Scholarly article on topic 'Energy and climate benefits of bioelectricity from low-input short rotation woody crops on agricultural land over a two-year rotation'

Energy and climate benefits of bioelectricity from low-input short rotation woody crops on agricultural land over a two-year rotation Academic research paper on "Earth and related environmental sciences"

CC BY-NC-ND
0
0
Share paper
Academic journal
Applied Energy
Keywords
{"Direct land use change" / "Eddy fluxes" / "Life cycle assessment" / "Energy ratio" / "GHG emissions"}

Abstract of research paper on Earth and related environmental sciences, author of scientific article — S. Njakou Djomo, O. El Kasmioui, T. De Groote, L.S. Broeckx, M.S. Verlinden, et al.

Abstract Short-rotation woody crops (SRWCs) are a promising means to enhance the EU renewable energy sources while mitigating greenhouse gas (GHG) emissions. However, there are concerns that the GHG mitigation potential of bioelectricity may be nullified due to GHG emissions from direct land use changes (dLUCs). In order to evaluate quantitatively the GHG mitigation potential of bioelectricity from SRWC we managed an operational SRWC plantation (18.4ha) for bioelectricity production on a former agricultural land without supplemental irrigation or fertilization. We traced back to the primary energy level all farm labor, materials, and fossil fuel inputs to the bioelectricity production. We also sampled soil carbon and monitored fluxes of GHGs between the SRWC plantation and the atmosphere. We found that bioelectricity from SRWCs was energy efficient and yielded 200–227% more energy than required to produce it over a two-year rotation. The associated land requirement was 0.9 m 2 kW h e - 1 for the gasification and 1.1 m 2 kW h e - 1 for the combustion technology. Converting agricultural land into the SRWC plantation released 2.8 ± 0.2tCO2e ha−1, which represented ∼89% of the total GHG emissions (256–272gCO2e kW h e - 1 ) of bioelectricity production. Despite its high share of the total GHG emissions, dLUC did not negate the GHG benefits of bioelectricity. Indeed, the GHG savings of bioelectricity relative to the EU non-renewable grid mix power ranged between 52% and 54%. SRWC on agricultural lands with low soil organic carbon stocks are encouraging prospects for sustainable production of renewable energy with significant climate benefits.

Academic research paper on topic "Energy and climate benefits of bioelectricity from low-input short rotation woody crops on agricultural land over a two-year rotation"

Contents lists available at SciVerse ScienceDirect

Applied Energy

journal homepage: www.elsevier.com/locate/apenergy

Energy and climate benefits of bioelectricity from low-input short ■. Cros£

rotation woody crops on agricultural land over a two-year rotation ^

S. Njakou Djomoa*, O. El Kasmioui3, T. De Grootea,b, L.S. Broeckxa, M.S. Verlinden3, G. Berhongaray3, R. Fichota, D. Zona a, S.Y. Dillen a,J.S. Kingc, I.A. Janssensa, R. Ceulemansa

a University of Antwerp, Department of Biology, Research Group of Plant and Vegetation Ecology, Universiteitsplein 1, B-2610 Wilrijk, Belgium b VITO, Unit Environmental Modelling, Boeretang 200, B-2400 Mol, Belgium

cNorth Carolina State University, Department of Forestry and Environmental Resources, Raleigh, NC 27695, USA

HIGHLIGHTS

• A full energy and GHG balance of bioelectricity from SRWC was performed.

• Bioelectricity was efficient; it reduced GHG by 52-54% relative to the EU non-renewable grid mix.

• Bioelectricity required 1.1 m2 of land kWh 1 land conversion released 2.8 ± 0.2 t CO2e ha 1

• SRWC reduced GHG emission when producing electricity during the 1st rotation period.

ARTICLE INFO ABSTRACT

Short-rotation woody crops (SRWCs) are a promising means to enhance the EU renewable energy sources while mitigating greenhouse gas (GHG) emissions. However, there are concerns that the GHG mitigation potential of bioelectricity may be nullified due to GHG emissions from direct land use changes (dLUCs). In order to evaluate quantitatively the GHG mitigation potential of bioelectricity from SRWC we managed an operational SRWC plantation (18.4 ha) for bioelectricity production on a former agricultural land without supplemental irrigation or fertilization. We traced back to the primary energy level all farm labor, materials, and fossil fuel inputs to the bioelectricity production. We also sampled soil carbon and monitored fluxes of GHGs between the SRWC plantation and the atmosphere. We found that bioelectricity from SRWCs was energy efficient and yielded 200-227% more energy than required to produce it over a two-year rotation. The associated land requirement was 0.9 m2 kWh^1 for the gasification and 1.1 m2 kWh^1 for the combustion technology. Converting agricultural land into the SRWC plantation released 2.8 ± 0.2tC02eha \ which represented ~89% of the total GHG emissions (256272 g C02e kWh^1) of bioelectricity production. Despite its high share of the total GHG emissions, dLUC did not negate the GHG benefits of bioelectricity. Indeed, the GHG savings of bioelectricity relative to the EU non-renewable grid mix power ranged between 52% and 54%. SRWC on agricultural lands with low soil organic carbon stocks are encouraging prospects for sustainable production of renewable energy with significant climate benefits.

© 2013 The Authors. Published by Elsevier Ltd. All rights reserved.

Article history: Received 5 November 2012 Received in revised form 4 May 2013 Accepted 7 May 2013

Keywords:

Direct land use change Eddy fluxes Life cycle assessment Energy ratio GHG emissions

1. Introduction

Renewable electricity represented 19.6% of the European Union (EU) grid mix power generation in 2009 [1]. Limited in natural resources, the EU imports large quantities of non-renewable fuels for

q This is an open-access article distributed under the terms of the Creative Commons Attribution-NonCommercial-No Derivative Works License, which permits non-commercial use, distribution, and reproduction in any medium, provided the original author and source are credited.

* Corresponding author. Tel.: +32 3 2652827; fax: +32 3 2652271. E-mail address: sylvestre.njakoudjomo@ua.ac.be (S. Njakou Djomo).

its electricity production. Shifting electricity production away from non-renewable fuels towards renewable energy sources could increase the diversity of the generation mix, reduce the import bills, and help to mitigate climate change [2,3].

Biomass has the potential to provide non-intermittent renewable base-load electricity and thus could contribute to meeting the EU's renewable energy targets in 2020 [4-6]. Within the biomass portfolio, short-rotation woody crops (SRWCs) with e.g. poplar (Populus) or willow (Salix) are candidates for large-scale application [7,8]. Compared to food crops SRWCs require low agri-chemical inputs and less fertile land. Wood chips from SRWC can be burned, gasified, or co-fired with coal to produce electricity. In

0306-2619/$ - see front matter © 2013 The Authors. Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.apenergy.2013.05.017

addition to the non-renewable electricity offsets, SRWC may also store carbon in agricultural soils [9,10], thus helping to reach the EU climate and renewable energy policy targets, whilst maintaining a reliable electricity system.

The greenhouse gas (GHG) performance of bioelectricity from SRWCs can also be affected by carbon stock changes due to land conversion from the previous land use. Converting agricultural lands to SRWC plantations may lead to losses of soil organic carbon (SOC) within the first two years following soil disturbance, although these changes are seldom statistically significant due to the high background variability in soil carbon stocks [10-12]. Such losses of carbon due to land use changes can compromise or even cancel the GHG saving benefits of bioenergy [13,14]. Also, biogenic methane (CH4) and nitrous oxide (N2O) emitted during crop production may outweigh the GHG benefits of SRWC-based bioelec-tricity [15]. Thus, an analysis of bioenergy impacts should consider its full life-cycle costs and benefits before policies aiming at large scale commercialization are adopted and implemented.

Much of the existing science on the energy and GHG performance of bioenergy has focused on liquid biofuels [16,17] with fewer studies investigating the energy and GHG balances of bio-electricity from SRWC [18-22]. The majority of these studies in turn have concentrated on CO2 emissions from fossil fuel combustion during management activities rather than biogenic GHG emissions from land use change. Direct land use change (dLUC) emissions have been particularly neglected [23], even though the initial loss in soil organic carbon (SOC) as well as emissions of CH4 and N2O from agricultural soils may be substantial [24]. Moreover, the accounting of farm labor inputs, and land requirement are missing in earlier studies. Furthermore, the lack of reliable measurements of GHG fluxes (CO2, CH4, and N2O) during the SRWC

Table 1

General inventory data for the production of short rotation woody crops. The columns from left to right denote the field activities, the implement used, tractor used, the operating rate, total fuel consumption, the area covered, and the material inputs.

production increases the degree of uncertainty of previous estimates.

Here we report and document quantitative data on the land requirement, energy yield and GHG offsets of bioelectricity production from SRWCs on former agricultural land. In order to obtain quantitative data on the land requirement, energy yield, and GHG offsets of bioelectricity from SRWC, we managed an industrial-sized SRWC plantation for bioelectricity production without supplemental irrigation or fertilization for two years. We included all energy and GHG emissions incurred during the production and conversion of biomass from SRWCs to bioelectricity.

2. Materials and methods

2.1. Site location, soil carbon, and plant material

An operational SRWC plantation was installed in Lochristi, Belgium (51°06'N, 3°51'E, 6.25 m asl). The long-term mean annual temperature was 9.5 °C and the average rainfall was 726 mm a-1 [25]. The soil texture in the top 30 cm was 86.8% sand, 11.4% clay, and 1.8% silt with a mean pH of 5.51 (Table S1). The region of the site is considered to be a sandy region with a poor drainage [26]. Historically, the site was cleared of the original forest in the early 20th century and has since been under agricultural land use, regularly plowed and fertilized at 200 kg N ha-1 for production of cereals (wheat and maize) and tuberous (potatoes) crops. Prior to deep plowing, we carried out detailed soil survey in March 2010 by analyzing soil samples taken at 110 locations, uniformly distributed over the agricultural land. Soils were sampled to a depth of 15 cm using core sampling. The conversion of the agricultural land

Activities Implement used Tractor used Operating rate Total fuel Total lubricant Coverage Input rates

Type Weight (kg) Type Weight (kg) Power (kW) (h ha-') consumption (l) consumption (l) (%) (unit ha-1)

Chemical HBS 800 Fendt V 7000 119 0.43 42 0.3 32 3.5 l

treatment 415

Deep plowing PF 820 Fendt V820 9000 157 0.95 105 2.1 32 -

Plowing CP 820 Fendt V820 9000 157 0.93 285 5.0 100 -

Flattening R 716 Fendt V415 7000 119 0.72 242 4.4 100 -

Planting LP 600 Massey F6480 5000 97 3.44 302 5.2 78 8000 cuttings

Application of HBS 800 Fendt 7000 119 0.21 54 0.9 78 0.3 l AZ500

PPEH V415

Application of CBS 200 Iseki TU 400 12 2.52 32 0.1 33 1 l Tomahawk

PEH 165

Application of CBS 200 Iseki TU 400 12 2.52 32 0.1 38 1 l Matrigon

PEH 165

Application of HBS 800 Fendt 7000 119 0.36 45 0.8 78 2.5 l Aramo

PEH V415

Mechanical ST 500 Fendt 5000 97 2.76 120 1.7 78 -

weeding V712

Mechanical GS - GS. FS 400 8 1.9 17.36 28 - 62 -

weeding

Mechanical GM - GM Rapid 237 14.6 6.11 45 - 62 -

weeding euro

Mechanical HDM - HDM 78 3.2 1.24 28 - 62 -

weeding

Manual - - - - - - - - 78 49.1 h

weeding

Harvesting E- 7000 JD 6920T 14000 110 1.66 710 1.0 78 -

harvester

The data were collected on-site. HBS, hardy bomb sprayer; HDM, heavy duty machine; GM, grass mulcher; CBS, custum build sprayer; LP, leek planter; R, roller; PF, plow 4 furrow; GS, grass strimmer; E-harvester, energy harvester; CP, chilser plow; ST, steketee; JD, John Deere; PPEH, pre-emergent herbice; PEH, post emergent herbicide. Deep plowing, plowing and flattening have been grouped into land preparation.

to SRWC plantation began on the 26th March 2010 with the application of glyphosate (3.5 l ha-1) to the soil, followed by deep plowing (up to 70 cm depth), and flattening before planting (Table 1). In April 2010, the SRWC plantation was established on 18.4 ha of this former agricultural land (Fig. S1). Twelve poplar and three willow genotypes representing different species and hybrids of Populus deltoides, P. maximowiczii, P. nigra and P. trichocar-pa and Salix viminalis, S. dasyclados, S. alba or S. schwerinii were planted at a density of 8000 cuttings ha-1. During the first months after planting, chemical, mechanical, and manual weed controls were performed as SRWCs are exposed to weed competition during the first growing season. No irrigation or fertilization was applied in this SRWC plantation. Before harvesting in 2012, we resampled the soil to 15 cm depth at 16 sampling points, with each point located less than 40 m from one of the eight selected sampling points of the initial soil survey of March 2010 [27]. All soil samples were oven-dried at 60 °C for 72 h and analyzed for soil carbon concentration in an elemental analyzer (Carlo Erba Instruments, Italy). The SOC stock was estimated by multiplying the SOC content of the first 15 cm by the bulk density of that soil layer (Table 2).

2.2. GHG flux data

Greenhouse gas flux measurements were carried out from June 2010 to December 2011 using the eddy covariance (EC) method. The EC system consists of a 3-D sonic anemometer, a closed-path C02/H20 analyzer (Li-700, Li-Cor Inc.), and closed-path N20/C0 (Los Gatos-908, Los Gatos Research), and CH4 (Los Gatos DLT-100, Los Gatos Research) analyzers mounted on a 5.8 m high micromete-orological flux tower, in the plantation (Fig. S1). Raw data were recorded at a 10 Hz sampling rate; momentum, energy, C02, N20, and CH4 fluxes were derived. The C02, N20 and CH4 fluxes were then converted to densities using a CR5000 data logger. Data processing was done following the generally accepted EC protocols [28] including among other a 2D coordinate rotations of wind components. The EC fluxes were calculated as the mean covariance between C02, N20, and CH4 concentrations and fluctuations in the vertical wind speed over 30 min after removing spikes in raw data and corrections for air density fluctuations [29]. Individual data points were removed when the following criteria were met: (i) for C02 and for the wind velocity components (u, v and w) when the standard deviation of the 30 min mean was higher than 10, for N20 when it was higher than 8, for H20 when it was higher than 1, for CH4 when it was higher than 0.3; (ii) when CH4and N20 minimum concentrations were less than zero, and (iii) when data points came from outside of the footprint of interest (wind direction between 50 °C and 250 °C) [30,31]. Gaps in data were filled using different techniques. For C02 data the method of Reichstein et al. [32] was applied. As no standard gap-filling method exists for CH4 and N20, fluxes of CH4 and N20 were linearly interpolated in periods with similar emission rates [33]. An overall annual GHG budget was computed by cumulating the net ecosystem exchange (NEE), N20, and CH4 over each year of the study. C02, CH4 and N20 were converted to C02 equivalent using the IPCC conversion factors [34]. A more detailed description of EC flux calculations can be found in Zona et al. [35].

2.3. Life cycle assessment

To identify and compare the GHG emissions of the investigated bioelectricity system to those of the EU non-renewable grid mix electricity generation (reference system), a life cycle assessment (LCA) was performed. In this analysis, the functional unit was 1 kWhe of bioelectricity. We included all relevant processes of bio-electricity production - from agrichemicals production, land preparation, planting, weeding, harvest and chipping, to the final conversion of chips to electricity - and all transportation needed within the system boundary (Fig. 1). Capital equipment was also included. Since irrigation and fertilization were not applied in this plantation, the unit processes of irrigation and fertilizer production and application were excluded from the system boundary. Moreover, the agricultural land in this study was not in conservation tillage; therefore, carbon was not being stored in the soil prior to its conversion to SRWC plantation. The foregone carbon sequestration (i.e., the ongoing carbon storage that is given up by devoting the agricultural land in this study to the production of SRWC for bioelectricity) was zero and therefore not included in the system boundary.

The system boundary of the EU non-renewable grid mix electricity generation included the extraction, transport, refining, storage, and conversion of non-renewable fuels to electricity (Fig. 1). Environmental impacts were based on the Impact 2002 + method [36], and were limited only to land requirement, energy balance, and GHG emissions of the bioelectricity production. LCA modeling was performed in Simapro 7.1 [37]. Furthermore, the energy ratio and GHG savings of the system were assessed.

2.3.1. Management input data

We inventoried all activities of the SRWC biomass production in the field (Fig. 1). Using a book keeping method we measured the amount of diesel and lubricant consumed to carry out each activity (e.g., plowing, weeding, planting, and harvesting). We also quantified the types and amount of chemicals used for weeding, as well as the amount of cuttings used in SRWC biomass production. Besides, we collected data on the lifespan, weight of implements and tractors used, and operation-time for each farming activity (Table 1). In addition, we collected data on the production of cuttings (Table 3). We also gathered data on vehicle types (truck or van), weights carried, and the distance traveled to transport farm materials (e.g. chemicals and cuttings) and tractors from the regional storage facilities to the SRWC plantation (Table S2). We further assumed the trucks or vans returned empty. Solar energy, which drives the build-up of SRWC biomass, was excluded from the system boundary. However, unlike in most studies, the human labor input for manual weeding was considered. To estimate the human energy input, we quantified the amount of person-hours of labor for manual weeding (Table 1), and multiplied it by the energy expended (1.9 MJh-1 [38]) by a male worker to carry out manual weeding. No attempt was made to include the human labor associated with manufacturing of farm equipment and agrichemicals. All agricultural input data were collected in the field and referred to the 2010-2012 operations.

Table 2

Soil carbon stocks at depth of 15 cm and change in carbon stock due to land conversion from agricultural land to SRWC plantation. The positive value of the relative change in S0C stock denotes a loss (significant at p < 0.001) in carbon.

Land use type Sampling depth (cm) Bulk density (kg m-3) Carbon conc. (kg C kg-1 soil) Total carbon (t C ha-1) AS0C2010-2012 (t C ha-1)

Agricultural (2010) (n = 8) 0-15 1298 ±169 0.015 ± 0.004 28.38 ± 7.07

SRWC plantation (2012) (n = 16) 0-15 1519 ± 59 0.009 ± 0.002 20.79 ±3.33 7.59 ± 7.81

S0C, soil organic carbon; SRWCs, short-rotation woody crops; AS0C2010-2012, change in soil organic carbon; n, number of samples.

Fig. 1. System boundary of the bioelectricity production as well as of the EU non-renewable grid mix electricity used in this study. The boxes represent unit processes (or activities) and the arrows refer to material and energy flows. The solid lines represent the system boundary. Land preparation includes: deep plowing (DP), plowing (P), and flattening (F).

Table 3

Inputs for the production of SRWC cuttings. The column from left to right denote the field activities, the implement used, tractor used, the operating rate, total fuel consumption. These data are based on 1 ha land use and 15,000 plants at the nursery. The annual average production of cuttings (only three harvests) is 153,300 cuttings per hectare.

Activities Implement used Tractor used Operating rate (h Total fuel consumption Input rates (unit

Types Weight Types Weight Power ha-1) (l) ha-1)

(kg) (kg) (kW)

Plowing 4 1800 Fendt 6000 104 2 24 -

Furrow

Flattening Roller 1200 Fendt 6000 104 1.5 18 -

Fertilising Sprayer 100 Lamborghini 1150 31 1 7 80 kg N

Fertilising Sprayer 100 Lamborghini 1150 31 1 7 100 kgP

Fertilising Sprayer 100 Lamborghini 1150 31 1 7 60 kg K

Fertilising Sprayer 1000 Fendt 6000 104 1 11 1000 kg CaCO3

Chemical weeding Sprayer 100 Lamborghini 1150 31 1 7 11 AZ 500

Chemical weeding Sprayer 100 Lamborghini 1150 31 1 7 11 Kerb50

Chemical weeding Sprayer 100 Lamborghini 1150 31 1 7 11 Basta

Mechanical - 130 Deutz 2150 45 2.5 20 -

weeding Agrocompacts

Manual weeding - - - - - - - 65 h

Note: The data were obtained from the Research Institute for Nature and Forest (INBO). SRWCs, short-rotation woody crops.

2.3.2. Data on energy conversion technologies and allocation method We assumed that woody biomass chips were used in combined heat and power (CHP) plants. Two existing CHP plants were modeled for converting SRWC chips to bioelectricity: (i) a gasification plant which gasified 35.5 kton a-1 of dried SRWC chips at 30% moisture to produce 40.2 GWhe a-1 at 27.5% efficiency and (ii) a combustion plant that burned 31.3 kton a-1 of dried SRWC chips at 30% moisture to produce 25.9 GWhe a-1 at 22% efficiency (Table S3). Bioheat is also produced during power generation in CHP plants (Table S3). Since bioheat has a positive economic value and displaces heat that would otherwise be supplied from other sources, inputs (e.g. land and en-

ergy use) and outputs (GHG emissions) need to be allocated between bioelectricity and bioheat. We used the exergy-based allocation to partition inputs/outputs between bioelectricity and bioheat. First, we assumed an ambient temperature of 10 °C (283 K) and a steam temperature of 120 °C (393 K) for both the gasification and combustion. Next, we estimated the Carnot factor as indicated in Table S4, and then calculated the bioheat exergy by multiplying this Carnot factor (i.e. 0.27) by the annual amount of bioheat produced by the gasification (83.9 GWh) and combustion (82.3 GWh) technologies, respectively. For bioelectricity, we assumed the exergy is equal to its

energy content and thus the share of bioelectricity of the total delivered exergy (Table S4).

2.3.3. Energy balance and GHG savings

All the collected data were normalized to the functional unit (i.e., 1 kWhe), entered into Simapro 7.1, and modeled into environmental inputs and outputs. Simulation results were then exported from Simapro 7.1 to an Excel spreadsheet where calculations of the energy balance and GHG savings were performed. We calculated the energy ratio by dividing the energy content of bio-electricity output (i.e., 1 kWhe = 3.6 MJe) by the sum of all fossil energy inputs needed to produce one unit of bioelectricity. To estimate the GHG emission savings, we first multiplied the emission factors of electricity from natural gas, coal, uranium, and oil derived from Ecoinvent [39] by the fraction of natural gas fired (29%), coal burning (32%), nuclear (35%), and oil fired power (4%) making up the EU non-renewable grid mix electricity in 2009 [40]. We then summed-up these products to obtain the GHG emission rates for the EU non-renewable grid mix electricity in 2009 (Table S5). Finally, we estimated the GHG emission savings by comparing the GHG emission rates for the SRWC-bioelectricity chain to that for the grid mix electricity in the EU in 2009.

3. Results

3.1. Biomass yield

The mean yield after two years of growth was 4 ton ha-1 a"1. Considerable tree mortality (~18%) was observed in the establishment year [41]. Because tree mortality was evenly distributed across the plantation, and established trees occupied the vacant spaces, no large gaps occurred. At the end of two years of growth, about 114 tons of biomass were harvested from 14.2 ha and transported to the bioelectricity plant. The chemical composition and the measured heating value (19.5 MJ kg"1) of the harvested SRWC chips from our plantation are summarized in Table S6.

3.2. Energy requirement and energy ratio

The total energy input to produce one unit of bioelectricity was 1.1 MJ kWh"1 for the gasification and 1.2 MJ kWh"1 for the com-

Fig. 2. Energy balance of the biomass gasification (left) and biomass combustion (right) technology investigated in this study. The black bars represent the energy output whereas the stacked bars represent the energy input items of each biomass conversion technology. Land preparation includes: deep plowing, plowing, and flattening. ER: energy ratio (output-input ratio). Data presented are based on a biomass yield of 4 odt ha"1 a"1, and a single two-year rotation which includes only one harvest.

bustion technology. The breakdown of the total energy input across the different components of the bioelectricity production is shown in Fig. 2. Land preparation was the activity that consumed the most energy (24%), followed in decreasing order by harvesting (20%), production of cuttings (18%), weeding (17%) and planting (10%). The contribution of cutting production to the total energy input was high because the production of SRWC cuttings covered only three harvests. Facility construction and transport were the activities that consumed the least amount of energy (Fig. 2). For both conversion technologies (gasification and combustion), the energy output was much higher than the total energy inputs to produce 1 kWhe of bioelectricity. The energy ratio (ratio of the energy content of 1 kWhe = 3.6 MJe of bioelectricity over the total energy input for its production) was 3 for the combustion and 3.3 for the gasification technology (Fig. 2). Thus, the bioelectricity from SRWC yielded 200-227% more energy than the energy invested in its production.

3.3. Cumulative dLUC emissions

Conversion of agricultural land to an SRWC plantation resulted in a loss of SOC of ~27.8 ± 9.6 ton CO2e ha"1 in the top 15 cm of soil over the two-year period (Table 2). The cumulative NEE measured by eddy covariance is shown in Fig. 3. High amounts of CO2 were released after the plantation establishment in 2010 and during the autumn-winter period, whereas much of the CO2 uptake by the SRWC canopy occurred during the growing season. Integrated over the measuring period, the cumulative NEE which also includes the loss of SOC was "0.8 ± 0.6 ton CO2e ha"1 (Fig. 3). This suggests that the SRWC plantation was a sink of CO2 despite the initial loss in SOC. No seasonal trends in N2O and CH4 fluxes were observed in this study during the entire measurement period. Most of the N2O emissions occurred in July-August 2010, and the cumulative amount changed very little thereafter. CH4 fluxes were very small throughout the measurement period (Fig. 3). The cumulative N2O and CH4 emissions were 2.39 ± 0.52 ton CO2e ha"1 and 1.12 ± 0.07 ton CO2e ha"1, respectively (Fig. 3). These positive cumulative fluxes more than offset the CO2 uptake, and turned the SRWC plantation from a net CO2 sink into a small source of GHGs. At the end of the study period, the cumulative dLUC GHG emissions, taking into account the CO2, N2O and CH4 fluxes, amounted to 2.8 ± 0.2 ton CO2e ha"1 (Fig. 3). This indicates that N2O and CH4 played an important role in GHG emissions associated with dLUC at our site. Thus, the soil N2O and CH4 fluxes may reduce the potential sink strength of SRWC plantations during the first two years of culture.

3.4. GHG emissions and savings of bioelectricity

The combustion of SRWC chips in an existing biomass-fired power station resulted in a total GHG emission of about 272 g CO2e kWh"1. The gasification of these chips showed a lower total GHG emission of ~256 g CO2e kWh"1 (Fig. 4). For both the gasification and the combustion technology, dLUC accounted for 89% of the total GHG emissions, while the emissions from all the other processes associated with bioelectricity production made-up the remaining fraction (11%) (Fig. 4). The GHG emission reduction compared to the EU non-renewable grid mix electricity was 52% for the combustion and 54% for the gasification technologies. The GHG savings even reached 67% and 69% for the combustion and gasification technologies, respectively, when the EU fossil fuels grid mix electricity (i.e., excluding nuclear power) was considered as a baseline (Fig. 4). Therefore, converting agricultural land into a low-input SRWC plantation did not negate the GHG emission benefits of bioelectricity production regardless of the conversion technology chosen and of the EU grid mix electricity displaced.

Fig. 3. Cumulative GHG fluxes over the measuring period (2010-2011). The NEE is represented by the black line, N20 fluxes by the dash-dot grey line, CH4 fluxes by the light grey, whereas the cumulative GHGs is represented by the dotted grey line. Positive values denote the loss from ecosystem and negative values denote uptake. GHGs: greenhouse gases, NEE: net ecosystem exchange, N20: nitrous oxide, CH4: methane.

Fig. 4. Greenhouse gas emissions relative to the EU non-renewable and fossil fuels grid mix electricity. The bars represent the total GHG emission of each bioelectricity production technology whereas the dotted- and solid lines above the bars represent the GHG emission of the EU non-renewable as well as fossil fuels (i.e., excluding nuclear power) grid mix electricity production respectively. Land preparation includes: deep plowing, plowing, and flattening. Data presented are based on a biomass yield of 4 odt ha-1 a-1, and a single two-year rotation which includes only one harvest.

3.5. Land requirement

The total land requirement for bioelectricity production was 1 m2 kWh-1 for the combustion and 0.9 m2 kWh-1 for the gasification technology (Fig. S2). For both the combustion and the gasification technology, the land requirement of the SRWC chips accounted for 95% of the total land requirement while the land needed for the production of cuttings used in the site establishment accounted for only 5% of the total land requirement (Fig. S2). Because of its high electrical efficiency, gasification reduced the total land requirement by 10% compared to combustion technology. This reduction suggests that conversion efficiency played a considerable effect on the land requirement.

4. Discussion

The biomass yield was 50-60% lower than the average attainable yield in Europe (8-10 ton ha-1 a-1) [42], which may be

explained by the young age of the plantation, the soil type, the low planting density, and possibly other factors such as weather, and weed pressure. It has been shown that SRWC stands invest more in their root systems and less in aboveground growth early in stand development [43,44]. Moreover, unlike in most European studies, neither fertilizer nor irrigation was used at our site. Given that the number of shoots per stool and stem diameter usually increase between the first and subsequent harvests [45], an increase in yield is likely in the second harvest of our SRWC plantation.

0ur bioelectricity production systems yielded 200-227% more energy than required for its production, which implies that bio-electricity from SRWCs grown on agricultural land is a valuable energy substitute, and producing more bioelectricity from SRWCs could displace non-renewable fuel imports, which would increase energy security. The high net energy yield of bioelectricity from SRWC in this study was attributable to the low inputs (no fertilization and no irrigation) during the feedstock production phase and the use of bioheat as a co-product of bioelectricity production. All recent studies showed that bioelectricity from SRWCs has a positive energy balance, and reported energy ratios ranged from 3 to 16 [46]. 0ur results, even though they are for a single two-year rotation with only one harvest, are consistent with other studies that show that bioelectricity has a positive energy balance. However, our estimated energy ratio (3-3.3) was at the lower end of this range. The main reason for the low energy ratio estimate is the low biomass yield during the first two years of tree growth.

The loss of S0C from the SRWC plantation may be attributed either to decreased organic inputs to the soil relative to decomposition early in the SRWC establishment [47] or to the effect of tillage during site preparation which renders more S0C vulnerable to decomposition and thus triggered the release of S0C [48]. Carbon dioxide fluxes showed high seasonal variability, driven primarily by the SRWC response to seasonal changes in temperature and soil moisture (Fig. 3). Nitrous oxide fluxes were high, in which a peak emission of ~60% of the annual N20 flux occurred after a single rainfall event in August 2010 (Fig. 3). The high N20 emissions were contrary to what we expected given that our SRWC plantation was unfertilized. A likely explanation is that decades of intensive fertilization and very high atmospheric N deposition - due to high ammonia and nitrogen oxides from dense livestock and traffic,

respectively, in Flanders - led to a high N content of the soils (9.3 ton N ha-1) some of which could not be fixed by plants and was converted to N2O during nitrification and denitrification processes. The observed fluxes also showed that our SRWC plantation was a source of CH4. These CH4 fluxes could be attributed to a high water table and low atmospheric evaporative demand at our site in winter, causing soil saturation, favoring methanogenesis and restricting the oxidation of CH4 by methanotrophs.

The cumulative dLUC GHG emissions occurring during the two-year rotation was 2.8 ± 0.2 ton CO2e ha-1 (Fig. 3), indicating that the SRWC plantation was a net source of GHGs due to low biomass yield, and low input from leaf litter and root turnover relative to soil carbon loss. However, it is likely that our SRWC plantation may become a net sink of GHGs in the longer term. In fact, Arevalo et al. [11] showed that at least four years were necessary for a SRWC plantation on croplands to reach its pre-plantation carbon level and become a net sink of GHGs. Our estimate of dLUC GHG emissions was much lower than that for conversion of grassland to a corn plantation (~12 ton CO2e ha-1) [49], and for establishment of fertilized SRWC on pastureland (7-11 ton CO2e ha-1) [50], and lower than for conversion of abandoned cropland to prairie biomass (~6 ton CO2e ha-1) [13]. Our estimate of dLUC GHG emissions was low because the SRWC plantation was established on agricultural land that contained depleted SOC pools due to repeated tillage. Thus, our dLUC estimate was limited only to emissions from soil disturbance during land preparation. The total GHG emissions of bioelectricity production (256-272 g CO2e -kWh-1) in this study were well above the maximum value (39132 g CO2e kWhe-1) reported in [46] because of the inclusion of dLUC GHG emissions in our system boundary. Also, differences in SRWC yields, assumptions about efficiencies of conversion technologies, as well as the allocation method used in this study partly explained the differences with previous analyses. When leaving out the contribution of dLUC (~89% of total GHG emissions), our estimate of GHG emissions from bioelectricity fell to 29-31 g CO2e -kWh-1. This latter range compared well with estimates reported in [46] since dLUC emissions were ignored in all articles reviewed in that study. On a kWhe basis, bioelectricity from SRWC on agricultural lands reduced emissions by at least 52-54% compared to the current EU non-renewable grid mix power (Fig. 4). Thus, despite entailing dLUC GHG emissions (Fig. 4), bioelectricity still provided immediate GHG benefits because SRWC were grown on agricultural lands and were converted to electricity using efficient technologies.

The EU has committed to producing 20% of consumed energy from renewable energy sources by 2020 [4]. To meet this target,

it has been projected that about 232 TWhe would come from biomass [51]. Considering that it takes about m2 kWh-1 electricity (Fig. S2), and assuming that electricity from cultivated woody biomass represents 15% of the projected amount [52], about 34,800 km2 of land (~2% of the EU's total utilized agricultural area) would be required to meet the EU 2020 bioelectricity target from cultivated woody biomass. Unless yield of food crops is increased on existing croplands to meet the growing food and feed demand, it may be difficult to devote ~2% of EU's utilized agricultural land to SRWC. However, yield increase in the future years has the potential to decrease the land requirement for SRWC [53,54].

A number of sensitivity analyses were carried out to assess the influence of some key inputs variables and assumptions and the robustness of the obtained results. The elasticity method (i.e., the ratio of the change in the results to the change in data) was used to perform the sensitivity analysis. When the biomass yield (4 ton ha-1 a-1) in this study was doubled, we found that the energy demand and GHG emissions of bioelectricity production were reduced significantly (Table 4). A sensitivity analysis on initial SOC content revealed that, even if SRWCs were grown on an agricultural land containing 20% more SOC than at our site, the overall GHG reduction would still be 44-47% relative to the current EU non-renewable grid mix power (Table 4). We also hypothesized a case where the electrical conversion efficiencies of both the gasification and combustion were reduced by 20%. In that scenario, we found that the land requirement would increase by 10% while the energy ratio and the GHG saving would decrease by a similar percentage. But bioelectricity would still provide energy and GHG benefits (Table 4). Finally, when the energy-based allocation method was adopted, the land requirement, energy demand, and GHG emissions of bioelectricity were strongly reduced for both conversion technologies (Table 4).

4.1. Limitations and cautions to the interpretation of results of this study

One limitation of the current study is that it only addresses the dLUC GHG emissions of bioelectricity. However, growing SRWC on agricultural lands for bioelectricity may trigger land conversion elsewhere in the world, releasing GHGs through indirect land use (iLUC) [14,55-57]. Therefore, a complete assessment needs to include both dLUC and iLUC. Another limitation is that this study considers only SRWC from tilled agricultural land. Grassland, non-tilled agricultural land as well as set-aside lands are currently a sink of GHGs [58,59]. Consequently, converting these lands to SRWC plantations would result in significant dLUC GHG emissions,

Table 4

Sensitivity analysis of key parameters on results.

Parameters Scenario Gasification Combustion

Land requirement (m2 kWh"1) Energy demand (MJ kWh,:1) GHG emissions (gCO2 kWhe:1) Land requirement (m2 kWhjT1) Energy demand (MJ kWhjT1) GHG emissions (gCO2kWhe:1)

Yield Increase 0.452 0.575 130.2 0.481 0.622 140.6

Electrical Increase 0.8 0.9 213.3 0.8 0.9 227.1

conversion +20%

efficiency Decrease 1.1 1.4 298.2 1.2 1.4 316.8

Initial carbon stock Increase na na 300.0 na na 318.2

Decrease na na 211.9 na na 225.4

Allocation method Energy 0.5 0.6 128.1 0.4 0.5 121.4

approach

NB, 100% decrease in yield makes no sense and was therefore not considered in sensitivity analysis; na, not applicable.

which in turn would result to little or no GHG savings. Thus, our study likely understates the disadvantage of bioelectricity production relative to facilities that would obtain SRWCs from set-aside or grasslands. Finally, our study assesses the life-cycle of existing bio-electricity technologies that currently have low to medium electrical efficiency. The adoption of advanced gasification/combustion technologies (i.e. ge p 35) changes the results of this analysis. Thus, the land requirement, energy demand, and GHG emissions reported here reflect today's average technologies. Despite these limitations, our study suggests that (i) in areas where SRWCs can be grown sustainably, even with low yields and one two-year rotation, there is a positive energy balance; and (ii) bioelectricity would contribute to GHG mitigation in the power sector if appropriate lands, feedstocks, and the correct conversion technologies were used, and if the SRWC plantation was maintained as a low-input system.

5. Conclusion

By combining field measurements and LCA approach we showed that low input SRWC plantations on agricultural lands for bioelectricity production resulted in immediate GHG savings relative to grid mix electricity. Consequently SRWCs that come from agricultural land with low carbon stocks are an encouraging prospect for sustainable production of renewable energy with significant climate benefits.

Acknowledgements

This research has received funding from the European Research Council as ERC Grant Agreement Nr. 233366 (POPFULL) and from a Marie Curie International Re-integration Grant (MIRG) as Grant Agreement Nr. 268257 under the EC 7th Framework Programme (FP7/2007-2013). Further funding was provided by the Flemish government through the Hercules Foundation as Infrastructure Contract ZW09-06 and through the Methusalem Program. OEK and SYD were researchers from the Flemish Research Foundation (FWO), and TDG received a grant from FWO and the Flemish VITO Institute. GB held a grant from the Erasmus-Mundus Consortium EADIC (Lot 16) financed by the European Union Mobility Programme (2009-1655/001-001). JSK was supported at UA by the International Francqui Foundation and by the Fulbright Program of the Commission for International Exchange. We gratefully acknowledge the excellent technical support of Joris Cools and the logistic support of Kristof Mouton at the field site. Finally, we thank Dr. Florian Gahbauer for correcting English language, and the two anonymous reviewers for their constructive comments and valuable suggestions on an earlier version of the manuscript.

Appendix A. Supplementary material

Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.apenergy.2013. 05.017.

References

[1] Eurelectric. Statistics and prospects for the European electricity sector. Report no. 2008-542-0001, Union of the Electricity Industry. Brussels (Belgium); 2011.

[2] Yan J. Biofuels in Asia. Appl Energy 2009;86:S1.

[3] Mobini M, Sowlati T, Sokhansanj S. Forest biomass supply logistics for a power plant using the discrete-event simulation approach. Appl Energy 2011;88:1241.

[4] EU. Directive 2009/28/EC of the European parliament and of the council of 23 April 2009 on the promotion of the use of energy from renewable sources and

amending and subsequently repealing directives 2001/77/EC and 2003/30/EC. O.J.O.T.E Union (ed. Union OJOTE), EU. Brussels (Belgium); 2009.

[5] Dornburg V, van Vuuren D, van de Ven G, Langeveld H, Meeusen M, Banse M, et al. Bioenergy revisited: key factors in global potentials of bioenergy. Energy Environ Sci 2010;3:258.

[6] Kalt G, Kranzl L Assessing the economic efficiency of bioenergy technologies in climate mitigation and fossil fuel replacement in Austria using a techno-economic approach. Appl Energy 2011;88:3665.

[7] Liberloo M, Luyssaert S, Bellasten V, Djomo SN, Lukac M, Calfapietra C, et al. Bioenergy retains its mitigation potential under elevated CO2. Plos One 2010;5:e11648,1.

[8] Srirangan K, Akawi L, Moo-Young M, Chou CP. Towards sustainable production of clean energy carriers from biomass resources. Appl Energy 2012;100:172.

[9] Garten CT. Soil carbon storage beneath recently established tree plantations in Tennessee and South Carolina. USA Biomass Bioenerg 2002;23:93.

[10] Grigal DF, Berguson WE. Soil carbon changes associated with short-rotation systems. Biomass Bioenerg 1998;14:371.

[11] Arevalo CBM, Bhatti JS, Chang SX, Sidders D. Land use change effects on ecosystem carbon balance: From agricultural to hybrid poplar plantation. Agric Ecosyst Environ 2011;141:342.

[12] Hansen EA. Soil carbon sequestration beneath hybrid poplar plantations in the North Central United-States. Biomass Bioenerg 1993;5:431.

[13] Fargione J, Hill J, Tilman D, Polasky S, Hawthorne P. Land clearing and the biofuel carbon debt. Science 2008;319:1235.

[14] Searchinger T, Heimlich R, Houghton RA, Dong FX, Elobeid A, Fabiosa J, et al. Use of US croplands for biofuels increases greenhouse gases through emissions from land-use change. Science 2008;319:1238.

[15] Crutzen PJ, Mosier AR, Smith KA, Winiwarter W. N2O release from agro-biofuel production negates global warming reduction by replacing fossil fuels. Atmos Chem Phys 2008;8:389.

[16] Schmer MR, Vogel KP, Mitchell RB, Perrin RK. Net energy of cellulosic ethanol from switchgrass. Proc Natl Acad Sci USA 2008;105:464.

[17] Hill J, Nelson E, Tilman D, Polasky S, Tiffany D. Environmental, economic, and energetic costs and benefits of biodiesel and ethanol biofuels. Proc Natl Acad Sci USA 2006;103:11206.

[18] Heller MC, Keoleian GA, Volk TA. Life cycle assessment of a willow bioenergy cropping system. Biomass Bioenerg 2003;25:147.

[19] Gasol CM, Gabarrell X, Anton A, Rigola M, Carrasco J, Ciria P, et al. LCA of poplar bioenergy system compared with Brassica carinata energy crop and natural gas in regional scenario. Biomass Bioenerg 2009;33:119.

[20] Adler PR, Del Grosso SJ, Parton WJ. Life-cycle assessment of net greenhouse-gas flux for bioenergy cropping systems. Ecol Appl 2007;17:675.

[21] Matthews RW. Modelling of energy and carbon budgets of wood fuel coppice systems. Biomass Bioenerg 2001;21:1.

[22] Styles D, Jones MB. Energy crops in Ireland: quantifying the potential life-cycle greenhouse gas reductions of energy-crop electricity. Biomass Bioenerg 2007;31:759.

[23] Tonini D, Astrup T. LCA of biomass-based energy systems: a case study for Denmark. Appl Energy 2012;99:234.

[24] Larson ED. A review of life-cycle analysis studies on liquid biofuel systems for the transport sector. Energy Sust Develop 2006;10:109.

[25] RMIB. Klimaat in de wereld, klimatogram Gent-Melle <http://www.meteo.be/ meteo/view/nl>.

[26] Van Ranst E, Sys C. Eenduidige legende voor de digitale bodemkaart van vlaanderen (schaal 1:20 000), Laboratorium voor bodemkunde, Ghent (Belgium); 2000 [in Dutch].

[27] Verlinden MS, Broeckx LS, Wei H, Ceulemans R. Soil CO2 efflux in a bioenergy plantation with fast-growing Populus trees - influence of former land use, inter-row spacing and genotype. Plant Soil 2013:1. http://dx.doi.org/10.1007/ s11104-013-1604-5.

[28] Aubinet M, Chermanne B, Vandenhaute M, Longdoz B, Yernaux M, Laitat E. Long term carbon dioxide exchange above a mixed forest in the Belgian Ardennes. Agric For Meteorol 2001;108:293.

[29] Webb EK, Pearman GI, Leuning R. Correction of flux measurements for density effects due to heat and water-vapor transfer. Quart J Royal Meteorol Soc 1980;106:85.

[30] Rannik U, Markkanen T, Raittila J, Hari P, Vesala T. Turbulence statistics inside and over forest: influence on footprint prediction. Bound-Layer Meteorol 2003;109:163.

[31] Foken T, Wichura B. Tools for quality assessment of surface-based flux measurements. Agric For Meteorol 1996;78:83.

[32] Reichstein M, Falge E, Baldocchi D, Papale D, Aubinet M, Berbigier P, et al. On the separation of net ecosystem exchange into assimilation and ecosystem respiration: review and improved algorithm. Glob Change Biol 2005;11:1424.

[33] Mishurov M, Kiely G. Gap-filling techniques for the annual sums of nitrous oxide fluxes. Agric. For. Meteorol. 2011;151:1763.

[34] IPCC. The physical science basis. Solomon S et al. editors. Cambridge University Press: Cambridge (United Kingdom); 2007.

[35] Zona D, Jassens IA, Aubinet M, Vicca S, Gioli B, Fichot R, et al. Fluxes of the greenhouse gases (CO2, CH4, and N2O) above a short-rotation poplar planation after conversion from agricultural land. Agricultural and Forest Meteorology 2013;169:100.

[36] Jolliet O, Margni M, Charles R, Humbert S, Payet J, Rebitzer G, et al. IMPACT 2002+: a new life cycle impact assessment methodology. Int J Life Cycle Assess 2003;8:324.

[37 [38

[43 [44 [45

Pre-Consultant. Simapro 7.1 LCA software. 7.1 ed. Amersfoort, The [48 Netherlands: Preconsultants B.V.; 2006.

Singh H, Mishra D, Nahar NM, Ranjan M. Energy use pattern in production [49 agriculture of a typical village in and zone India: part II. Energy Convers Manage 2003;44:1053.

Frischknecht R, Jungbluth N, Althaus H, Bauer C, Doka G, Dones R, et al. [50

Implementation of life cycle impact assessment methods. Dubendorf,

Switzerland: Swiss Centre for Life Cycle Inventory; 2007.

EEA. Share of electricity production by fuel type in 2008. European

environment Agency, ENER027_Fig03_table_2010.xls; 2012 <http:// [51

www.eea.europa.eu/data-and-maps/figures/share-of-electricity-production-

by-5>.

Broeckx LS, Verlinden MS, Ceulemans R. Establishment and two-year growth

of a bio-energy plantation with fast-growing Populus trees in Flanders [52

(Belgium): effects of genotype and former land use. Biomass Bioenerg

2012;42:151. [53

Don A, Osborne B, Hastings A, Skiba U, Carter MS, Drewer J, et al. Land-use

change to bioenergy production in Europe: implications for the greenhouse [54

gas balance and soil carbon. Glob Change Biology Bioenerg 2012;4:372.

Kern CC, Friend AL, Johnson JMF, Coleman MD. Fine root dynamics in a [55

developing Populus deltoides plantation. Tree Physiol 2004;24:651.

Pregitzer KS, Zak DR, Curtis PS, Kubiske ME, Teeri JA, Vogel CS. Atmospheric

CO2, soil-nitrogen and turnover of fine roots. New Phytol 1995;129:579. [56

Al Afas N, Marron N, Van Dongen S, Laureysens I, Ceulemans R. Dynamics of

biomass production in a poplar coppice culture over three rotations (11 years).

For Ecol Manag 2008;255:1883. [57

Djomo SN, El Kasmioui O, Ceulemans R. Energy and greenhouse gas balance of

bioenergy production from poplar and willow: a review. Glob Change Biol [58

Bioenerg 2011;3:181.

Paul KI, Polglase PJ, Nyakuengama JG, Khanna PK. Change in soil carbon [59 following afforestation. For Ecol Manage 2002;168:241.

Kim H, Kim S, Dale BE. Biofuels, land use change, and greenhouse gas emissions: some unexplored variables. Environ Sci Technol 2009;43:961. Gelfand I, Zenone T, Jasrotia P, Chen J, Hamilton SK, Robertson GP. Carbon debt of conservation reserve program (CRP) grasslands converted to bioenergy production. Proc Natl Acad Sci USA 2011;108:13864.

Nikiema P, Rothstein DE, Miller RO. Initial greenhouse gas emissions and nitrogen leaching losses associated with converting pastureland to short-rotation woody bioenergy crops in Northern Michigan. USA Biomass Bioenerg 2012;39:413.

Beurskens L, Hekkenberg M. Renewable energy projections as published in the National Renewable Energy Actions Plans of the European Member States. Energy Center of the Netherlands (ECN). Report. ECN-E-10-069. Petten. The Netherlands; 2011. 244 pp.

EREC, Renewable energy technology roadmap 20% by 2020. REH: Brussels (Belgium); 2008.

Djomo SN, Ceulemans R. A comparative analysis of the carbon intensity of biofuels caused by land use changes. Glob Change Biol Bioenerg 2012;4:392. Mathews JA, Tan H. Biofuels and indirect land use change effects: the debate continues. Biofuels Bioproducts Biorefining-Biofpr 2009;3:305. Lapola DM, Schaldach R, Alcamo J, Bondeau A, Koch J, Koelking C, et al. Indirect land-use changes can overcome carbon savings from biofuels in Brazil. Proc Natl Acad Sci USA 2010;107:3388.

Havlik P, Schneider UA, Schmid E, Boettcher H, Fritz S, Skalsky R, et al. Global land-use implications of first and second generation biofuel targets. Energy Policy 2011;39:5690.

Melillo JM, Reilly JM, Kicklighter DW, Gurgel AC, Cronin TW, Paltsev S, et al. Indirect emissions from biofuels: how important? Science 2009;326:1397. Paustian K, Six J, Elliott ET, Hunt HW. Management options for reducing CO2 emissions from agricultural soils. Biogeochemistry 2000;48:147. Bruce JP, Frome M, Haites E, Janzen H, Lal R, Paustian K. Carbon sequestration in soils. J Soil Water Conserv 1999;54:382.