Scholarly article on topic 'Sequestration of C in soils under Miscanthus can be marginal and is affected by genotype-specific root distribution'

Sequestration of C in soils under Miscanthus can be marginal and is affected by genotype-specific root distribution Academic research paper on "Biological sciences"

Share paper
OECD Field of science
{" Miscanthus genotype" / Tuft-forming / "Root length density" / "Soil carbonate" / "Organic carbon sequestration" / δ13C / Decarbonizing}

Abstract of research paper on Biological sciences, author of scientific article — Goetz M. Richter, Francesco Agostini, Marc Redmile-Gordon, Rodger White, Keith W.T. Goulding

Abstract Miscanthus is a low input energy crop suitable for low fertility marginal arable land and thought to provide carbon sequestration in soil. We analysed a long-term field experiment (14-year) to determine whether differences in genotype, growth habit, and root distribution affected soil carbon spatially under different Miscanthus genotypes. Soil cores were taken centrally and radially to a depth of 1m, and divided into six vertical segments. Total root length (TRL), root dry matter (RDM) and δ13C signature of soil organic carbon (SOC) were measured directly, and root length density (RLD), fractions of Miscanthus-derived soil organic C (SOCM), and residual soil carbon (SOCorig) were calculated. Genotype was found to exhibit a statistically significant influence on spatial allocation of SOC. Grouping varieties into ‘tuft-forming’ (T) and ‘non-tuft-forming’ (NT) phenotypes revealed that respective groups accumulated similar amounts of RDM over 14 years (11.4±3.3 vs. 11.9±4.8Mgha−1, respectively). However, phenotype T allocated more carbon to roots in the subsoil than NT (33% vs. 25%). Miscanthus genotypes sequestered between 4.2 and 7.1gC4-SOCkg−1 soil over the same period, which was more than the average loss of C3-derived SOC (3.25gkg−1). Carbon stocks in the ‘A horizon’ under Miscanthus increased by about 5Mgha−1 above the baseline, while the net increase in the subsoil was marginal. Amounts of Miscanthus root C in the subsoil were small (1.2–1.8MgCha−1) but could be important for sustainable sequestration as root density (RLD) explained a high percentage of SOCM (R 2 =0.66).

Academic research paper on topic "Sequestration of C in soils under Miscanthus can be marginal and is affected by genotype-specific root distribution"

Contents lists available at ScienceDirect

Agriculture, Ecosystems and Environment

journal homepage:

Agriculture Ecosystems & Environment

Sequestration of C in soils under Miscanthus can be marginal and is affected by genotype-specific root distribution

Goetz M. Richter*, Francesco Agostini, Marc Redmile-Gordon, Rodger White, Keith W.T. Goulding

Department of Sustainable Soils and Grassland Systems, Rothamsted Research, Harpenden AL5 2JQ United Kingdom




Article history:

Received 21 August 2014

Received in revised form 6 November 2014

Accepted 10 November 2014

Available online 5 December 2014


Miscanthus genotype Tuft-forming Root length density Soil carbonate

Organic carbon sequestration S13C


Miscanthus is a low input energy crop suitable for low fertility marginal arable land and thought to provide carbon sequestration in soil. We analysed a long-term field experiment (14-year) to determine whether differences in genotype, growth habit, and root distribution affected soil carbon spatially under different Miscanthus genotypes. Soil cores were taken centrally and radially to a depth of 1 m, and divided into six vertical segments. Total root length (TRL), root dry matter (RDM) and 813C signature of soil organic carbon (SOC) were measured directly, and root length density (RLD), fractions of Miscanthus-derived soil organic C (SOCM), and residual soil carbon (SOCorig) were calculated. Genotype was found to exhibit a statistically significant influence on spatial allocation of SOC. Grouping varieties into 'tuft-forming' (T) and 'non-tuft-forming' (NT) phenotypes revealed that respective groups accumulated similar amounts of RDM over 14 years (11.4±3.3 vs. 11.9±4.8Mgha 1, respectively). However, phenotype T allocated more carbon to roots in the subsoil than NT (33% vs. 25%). Miscanthus genotypes sequestered between 4.2 and 7.1 gC4-SOCkg 1 soil over the same period, which was more than the average loss of C3-derived SOC (3.25 g kg 1). Carbon stocks in the 'A horizon' under Miscanthus increased by about 5 Mg ha 1 above the baseline, while the net increase in the subsoil was marginal. Amounts of Miscanthus root C in the subsoil were small (1.2-1.8MgCha 1) but could be important for sustainable sequestration as root density (RLD) explained a high percentage of SOCM (R2 = 0.66). © 2014 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license


1. Introduction

Miscanthus is a favored perennial feedstock for bioenergy in subtropical and temperate regions due to its high potential productivity (Heaton et al., 2010; Lewandowski et al., 2003) and benefits with regard to the carbon and greenhouse gas balance (Dondini et al., 2009; Hillier et al., 2009). Domestication of these perennials is in its infancy and genotypes may be found or bred that suit a wider range of ecological conditions and maximize efficiency of carbon sequestration (Clifton-Brown et al., 2008; Karp and Shield, 2008; Yan et al., 2012). The increasing interest in Miscanthus (e.g. Stewart et al., 2009; Yan et al., 2012; Sang and Zhu, 2011) should be accompanied by the exploration of the carbon

Abbreviations: AGB, above ground (dry) biomass (Mgha-1); BGB, below ground (dry) biomass (Mgha-1); G, gap; P, plant; RLD, root length density (cmcm~3); RDM, root dry matter (g m~2); RD, root diameter (mm); SOC, soil organic carbon (%); TRL, total root length (kmm~2); T, tuft forming Miscanthus genotypes; NT, non-tuft forming Miscanthus genotypes.

* Corresponding author. Tel.: +44 1582 763 133x2593; fax: +44 1582 469 036.

E-mail address: (G.M. Richter).

budgets of other genotypes in addition to commercially grown Miscanthus x giganteus. This would clarify whether contrasting Miscanthus phenotypes (growth habit, rooting pattern) act as an effective sink (Qin et al., 2011) or even a source (Yazaki et al., 2004) of atmospheric carbon. Key considerations in determining the soil organic carbon (SOC) balance require measurement of the C fraction deposited into the subsoil, which is less likely to be remobilized than C deposited in the surface horizon (Kell, 2011; Lockwell et al., 2012). Measurements of 13C abundance can also be used to indicate the stability of these inputs in the surface and subsoil under commercially grown Miscanthus x giganteus (e.g. Zimmermann et al., 2013). Existing studies of the genotype effect focus on carbon near the surface (Zatta et al., 2014) which ignores the potentially beneficial effect of deep roots as a mechanism to sequester carbon (Kell, 2011).

It is of further interest how contrasting growth forms, e.g. phenotypes (tuft or non-tuft) and carbon allocation patterns, e.g. different above- and belowground biomass allocation (AGB and BGB, respectively), and root densities, affect SOC. An integrative comparison of genotypes can inform about the relationships between productivity, carbon partitioning and

0167-8809/© 2014 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (

carbon sequestration characteristics, including vertical and lateral root distribution in response to rhizome form and size. This may have practical implications, such as elucidating the potential for increasing sequestration by breeding or selecting varieties with deep roots (Kell, 2011).

The relative contributions of AGB and BGB are easily confounded (Cotrufo et al., 2010) with annual litter inputs from M. x giganteus being between 1.5 and 7 Mg ha-1 yr-1 (Beuch et al., 2000; Kahle et al., 2002; Amougou et al., 2011). The contribution of roots to SOC is thought to be significantly greater than that of litter in grassland (Gill et al., 2002) and woody ecosystems (Rasse et al., 2005). Clifton-Brown et al. (2007) estimated that the C sequestered from Miscanthus into SOC after 15 years was equal to 10% of the BGB assuming a total input of 20 Mg dry weight ha-1, which contributed 14% to the total SOC in the first 10 cm layer. In deep soils the Miscanthus root-fraction was shown to accumulate initially at the net rate of >2 Mg ha-1 yr-1, which then decreased to about 1 Mg ha-1 yr-1 as a result of >3 Mg ha-1 growth and >2 Mg ha-1 decomposition (Neukirchen et al., 1999). In the present work, we aim to characterise the distribution of Miscanthus-derived SOC (SOCM) throughout the soil profile with particular attention to contrasts between individual genotypes from the main phenotypic growth forms (tuft vs. non-tuft; see definition below), and relate these differences to measurements of root distribution.

Starting from a solely C3-cropped site we use the Miscanthus induced change in 813C signature to distinguish between the original C3-based organic carbon (SOCorig) and SOCM under contrasting genotypes. From these quantities we then estimate C sequestration throughout the soil profile (depth 1 m) and relate this to the rooting and growth patterns of these genotypes on a marginal arable soil under low nitrogen (N) input and climatic conditions typical of the site at Rothamsted, UK.

2. Materials and methods

2.1. Field experiment and genotypes

The field experiment used in this study was established in 1997 as part of the European Miscanthus Improvement (EMI) program conducted at five locations in Europe (Clifton-Brown et al., 2001). The EMI field trial in England was established on a long-term arable field at Rothamsted Farm (51.81 N-0.358 E) on a silty clay loam with sandy inclusions (Batcombe-Carstens series; chromic luvisol or aquic paleudalf). C3 annual cereals and break-crops were grown exclusively on both the Miscanthus (Long Hoos III) and reference arable sites (Long Hoos IV) and conventionally tilled for 50 years or more (Johnston et al., 1981). The reference had remained under continuous arable management for all years since the Miscanthus was planted. The N input to a mixed arable crop rotation averaged 141 (range 80-190) kgNha-1yr-1. The Mis-canthus genotypes were planted as micro-propagated plantlets in 5 m x 5 m plots at a density of two plants per square meter in late May 1997. The trial had a fully randomised block design with three

replicates. Plants had been drip irrigated (+273 mm above the natural rainfall) during the first year. Details of fertiliser applications and management can be found in Riche et al. (2008). Over 14 years approximately 50 kg N ha-1 yr-1 was applied to support increasing annual yields between 4.8 and 15.9Mgha-1yr-1 which then declined and accumulated totals of 100-123 Mg ha-1 (Table 1).

Out of the 15 genotypes included in the EMI program we selected five genotypes that represent four genetic groups (Table 1): (1) M. x giganteus (Gig-1) is a vigorous natural hybrid of Miscanthus sinensis and Miscanthus sacchariflorus, widely grown commercially in UK and Europe, (2) M. sacchariflorus (Sac-5) is also grown in central Europe, originally obtained from Japan in 1992, (3) and (4) are two genotypes (SinH-6 and SinH-9) from the M. sinensis hybrid collection, which are characterised by a higher leaf fraction (36-62%) and yield reduction under drought, (5) Sin-11, a M. sinensis from Japan, which showed the least yield variation among the chosen genotypes.

These genotypes can also be grouped according to their aboveground growth habit or rhizomes (Lewandowski et al., 2003). M. sacchariflorus has broad, thick-stemmed rhizomes which creep laterally from where shoots develop out of internodal buds (non-tuft, NT phenotype) while rhizomes of M. sinensis genotypes are much smaller, do not exhibit the lateral creeping habit and aboveground shoots form dense centralised tufts made out of thinner stems (tuft, T phenotype). The annual dry matter allocation to rhizomes was estimated from earlier whole plant analysis and excavations. Based on the much larger fraction of rhizome accumulated under NT than T genotypes (>30 vs. <10Mgha-1, respectively; Table 1) one could consider this an important phenotypic trait. The rhizome fraction ranged from 23% for Sac-5 to between 6 and 11% of total accumulated yield for the M. sinensis genotypes. The hybrid, M. x giganteus, allocates circa 15% of the C to intermediate rhizomes, which creep less than M. sacchariflorus (Table 1). For investigating the effect of this phenotype contrast we grouped these into tuft forming (T; Sin-H6, Sin-H9, Sin-11) and non-tuft forming groups (NT; Gig-1 and Sac-5).

2.2. Soil sampling and preparation

A corer with an inner sleeve that could be dismantled longitudinally (diameter 70 mm; length 1 m) was driven into the soil using a hydraulic jackhammer and extracted using a tripod ratchet. Two cores were taken from each plot to a depth of 100 cm, one central to the original planting site (P) and one between plants, in the gap (G) situated midway between plants (32.5 cm from position P). Cores were wrapped in polythene and stored at -18 °C pending root and soil analyses. A further three random cores were taken from the adjoining arable reference site, approximately 10 m from edge of the EMI trial as reference points for 813C and total C (hereon termed 'Reference Arable'). An equivalent soil mass (ESM) of the A horizon (0-30 cm) of the Miscanthus plots was found in the 0-26 cm layer of the Reference Arable soil (due to Miscanthus

Table 1

Growth characteristics of Tuft- and Non-Tuft (T, NT) forming Miscanthus genotypes used in the root and carbon analysis; harvested yields (min, max) and cumulative production (dry matter) over 14 years after planting; litter residues and rhizome dry matter accumulated at time of hand harvest (April 2011). Fractional area of tuft derived from circumference (m2 m~2).

Genotype Phenotype Tuft size (-) Yield (Mg ha-Min 1yr-1) Max Culmuative biomass (Mgha 1) Harvest Litter Rhizome

M. giganteus Gig-1 NT 5.6 15.9 120 9.3 33

M. sacchariflorus Sac-5 NT 7.5 15.4 123 8.5 29

M. sinensis hybr Sin-H6 T 0.56 6.5 13.9 114 3.3 9

M. sinensis hybr Sin-H9 T 0.44 4.8 15.3 114 8.3 7

M. sinensis Sin-11 T 0.46 6.1 10.8 101 8.9 11

reducing the bulk density, especially in the top 10 cm). In addition to the reference samples, archived samples from the site were retrieved for the period prior to planting the Miscanthus (1988) and analysed to obtain the baseline SOC content. This archived material always used to be sampled to a depth of 0-23 cm because this represented former ploughing depth.

The cores were cut into three sections for each horizon composed of topsoil (A; depths 0-10, 10-20, 20-30 cm) and subsoil (B; 30-50, 50-70, 70-100 cm), respectively. Due to high soil moisture at sampling, the cores were variably compressed, two thirds between 0 and 5%, and only three cores were exceeding 10% compression. A proportional adjustment was made for all sections of the compressed core before division. These sections were then divided into approximately equal half-cores and each half was weighed. One half-sample was kept for root washing while the other half was air-dried for the determination of soil moisture and chemical analysis (see below). Stones were removed from both sub-samples. Dry bulk densities were determined from each segment using the volume and stone-free dry matter content to estimate the carbon content at each depth. The air-dried soils were gently separated from visible organic matter, litter, roots, rhizomes and large stones (flint, chalk). The soil was then sieved (<2mm) and crushed using a disk mill (TEMA Machinery Ltd.).

2.3. Root extraction and characterisation

The root core half section was placed in a bowl of warm water to gently tease the soil apart from Miscanthus structures, and then to carefully separate plant roots from rhizomes and litter debris. Large roots were collected on a fine sieve (0.25 mm) to enable soil dispersion and small soil particles to be removed, and roots were then placed straight into a water-filled glass jar. Rhizomes and aboveground plant materials were removed from the top section of each core. Once the visible roots had been removed, the content of the bowl containing the fine roots was poured onto the sieve and rinsed thoroughly with water to remove any soil. Roots were subjected to a second rinse if the sample was not clean (i.e. containing chaff or clay) and then combined and stored in aqueous 10% ethanol in plastic bottles and kept in a dark cool room (4° C) before scanning.

2.4. Root scans and image analysis

Root samples were spread on an A4 plexiglass water tray of the WinRhizo flatbed scanner (Epsom STD 4800). Root length and diameter were quantified using the WinRhizo Pro (2008) software package, applying a standard set of acquisition parameters for black and white (grey shades) for the root length and diameter classes. Scans were saved as tiff files pending further image analysis. After scanning was complete, the roots were dried at 40 °C to determine the root dry matter (RDM; g). This was then converted to RDM per depth increment (RDM(z); Mgha-1) using the respective stone-free dry bulk density of the cores (pSF). Root length density (RLD) was calculated from the measured total root length (TRL) and the volume of the stone-free soil in each segment.

2.5. Carbonate and its removal

Carbonates in the soil from underlying chalk or added lime interfere with the determination of 813C of soil organic matter (SOM) because they exhibit a 813C signature close to that of PDB (around 0m). Therefore, carbonates were removed before isotope analysis by acid treatment. To avoid loss of soluble organic C by acid washing, Harris et al. (2001) proposed to expose moistened soil to concentrated HCL vapor (12 M) for 6-8 h before isotope ratio mass spectrometry (IRMS). However, this popular technique was found

to deposit strongly acidic residues that remained even after repeated application of vacuum (not shown). The residual HCL appeared to interfere with the analyses and damage the mass spectrometer. We therefore applied the following method, similar to that developed for the removal of carbonates from coastal sediments (Komada et al., 2008), which combines the advantages of utilising an invasive aqueous phase without losing soluble organic C or accumulating problematic quantities of HCL.

Subsamples of 20.00 ± 0.20 mg milled soil were weighed into A-foil capsules (9 mm) and placed in a random arrangement on microtitre plates (F96 Maxisorp, Nunc, Denmark) allowing adequate space between each sample. To each capsule, sufficient aqueous solution of trace analysis grade HCl (1 M) was added to bring the soils approximately to field capacity (in this case 35 mL). The plate was subsequently placed in an empty, carbon-free desiccator (i.e. no silicone grease) for 30 min to allow the acid to permeate throughout the sample. The desiccator was then fitted with a Viton seal and evacuated for 2-3 min until equilibrium was reached, then left at this pressure for 1 h to allow HCL to permeate throughout the sample and reduce the likelihood of trapped air. After carefully and slowly returning the desiccator to atmospheric pressure, a further 35 mLof HCL was added, before transferring to a clean oven set to 40 °C for 1 h (or until dry). A final mobilisation of 35 mL de-ionised water was applied. Samples were returned to the oven to dry at 40°C after which the Ag-foils were closed and analysed by IRMS.

For quantifying inorganic carbon (IC) an excess of hydrochloric acid (7 ml of 4 M HCL) was applied to 5 g of soil and the resulting CO2 measured using a pressure calcimeter. Soil organic carbon (SOC) was calculated from the difference between total C and IC.

2.6. Stable carbon isotope analysis

The isotopic composition (813C) and %C of each sample was measured using IRMS using an Anca SL20/20 Stable Isotope Analyser (Sercon, UK). The precision of the IRMS for the soil (internal standard) and plant (external standard: wheat flour) material was between 1 and 2% with regard to carbon and <1% for the isotope analysis. The 13C abundance was expressed as delta depletion (813Cm) from the international standard, Pee Dee Belemnite (PDB):

8 Cm =

Rsample Rr<

X 1000

where Rsample is the isotope ratio, 13C/12C, of the sample and Rreference is that of PDB.

2.7. Contribution of Miscanthus-derived soil carbon to SOC

The SOCM fraction of total SOC can be estimated for each soil layer by the balance of the measured 813C between the original SOC (SOCorig; reference values) and the signature of C4 Miscanthus inputs.



;]) = ( 813CSOCong X

+ (8 X [SOCm])

Although a small additive isotope fractionation effect is expected due to heterotrophic microbial respiration depleting SOC of lighter isotopes, in general this appears to be balanced by microbial preference for more 13C enriched substrates (Santruckova et al., 2000; Wynn and Bird, 2007). The fraction of soil carbon derived from Miscanthus,FM, in the ratio of C4 and total SOC was thus calculated as:


(8 Cobserved 8 CSOCorig)

(813 Cc4

8 Csoc

Table 2

Log-transformed (in brackets) and back-transformed mean RLD's and RDM's for each horizon depth section (A: 0-10,10-20, 20-30cm; B: 30-50, 50-70 and 70-100 cm). Statistically significant two-way interactions between the effects of horizon and position with phenotype (A/B vs. NT/T and P/B vs. NT/T, respectively) are presented for RLD, and a three dimensional factor interaction for RDM between horizon, sample position and growth type (A/B x P/G x NT/T), respectively.

Phenotype NT

Phenotype T

Wald statistic

RLD transformed mean (cm cm-3) Horizon A Horizon B Position P Position G RDM transformed mean (Mgha-1) Position P Horizon A

Horizon B

Position G Horizon A

Horizon B

7.64 (2.03) 1.25 (0.22) 3.35 (1.21) 2.84 (1.04)

2.73 (1.00) 0.45 (-0.81) 1.84 (0.61) 0.31 (-1.19)

5.71 (1.74) 1.48 (0.39) 3.84 (1.35) 2.20 (0.79)

3.01 (1.10) 0.44 (-0.83) 1.35 (0.30) 0.44 (-0.83)

(0.179) (0.179)


5.18 3.79

0.023 0.052

From this the contribution of Miscanthus-derived carbon (MgCha-1) was estimated using the profile respective depth and stone-free bulk densities. The d13C values of nearby reference arable samples had a mean 813C of -28.16m (range -28.36 to -27.75m PDB) and did not show any systematic variation or statistical significance over depth (p = 0.29; F = 1.4649). A single reference point (813C = -28.16) was thus used from which estimates of Miscanthus-derived SOC accumulation were estimated. The 813C4 marker value for Miscanthus used in this study was -11.7m 813C PDB (close to the -11.8m used by Dorodnikov et al. (2007)).

2.8. Statistical analysis

All statistical comparisons were made using GenStat 14 (Payne et al., 2011). Residual maximum likelihood (REML) methods were applied in preference to ANOVA as samples were not equally replicated (2 replicates Sin-H6, Sin-H9; 3 replicates Sin-11, Sac-5, Gig-1). Variables were transformed where necessary after examining the residual diagnostic plots for homogeneity of variance; the natural logarithm was selected as the most appropriate transformation. We present the transformed and back transformed data in Table 2 only for the variables with significant effects between interacting factors; all statistical comparisons were made using the transformed data. Natural means are presented in Figs. 3 and 4.

3. Results

3.1. Root distribution in the profile

Roots of all Miscanthus genotypes were up to an order of magnitude more abundant in the A (topsoil) than in the B horizon (subsoil) but varied greatly within each phenotype group (Fig. 1). Statistical significance of the differences between varieties or phenotypes (T vs. NT) was reduced due to spatial variability. Mean log-transformed RLD showed a significant two-way interaction between phenotype and vertical distribution (p = 0.023; Table 2). Phenotype'T' contributed much less RLD to the A horizon but more to the B horizon than NT phenotypes. Additionally, tufted varieties generally showed higher RLD in the plant position than did NT, and less in the gap than NT (Table 2). This interaction was just outside the 5% confidence limit (p = 0.052). A 3-way interaction between horizon, position and phenotype was found for root dry matter (RDM; p = 0.04), which reflects the strong contrast in spatial distributions in RDM between phenotypes (most notably in the A horizon gap).

Table 3 gives horizon TRL and RDM data on the natural scale. In the A horizon there was typically more than twice the RDM directly

under plants (P) of all genotypes except for M. giganteus, where RDM appears to be evenly distributed (having equally colonised the gap). However, within the T phenotype the high contrast between allocation of RDM to the G vs. P position of Sin-11 and Sin-H9 (RDMB%) was not seen with Sin-H6 (Table 3). Sin-H6 seems to be an exception to all genotypes (Table 1). Nevertheless, the

Fig. 1. Back-transformed mean root length density (RLD) under plant (a) and in gap (b). See Table 2 for SED.

Table 3

Total root length (TRL; kmm-2), and root dry matter (RDM; Mgha-1) per horizon (subscript A vs. B), genotype, and sample position (G vs. P).



TRLa (kmm-2)

TRLB (kmm-2)

TRL (km m-2)

RDMa (Mgha-1)

RDMB (Mgha-1)

RDMb% (%)

Gig-1 (NT) Sac-5 (NT)

26.8 25.4 30.3

10.3 7.4 11.2 9.0

34.2 36.7

11.35 10.10 7.85 4.71

3.34 2.78 3.30 2.17

22.8 21.6 29.6 31.6

Sin-11 (T) Sin-H6 (T) Sin-H9 (T)

32.1 15.6 18.1

20.2 34.5 12.9

15.8 9.5 15.4 10.0 14.0 7.7

47.9 25.0 33.5 30.2 48.5 20.7

14.49 4.59 6.88 4.38 12.97 5.64

3.93 3.43 3.18 2.59 4.02 4.24

42.8 31.7 37.1 23.7

Phenotype mean NT

38.8 36.7 44.0 25.3

25.7 25.1 24.0 41.5

T 'phenotype mean' allocation of RDM to the B horizon (RDMB%) was still nearly double that of NT (41.5% compared to 25.1%; Table 3).

3.2. 813 of soil organic matter (SOM).

The d13C signatures in almost all soil samples under Miscanthus were less negative than those found in the arable reference profile. Miscanthus cover had increased d13C up to -25.39 in the B, and -16.37m in the A horizons, respectively (Fig. 2). Statistical analysis of d13C changes indicated the main effect was depth (F5,78.1 = 178.45; p < 0.001) followed by an interaction between depth and genotype (F20782 = 1.76; p = 0.040). The statistics demonstrate that Miscanthus genotype is a relevant factor in SOC distribution over depth. Although the sampling position effect (P vs. G) was found to lie just outside the 5% confidence level (F1,8 = 5.25; p = 0.051) mean d13C over depth are presented for samples taken under both the original planting (P) and gap (G) positions, respectively (Fig. 2).

3.3. Changes in SOCorig and SOCM concentration

The SOCorig concentration in the A horizon under Miscanthus had declined to an average of 13.25 g kg-1 relative to the reference soil (17.91 gkg-1) but less from what had been measured in the baseline from the archive (16.4gkg-1; Fig. 3a). In the B horizon, there appears to be little change in the SOCorig (Fig 3b). By comparison with archive SOC values (Fig. 3a) it can be seen that Miscanthus more than compensated for losses in SOCorig through inputs of SOCM. For statistical comparison between genotypes, transformed and back transformed concentration data are presented in Table 4 (similarly in Table 5 for SOC stock change).

3.4. Net change of SOC stocks

Analysis of phenotype (T vs. NT) showed no statistically significant effect on SOCM stock (Mgha-1; p >0.05). However, as with d13C, there was an important interaction between horizon and genotype (F4,14.5 = 2.94; p = 0.016; Table 5). Comparison of back-transformed means show subsoil SOCM tended to be greatest under Sin-H6 (2.20Mgha-1) and lowest under Sin-H9 (0.39Mgha-1). Gig-1 showed the greatest residual contribution (back-transformed) to SOCM in the A horizon (16.64Mgha-1) while Sin-11 residual contributions were low (9.84Mgha-1).

Determining the change in SOC storage over area is complicated by the large variation of bulk density due to the variation in stone,

root, and rhizome content. Mean stone-free bulk densities in the A-horizon directly under the plants were significantly smaller than in the gap (1.10 and 1.40; SED ± 0.077 Mgm-3). The difference was mainly attributable to the T-phenotype, where roots and rhizomes displaced the soil and raised the soil surface. No significant differences were found for the B horizons (1.30 and 1.33 ± 0.077 Mgm-3). Accordingly, statistical analysis of SOCM concentration (gkg-1) showed a stronger interaction between horizon and

Fig. 2. 813C signature of SOM as affected by Miscanthus genotype under plant (a) and in gap position (b). Data given on the natural scale (F-statistics for REML are

provided in text).

Table 5

Log-transformed (in brackets) and back-transformed SOCM stock in horizon A vs. B (per unit area). SOCM distribution was affected by a statistically significant two-way interaction between genotype and horizon (F4,14.5 = 4.40; p = 0.016).

Fig. 3. Mean SOCorig (grey bars) and SOCM (black bars) concentrations in the A horizon (a) and B horizon (b). Carbon concentration measured in the equivalent soil mass (ESM) corresponding to 0-30 cm for Miscanthus and 0-26 cm for Reference Arable. Depths of B horizons correspond to 30-100cm for Miscanthus and to

26-96 cm for Reference Arable. SOC in Reference Arable (—) and archive (......);

archive SOC for B horizon not available. Positive bars indicate standard error for SOCM, negative bars for SOCorig.

genotype (F413 5 = 7.46; p = 0.002; Table 4) than estimates over area (Table 5). Again however, there was no statistically significant effect of phenotype (growth-form) upon SOC stocks.

Fig. 4a and b show natural mean stock estimates over area for SOCorig and SOCM combined (inclusive of bulk density effects). Due to high residual variability there were no statistically significant differences in transformed SOCorig data (for statistical comparison

Table 4

Log-transformed (in brackets) and back-transformed mean SOCM concentrations by horizon (gkg-1). SOCM distribution was affected by a statistically significant two-way interaction between genotype and horizon (F413 5 = 7.46; p = 0.002).

Genotype Phenotype class SOCm (gkg-1)

Horizon A Horizon B

Gig-1 NT (1.89) 6.63 (-2.19) 0.11

Sac-5 NT (1.41)4.09 (-1.72) 0.18

Sin-11 T (1.37) 3.92 (-1.80) 0.17

Sin-H6 T (1.64) 5.17 (-1.10) 0.33

Sin-H9 T (1.59) 4.92 (-3.31) 0.04

s.e.d (0.387)

Genotype Phenotype class SOC m (Mgha-1)

Horizon A Horizon B

Gig-1 NT (2.81) 16.64 (-0.07) 0.93

Sac-5 NT (2.59) 13.26 (0.39) 1.48

Sin-11 T (2.29) 9.84 (0.39) 1.48

Sin-H6 T (2.74) 15.52 (0.79) 2.20

Sin-H9 T (2.75) 15.71 (-0.93) 0.39

s.e.d (0.364)

of SOCM between genotypes see Table 5). Strictly speaking, no statistical calculation can be applied for the stock change in SOCorig by reference to the arable plot, as the reference arable is not within the randomised block design. However, Miscanthus inputs (SOCM) appeared to compensate for losses in SOCorig in all cases except with Sin-11 (Fig. 4a). This was not seen on a concentration basis

Fig. 4. Total SOCorig (grey bars) and SOCM (black bars) stocks in the A horizon (a), and B horizon (b). Carbon stock estimated on the basis of equivalent soil mass (ESM) corresponding to 0-30 cm for Miscanthus and 0-26 cm for Reference Arable. Depths of B horizons correspond to 30-100cm for Miscanthus and to 26-96cm for

Reference Arable. SOC in Reference Arable (—) and archive (......); archive SOC for

B horizon not available. Positive bars indicate standard error for SOCM, negative bars for SOCorig.

(Fig. 3a), indicating that the A horizon of Sin-11 had a lower average bulk density. Comparison indeed showed its bulk density to be the lowest among all genotypes (1.12 Mgm~3 vs. 1.19, 1.23, 1.20 and 1.35 Mgm~3 for Gig-1, Sac-5, Sin-H6 and Sin-H9, respectively).

4. Discussion

The results presented describe root distribution for a range of fundamentally different Miscanthus genotypes of the two major phenotypes (T/NT-growth forms) and the impact of their long-term cultivation on the SOC of a silty clay loam in England. Similar experiments exist in Germany (Gauder et al., 2012) and Denmark (Jorgensen et al., 2003) but from those studies no data for roots and Miscanthus-derived SOC have been published. Our data for RLD, root biomass, and their distribution down the profile are consistent with those reported for M. x giganteus, the only commercially grown genotype (Neukirchen et al., 1999; Amougou et al., 2011; Monti and Zatta, 2009; Dohleman et al., 2012). Our research illustrates the spatial heterogeneity (horizontal and vertical) of rooting and carbon allocation. Analysis also shows that SOC enrichment is more closely correlated with RLD than root biomass, which is consistent with findings that link carbon inputs to root exudates and rhizo-depositions from finer roots (Kanova et al., 2010; Techer et al., 2011).

4.1. Is a certain rooting pattern a desirable trait?

The similarity in RDM distribution irrespective of position with M. x giganteus (Table 3) suggests that the mature stand had completely colonised the A horizon (only a 13% difference between horizontal sampling points 32.5 cm apart). For a young M. x giganteus stand RLD is considerably lower in the gap between plants than in the plant center (Neukirchen et al., 1999). A recent comparison of different M. sacchariflorous x M. sinensis crosses found genotypes that spread their root system less than M. x giganteus and exhibited a growth form that was still very much akin to the T phenotype (Zatta et al., 2014). However, their data do not distinguish between root and rhizome fractions of the BGB, although RDM traits could be approximated from biomass in their deeper horizon (15-30cm).

RDM allocation to the subsoil (B horizon) is thought to be a desirable trait (Kell, 2011) but rarely studied as sampling poses a challenge in terms of temporal and spatial variability (Neukirchen et al., 1999). Here, C allocation to the B horizon was found to be lower in the NT than in the T phenotype (25 vs. 33%), presumably because of the high allocation of C into rhizomes and associated roots. T types spread laterally to the B horizon in the gap (Table 3). Spatial variability between replicates was large, and RLD did not show a statistically significant interaction between phenotype and sampling position whilst it exists for RDM (Table 2).

The T phenotypes included genotypes with contrasting AGB traits which could be mirrored in BGB accumulation. Sin-11 and Sin-H9 showed high proportional allocation of RDM to the B horizon (>40%), whilst Sin-H6 did not follow the same pattern (Table 3; neither for RDMA nor RDMB). Litter accumulation was also low for this genotype (Table 1) but a causal link between these observations for Sin-H6 is unknown. The concomitant high SOCM and low residual litter, however, suggest potentially high SOC accumulation from more rapidly decomposing leaf and root litter. Parameters of Miscanthus residues differ (see turnover of root and leaf litter; Amougou et al., 2011) but our results indicate a wide range for genotype-specific parameters. The relationship between RLD and Miscanthus-derived SOC (Fig. 5) identifies genotype Sin-H6 as an example showing greater SOCM accumulation per unit root length. This supports the hypothesis that faster turnover of

roots and litter increase SOCM. It would therefore be of great interest to characterise the biochemical composition of roots and leaf residues to explore the reasons behind the observed contrast.

The genotype-specific variation of carbon allocation was previously observed in controlled conditions (Clifton-Brown et al., 2001). Data from the field are the products of greater functional complexity of C turnover, affected by temperature (Beuch et al., 2000; Magid et al., 2004; Amougou et al., 2011) and soil hydrology, especially in the A horizon. The contrasting carbon allocation at depth for Sin-11 and Sin-H6 to the B horizon (Table 4) could reflect the different responses to water stress shown in laboratory experiments: In contrast to M. x giganteus and M. sacchariflorus, the M. sinensis type (Clifton-Brown and Lewandowski, 2000) showed higher drought-tolerance possibly due to increased RLD at depth.

The high proportional RDM allocation to the B horizon and lateral spread under Sin-11 (Table 3) suggest that root density was the primary factor influencing the accumulation of SOCM (1.48 Mgha-1; Table 4). In contrast, Sin-H9 resulted in the lowest quantity of SOCM in spite of high absolute and relative RDM to the B horizon. The fact that there was a statistically significant interaction of SOCM allocation to soil horizon with individual genotype (Table 4), but not growth form (T vs. NT) indicates that other factors affect the SOCM distribution between soil profiles. Production of root exudates and other rhizo-depositions can be of similar magnitude as roots (Rasse et al., 2005; Techer et al., 2011). These root-associated biochemical factors may explain the differences between the effects of roots on SOCM and SOCorig in the B horizon for genotypes Sin-11 and Sin-H9. Priming of existing SOM in response to the input of easily decomposable organic substrate is a transient phenomenon (Kuzyakov, 2002). These potentially accelerate the degradation of SOCorig as observed for Miscanthus phenotypes (Figs. 3 and 4) as suggested by Zatta et al. (2014). In this context it is controversial how the addition of mineral N affects SOC turnover (e.g. Wang et al., 2014; Foereid et al., 2004; Amougou et al., 2011). The differences we found for the impact of RLD on SOCM under different genotypes need further research, also with regard to the degradation of SOCorig.

4.2. Miscanthus effects on carbon stocks

The comparison of SOC over area (Table 5; Mgha-1) is affected by the change in bulk density (pSF) over time. The change in pSF affects sampling depth because soil cores taken to a fixed depth will not access the same mineral soil as before expansion (Ellert and Bettany, 1995). Such artefacts have been discussed by

Fig. 5. Miscanthus derived soil carbon (SOCM) as % of total SOC versus root length density (RLD) under various Miscanthus genotypes; "A" Sin-H6, "*" all genotypes.

Palm et al. (2014) in the context of tillage effects on soil C stocks, proposing Equivalent Soil Mass (ESM) sampling. In practical terms, an ESM controlled sampling strategy is only possible once the bulk density has been measured; however, pre-emptive sampling of every plot would be prohibitively expensive.

In the present study the SOC quantities over area incorporate variation of pSF, thus, we also compared SOC change on a concentration basis (gkg-1; Table 4 and Fig. 3). In this way, plot variability in pSF was avoided and genotype-related effects were identified with greater statistical confidence, at p = 0.002 (Table 4) as opposed to p = 0.016 (Table 5). Measurement on a concentration basis also provided more precise estimates of inputs (Fig. 3) as anticipated by Lee et al. (2009). Figs. 3a and 4a show that, while SOCorig declines over 14 years of Miscanthus cropping from its baseline (archive sample), the total SOC increased (~5Mgha-1) above the baseline and was, in spite of much lower N inputs, similar to the high N reference arable. In contrast, any SOCorig decline or net SOC increase in the B horizon (Figs. 3b and 4b) was not measurable. This is in accordance with the premise that deep C is more stable due to closer organo-mineral interactions (Schrumpf et al., 2013).

Three distinct groups of Miscanthus can be separated according to their effects on the A horizon (Fig. 3a): (1) M. giganteus, which contributes the greatest SOCM with accompanied greatest loss of SOCorig; (2) M. sacchariflorus with lowest contributions of SOCM and greatest retention of C3 originating SOC; (3) M. sinensis with intermediate and similar effects on respective SOC concentrations. Recent analysis showed that the average fertiliser application of 50 kg N ha-1 yr-1 as ammonium nitrate to M. x giganteus, applied in the present study, would limit biomass production (Shield et al., 2014). This would be even more relevant for M. sinensis genotypes as these have only a very small rhizome system (Table 1), which would limit recycling of N within the plant (Amougou et al., 2011). It is hypothesised here that under conditions of N limitation the production of labile Miscanthus root exudates increases to support a microbial community capable of mineralising soil organic N and thus depleting SOC. Similar ideas have been discussed by Kuzyakov (2002) but empirical studies are now needed to investigate this and quantify its potential impact upon calculations of energy budgets for carbon crops.

4.3. Meta-contributions to SOC sequestration

SOCorig varied greatly between plots and genotypes and no statistically significant differences were observed. However, any differences in expansion of the soil profile due to rhizome and root growth will physically protect SOC. Lower temperature and oxygen, protection from freeze-thaw and drying and wetting will all affect microbial activity and contribute to the preservation of SOC (Balesdent et al., 2000). Sequestration of C in soil purely through burying is a commonly overlooked mechanism (Chaopricha and Marin-Spiotta, 2014). In temporal perennials its sustainability will depend on the form of reversion to arable. Nevertheless, the expanding surface horizons (burying subsoil) could be a useful indirect trait to develop otherwise shallow soils on marginal land and compliment the established benefits from protection against soil erosion (e.g. Kroumov and Blagoeva, 2011). Interestingly, the SOCorig 'burying effect' could be of potentially greater contribution to sequestration of C than the direct allocation of Miscanthus C at depth because SOCorig is likely to be more closely associated with the mineral fraction and more stable than recent C inputs in subsoils (Schrumpf et al., 2013). Furthermore, turnover rates of SOC are inversely related to concentration (Don et al., 2013) which is inherently low in the subsoil. The apparent uncertainty about changes in SOCorig in the subsoil (Figs. 3b and 4b) demand further evidence to support these concepts.

5. Conclusions

We found variation in root distribution between genotypes, with 'T-phenotype' (tuft growing genotypes) allocating more biomass -relatively (33 vs. 25%) and absolutely (3.6 vs. 2.9 Mg ha-1) - to roots at greater depth than NT phenotypes. Analysis of SOC concentration and isotope composition revealed allocation patterns for SOCM to be significantly different with respect to depth and genotype (Table 4). Within the T phenotype, the higher dry matter allocation and lateral spread of roots in the subsoil observed for Sin-11 and Sin-H9 (Table 3) were not seen with Sin-H6, indicating high diversity within this phenotype.

These results reveal a statistically significant link between RLD distribution and newly derived SOCM which supports the premise that, for C sequestration, it is important to consider the effect of all carbon inputs, including short-lived rhizo-depositions. Furthermore, a subgroup of higher SOCM values associated with low root volumes in M. sinensis could point to higher root turnover in some species (Sin-H6), a hypothesis to be followed up in research that integrates N and C turnover.

In view of the limiting quantity of N-fertiliser, the accumulation of SOCM in this trial is likely to be lower than would occur under optimum conditions for biomass production. The net SOC stock increase under low N input (50kgNha-1yr-1) Miscanthus production (< 10 Mg ha-1 yr-1) was small (<0.5 Mg SOC ha-1 yr-1). Although its C stock was similar to the SOC in a high input arable reference soil after 14 years of cultivation, its net carbon gain (CO2 eq.) is by far superior when accounting for the respective average N fertiliser inputs (50 vs. 141 kgN ha-1).

Future work could expand on the contribution of root exudates and (D)OC from the litter to SOCM and whether the effects of chemical and physical properties on decomposition can be disentangled (texture and aggregation of the soil associated with higher SOCM/root ratios). The measured SOCM enrichment and SOCorig decline as well as biomass production, litter and root accumulation will be instrumental in the calibration and validation of models such as RothC for simulating soil C sequestration under Miscanthus.


This work was partially funded by the UK Natural Environment Research Council (NERC) - (NE/H010742/1; Carbo-Biocrop), the UK Biotechnology and Biological Sciences Research Council (BBSRC) and the Institute's Strategic Programme Grant "Cropping Carbon" at Rothamsted. We would like to thank Alexandra Barker for her dedication to wash numerous root samples out of the most difficult soil, a flinty silty clay loam. Special thanks to Anne Duffy and her team, including Martina Masna, for careful preparation and chemical analysis of the soil and root samples. This paper is based on a talk given at the "Root to the Future" Conference of the International Society of Root Research, Dundee, 26-29 June 2012.


Amougou, N., Bertrand, I., Machet, J.M., Recous, S., 2011. Quality and decomposition in soil of rhizome, root and senescent leaf from Miscanthus x giganteus, as affected by harvest date and N fertilization. Plant Soil 338, 83-9 . Balesdent, J., Chenu, C., Balabane, M., 2000. Relationship of soil organic matter

dynamics to physical protection and tillage. Soil Till. Res. 53, 215-230. Beuch, S., Boelcke, B., Belau, L., 2000. Effect of the organic residues of

Miscanthus x giganteus on the soil organic matter level of arable soils. J. Agron. Crop Sci. 184,111-119. Chaopricha, N.T., Marin-Spiotta, E., 2014. Soil burial contributes to deep soil organic

carbon storage. Soil Biol. Biochem. 69, 251 -264. Clifton-Brown, J.C., Lewandowski, I., 2000. Water use efficiency and biomass partitioning of three different Miscanthus genotypes with limited and unlimited water supply. Ann. Bot. 86,191-200. Clifton-Brown, J.C., Lewandowski, I., Andersson, B., Basch, G., Christian, D.G., Kjeldsen, J.B., Jorgensen, U., Mortensen, J.V., Riche, A.B., Schwarz, K.U., Tayebi, K.,

Teixeira, F., 2001. Performance of 15 Miscanthus genotypes at five sites in Europe. Agron. J. 93, 1013-1019.

Clifton-Brown, J.C., Breuer, J., Jones, M.B., 2007. Carbon mitigation by the energy crop, Miscanthus. Glob. Change Biol. 13, 2296-2307.

Clifton-Brown, J., Chiang, Y.-C., Hodkinson, T.R., 2008. Miscanthus: Genetic resources and breeding potential to enhance bioenergy production. In: Vermerris, W. (Ed.), Genetic Improvement of Bioenergy Crops. Springer, New York.

Cotrufo, M.F., Conant, R.T., Paustian, K., 2010. Soil organic matter dynamics: land use, management and global change. Plant Soil 338, 1 -3.

Dohleman, F.G., Heaton, E.A., Arundale, R.A., Long, S.P., 2012. Seasonal dynamics of above- and below-ground biomass and nitrogen partitioning in Miscanthus x giganteus and Panicum virgatum across three growing seasons. Glob. Change Biol. Bioenergy 4, 534-544.

Don, A., Rodenbeck, C., Gleixner, G., 2013. Unexpected control of soil carbon turnover by soil carbon concentration. Environ. Chem. Lett. 11, 407-413.

Dondini, M., Van Groenigen, K.J., Del Galdo, I., Jones, M.B., 2009. Carbon

sequestration under Miscanthus: a study of (13)C distribution in soil aggregates. Glob. Change Biol. Bioenergy 1, 321-330.

Dorodnikov, M., Fangmeier, A., Kuzyakov, Y., 2007. Thermal stability of soil organic matter pools and their delta C-13 values after C3-C4 vegetation change. Soil Biol. Biochem. 39, 1173-1180.

Ellert, B.H., Bettany, J.R., 1995. Calculation of organic matter and nutrients stored in soils under contrasting management regimes. Can. J. Soil Sci. 75, 529-538.

Foereid, B., de Neergaard, A., Hogh-Jensen, H., 2004. Turnover of organic matter in a Miscanthus field: effect of time in Miscanthus cultivation and inorganic nitrogen supply. Soil Biol. Biochem. 36,1075-1085.

Gauder, M., Graeff-Hoenninger, S., Lewandowski, I., Claupein, W., 2012. Long-term yield and performance of 15 different Miscanthus genotypes in southwest Germany. Ann. Appl. Biol. 160, 126-136.

Gill, R.A., Burke, I.C., Lauenroth, W.K., Milchunas, D.G., 2002. Longevity and turnover of roots in the shortgrass steppe: influence of diameter and depth. Plant Ecol. 159, 241-251.

Harris, D., Horwath, W.R., van Kessel, C., 2001. Acid fumigation of soils to remove carbonates prior to total organic carbon or carbon-13 isotopic analysis. Soil Sci. Soc. Am. J. 65, 1853-1856.

Heaton, E.A., Dohleman, F.G., Miguez, A.F., Juvik, J.A., Lozovaya, V., Widholm, J., Zabotina, O.A., McIsaac, G.F., David, M.B., Voigt, T.B., Boersma, N.N., Long, S.P., 2010. Miscanthus: A Promising Biomass Crop. Adv. Bot. Res. 56, 75-137.

Hillier, J., Whittaker, C., Dailey, A.G., Aylott, M., Casella, E., Richter, G.M., Riche, A., Murphy, R., Taylor, G., Smith, P., 2009. Greenhouse gas emissions from four bioenergy crops in England and Wales: integrating spatial estimates of yield and soil C balance in life cycle analyses. Glob. Change Biol. Bioenergy 1, 267-28 .

Johnston, A.E., Poulton, P.R., McEwen, J., 1981. Soils of Rothamsted Farm England UK. Carbon and nitrogen content of the soils and the effect of changes in crop rotation and manuring on soil pH phosphorus potassium and magnesium. Rothamsted Exp. Station Report. Part 2 5-20.

Jorgensen, U., Mortensen, J., Kjeldsen, J.B., Schwarz, K.U., 2003. Establishment, development and yield quality of fifteen Miscanthus genotypes over three years in Denmark. Acta Agric. Scand. B 53, 190-199.

Kahle, P., Belau, L., Boelcke, B., 2002. Effects of 10 years of Miscanthus cultivation on different properties of mineral soil in North-east Germany. J. Agron. Crop Sci. 188, 43-50.

Kanova, H., Carre, J., Vranova, V., Rejsek, K., Formanek, P., 2010. Organic compounds in root exudates of Miscanthus x Giganteus greef et deu and limitation of microorganisms in its rhizosphere by nutrients. Acta Univ. Agric. Silvic. Mendelianae Brun. 58, 203-208.

Karp, A., Shield, I., 2008. Bioenergy from plants and the sustainable yield challenge. New Phytol. 179, 15-32.

Kell, D.B., 2011. Breeding crop plants with deep roots: their role in sustainable carbon, nutrient and water sequestratio. Ann. Bot. 108, 407-418.

Komada, T., Anderson, M.R., Dorfmeier, C.L., 2008. Carbonate removal from coastal sediments for the determination of organic carbon and its isotopic signatures, delta C-13 and Delta C-14: comparison of fumigation and direct acidification by hydrochloric acid. Limnol. Oceanogr. Methods 6, 254-262.

Kroumov, V., Blagoeva, V., 2011. Soil protection and energy efficiency of

Miscanthus x giganteus grown in conditions of organic agriculture in the area of Samokov Bulgaria. Pochvoznanie, Agrokhimiya i Ekologiya 45, 51 -56.

Kuzyakov, Y., 2002. Review: factors affecting rhizosphere priming effects. J. Plant Nutr. Soil Sci. 165, 382-396.

Lee, J., Hopmans, J.W., Rolston, D.E., Baer, S.G., Six, J., 2009. Determining soil carbon stock changes: simple bulk density corrections fail. Agric. Ecosyst. Environ. 134,


Lewandowski, I., Scurlock, J.M.O., Lindvall, E., Christou, M., 2003. The development and current status of perennial rhizomatous grasses as energy crops in the US and Europe. Biomass Bioenergy 25, 335-361.

Lockwell, J., Guidi, W., Labrecque, M., 2012. Soil carbon sequestration potential of willows in short-rotation coppice established on abandoned farm lands. Plant Soil 360, 299-318.

Magid, J., Luxhoi, J., Lyshede, O.B., 2004. Decomposition of plant residues at low temperatures separates turnover of nitrogen and energy rich tissue components in time. Plant Soil 258, 351 -365.

Monti, A., Zatta, A., 2009. Root distribution and soil moisture retrieval in perennial and annual energy crops in Northern Italy. Agric. Ecosyst. Environ. 132,


Neukirchen, D., Himken, M., Lammel, J., Czyionka-Krause, U., Olfs, H.W., 1999. Spatial and temporal distribution of the root system and root nutrient content of an established Miscanthus crop. Eur. J. Agron. 11, 301-309.

Palm, C., Blanco-Canqui, H., DeClerck, F., Gatere, L., Grace, P., 2014. Conservation agriculture and ecosystem services: an overview. Agric. Ecosyst. Environ. 187, 87-105.

Payne, R.W., Harding, S.A., Murray, D.A., Soutar, D.M., Baird, D.B., Glaser, A.I., Welham, S.J., Gilmour, A.R., Thompson, R., Webster, R., 2011. The Guide to GenStat Release 14. VSN International, Hemel Hempstead, UK.

Qin, Z., Zhuang, Q., Zhu, X., Cai, X., Zhang, X., 2011. Carbon consequences and agricultural implications of growing biofuel crops on marginal agricultural lands in China. Environ. Sci. Technol. 45,10765-10772.

Rasse, D.P., Rumpel, C., Digna, M.F., 2005. Is soil carbon mostly root carbon? Mechanisms for a specific stabilisation. Plant Soil 269.

Riche, A.B., Yates, N.E., Christian, D.G., 2008. Performance of 15 different Miscanthus species and genotypes over 11 years. Aspects Appl. Biol. 90, 207-212.

Sang, T., Zhu, W., 2011. China's bioenergy potential. Glob. Change Biol. Bioenergy 3, 79-90.

Santruckova, H., Bird, M.I., Lloyd, J., 2000. Microbial processes and carbonisotope fractionation in tropical and temperate grassland soils. Funct. Ecol. 14, 108-114.

Schrumpf, M., Kaiser, K., Guggenberger, G., Persson, T., Kogel-Knabner, I.,

Schulze, E.D., 2013. Storage and stability of organic carbon in soils as related to depth, occlusion within aggregates, and attachment to minerals. Biogeosciences 10, 1675-1691.

Shield, I.F., Barraclough, T.J.P., Riche, A.B., Yates, N.E., 2014. The yield and quality response of the energy grass Miscanthus x giganteus to fertiliser applications of nitrogen, potassium and sulphur. Biomass Bioenergy 68,185-194.

Stewart, J.R., Toma, Y., Fernandez, F.G., Nishiwaki, A., Yamada, T., Bollero, G., 2009. The ecology and agronomy of Miscanthus sinensis, a species important to bioenergy crop development, in its native range in Japan: a review. Glob. Change Biol. Bioenergy 1,126-153.

Techer, D., Laval-Gilly, P., Henry, S., Bennasroune, A., Formanek, P., Martinez-Chois, C., D'Innocenzo, M., Muanda, F., Dicko, A., Rejsek, K., Falla, J., 2011. Contribution of Miscanthus x giganteus root exudates to the biostimulation of PAH degradation: an in vitro study. Sci. Total Environ. 409, 4489-4495.

Wang, Q.K., Wang, Y.P., Wang, S.L., He, T.X., Liu, L., 2014. Fresh carbon and nitrogen inputs alter organic carbon mineralization and microbial community in forest deep soil layers. Soil Biol. Biochem. 72,145-151.

Wynn, J.G., Bird, M.I., 2007. C4-derived soil organic carbon decomposes faster than its C3 counterpart in mixed C3/C4 soils. Glob. Change Biol. 13, 2206-2217.

Yan, J., Chen, W., Luo, F., Ma, H., Meng, A., Li, X., Zhu, M., Li, S., Zhou, H., Zhu, W., Han, B., Ge, S., Li, J., Sang, T., 2012. Variability and adaptability of Miscanthus species evaluated for energy crop domestication. Glob. Change Biol. Bioenergy 4, 49-60.

Yazaki, Y., Mariko, S., Koizumi, H., 2004. Carbon dynamics and budget in a Miscanthus sinensis grassland in Japan. Ecol. Res. 19, 511-520.

Zatta, A., Clifton-Brown, J., Robson, P., Hastings, A., Monti, A., 2014. Land use change from C3 grassland to C4 Miscanthus: effects on soil carbon content and estimated mitigation benefit after six years. GCB Bioenergy 6, 360-370.