18, 2000 to October 31, 2012, and subset size: 0.5 x 0.5 km. Randerson, J.T., Liu, H., Flanner, M.G., Chambers, S.D., Jin, Y., Hess, P.G., Pfister, G., Mack, M.C., Treseder, K.K., Welp, L.R., Chapin, F.S., Harden, J.W., Goulden, M.L., Lyons, E., Neff, J.C., Schuur, E.A.G., Zender, C.S., 2006. The impact of boreal forest fire on climate warming. Science 314, 1130e1132. Sala, O.E., Chapin III, F.S., Armesto, J.J., Berlow, E., Bloomfield, J., Dirzo, R., HuberSanwald, E., Huenneke, L.F., Jackson, R.B., Kinzig, A., Leemans, R., Lodge, D.M., Mooney, H.A., Oesterheld, M., Poff, N.L., Sykes, M.T., Walker, B.H., Walker, M., Wall, D.H., 2000. Global biodiversity scenarios for the Year 2100. Science 287, 1770e1774. UNEP e United Nations Environment Programme, 1992. Convention on Biological Diversity. Text and annexes. UNEP, Geneva. Van der Voet, E., Lifset, R.J., Luo, L., 2010. Life-cycle assessment of biofuels, convergence and divergence. Biofuels 1 (3), 435e449. ~o, M., Bringezu, S., Hermann, B., Patel, M.K., Weiss, M., Haufe, J., Carus, M., Branda 2012. A review of the environmental impacts of biobased materials. J. Ind. Ecol. http://dx.doi.org/10.1111/j.1530-9290.2012.00468.x. WWF Wildfinder. Available at: http://worldwildlife.org/science/wildfinder/, (accessed November 2012).