Report

Land Clearing and the Biofuel Carbon Debt

See allHide authors and affiliations

Science  29 Feb 2008:
Vol. 319, Issue 5867, pp. 1235-1238
DOI: 10.1126/science.1152747

Abstract

Increasing energy use, climate change, and carbon dioxide (CO2) emissions from fossil fuels make switching to low-carbon fuels a high priority. Biofuels are a potential low-carbon energy source, but whether biofuels offer carbon savings depends on how they are produced. Converting rainforests, peatlands, savannas, or grasslands to produce food crop–based biofuels in Brazil, Southeast Asia, and the United States creates a “biofuel carbon debt” by releasing 17 to 420 times more CO2 than the annual greenhouse gas (GHG) reductions that these biofuels would provide by displacing fossil fuels. In contrast, biofuels made from waste biomass or from biomass grown on degraded and abandoned agricultural lands planted with perennials incur little or no carbon debt and can offer immediate and sustained GHG advantages.

Demand for alternatives to petroleum is increasing the production of biofuels from food crops such as corn, sugarcane, soybeans, and palms. As a result, land in undisturbed ecosystems, especially in the Americas and Southeast Asia, is being converted to biofuel production as well as to crop production when existing agricultural land is diverted to biofuel production. Such land clearing may be further accelerated by lignocellulosic biofuels, which will add to the agricultural land base needed for biofuels, unless those biofuels are produced from crops grown on abandoned agricultural lands or from waste biomass.

Soils and plant biomass are the two largest biologically active stores of terrestrial carbon, together containing ∼2.7 times more carbon than the atmosphere (1). Converting native habitats to cropland releases CO2 as a result of burning or microbial decomposition of organic carbon stored in plant biomass and soils. After a rapid release from fire used to clear land or from the decomposition of leaves and fine roots, there is a prolonged period of GHG release as coarse roots and branches decay and as wood products decay or burn (24).

We call the amount of CO2 released during the first 50 years of this process the “carbon debt” of land conversion. Over time, biofuels from converted land can repay this carbon debt if their production and combustion have net GHG emissions that are less than the life-cycle emissions of the fossil fuels they displace. Until the carbon debt is repaid, biofuels from converted lands have greater GHG impacts than those of the fossil fuels they displace. For crops with nonbiofuel coproducts (e.g., palm kernel oil and meal, soybean meal, or distillers' dry grains), we partition the carbon debt into a “biofuel carbon debt” and a “coproduct carbon debt” based on the market values of the biofuel and its coproducts (5).

We calculate how large biofuel carbon debts are, and how many years are required to repay them, for six different cases of native habitat conversion: Brazilian Amazon to soybean biodiesel, Brazilian Cerrado to soybean biodiesel, Brazilian Cerrado to sugarcane ethanol, Indonesian or Malaysian lowland tropical rainforest to palm biodiesel, Indonesian or Malaysian peatland tropical rainforest to palm biodiesel, and U.S. central grassland to corn ethanol (5) (table S1). These cases illustrate current impacts of biofuels on habitat conversion. Indonesia and Malaysia account for 86% of global palm oil production (6). Accelerating demand for palm oil is contributing to the 1.5% annual rate of deforestation of tropical rainforests in these nations (7). An estimated 27% of concessions for new palm oil plantations are on peatland tropical rainforests, totaling 2.8 × 106 ha in Indonesia (7). Brazilian Cerrado is being converted to sugarcane and soybeans, and the Brazilian Amazon is being converted to soybeans (810). Grassland in the United States, primarily rangeland or former cropland currently retired in conservation programs, is being converted to corn production. Rising prices for corn, wheat, and soybeans could cause a substantial portion of the 1.5 × 107 ha of land currently in the U.S. Conservation Reserve Program to be converted to cropland (11).

We estimated carbon debts by calculating the amount of CO2 released from ecosystem biomass and soils. Our analyses account for the amount of plant carbon released as CO2 through decomposition and combustion, the amount converted to charcoal (charcoal is not part of the carbon debt because it is recalcitrant to decomposition), and the amount incorporated into merchantable timber and other long-lived forestry products, which have a half-life of about 30 years (3, 12). Changes in carbon stores caused by land conversion and biofuel production, mainly from accelerated decomposition, were based on evaluation and synthesis of published studies in the relevant ecosystems (5). Our estimate of the carbon debt is conservative because timber products continue to decay after 50 years, but this time frame captures most of the carbon debt in systems with mineral soils.

Our results show that converting native ecosystems to biofuel production results in large carbon debts (Fig. 1A). We attribute 13, 61, and 17% of this carbon debt to coproducts for palm, soybeans, and corn, respectively (Fig. 1B) (5). The carbon debts attributed to biofuels (quantities of Fig. 1A multiplied by the proportions of Fig. 1B) would not be repaid by the annual carbon repayments from biofuel production (Fig. 1C and table S2) for decades or centuries (Fig. 1D). Converting lowland tropical rainforest in Indonesia and Malaysia to palm biodiesel would result in a biofuel carbon debt of ∼610 Mg of CO2 ha–1 that would take ∼86 years to repay (Fig. 1D). Until then, producing and using palm biodiesel from this land would cause greater GHG release than would refining and using an energy-equivalent amount of petroleum diesel. Converting tropical peatland rainforest to palm production incurs a similar biofuel carbon debt from vegetation, but the required drainage of peatland causes an additional sustained emission of ∼55 Mg of CO2 ha–1 yr–1 from oxidative peat decomposition (5) (87% attributed to biofuel; 13% to palm kernel oil and meal). After 50 years, the resulting biofuel carbon debt of ∼3000 Mg of CO2 ha–1 would require ∼420 years to repay. However, peatland of average depth (3 m) could release peat-derived CO2 for about 120 years (7, 13). Total net carbon released would be ∼6000 Mg of CO2 ha–1 over this longer time horizon, which would take over 840 years to repay. Soybean biodiesel produced on converted Amazonian rainforest with a biofuel carbon debt of >280 Mg of CO2 ha–1 would require ∼320 years to repay as compared with GHG emissions from petroleum diesel. The biofuel carbon debt from biofuels produced on converted Cerrado is repaid in the least amount of time of the scenarios that we examined. Sugarcane ethanol produced on Cerrado sensu stricto (including Cerrado aberto, Cerrado densu, and Cerradão), which is the wetter and more productive end of this woodland-savanna biome, would take ∼17 years to repay the biofuel carbon debt. Soybean biodiesel from the drier, less productive grass-dominated end of the Cerrado biome (Campo limpo and Campo sujo) would take ∼37 years. Ethanol from corn produced on newly converted U.S. central grasslands results in a biofuel carbon debt repayment time of ∼93 years.

Fig. 1.

Carbon debt, biofuel carbon debt allocation, annual carbon repayment rate, and years to repay biofuel carbon debt for nine scenarios of biofuel production. Means and SDs are from Monte Carlo analyses of literature-based estimates of carbon pools and fluxes (5). (A) Carbon debt, including CO2 emissions from soils and aboveground and belowground biomass resulting from habitat conversion. (B) Proportion of total carbon debt allocated to biofuel production. (C) Annual life-cycle GHG reduction from biofuels, including displaced fossil fuels and soil carbon storage. (D) Number of years after conversion to biofuel production required for cumulative biofuel GHG reductions, relative to the fossil fuels they displace, to repay the biofuel carbon debt.

Our analyses suggest that biofuels, if produced on converted land, could, for long periods of time, be much greater net emitters of greenhouse gases than the fossil fuels that they typically displace. All but two—sugarcane ethanol and soybean biodiesel on Cerrado—would generate greater GHG emissions for at least half a century, with several forms of biofuel production from land conversion doing so for centuries. At least for current or developing biofuel technologies, any strategy to reduce GHG emissions that causes land conversion from native ecosystems to cropland is likely to be counter-productive.

We also evaluated the possibility that U.S. cropland that has been retired from annual crop production and planted with perennial grasses may have a short payback time when converted to corn ethanol production, because these systems have already lost a substantial portion of their carbon stores. However, after abandonment from cropping, perennial systems gradually recover their carbon stores. For U.S. central grassland on cropland that has been enrolled in the U.S. Conservation Reserve Program for 15 years, we found that converting it to corn ethanol production creates a biofuel carbon debt that would take ∼48 years to repay (Fig. 1D).

If biofuels are to help mitigate global climate change, our results suggest that they need to be produced with little reduction of the storehouses of organic carbon in the soils and vegetation of natural and managed ecosystems. Degraded and abandoned agricultural lands could be used to grow native perennials for biofuel production (14, 15), which could spare the destruction of native ecosystems and reduce GHG emissions (Fig. 1). Diverse mixtures of native grassland perennials growing on degraded soils, particularly mixtures containing both warm-season grasses and legumes, have yield advantages over monocultures (14, 1618), provide GHG advantages from high rates of carbon storage in degraded soils (14, 19), and offer wildlife benefits (20). Monocultures of perennial grasses and woody species also can offer GHG advantages over food-based crops, especially if they are sufficiently productive on degraded soils (21), as can slash and thinnings from sustainable forestry, animal and municipal wastes, and crop residue (22).

Additional factors may influence biofuel impacts on GHG emissions. First, biofuel production can displace crops or pasture from current agricultural lands, indirectly causing GHG release via conversion of native habitat to cropland elsewhere (23). Second, improvements in biofuel production could reduce payback times (24, 25). Third, if land cleared for biofuel production had been accruing carbon (we assumed lands were at steady state), the debt would be increased by the loss of this future storage. Fourth, greater biofuel production might decrease overall energy prices, which could increase energy consumption and GHG release (26, 27).

Biofuel production that causes land clearing and GHG release may be favored by landowners who receive payments for biofuels but not for carbon management. Our results suggest that, in order to incorporate the costs of carbon emissions accurately, policy approaches to GHG emission reductions must be extended to include the net GHG emission or sequestration from land-use change. Indeed, the recently enacted U.S. Energy Independence and Security Act of 2007 specifies reductions in life-cycle GHG emissions, including land-use change, relative to a fossil fuel baseline. Moreover, it is important that international policy negotiations to extend the Kyoto Protocol beyond 2012 address emissions from land-use change due to increased demand for biofuels (28, 29).

Our results demonstrate that the net effect of biofuel production via clearing of carbon-rich habitats is to increase CO2 emissions for decades or centuries relative to the emissions caused by fossil fuel use. Conversely, biofuels from perennials grown on degraded cropland and from waste biomass would minimize habitat destruction, competition with food production, and carbon debts, all of which are associated with direct and indirect land clearing for biofuel production.

Supporting Online Material

www.sciencemag.org/cgi/content/full/1152747/DC1

Materials and Methods

Tables S1 and S2

References

References and Notes

View Abstract

Navigate This Article