• Open Access

The impact of extensive planting of Miscanthus as an energy crop on future CO2 atmospheric concentrations


  • 2This article [The impact of extensive planting of Miscanthus as an energy crop on future CO2 atmospheric concentrations] was written by [J. K. Hughes*, A. J. Lloyd1†, C. Huntingford†, J. W. Finch† and R. J. Harding†] of [*Hadley Centre Met Office, FitzRoy Road, Exeter EX13PB, UK, Centre for Ecology and Hydrology, Wallingford, Oxfordshire OX108BB, UK]. It is published with the permission of the controller of HMSO and the Queen's Printer for Scotland.

J. Hughes, e-mail: john.hughes@metoffice.gov.uk


A process-based model of the energy crop Miscanthus×giganteus is integrated into the global climate impact model IMOGEN, simulating the potential of large-scale Miscanthus plantation to offset fossil fuel emissions during the 21st century. This simulation produces spatially explicit, annual projections of Miscanthus yields from the present day to the year 2100 under an SRES A2 anthropogenic emissions scenario and includes the effects of climate change. IMOGEN also simulates natural vegetation and soil carbon storage throughout the 21st century. The benefit of Miscanthus cultivation (avoiding fossil fuel emissions of CO2) is then compared with the cost of displacing natural vegetation (carbon emissions from vegetation and soil). The time taken for these effects to cancel out, the pay-back time, is calculated regionally. The effects of large-scale Miscanthus plantation are then integrated globally to produce an estimate of atmospheric CO2 concentrations throughout the 21st century. Our best estimate of the pay-back time for Miscanthus plantation is 30 years. We project a maximum possible reduction in atmospheric CO2 of 323 ppmv by the end of 21st century, with a reduction of 162 ppmv corresponding to the best estimate scenario.


It is now established that the global climate is warming due to fossil fuel combustion. Atmospheric carbon dioxide (CO2) while not the only ‘greenhouse gas’ is the dominant species driving this climate change and has increased by over 100 ppmv since preindustrial times to 385 ppmv at present, a concentration exceeding anything experienced on this planet in the past 800 000 years (Lüthi et al., 2008). Further increases in atmospheric CO2 are almost certain and associated with it a commitment to further climate change at a level considered dangerous.

Energy crops have been promoted as a potential way to mitigate future CO2 emissions. The concept is relatively simple – as plants grow they draw-down atmospheric CO2 through photosynthesis. If they are then used to generate energy, the CO2 released will have previously been taken out of the atmosphere in the recent past, unlike mined fossil fuels and will not raise atmospheric CO2 concentrations. Biomass and agricultural waste are already burnt to provide energy and globally contribute an estimated 10% of primary energy consumption (direct burning to heat homes and cook food; IEA, 2006, p. 420). Most of this burning occurs in developing countries. Globally, dedicated energy crops currently contribute only a small proportion to the total energy produced from biomass each year; however, this proportion is set to increase (Sims et al., 2006). It is predicted that global biofuel production will quadruple within the next 15–20 years (IEA, 2004, pp. 167–168; Fairless, 2007; Himmel et al., 2007), due primarily to rising oil prices and government policies encouraging the increased production of ethanol and biodiesel (IEA, 2004, pp. 167–168; Gibbs et al., 2008). Future energy scenarios presented by Sims et al. (2006), using the IMAGE 2.2 integrated assessment model, explored the impact of increased replacement of energy derived from fossil fuels with those derived from energy crops and have estimated that energy crops could produce between 2 and 22 EJ yr−1 by 2025. In comparison, current global energy consumption is estimated at 470 EJ (11 204 Mtoe; IEA, 2006, p. 66). Extensive biofuel production is central to the more aggressive mitigation scenarios that will be used to drive simulations for the CMIP5 intercomparison project (Sims et al., 2006; Wise et al., 2009; http://www.cmip-pcmdi.llnl.gov/cmip5/), in preparation for the fifth IPCC assessment report (AR5; http://www.ipcc.ch/).

The use of bioenergy systems is now being promoted by governments, for example, the UK government has published its long-term energy policy (DEFRA, 2003) which includes support for energy crops (e.g. see sections 4.49 and 4.50, for example) and has identified biomass-derived energy as having a ‘central role to play in meeting the EU target of 20% renewable energy by 2020’ (DEFRA, 2007, p. 5). Europe and the United States have each carried out trials on both woody and herbaceous crops as potential energy crops. Under the woody group are tree species (Salix, Populus spp.) and within the herbaceous group are traditional row crops (Zea mays, Triticum spp.) and the perennial rhizomatous grasses (Arundo, Panicum, Miscanthus spp.).

There are, however, also important concerns about extensive planting of energy crops. Extensive energy crop plantations could reduce the land available for food crops, which for some countries would have a significant impact on their food security (a broad concept encompassing access, availability, quality and utilization of food – see Ewing & Msangi, 2009, for a review). In the future, with increasing demands of population growth and a global trend to more meat-based diet, there is actually a requirement to expand agricultural production thus exacerbating the competition for land. More extensive planting of energy crops could also be at the expense of other natural ecosystems. This may be especially true in tropical regions, such as Brazil, Malaysia and Indonesia as it has been suggested that United States and European Union states are unlikely to have the land base needed to achieve possible future bioenergy targets (IEA, 2006; UN, 2007; Gibbs et al., 2008) and as a consequence may rely on developing nations to provide these resources; however, this may not be the case and Perlack et al. (2005) suggest that the United States has the land resources to meet currently mandated bioenergy goals, as well as likely future mandates, with the cultivation of dedicated energy crops on arable land. Major changes to natural ecosystems, such as tropical forest losses due to either food or energy crop use, would obviously have a major impact on local carbon stores, biodiversity, water use, indigenous people and even tourism (Haughton et al., 2009; Balmford et al., 2009). However, given the trend in rising agricultural yields, it is largely unknown whether the land requirements of energy crop cultivation could be met by current arable land in the future (Ewers et al., 2009). For a more detailed review of potential impacts from energy crops, refer to Powlson et al. (2005).

The removal of natural vegetation to make space for energy crops could release large pulses of CO2 into the atmosphere (especially through deforestation; see Danielson et al., 2009 for an oil-palm example in south-east Asia). Including this displacement of natural carbon is, of course critical in accurate assessment of the impact of energy crop cultivation (Searchinger et al., 2009). One way of quantifying this effect is through the calculation of the ‘pay-back time,’ which may be defined as the time when terrestrial gain through CO2 draw-down outweighs the initial terrestrial carbon loses due to replacement of natural vegetation by this crop, ignoring carbon storage in energy crop plantations (Gibbs et al., 2008). In fact, Miscanthus plantations have been shown to increase soil carbon storage, in existing arable land (Dondini, et al., 2009a, b; Davis et al., 2010); however, we were not able to include this in our simulations due to the modelling framework. This additional carbon storage in Miscanthus plantations would obviously enhance the climate change mitigation potential of Miscanthus. Deforestation (mainly for agriculture) currently contributes an estimated 12% of global anthropogenic CO2 emissions (van der Werf et al., 2009). Considering energy crops specifically, Fargoine et al. (2008) estimated that natural carbon displacement could release between 17 and 420 times more than annual draw-down of CO2, i.e. the range in pay-back time is 17–420 years for different energy crops, the range depending on the energy crop type and the natural biome displaced. The upper limit of 420 years is, for example, for the production of Palm biodiesel displacing peatland rainforest. Searchinger et al. (2008), investigated the pay-back time of corn-based ethanol and estimated a value of 167 years, assuming that ethanol production displaces natural vegetation. Searchinger et al. (2008) used land-use conversion data from the 1990s and the GREET emissions model applied globally. Any pay-back time of the order of a century of more would mean that large-scale cultivation would actually lead to increased atmospheric CO2 during the 21st century and would have negligible or even negative immediate climate change mitigation value. Righelato & Spracklen (2007) compared a number of biofuel crops and concluded that all acted to increase CO2 concentrations over a 30-year time scale (i.e. the pay-back time was greater than 30 years).

The energy crop considered here is Miscanthus×giganteus (hereafter referred to as Miscanthus), a hybrid sterile perennial rhizomatous grass species originating from Asia. It is highly efficient in its use of available nutrients (Beale & Long, 1997), water (Beale et al., 1999) and solar radiation (Beale & Long, 1995) making it an attractive energy crop. This efficiency is primarily due to the fact that Miscanthus uses the C4 photosynthetic pathway. More specifically, plants that use the C4 pathway are able to utilize more of the light energy available to them and convert this into stored energy more efficiently than C3 plants (Mooney & Ehleringer, 1997). In the United Kingdom, for example, Miscanthus achieves 37% higher photosynthetic uptake than native C3 plants (Beale & Long, 1995) and, in Europe, can attain yields as high as 1.7 kg C m−2 yr−1 (Clifton-Brown et al., 2001). In addition, Miscanthus is naturally more adapted to cooler climates than most other C4 plants (the native range spans subtropical to subarctic Asia; Numata, 1979) and hence does not show impairment of essential photosynthetic apparatus under cooler temperatures (Beale et al., 1996; Naidu et al., 2003). Further, Miscanthus has a rhizome system that enables nutrients to be cycled seasonally between the above- and belowground portions of the plant, thereby minimizing the need for external input of fertilizers (Mclaughlin & Walsh, 1998). This is particularly important as Crutzen et al. (2008) demonstrated that N2O release from other types of energy crop cultivation could offset the CO2 mitigation benefit of the crop. These properties make Miscanthus an attractive energy crop. The resultant feedstock tends to have a low mineral content (in terms of chlorine, nitrogen, sulphur and ash) and therefore releases little pollution when combusted (Lewandowski & Kicherer, 1997; Heaton et al., 2004). In addition, Miscanthus×giganteus is a sterile hybrid and therefore unlikely of becoming an invasive species, unlike Miscanthus sinensis in United States.

To generate more accurate pay-back times for Miscanthus, a process-based model of Miscanthus is integrated into the IMOGEN global climate impacts model, producing regional estimates of Miscanthus whole plant net primary productivity (NPP, photosynthesis−plant respiration, see Cox et al., 1999) throughout the 21st century (including climate change). In addition to this, we apply a plantation scenario in regions where Miscanthus can produce an economically viable yield. In this sensitivity study, we assume 5% of the land area is converted to Miscanthus plantation in these regions. In order to calculate pay-back times, it is assumed that Miscanthus plantation displaces natural vegetation without discrimination to vegetation type (e.g. without preferentially cultivating existing grassland areas). The pay-back time itself is largely independent of the area converted to biofuel production since 10% extra forest cleared will result in 10% extra Miscanthus planted and 10% extra annual draw-down. The resulting trajectory of atmospheric CO2 concentration is, however, sensitive to this value and in the real world local feedbacks from land cover change would be expected to affect Miscanthus yields, but are not included here.

Materials and methods


Central to better estimates of pay-back time for particular energy crops are good model predictions of whole plant NPP and the ability of crops to draw-down CO2. This needs a model which is credible for present-day conditions and capable of making robust predictions in a changing climate. The Met Office Surface Exchange Scheme (Cox et al., 1998, 1999) is a land surface model designed for implementation in general circulation models (GCMs), but which is amenable to parameterization and analysis at single points. Initially called MOSES, this land surface model has now been released as a stand-alone community land surface scheme, JULES (http://www.jchmr.org/jules/). JULES calculates land-atmosphere fluxes of momentum, heat and in particular carbon dioxide, and typically at subdaily timesteps (30 min in the current study). JULES also predicts changes in vegetation structure and composition, driven by simulated carbon fluxes. JULES includes the TRIFFID dynamic global vegetation model, which represents vegetation as five plant functional types (PFTs): broadleaf trees, needleleaf trees, shrubs, C3- and C4-grasses. JULES includes carbon storage in vegetation and the underlying soil carbon storage. Soil carbon storage is the balance between simulated litterfall and microbial respiration. The C4 photosynthesis model used as standard in JULES is the same model (Collatz et al., 1992) as was used by Miguez et al. (2009) to develop their own model of Miscanthus. JULES does not yet include a representation of agricultural land, so it is not possible to consider Miscanthus plantations in relation to existing agriculture.

We use the IMOGEN analogue model to make global predictions using JULES. IMOGEN is an emulator of the Hadley Centre GCM, HadCM3. It is a computationally efficient means of exploring climate change CO2 scenarios and amenable as a climate impacts assessment tool. IMOGEN simulates climate globally on a grid of 2.5° latitude by 3.75° longitude. It includes a basic representation of atmosphere and ocean carbon storage, such that the global carbon cycle is ‘closed;’ however, it does not include hydrological feedbacks from the land surface, for example. Climate change simulations can then be driven by prescribed anthropogenic emissions, such as the SRES A2 emissions scenario (Nakicenovic & Swart, 2000). This allows climate simulations to be made including different feedbacks on atmospheric CO2. IMOGEN is described in detail in Huntingford & Cox (2000), Huntingford et al. (2004) and (2009).

  • Because in IMOGEN atmospheric CO2 concentration is calculated as the balance of different fluxes, reducing fossil fuel emissions by using Miscanthus is mathematically equivalent to subtracting a fraction of Miscanthus yields from the original A2 (no Miscanthus) CO2 trajectory. In this calculation, the fraction of Miscanthus yield that can be subtracted off the original A2 CO2 trajectory represents the ratio of fossil fuel emissions avoided by using biofuels to whole plant NPP. We consider this fraction to include the following effects:
  • The fraction of available energy contained in the crop, after accounting for energy costs of cultivation and processing.
  • The impacts of pests on yields.
  • The type of fossil fuel it replaces.
  • The possibility to associate carbon capture and storage technology (CCS) to biomass plants.

Unfortunately, a detailed life-cycle analysis is beyond the scope of the present study and we will therefore consider three different efficiency fractions: 50% (default), 25% and 100%.

We consolidated realistic literature parameter values for the representation of Miscanthus in JULES, using an existing C4 grass PFT. Using this representation of Miscanthus, IMOGEN is used to predict where Miscanthus is viable, taking into account changing climate throughout the 21st century and Miscanthus yields. For these regions, we then assume a substantial increase in Miscanthus cultivation, up to a prescribed fraction of 5% coverage. Regions of viable Miscanthus cultivation are recalculated every year (see later discussion for details of the methodology); however, by 2050 these regions are stable and do not change significantly after this. Although we include dynamic vegetation, in the simulations vegetation does not return to areas no longer suitable for Miscanthus plantation areas, or later (post 2050) displacement of natural carbon caused by expansion of viable Miscanthus distributions. Through the closed carbon cycle of the IMOGEN system, combined with a PFT representing Miscanthus, we have the modelling tools available to understand how extensive planting of this energy crop could affect future atmospheric CO2 concentrations.

Background to calibration of JULES model for Miscanthus

Miscanthus is a C4 grass with the ability to photosynthesize at low temperatures and originates in Asia (Numata, 1979; Beale & Long, 1995; Beale et al., 1996; Naidu et al., 2003). The modifications used here are based on literature values and also observed measurements of Miscanthus growing in the United Kingdom at a site in Lincolnshire (53°19.059N, 00° 35.640W, approximately at sea level), such as canopy height and leaf area index (LAI; Table 1). The C4 grass PFT is therefore calibrated to be specific to Miscanthus rather than using the generic C4 plant characteristics from Cox (2001). It was also necessary to calculate an allometric coefficient (awl) which relates stem biomass to leaf area (Enquist et al., 1998). To determine this, JULES was run using meteorological data measured at the Lincolnshire site and an ‘awl’ value chosen such that Miscanthus in JULES simulates a canopy height of 3 m and LAI of 3.0 (dimensionless) within 100 days of growth (quantities observed in Lincolnshire site). All other changes to the generic C4 PFT were derived from the LAI and canopy height values and also by comparing model predictions of latent heat and sensible heat fluxes against observations (Finch et al., 2004).

Table 1.   Plant functional type (for C4-grass, Miscanthus) vegetation parameters used in JULES.
Initial parameterValueSource
  1. * field measured data, ** calibrated via model runs to achieve specific output. For general definitions of these parameters see Cox et al. (1999). This table and the standard JULES model (available through http://www.jchmr.org/jules/) allow our model to be replicated.

Maximum LAI3.00*
Minimum LAI0.6*
Quantum efficiency (α), mol CO2−mol PAR photons0.067Beale et al. (1996)
Allometric coefficient (awl), kg carbon m−20.014**
Woody biomass as a multiple of live stem (a_ws)0.90Finch et al. (2004)
PAR extinction coefficient (kpar), m2 leaf m−2 ground0.30Finch et al. (2004)
Growth respiration fraction (r_grow)0.10Finch et al. (2004)
Temperature at which leaves dropped (tleaf_of) K280.15Finch et al. (2004)
Lower temperature for photosynthesis (tlow) K (not an absolute lower limit).281.0Finch et al. (2004)
Root depth (rootd_ft), m1.70Finch et al. (2004)
Rate of change of canopy capacity with LAI (dcatch_lai), kg m−20.20Finch et al. (2004)

Worldwide, a number of values for the maximum LAI and associated canopy extinction coefficient have been reported, ranging from 3 up to 10, thus the UK-specific values used here are conservative. As a sensitivity test, the global model was run with a range of LAI values (3.0, 6.0 and 10.0) and PAR extinction coefficients (0.3, 0.5 and 0.68) found in the literature (Bullard et al., 1997; Clifton-Brown et al., 2000; Li et al., 2003; Price et al., 2004; Danalatos et al., 2007). In fact, this range of values made little difference to the overall carbon exchanges and atmospheric CO2 concentrations and for the rest of our analysis we have used the UK-specific values. It is important to note, however, that IMOGEN does not include hydrological feedbacks from the land surface and this insensitivity may not occur in model realistic climate models. Insensitivity to the upper limit set on LAI does not, however, undermine predictions of NPP (Table 1).

Implementation of IMOGEN impacts modelling system

IMOGEN simulations all require a spin-up period to establish the appropriate stores of terrestrial carbon (in vegetation and soils) for a preindustrial period (taken as year 1860). Spin-up is achieved by running IMOGEN repeatedly with a mean preindustrial climate, until equilibrium is reached (for details see Huntingford et al., 2009). IMOGEN is then run to the end of the 21st century using a SRES A2 emissions scenario.

Here we present a pair of IMOGEN simulations which are then combined to produce analysis of the global potential of Miscanthus. The first simulation, ‘control’, is a standard IMOGEN simulation, including only natural vegetation. In the second simulation only the Miscanthus PFT is included (i.e. no natural vegetation). This will be referred to as the ‘viability’ simulation. This viability simulation is subjected to the atmospheric CO2 concentration calculated interactively in the control simulation. Both of these simulations start from the same preindustrial state.

The viability simulation produces annual estimates of Miscanthus photosynthesis with global coverage until 2100. We classify regions as viable if the predicted NPP is >0.4 kg C m−2 yr−1 (which is equivalent to a yield of 9 tons dry weight ha−1 yr−1, Clifton-Brown et al., 2007, using the dry weight to carbon conversion provided in Michel et al., 2006). A factor of approximately two converts dry matter to carbon (Michel et al., 2006) and 10 from t ha−1 to kg m−2. NPP rates in regions above this threshold are stored as it is assumed that in these regions Miscanthus has been planted. The distribution of viable sites is also stored (which defines where Miscanthus could be planted).

After completing the two simulations described above, we are able to quantify the climate mitigation potential of Miscanthus plantation. In order to combine the results we assume a plantation scenario of linear increase in cultivation from 0% to 5% coverage in plantation regions between 2010 and 2040; i.e. in regions classified as viable at 2010, have zero coverage of Miscanthus and by 2040 there is a local coverage of 5%Miscanthus if still viable. The impact on atmospheric CO2 will have three components (1) CO2 released into the atmosphere from loss of natural vegetation (and associated changes in soil carbon), (2) loss of natural vegetation CO2 draw-down within areas of Miscanthus plantation, (3) CO2 draw-down of Miscanthus in regions where viable, that contribute to offset fossil fuel emissions. These effects are integrated globally and then applied to the global CO2 trajectory, effectively linearizing the system. It is at this stage that the assumption of NPP usage efficiency is applied, with 50% of NPP offsetting fossil fuel emissions used as default.


From the control simulation, Fig. 1 shows the terrestrial carbon stores for the response of natural vegetation and soils to transient climate change driven by the SRES A2 emission scenario in year 2050. There is no prescribed land-use change. Total global carbon storage is 2003 GtC, with 680 GtC in the vegetation (Fig. 1a) and 1323 GtC in the soil (Fig. 1b). Although this includes climate-driven changes relative to the present day, the global value for carbon storage is realistic. Sarmiento & Gruber (2006), estimate terrestrial carbon storage during the 1990s at 2300 GtC. In general modelled soil carbon content is similar to published estimates, for example 5–10 kg C m−2 in the tropics and 10–30 kg C m−2 in mid-latitudes (Batjes, 1996). IMOGEN predicts the loss of the Amazon rainforest by 2050, due to climate change. IMOGEN however underestimates the very high carbon storages in boreal and tropical wetlands. This underestimate is not surprising because the version of JULES used here does not include a representation of organic soils and so cannot simulate the very high accumulation rates observed in these regions.

Figure 1.

 Terrestrial carbon stores for the response of natural vegetation and soils to transient climate change driven by the SRES A2 emission scenario in year 2050, for (a) vegetation carbon, (b) soil carbon (kg C m−2).

Figure 2 shows global predictions of Miscanthus NPP from the viability simulation. The calculated values range from 0.5 kg C m−2 yr−1 in boreal region to between 1 and 2 kg C m−2 yr−1 in mid-latitudes and 3 and 5 kg C m−2 yr−1 in the tropics. The mid-latitude values are consistent with observed yields in Germany (Kahle et al., 2002). They are also consistent with observed yields of between 0.4 and 1.6 kg C m−2 yr−1 for Europe (Clifton-Brown et al., 2001) and North America (Heaton et al., 2008). NPP is used to determine where Miscanthus could potentially be grown, based on the lower limit of 0.4 kg C m−2 yr−1. Regions where Miscanthus is not viable (i.e. NPP is below this value) are displayed as white in Fig. 2. Based on a threshold of 0.4 kg C m−2 yr−1, almost all regions of the world have the potential to grow extensive Miscanthus plantations. Alternative limits can be derived from literature, such as 0.53 kg C m−2 yr−1 (Powlson et al., 2005) and 1.16 kg C m−2 yr−1 (Clifton-Brown et al., 2007). A value of 0.53 kg C m−2 yr−1 would not significantly affect the distribution of viable Miscanthus but a value of 1.16 kg C m−2 yr−1 would mean that the boreal cultivation of Miscanthus is not viable.

Figure 2.

 Miscanthus whole plant Net Primary Productivity (NPP) at 2050 (kg C m−2 yr−1).

Throughout the 21st century, predicted regions of viable Miscanthus are relatively stable. However, a key change in plantation viability that occurs around year 2050 in these simulations is for the Amazon region, where IMOGEN predicts after this date almost no vegetation (including Miscanthus) can grow. This mirrors the concerns regarding potential Amazon rainforest ‘die-back’, first identified by White et al. (1999) and subsequently investigated in more detail in Cox et al. (2000), Cox et al. (2004) and Harris et al. (2008). While this feature is not unique to Miscanthus the predicted inability of Miscanthus to grow in central Amazonia by 2050 should be noted, though this could be managed through irrigation.

It is now possible to make a first estimate of local pay-back times. The values are found for 2050 by calculating the number of years of NPP (Fig. 2) required to balance the loss of natural terrestrial carbon stores (Fig. 1) and are displayed in Fig. 3, assuming that 50% of NPP is available to offset fossil fuel emissions. In other words, this is the time, for each geographical position, where the transition to Miscanthus planting has drawn-down at least enough atmospheric CO2 to balance loses to the atmosphere through displacement of natural vegetation. There is a clear continental-scale structure with high latitudes having pay-back times of 20–50 years and tropical continents having values of <30 years. The global mean pay-back time over viable regions is 30 (29.8) years. The higher pay-back times in boreal regions and predicted low Miscanthus NPP values (see earlier discussion) is worth noting.

Figure 3.

 Global distributions of pay-back times at 2050 (years), assuming 50% of NPP is available to offset fossil fuel emissions.

Figure 4 displays the globally integrated effects of Miscanthus, quantifying the overall influence of plantation on future atmospheric CO2 concentrations. In Fig. 4, we present the total modelled Miscanthus draw-down (red curve) achieved by extensive plantation (i.e. up to 5% where viable). Also presented is the loss of CO2 to the atmosphere through displacement of natural vegetation – shown for the period years 2010–2040, both individually for soil and vegetation and the combined total. To place these fluxes in perspective the SRES A2 emissions are included (historical and projected future emissions). All fluxes are given in units of ppmv per year.

Figure 4.

 Magnitude of different carbon fluxes compared. These include the SRES A2 emissions scenario used, the globally integrated draw down of Miscanthus, emissions of total terrestrial carbon displaced by Miscanthus cultivation (CV+CS), emissions from displaced soil carbon (CS) and from displaced natural vegetation carbon (CV).

Figure 5 shows the effects of extensive Miscanthus planting on future trajectories of atmospheric CO2 concentration. The red curve is the projected CO2 trajectory including the effects of Miscanthus, while the black line is the default (i.e. no Miscanthus) A2 trajectory. The point in Fig. 5a) at which the curves cross (the year 2046) is our best estimate of the point beyond which extensive biofuel planting has had a beneficial effect on atmospheric CO2 concentration. This differs from the pay-back time in that it includes the gradual increase of plantation regions between 2010 and 2040 and climate driven changes in NPP. Figure 5b provides a basic sensitivity analysis of the default scenario (red curve). In addition to the default 50% scenario in Fig. 5b, we present values for 100% and 25% offset fractions. When the 100% offset fraction is used CO2 falls below the A2 scenario curve by 2032, while for the 25% value the resulting CO2 trajectory remains above the A2 scenario curve until 2065. For the 100% offset fraction the global pay-back time is 15 years, while for the 25% offset fraction this value is 60 years.

Figure 5.

 (a) Integrated effects of Miscanthus plantation on the atmospheric CO2 trajectory. Black solid line is the basic A2 emissions scenario trajectory. The red line is the result of displaced carbon and Miscanthus drawdown (assuming 50% of NPP is available to offset fossil fuel emissions). (b) Predicted atmospheric CO2 profiles, exploring a number of alternative scenarios.

The benefit of avoiding some of the natural carbon displacement (an assumed reduction of 50%) is illustrated by the dark blue curve, with a net reduction in CO2 from the year 2021 onwards. The effect of completely avoiding any natural carbon displacement (the most optimistic scenario) is shown by the green curve. Realizing this scenario would require many different techniques, such as storage of wood, active management of soil carbon storage and using marginal land. In this case, by the end of the 21st century atmospheric CO2 is 323 ppmv lower than the SRES A2 and the time at which atmospheric CO2 exceeds 550 ppmv, for example, is postponed from 2042 until 2069.


With growing international concern over climate change, a broad range of mitigation strategies are now being proposed. Large-scale energy crop cultivation is a well-established consideration. The fundamental principle of energy crops is that the CO2 released when burnt was recently taken up from the atmosphere and therefore does not increase atmospheric CO2 concentrations. Our simulations suggest that by replacing fossil fuel consumption with Miscanthus could lower CO2 at the end of the current century by up to 323 ppmv (from the 100% efficiency scenario, avoiding natural carbon displacement). As with all proposed large-scale mitigation strategies, it is important to consider all aspects. A commonly cited problem with energy crops is that they would use land that could otherwise produce food (Haughton et al., 2009). Secondly, if natural vegetation is cleared this would release CO2 into the atmosphere (Searchinger et al., 2008). Including this displaced carbon, cultivation of energy crops could actually lead to a net increase in atmospheric CO2. When natural carbon displacement is included we predict that Miscanthus cultivation would only result in net lower CO2 levels after the year 2046, assuming a linear increase in plantation area between 2010 and 2040 for viable regions. In terms of the pay-back time metric (which assumes instantaneous displacement of natural carbon), we calculate a global mean value of 30 years. Our estimated pay-back time lies within the range calculated by Fargoine et al. (2008), 17–420 years, for other energy crops and specific biomes displaced. This result is highly sensitive to the fraction of NPP that is assumed to offset fossil fuel emissions. The default scenario uses a fraction of 50%, while values of 100% and 25% were also explored (giving pay-back times of 15 and 60 years).

Estimates of land potentially available for energy cropping vary widely, for example Hoogwijk et al. (2003) estimate that up to 66% of agricultural land could be available for biomass production (this assumes a moderate diet and high input agriculture). In addition Campbell et al. (2008) and others have estimated there is 385–472 million ha of abandoned agricultural land globally (just under 10% of total agricultural land). In terms of energy supplied from this available land, Hoogwijk et al. (2003) estimate that conversion of 50% of agricultural land could provide up to 988 EJ yr−1, compared with current annual primary energy use of 470 EJ (IEA, 2006, p. 66).

The current study is highly idealized, however, it is useful in that it illustrates the potential of Miscanthus and makes clear the magnitude of some of the carbon costs involved in its plantation. We do not consider alternative energy crops or arable land or the need for Miscanthus to be replanted. We have not considered the effects of extreme Meteorological conditions. Hastings et al. (2009), for example, predicted that Miscanthus productivity would be limited by frost and drought in Europe. Wetland carbon stores are not included, displacement of which could be expected to dramatically increase local pay-back times (e.g. in Indonesia and boreal wetlands). Although the Miscanthus model used here includes mechanisms to include the effects of changing climate it is important to note that the high NPP values in the tropics have not been verified by observations. The maximum CO2 reduction possible with Miscanthus cultivation is dependent on fractional coverage and is proportional to the fraction adopted; hence for a maximum local coverage of 1% a reduction of 65 ppmv is possible if natural vegetation displacement can be avoided. Although in very early stages of development, higher fractional coverage when combined with power stations with carbon capture and storage technology could theoretically provide a route to negative emissions if required. It is also important to appreciate that with energy crops once natural vegetation has been displaced the carbon loss is near-instantaneous, but the energy crop offset requires sustained commitment. Because our simulations provide regional estimates, future plantation scenarios can take this into account.


John Hughes was supported by the Joint DECC and Defra Integrated Climate Programme – DECC/Defra (GA01101). The manuscript also benefited greatly from suggestions by Olivier Boucher, Chris Jones and three anonymous reviewers.