at SciVerse ScienceDirect

Soil Biology & Biochemistry 56 (2013) 40e48

Contents lists available

Soil Biology & Biochemistry

journal homepage: www.elsevier .com/locate/soi lbio

Litter quality assessed by solid state 13C NMR spectroscopy predicts decay ratebetter than C/N and Lignin/N ratios

Giuliano Bonanomi a,*, Guido Incerti a, Francesco Giannino b, Antonio Mingo a, Virginia Lanzotti c,Stefano Mazzoleni a

aDipartimento di Arboricoltura, Botanica e Patologia Vegetale, University of Naples Federico II, via Università 100, 80055 Portici, Naples, ItalybDipartimento di Ingegneria Agraria e Agronomia del Territorio, University of Naples Federico II, via Università 100, 80055 Portici, Naples, ItalycDipartimento di Scienza degli Alimenti, University of Naples Federico II, via Università 100, 80055 Portici, Naples, Italy

a r t i c l e i n f o

Article history:Received 2 August 2011Received in revised form17 February 2012Accepted 6 March 2012Available online 29 March 2012

Keywords:C-cycleC stocksDecompositionLitterbagLitter qualityPrincipal component regressionProximate cellulose and lignin

* Corresponding author. Tel.: þ39 081 2539015; faxE-mail address: giuliano.bonanomi@unina.it (G. B

0038-0717/$ e see front matter � 2012 Elsevier Ltd.doi:10.1016/j.soilbio.2012.03.003

a b s t r a c t

Predictions of litter decomposition rates are critical for modelling biogeochemical cycling in terrestrialecosystems and forecasting organic carbon and nutrient stock balances. Litter quality, besides climaticconditions, is recognized as a main factor affecting decay rates and it has been traditionally assessed bythe C/N and lignin/N ratios of undecomposed materials. Here, solid state 13C NMR spectroscopy andproximate chemical analysis have been used to characterize litter organic C in a litterbag experimentwith 64 different litter types decomposing under controlled conditions of temperature and watercontent. A statistical comparative analysis provided evidence that C/N and lignin/N ratios, showingdifferent trends of correlation with decay rates at different decomposition stages, can be used to describethe quality of undecomposed litter, but are unable to predict mass loss of already decomposed materials.A principal component regression (PCR) model based on 13C NMR spectra, fitted and cross-validated byusing either two randomly selected sets of litter types, showed highly fitting predictions of observeddecay rates throughout the decomposition process. The simple ratio 70e75/52e57 corresponding to O-alkyl C of carbohydrates and methoxyl C of lignin, respectively, showed the highest correlation withdecay rate among different tested parameters. These findings enhance our understanding of litter quality,and improve our ability to predict decomposition dynamics. The 13C NMR-based 70e75/52e57 ratio isproposed as an alternative to C/N and lignin/N ratios for application in experimental and modelling work.

� 2012 Elsevier Ltd. All rights reserved.

1. Introduction

Leaf litter decomposition is a critical process for biogeochemicalcycling in terrestrial ecosystems (Attiwill and Adams, 1993) withslow decomposition rates enhancing the accumulation of organiccarbon and nutrient stocks. Litter decomposition rates are stronglyaffected by climatic variables (Aerts, 1997) and litter quality (i.e. thesusceptibility of the substrate to decomposer attack; Swift et al.,1979; Rovira and Vallejo, 2007), which interact and prevail acrossdifferent spatial scales (Liski et al., 2003). Temperature and wateravailability are considered the most important factors acting atglobal and regional scale (Aerts, 1997), while at local scale, withnearly uniform climate, litter decay rate is mostly affected by litterquality (Meentemeyer, 1978).

The definition of litter quality in terms of organic chemicalcomposition (Swift et al., 1979) is operationally difficult because

: þ39 0817760104.onanomi).

All rights reserved.

litter material hold a multitude of organic compounds withdifferent susceptibility to decomposition (e.g. lignin, tannins,cellulose, organic acids, aminoacids, simple sugars) whose relativefractions, together with several inorganic elements (e.g. N, P, S),vary with the decay stage (Berg and McClaugherty, 2008). Duringthe last decades a substantial effort has been made in search ofeffective indicators of litter quality, capable to provide reliablepredictions of litter decay rate (Cornelissen and Thompson, 1997;Cornwell et al., 2008). Traditional approaches have been based onthe assessment of selected litter characteristics to identify param-eters or indexes correlated with decay rates, and thus useful forpredictive purposes (Meentemeyer, 1978; Melillo et al., 1982; Coqet al., 2010). Several papers reported that lignin content, one ofthe most abundant biopolymers resistant to decomposition, wasnegatively correlated with decay rate, especially in the cases ofroot and wood plant tissues included in the sample of analysedlitters (Fogel and Cromack, 1977; Meentemeyer, 1978). Consistentnegative correlations with litter decay rates have been also re-ported for carbon-to-nitrogen content (C/N) (Taylor et al., 1989) and

G. Bonanomi et al. / Soil Biology & Biochemistry 56 (2013) 40e48 41

lignin-to-nitrogen content (lignin/N) ratios (Melillo et al., 1982). Indetail, Taylor et al. (1989) reported that C/N, for substrates withlimited lignin content (e.g. understory herbs) predicted litter decayrate better than lignin/N. In any case, both C/N and lignin/N ratioshave been extensively used in most C-cycle models, such asCENTURY (Parton et al., 1994), BIOME-BGC (Hunt et al., 1996) andLinkages (Pastor and Post, 1986), as descriptors of litter qualitycontrolling mass loss rate (Burke et al., 2003; Adair et al., 2008).

However, litter quality continuously changes during decompo-sition being function of the relative susceptibility to breakdown ofits organic components, with rapid degradation of labile simplesugars, and selective preservation of less degradable tannins andlignin (Rovira and Vallejo, 2007). Berg and Matzner (1997) devel-oped a three-phase model with labile organic compounds andnutrients (i.e. N and P) controlling decay rate up to 30e40% of massloss, and lignin content becoming progressively more importantafterwards. The model has been successfully tested on a range oflitter types including hardwoods and coniferous plants (Berg andMcClaugherty, 2008). Further modelling attempts to take intoaccount progressive chemical changes during decomposition havebeen implemented (Liski et al., 2005; Adair et al., 2008; Incertiet al., 2011), with organic matter quality described by litter masspartitioning into different pools, exponentially decaying atdifferent rates according to first-order kinetics (Olson, 1963).

In spite of research efforts and of positive evidences, bothparameters and indexes currently used to describe litter quality andchanges during decomposition are not free of uncertainty. Forinstance, Berg and McClaugherty (2008) suggested that using C/Nratio to predict leaf decay rate throughout the decompositionprocess should be avoided, because, irrespectively of its initialvalue, C/N progressively decreases as C is lost by respiration and Nis immobilized in microbial biomass (Bonanomi et al., 2010).Recently, Goebel et al. (2011) reported that decomposition of fineroots of four temperate tree species were opposite to what mightbeen expected from C/N ratio. Moreover, Hättenschwiler et al.(2011) reported that both C/N and lignin/N ratios poorly predictlitter decomposition rate of many tree species of lowland Amazo-nian forest.

In the last decade, chemical throughput methods as pyrolysis-gas chromatography/mass spectrometry (Huang et al., 1998), nearinfrared reflectance spectroscopy (Gillon et al., 1999) and 13C-cross-polarization magic angle spinning (CPMAS) nuclear magneticresonance (NMR) spectroscopy (Kögel-Knabner, 2002) have beenapplied to characterize organic matter at molecular level. Inparticular, 13C-CPMAS NMR has been proven useful to providea description of the total organic chemical composition of complexmatrices, such as plant litter (Kögel-Knabner, 2002), allowing toobtain the resonance signals of all the carbons of the analyzedsamples. Since the chemical shifts of different C atoms depend ontheir molecular environment, important information about theirchemical type and the nature and number of substituents allowsthe attribution of observed carbons to a particular class of organiccompounds. Then, by analysing litter samples at different decom-position stage, the changes of different classes of organic C corre-sponding to different levels of litter decay can be assessed. Severalstudies based on 13C-CPMAS NMR data reported a continuouschemical shift during litter decomposition, with a progressivedecrease of carbohydrates and increase of lipids and lignin (Prestonet al., 2009; Bonanomi et al., 2011). As a consequence, some indexesof organic matter stability have been proposed, such as the ratioshydrophobic to hydrophilic C (HB/HI) and alkyl to O-alkyl C (alkyl/O-alkyl), both consistently increasing during decomposition(Almendros et al., 2000; Preston et al., 2009; Ono et al., 2011), andthe ratio carbohydrate to methoxyl C (CC/MC), steeply droppingdown at the initial stages of the decay process (Mathers et al., 2007;

Bonanomi et al., 2011). More recently, Bonanomi et al. (2011)proposed a new index of organic matter stability based on 13C-CPMAS NMR that was highly correlated with litter phytotoxicity.Such index, calculated as the direct ratio between two restrictedregions of NMR spectra (i.e. 70e75/52e57) included major contri-bution of C2, C3, and C5 signals of carbohydrates (70e75 ppm) aswell as methoxyl C signal of lignin (52e57 ppm).

These evidences, in contrast to traditional C/N and lignin/Nindexes, provided a promising possibility of using 13C-CPMAS NMRdata to operationally define litter quality but, so far, no systematiccomparison between these different approaches has been reported.Instead, this has been the general objective of this study. To thispurpose, sixteen different plant litter types, decomposing inmicrocosm under optimal controlled conditions to avoid limitingeffects of water availability and temperature on litter decay rates(Taylor et al., 1989; Gholz et al., 2000), were analysed in a litter-bagexperiment. A previous report, based on the same litter decom-position experiment, showed the possibility of monitoring by 13C-CPMAS NMR the quality changes of organic C during the decom-position process thus allowing the prediction of the litter inhibitoryeffects on plant growth (Bonanomi et al., 2011). Specific objectivesof this second contribution were:

(1) to test the relationship between litter decay rate and the C/Nand lignin/N ratios from the decomposingmaterials at differentdecay stages;

(2) to analyse by 13C-CPMAS NMR the organic C dynamics relatedto litter decomposition;

(3) to assess the possibility of using empirical indices and statis-tical models based on 13C-CPMAS NMR data to predict litterdecomposition rate;

(4) to compare 13C-CPMAS NMR data and C/N and lignin/N ratios,as functional indicators of litter quality and predictors of litterdecay rates.

2. Materials and methods

The current work is based on a previous litterbag decompositionexperiment focused on phytotoxic effects of plant litter (Bonanomiet al., 2011).

2.1. Plant material collection

A pool of plant species representing awide range of litter qualitywere selected from vegetation types of Mediterranean andtemperate environments (Campania Region, Southern Italy)including two grasses (Ampelodesmos mauritanicus, Festuca dry-meia), one forb (Medicago sativa), two evergreen shrubs (Arbutusunedo, Coronilla emerus), one vine (Hedera helix), three evergreentrees (Picea excelsa, Pinus halepensis, Quercus ilex), and sevendeciduous trees (Castanea sativa, Fagus sylvatica, Fraxinus ornus,Populus nigra, Quercus pubescens, Robinia pseudoacacia, Salix alba).Three species are nitrogen-fixing (M. sativa, C. emerus,R. pseudoacacia). For each species, a number >20 of individualsfrom natural communities was randomly selected at the samplingsites, and freshly abscised leaves were collected by placing netsunder the plants, dried (40 �C until constant weight was reached)and then stored at room temperature.

2.2. Litter decomposition experiment

Decomposition experiments were carried out according to thelitterbag method (Berg and McClaugherty, 2008) in microcosmsplaced in a growth chamber to simulate field decomposition

G. Bonanomi et al. / Soil Biology & Biochemistry 56 (2013) 40e4842

conditions. Microcosms were kept under controlled temperature(18� 2 �C night and 25� 2 �C day) and water (watered every sevendays to field capacity with distilled water). Terylene litterbags(20 � 20 cm, mesh size 2 mm) were filled with 6 g of dry leaf litterand placed inside trays (30 cm deep,100 cm for each side). A total of384 litterbags (16 species � 3 sampling dates � 8 replicates) wereharvested after 30, 90 and 180 days of decomposition. Bags weredried at the laboratory (40 �C until constant weight was reached)and the remaining material weighted afterwards (for details seeBonanomi et al., 2011).

2.3. Litter chemical analyses and 13C-CPMAS NMR characterization

Total C and N content in litters from the litterbag experimentwere determined by flash combustion of microsamples (5 mg oflitter) in an Elemental Analyser NA 1500 (Carlo Erba Strumenta-zione, Milan, Italy). The content of acid-detergent hydrolysablefibre (thereafter indicated as labile C), proximate cellulose andlignin were assessed following the procedure described by Gessner(2005). Briefly, labile C was determined by mild acid hydrolysiswith 0.5 M H2SO4 added with the detergent cetyl-trimethylammonium (CTAB, 20 g l�1), proximate cellulose wasdetermined as hydrolysable fraction after drastic sulphuric acidtreatment (loss due to 72% H2SO4 for 3 h), and proximate lignin asthe unhydrolyzable fraction (loss upon ignition after the abovementioned H2SO4 treatment). It is important to note that the ligninassessed in this way does not correspond to pure lignin because itmay include some other hydrolysis-resistant organic structures,such as cutin, waxes, and condensed tannins at varying proportions(Berg and McClaugherty, 2008; Preston et al., 2009). All carbonfractions are presented as ash-free dry mass.

All litters types were characterized by 13C-CPMAS NMR in solidstate under the same condition, thus allowing a quantitativecomparison among NMR spectra. Litter samples were analysed atfour different times, corresponding to leaves decomposition for 0,30, 90 and 180 days (total of 64 litter types). A spectrometer (BrukerAV-300, Bruker Instrumental Inc, Billerica, MA, USA) equipped witha magic angle spinning (MAS) probe with wide-bore of 4 mm wasused, set up with MAS of 13,000 Hz of rotor spin, a recycle time of1 s, a contact time of 1 ms, an acquisition time of 20 ms, and 2000scans (for details see Bonanomi et al., 2011).

Spectral regions have been selected and C types identified asreported in Bonanomi et al. (2011), following previous referencestudies (Kögel-Knabner, 2002; Mathers et al., 2007; Pane et al.,2011). The following seven regions and C types were considered:0e45 p.p.m ¼ alkyl C; 46e60 ppm ¼methoxyl C; 61e90 ppm ¼ O-alkyl C; 91e110 ppm ¼ di-O-alkyl C; 111e140 ppm ¼ H- and C-substituted aromatic C; 141e160 ppm O-substituted aromaticC (phenolic and O-aryl C); 161e190 ppm carboxyl C. Four indexes,which are considered robust indicators of the degree of litterdecomposition, were calculated: i. the hydrophobicity index (HB/HI); ii. the alkyl C/O-alkyl C ratio (0e45/61e110), and; iii. the O-alkyl C/methoxyl and N-alkyl C ratio (61e90/46e60; thereafterindicated as CC/MC) (Spaccini et al., 2000; Almendros et al., 2000;Mathers et al., 2007); iv. finally, a new index recently proposed byBonanomi et al. (2011) was calculated as the ratio between tworestricted regions of O-alkyl C (70e75 ppm) and methoxyl and N-alkyl C (52e57 ppm), respectively (i.e. 70e75/52e57).

2.4. Data analysis

Dynamics of litter mass remaining, changes of N, labile C frac-tion, cellulose and lignin content, C/N ratio and lignin/N ratio werestatistically evaluated by General Linear Models (GLM), consideringplant species (16 levels) and decomposition time (continuous

variable), as well as their interaction, as independent factors.Following the same approach, additional GLM models were testedfor the seven 13C-CPMAS NMR resonance regions selected fromreference literature (Spaccini et al., 2000; Almendros et al., 2000;Kögel-Knabner, 2002) and described above. For all GLMs, signifi-cant differences between groups were tested using Tukey’s HSDpost hoc test.

A previous comparative analysis of existing decompositionequations (Rovira and Rovira, 2010), showed that different decom-position datasets can be best fitted by different equations. Testing ofsimple-exponential equation (Olson, 1963), double-exponential(Gillon et al., 1994), single-compartment with variable rate (Yangand Janssen, 2000), and composite-exponential (Rovira andRovira, 2010) models were performed on our datasets. Results areshown in supplementary material (Table S1). Considering that thesample size was not adequate for the application of best fittingmodels, requiring a higher number of parameters, we applied litternegative exponential decayconstant (k) calculated according toBergand McClaugherty (2008) by applying the Olson (1963) model. Themodel equationwasMt¼M0 e

�kt, whereM0 is the initial litter mass,Mt is themass remaining after a certain time t, and k is the decay rateconstant. Three consecutive exponential models have been appliedto three different decomposition periods (0e30, 30e90 and 90e180days). Both univariate and multivariate statistics were used toaddress the relationship between litter decay rate (k) and litterchemical parameters (i.e. N, labile, cellulose and lignin content, C/Nratio, lignin/N ratio, data from 13C-CPMAS NMR spectra). Simplecorrelation was extensively tested between litter decay rate andeach parameter, including all 13C-CPMAS NMR spectral signals.Using a multivariate approach, principal component regression(PCR) was performed. Such analysis consists of a multiple linearregression (MR) inwhichprincipal component analysis (PCA) is usedto calculate independent variables (Jolliffe, 2002) to avoid multi-collinearity (i.e. a significant correlation between two or morepredictor variables in a multi-regressive model). A multi-regressivemodel was tested, in which the seven spectral regions wereconsidered as predictive variables for litter decay rate. Based ona matrix of litter types and spectral regions, independent linearcombinations of the predictors were calculated by PCA. The corre-lation between litter decay rate (k) and each principal componentwas calculated using k values and the factorial scores of littersamples. Following the approach suggested by Legendre andLegendre (1998) for supplementary variables, litter decay rate wasalso plotted as a loading vector on the bi-dimensional PCA spaceeven if it was not used to compute the eigenvalues of the sameordination space. Finally, the four orthogonal PCA axes most corre-lated with litter inhibitory effects were used as independentpredictive variables in the multiple regression model. The modelwas fitted using data from 30 litter samples obtained by randomlyselecting 10 plant species and 3 decomposition stages from thecomplete dataset. Thereafter, the model was validated on littersamples of the remaining 6 plant species, by comparing predicted vsobserved litter decay rate.

3. Results

3.1. Dynamics of litter decomposition

Littermass loss largely varied among different plant species, andfor each species over time (Fig. 1A), with decay rates (k) rangingbetween 18.8 and 0.31. Mass loss was initially extremely rapid forC. emerus, H. helix, and F. ornus, and slower for Q. ilex and P. excelsa,showing intermediate values for other species (Fig. 1A). However,as decomposition proceeded, mass loss declined for all litter types,with rates decreasing over time during the decay process (Fig. 1A).

0 30 60 90 120 150 180

Days of decomposition

A

B

C

)%(t netnoc

N

Lignin/NC/N

0

30

25

20

15

10

5

N/C,

N/ningiL)

%(gninia

merssa

M

0

2.5

2

1.5

1

0.5

)%(tnetnoc

ning iL

Lignin (%)N (%)

0

50

40

30

20

10

0

25

50

75

100

Fig. 1. Dynamics of litter mass remaining (A), nitrogen and lignin content (B), and C/Nand lignin/N ratios (C) during 180 days of decomposition. Data refer to mean of 8replicates for each species (A), and mean � SD of all species (B and C). Plant species in(A) are: Ampelodesmos mauritanicus ( ), Arbutus unedo ( ), Castanea sativa( ), Coronilla emerus ( ), Hedera helix ( ), Festuca drymeia ( ),Fraxinus ornus ( ), Fagus sylvatica ( ), Medicago sativa ( ), Picea excelsa( ), Pinus halepensis ( ), Populus nigra ( ), Robinia pseudoacacia ( ),Quercus ilex ( ), Q. pubescens ( ), Salix alba ( ).

G. Bonanomi et al. / Soil Biology & Biochemistry 56 (2013) 40e48 43

Percent content of N and lignin in decaying materials progres-sively increased during the 180 days of decomposition (Fig. 1B).Both parameters were significantly different among litter types(Table 1). Highest initial N content was found in nitrogen-fixersC. emerus (3.53%) and M. sativa (3.94%), and lowest inA. mauritanicus (1.19%) and Q. ilex (1.34%). Percent lignin contentranged from values slightly above 5% of undecomposed herb andvine materials (C. emerus and H. helix) to values over 40% of littersfrom trees after 180 days of decomposition, with maximum(50.13%) observed for Q. ilex. However, variations over time werestatistically significant only in the case of lignin, but not for Nconcentration (Table 1).

Undecomposed litters showed very different values of C/N ratio,ranging between 40 (A. mauritanicus) and 11 (M. sativa) withmean� standard errorof 22.9�7.0.High interspecific variabilitywasalso found in the case of lignin/N ratio,withmean� standard error of

9.5 � 5.1. As decomposition proceeded, the C/N and lignin/N ratiosshowed contrasting dynamics, with a decrease and a slight increase,respectively (Fig.1C). Both these trendswere significantly affected bylitter type anddecomposition time (Table 1). A significant interactiveeffect of litter type and decomposition time was found for the ratiolignin/N (Table 1) indicating that the increase of such parameterwithdecomposition time was significantly different among the consid-ered litter types, being more pronounced in materials with highlignin/N ratio, and less evident in litters with lower lignin/N.

13C-CPMAS NMR spectra showed remarkable and consistentchanges of organic C components in litters, with all spectral regionssignificantly affected by litter type (Table 1), and major chemicalshifts observed for the initial decomposition period (Fig. 2). Thealkyl-C region (0e45 ppm), characteristic of lipid waxes and cutins,significantly increased during the first 30 days of decomposition(Fig. 2), and much less later on. The methoxyl and N-alkyl C region(46e60 ppm) showed a progressive increase, with significantdifferences between initial, intermediate, and final values (Fig. 2).The O-alkyl-C region (61e90 ppm), mainly associated with sugarsand polysaccharides, steeply decreased at the initial decompositionstage, then showing a continuous but not significant reduction(Fig. 2). The di-O-alkyl-C region (91e100 ppm) showed a similarpattern, with significant reduction during the first 30 days ofdecomposition, but with a less pronounced reduction (Fig. 2). Otherregions, including H- and C-substituted aromatic C, O-substitutedaromatic C, and carboxyl C, though significantly affected by littertype (Table 1), were not affected by significant changes duringdecomposition (Fig. 2, Table 1). Both the reduction of the O-alkyl-Cand the increase of the alkyl-C regions were more evident for fastdecomposing materials (e.g. leaf litter of H. helix, C. emerus, F. ornus)than for slow decaying ones (e.g. P. excelsa, Q. pubescens,F. sylvatica). The degree of hydrophobicity (HB/HI) and the ratioalkyl-C/O-alkyl-C showed a significant increase during decompo-sition. Finally, the CC/MC and the 70e75/52e57 ratios rapidlydecreased in the first 30 days and then showed a further progres-sive reduction (see Bonanomi et al., 2011).

3.2. Relationships between litter decay rate and traditional indicesof litter quality

The ratios C/N and lignin/N from undecomposed materialsshowed negative correlations with litter decay rate in the earlydecomposition stage (0e30 days, Fig. 3). However, different resultswere obtained considering C/N and lignin/N from already decom-posing materials, which showed non-significant correlation withdecay rates both at intermediate (30e90 days) and late (90e180days) decomposition stages (Fig. 3). Interestingly, in these lattercases, Pearson’s r values were almost null for lignin/N, and evenshifted towards positive values in the case of C/N (Fig. 3).

Concerning other elemental and proximate chemical parame-ters, different patterns were observed. Labile C and N contentshowed significant positive correlations with decay rate only at theearly decomposition stage (0e30 d). Lignin and cellulose contentshowed the opposite trend, with significant and non-significantnegative correlation, respectively, at the early decomposition stage(0e30 d; see supplementary material, Table S2). However, none ofthe considered parameters was significantly related, either posi-tively or negatively, with decay rate of already decomposingmaterials (Table S2).

3.3. Predictivity of the 13C NMR-based PCR model of litter decay rate

Principal component analysis (PCA) provided a satisfactoryordination of the 13C-CPMAS NMR data across litter types (Fig. 4),with the first four eigenvalues accounting for 98.5% (54.9%, 24.9%,

Table 1Summary of the GLM testing (type III Sum of Squares, in bold values significant at P < 0.05) for effects of plant species (Sp) and decomposition time (T) on litter mass loss,elemental and proximate chemical parameters, and 13C-CPMAS NMR spectral regions.

d.f SS MS F p d.f SS MS F p

Mass remaining Alkyl CSpecies 15 3067.4 204.49 3.17 0.003 Species 15 8.89 0.59 2.487 0.016Time 1 1623.2 1623.2 25.17 <0.001 Time 1 2.72 2.72 11.416 0.002Sp � T 15 249.2 16.61 0.26 0.996 Sp � T 15 2.33 0.16 0.651 0.809Residuals 32 1031.7 64.48 Residuals 32 7.39 0.24N content Methoxyl CSpecies 15 12.67 0.85 5.80 <0.001 Species 15 0.43 0.03 3.78 <0.001Time 1 0.28 0.28 1.91 0.177 Time 1 1.14 1.14 148.28 <0.001Sp � T 15 3.17 0.21 1.45 0.184 Sp � T 15 0.15 0.01 1.30 0.256Residuals 32 4.67 0.15 Residuals 32 0.25 0.01Labile C O-alkyl CSpecies 15 3335.1 222.3 7.36 <0.001 Species 15 7.45 0.50 3.331 0.002Time 1 1602.7 1602.7 53.05 <0.001 Time 1 7.84 7.84 52.621 <0.001Sp � T 15 481.2 32.2 1.06 0.426 Sp � T 15 1.83 0.12 0.817 0.653Residuals 32 966.9 30.2 Residuals 32 4.77 0.15Cellulose Di-O-alkyl CSpecies 15 849.5 56.6 5.42 <0.001 Species 15 0.76 0.05 3.477 0.002Time 1 254.0 254.0 24.30 <0.001 Time 1 0.42 0.42 28.652 <0.001Sp � T 15 458.4 30.6 2.92 0.005 Sp � T 15 0.18 0.01 0.852 0.619Residuals 32 334.6 10.5 Residuals 32 0.46 0.01Lignin H- and C-substituted aromatic CSpecies 15 1327.9 88.5 3.31 0.002 Species 15 0.96 0.06 3.062 0.004Time 1 2130.6 2130.6 79.75 <0.001 Time 1 0.14 0.14 2.765 0.106Sp � T 15 630.8 42.1 1.57 0.137 Sp � T 15 0.36 0.02 1.145 0.360Residuals 32 854.9 26.7 Residuals 32 0.67 0.02C/N ratio O-substituted aromatic CSpecies 15 881.8 58.8 5.93 <0.001 Species 15 0.28 0.02 3.063 0.004Time 1 629.7 629.7 63.53 <0.001 Time 1 0.00 0.00 0.047 0.830Sp � T 15 219.9 14.7 1.48 0.172 Sp � T 15 0.11 0.01 1.165 0.345Residuals 32 317.2 9.9 Residuals 32 0.20 0.01Lignin/N ratio Carboxyl CSpecies 15 652.9 43.5 13.13 0.001 Species 15 0.65 0.04 6.334 <0.001Time 1 267.3 267.3 80.64 <0.001 Time 1 0.08 0.08 2.997 0.093Sp � T 15 130.7 8.7 2.63 0.011 Sp � T 15 0.17 0.01 1.614 0.125Residuals 32 106.1 3.3 Residuals 32 0.22 0.01

G. Bonanomi et al. / Soil Biology & Biochemistry 56 (2013) 40e4844

11.5%, and 7.2%) of the total variance, respectively. The multipleregression analysis based on the results of PCA, aimed at assessingthe relationship between 13C-CPMAS NMR data and litter decayrate, provided a highly significant linear model (Table 2). Themodel, once applied to the validation dataset, provided very highlyfitting predictions of litter decay rates (predicted vs observedvalues: r ¼ 0.91) (Fig. 4).

0.2

0

0.1

0.3

0.4

0.5

0 0 0 0 0 0 5

Rel

ativ

e ab

unda

nce

0 d30 d90 d180 d

a a a aa a a a

a a a a abbb

bb b

a

a

a

b b

b b b

c

Car

boxi

licC

(1

61-1

90 p

pm)

O-s

ubst

itute

d ar

omat

ic C

(141

-160

ppm

)

H-C

-sub

stitu

ted

arom

atic

C(1

11-1

40 p

pm)

di-O

-alk

yl C

(91-

110

ppm

)

O-a

lkyl

C(6

1-90

ppm

)

Met

hoxy

l C(4

6-60

ppm

)

Alky

l C(0

-45

ppm

)

Fig. 2. Variation of seven main classes of organic C assessed by 13C-CPMAS NMRspectroscopy in 16 plant litters at different decomposition stages in controlledconditions. Data refer to mean values (n ¼ 16) � SD. Different letters indicate time-dependent significant differences (Tukey’s HSD test at P < 0.05).

3.4. Relationship between litter decay rate and dynamics of selectedorganic C types

Litter decay rate was not significantly related with three 13C-CPMAS NMR regions (Carboxyl C, O-substituted aromatic C, and H-and C-substituted aromatic C, with r ¼ 0.08, �0.07, and �0.19,respectively). A significant negative correlation was found in twocases, including alkyl C (r ¼ �0.38, P < 0.01) and methoxyl and N-alkyl C regions (r¼�0.46, P< 0.001), whereas a significant positivecorrelation was found for the O-alkyl C (r ¼ 0.48, P < 0.001) and di-O-alkyl (r¼ 0.37, P< 0.01) regions. Concerning indices derived fromthe literature (Spaccini et al., 2000; Almendros et al., 2000; Kögel-Knabner, 2002; Bonanomi et al., 2011) litter decay rate was nega-tively correlatedwith the Alkyl C/O-alkyl C (r¼�0.35, P< 0.05) andthe hydrophobic/hydrophilic (HB/HI, r ¼ �0.46, P < 0.001) ratios,and positively with the carbohydrate C/methoxyl C (CC/MC) ratio(r ¼ 0.51, P < 0.001). An extensive correlation analysis of all signalsof the 13C-CPMAS NMR spectra from litter types showed severalrestricted regions significantly related to litter decay rate. Thestrongest positive and negative coefficients were found for tworestricted fractions of the O-alkyl C (70e75 ppm) and the methoxyland N-alkyl C (52e57 ppm) regions, respectively, with higherabsolute values compared to the corresponding wide regions.A simple empirical index calculated as the ratio between the rela-tive abundance of the two restricted regions (i.e. 70e75/52e57;Bonanomi et al., 2011) showed the most consistent correlationpattern with litter decay rate among all the parameters tested inthis study (Fig. 5). This was evident considering spectral data fromundecomposed materials and short-, middle-, and long-term decay

5

N/ningiLN/C

30 - 90 days

r=0.35; P=0.177

0

1

2

3

4

5

0 10 20 30 40

0

1

2

3

4

0

1

2

3

4

5

0 5 10 15 20 25

0

1

2

3

4

5

0

5

10

15

20

Litte

r dec

ay ra

te (k

)

r=-0.32; P=0.217

0 - 30 days

90 - 180 days

r=0.20; P=0.466

r=-0.67; P=0.004

r=0.05; P=0.859

r=-0.09; P=0.739

Fig. 3. Correlation (Pearson’s r coefficient and associated P-value) between decay rate (k) and C/N (left) and lignin/N (right) ratios of 16 plant species litters. Top to bottom figuresrefer to early, intermediate, and late decomposition stages.

G. Bonanomi et al. / Soil Biology & Biochemistry 56 (2013) 40e48 45

rates, as well as already decomposing materials, at differentdecomposition stages (Fig. 5). In fact, for all datasets, the correlationwas positive and significant.

4. Discussion

C/N and lignin/N ratios are known as predictors of litter decayrate. In this work we found that this is true only in the case ofundecomposed materials. In fact, such indices were not signifi-cantly associated to litter decay rates when applied to alreadydecomposed materials, both showing contrasting correlation withmass loss in the different phases of the decomposition process. Onthe other hand, the proposed approach based on 13C-CPMAS NMRanalysis provided consistent improvement of litter decay ratepredictions, by the complex, but highly fitting multivariate modelbased on PCR. Interestingly, the simple empirical index 70e75/52e57, differently from the traditional indices, maintaineda consistent positive correlation with litter decay rate during alldecomposition phases, thus providing a simpler straightforwardapplication of NMR spectral data.

Our findings can be explained as follows: the progressivereduction of litter decay rate during decomposition indicatesa progressive deterioration of litter biochemical quality likelyrelated to: (i) the fast decay of labile C fraction, progressivelyreducing the relative content of more easily degradable matter; (ii)the progressive decrease of available nutrients; (iii) the slowdecomposition of resistant C fractions, with consequent

accumulation of persistent organic compounds (e.g. lignin). In thiscontext, C/N and lignin/N ratios seem unable to describe all theobserved changes of litter quality, with limited predictivity ofdecomposition dynamics related to changes of N concentration andC quality observed during the decay process. The increase of Nconcentration during decomposition of the 16 plant species litter isconsistent with previous findings from a range of environmentsand litter types (Berg and McClaugherty, 2008). However, it is alsoknown that N content has a dual, contrasting effect on litterdecomposition, enhancing the decay rate during the early stages(up to 30e40% of mass loss) and limiting it thereafter (Berg andMatzner, 1997). At the beginning of the process, high N contentsustains large microbial populations that rapidly consume labile Ccompounds, resulting in a high mass loss rate. In contrast, high Nconcentration at the later stages inhibits mass loss by favouring theformation of recalcitrant chemical complexes with lignin, thoughthe microbiological and biochemical mechanisms have not yet fullyclarified (Hatakka, 2001). In fact, Berg et al. (1996) reported a higherdecomposition limit (i.e. the amount of mass remaining at whichthe decay rate approaches zero) in nitrogen-rich litters withinitially fast decay, compared to more lignified plant tissues,decomposing at an initially lower, but less variable rate. Theobserved pattern of correlation between litter N content and massloss rate, with opposite signs at different phases of decomposition,is consistent with the foregoing rationale, and partially explains thereasons why C/N ratio fails in describing litter quality during thewhole decomposition process. Besides, C/N ratio has a further,

-1.0 -0.5 0.0 0.5 1.0-1.0

-0.5

0.0

0.5

1.0

PC I (scores)-4 -2 0 2 4

-4

-2

0

2

4

)sgnidaol(IICP

Observed litter decay rate ( )k

0 5 10 15 20

0

5

10

15

20

(etar

yacedrett ildetcide rP

)k

)serocs(IICP

PC I (loadings)

Fig. 4. Principal component regression (PCR) between 13C-CPMAS NMR spectralregions and decay rate (k) of litter types. Top: loading vectors of spectral regions. Litterdecay rate (k) is also plotted as a supplementary variable following Legendre andLegendre (1998). Middle: trajectories of plant species litters between 0 and 180 daysof decomposition, based on factorial scores. Litters are symbolized according to theseinclusion in datasets used for either fitting (open circles) or validation (filled circles) ofthe regression model (see also Table 2). Bottom: predicted and observed decay rates forlitters of the validation dataset.

Table 2Multiple regression between decay rate (k) and linear combinations of 13C-CPMASNMR spectral regions of 30 litter types (10 plant species at three stages of decom-position). Independent predictive variables in the model (multiple R ¼ 0.818,F ¼ 12.656, P < 0.001) correspond to the first four principal components (PC) in theordination of the 13C-CPMAS NMR regions (see also Fig. 3). For each predictive factorin the models, estimates for regression coefficient (B: raw value; b: standardized),associated standard errors (SEs), and statistical significance are reported. The modelequation is: k ¼ 0.608 * PC I � 1.005 * PC II þ 0.250 * PC III � 0.261 * PC IV þ 3.103.

Independent variables B B SE ba b SE t P

Intercept 3.103 0.282 11.02 <0.001PC I 0.608 0.139 0.505 0.115 4.38 <0.001PC II �1.005 0.188 �0.628 0.118 �5.34 <0.001PC III 0.250 0.454 0.065 0.118 0.55 0.587PC IV �0.261 0.455 �0.066 0.115 �0.57 0.571

a The b coefficients are obtained by first standardizing all the variables, thusallowing comparison of the relative contribution of each independent variable in theprediction of the dependent variable.

G. Bonanomi et al. / Soil Biology & Biochemistry 56 (2013) 40e4846

intrinsic limit: this index, by including the amount of total organicC, does not take into account the biochemical dynamics affectingthe C quality during the decomposition process. By analogy with C/N, also lignin/N ratio, showing negative correlationwith litter decayrate at early stages of decomposition, and either positive or notsignificant thereafter, likely undergoes the contrasting effects of Non litter decay rate. Consistently, the simple lignin content, devoidof information on N, showed higher correlation with litter decayrate compared to lignin/N ratio. Interestingly, labile C and Ncontent, which are auto-correlated, showed similar relationshipwith decay rate with positive correlation at early stages, but notsignificant association at later decomposition stages. Again, theseresults could be due to the contrasting effects of nitrogen, oreventually to the accumulation of C compounds identified as labileby proximate analyses, but yet resistant to decomposition. Thisinconsistency may reflect the limitations of proximate analyticalmethods, which cluster, as mentioned before, a range of differentcompounds (Gessner, 2005). The amount of these compounds oftenincreases during the decomposition process (Berg andMcClaugherty, 2008). For instance, in a long-term study, Prestonet al. (2009) found that, during decomposition, lignin alone couldnot account for the observed increase of acid unhydrolyzablefractions. Other studies will investigate further the characterizationof different litter types by pyrolysis-gas chromatography/massspectrometry analysis (Huang et al., 1998) and differential ther-mogravimetry calorimetry (Rovira et al., 2008).

13C-CPMAS NMR has been used to assess the suppressivecapability of organic matter to soilborne pathogens (Pane et al.,2011), and to describe litter biochemical quality (Preston et al.,2009; Ono et al., 2011) with the purpose of predicting litter massloss (Almendros et al., 2000). Our statistical model based on

20

0

5

10

15

0 2 4 6 8 10

Litte

r dec

ay ra

te (k

)

70-75 / 52-57

r=0.67; P<0.01

Fig. 5. Correlation (Pearson’s r coefficient and associated P-value) between litter decayrate (k) and the empirical index 70e75/52e57, based on restricted fractions of the O-alkyl C (61e90 ppm) and methoxyl and N-alkyl C (46e60 ppm) 13C NMR spectralregions. The linear equation was: k ¼ 1.55 * (70e75/52e57) � 3.10.

G. Bonanomi et al. / Soil Biology & Biochemistry 56 (2013) 40e48 47

principal component regression of spectral regions provided highlysignificant predictions of litter decay rate (Fig. 5, Table 2), showingthe possibility to predict decay rate in controlled conditions basedon organic C data only. This is in accordance to recent process-based modelling of litter decomposition (Incerti et al., 2011) inwhich reliable simulations of litter decay rate from field observa-tions are provided based on organic C pools and climatic input data,irrespective of amount and timing of N amendments, which insteadare required as input parameters in most of current C-cycle models(e.g. Del Grosso et al., 2009). Both experimental and modellingevidences are based on datasets from Mediterranean ecosystems,and are yet to be investigated under different environmentalconditions. However, limited to the tested operational conditions,our PCR model could be suggested as a suitable tool for modellingapplications, requiring a reliable empirical definition of litterquality and decay rate. On the other hand, our multivariate resultssupport the multiple chemical nature of litter decompositiondynamics. Consistently with previous studies (Preston et al., 2009;Ono et al., 2011), we observed a general decrease of the relativeabundance of O-alkyl C in litter materials, and a progressiveincrease of alkyl C and methoxyl C, as decomposition wasproceeding. The rates of these biochemical changes were particu-larly high during the first 30 days of decomposition, probably inrelation to the not limiting environmental conditions used in themicrocosm experiment. Litter biochemical quality assessed by 13C-CPMAS NMR was more predictive of decay rate compared to theindexes derived from classic elemental and proximate chemicalanalyses. However, at a first glance, wide regions of the 13C-CPMASNMR spectrum (i.e. O-alkyl C, methoxyl C, aliphatic-alkyl C)and indexes derived from literature (alkyl C/O-alkyl C, HB/HI, andCC/MC ratios, Spaccini et al., 2000; Almendros et al., 2000; Kögel-Knabner, 2002) did not provide a satisfactory result, showingsignificant correlation with litter decay rate, but with worseperformance than labile C and lignin indicator. This was surprisingconsidering that the alkyl C/O-alkyl C and CC/MC ratios are bothconsidered robust indicators of the degree of litter decomposition(Almendros et al., 2000; Mathers et al., 2007; Ono et al., 2011). Thealkyl-C/O-alkyl-C ratio, negatively correlated with litter decay rate,increased as decomposition proceeded due to both the relativedecrease of carbohydrates (O-alkyl-C) and the relative increase ofthe hydrophobic by-products of decomposition (alkyl C), whichinclude fungal and microbial spoilage. On the other hand, therelative increase of lignin C (methoxyl and N-alkyl C) and thecontemporary decrease of carbohydrate C (O-alkyl C), resulted ina rapid decrease of the CC/MC ratio, which indeed was positivelycorrelated with litter decay rate. Further analyses led to the iden-tification of restricted 13C-CPMAS NMR spectral regions whichallowed a better understanding of observed litter mass losspatterns. First, we identified two restricted fractions of the O-alkylC and methoxyl C regions (70e75 ppm and 52e57 ppm, respec-tively) by excluding poorly correlated peaks from the correspond-ing wider regions. Then, the simple ratio of such restricted regions(70e75/52e57) resulted more highly correlated with litter decayrate. The two restricted O-alkyl C and methoxyl C regions can besuggested as corresponding to crucial components of litterbiochemical quality: the C2, C3, and C5 of carbohydrates, and themethoxyl C of lignin (Kögel-Knabner, 2002; Bonanomi et al., 2011).This index, in part, is comparable with the holocellulose/lignin ratioproposed by McClaugherty and Berg (1987). Interestingly, thisindex has also been proven significantly correlated with litterphytotoxic effects (Bonanomi et al., 2011). On this base, we suggestthat this new index could be used to characterize the progressivedegradation of litter biochemical quality occurring during decom-position processes, with such changes ascribed to the progressivedecrease of labile C coupled with the increase of lignin-related C.

5. Conclusions

Our study provided clear-cut evidence that the use of C/N andlignin/N ratios as descriptors of litter quality is limited to unde-composed materials, with the same indices unable to predict massloss of already decomposed litter. In contrast, an approach basedon 13C-CPMAS NMR analysis led to more predictive descriptions oforganic matter quality throughout the whole investigateddecomposition process. We showed that the relationship betweenlitter quality and its decay rate can be satisfactorily predicted bymultivariate statistical models based on 13C-CPMAS NMR spectralregions. Besides, a new index of litter biochemical quality (70e75/52e57) provided a consistent improved correlation with litterdecay rate, compared to both proximate chemical analyses andother 13C-CPMAS NMR indexes from literature. We are aware thatour experiment was based on a not exhaustive number of littertypes, decomposing under optimal conditions of temperature andwater content. Consequently, further studies should investigatethe consistency of the proposed indexes under more limitingconditions, extending the analysis to materials from otherecosystems (e.g. agro-ecosystems, grasslands, boreal and tropicalforests). A practical limit of our study might be related to thelimited availability of solid state 13C NMR spectroscopy, which is,so far, accessible to few laboratories. However, we showed thatthis method provides good insights on the chemical dynamics ofdecomposition processes. Further comparative studies between13C NMR spectroscopy and other analytical methods will eventu-ally support the identification of other indicators for litter masspredictions.

Acknowledgements

The 13C-CPMAS NMR measurements were performed at theCERMANU-Interdepartmental Research Centre, University ofNapoli Federico II.

Appendix. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at doi:10.1016/j.soilbio.2012.03.003.

References

Adair, E.C., Parton, W.J., Del Grosso, S.J., Silver, W.L., Harmon, M.E., Hall, S.A.,Burke, I.C., Hart, S.C., 2008. Simple three-pool model accurately describespatterns of long-term litter decomposition in diverse climates. Global ChangeBiology 14, 2636e2660.

Aerts, R., 1997. Climate, leaf litter chemistry and leaf litter decomposition interrestrial ecosystems: a triangular relationship. Oikos 79, 439e449.

Almendros, G., Dorado, J., González-Vila, F.J., Blanco, M.J., Lankes, U., 2000. 13C NMRassessment of decomposition patterns during composting of forest shrubbiomass. Soil Biology and Biochemistry 32, 793e804.

Attiwill, P.M., Adams, M.A., 1993. Nutrient cycling in forests. New Phytologist 124,561e582.

Berg, B., Johansson, M., Ekbohm, G., McClaugherty, C., Rutigliano, F., Virzo DeSanto, A., 1996. Maximum decomposition limits of forest litter types:a synthesis. Canadian Journal of Botany 74, 659e672.

Berg, B., Matzner, E., 1997. The effect of N deposition on the mineralization of C fromplant litter and humus. Environmental Review 5, 1e25.

Berg, B., McClaugherty, C., 2008. Plant Litter: Decomposition, Humus Formation andCarbon Sequestration, second ed. Springer-Verlag, Berlin, Heidelberg.

Bonanomi, G., Incerti, G., Antignani, V., Capodilupo, M., Mazzoleni, S., 2010.Decomposition and nutrient dynamics in mixed litter of Mediterranean species.Plant and Soil 331, 481e496.

Bonanomi, G., Incerti, G., Barile, E., Capodilupo, M., Antignani, V., Mingo, A.,Lanzotti, V., Scala, F., Mazzoleni, S., 2011. Phytotoxicity, not nitrogen immobi-lization, explains plant litter inhibitory effects: evidence from solid-state 13CNMR spectroscopy. New Phytologist 191, 1018e1030.

Burke, I.C., Kaye, J.P., Bird, S.P., Hall, S.A., McCulley, R.L., Sommerville, G.L., 2003.Evaluating and testing models of terrestrial biogeochemistry: the role oftemperature in controlling decomposition. In: Canham, C.D., Lauenroth, W.K.

G. Bonanomi et al. / Soil Biology & Biochemistry 56 (2013) 40e4848

(Eds.), Models in Ecosystem Science. Princeton University Press, Princeton, NJ,pp. 225e253.

Coq, S., Souquet, J.M., Meudec, E., Cheynier, V., Hättenschwiler, S., 2010. Interspecificvariation in leaf litter tannins drives decomposition in a tropical rain forest ofFrench Guiana. Ecology 91, 2080e2091.

Cornelissen, J.H.C., Thompson, K., 1997. Functional leaf attributes predict litterdecomposition rate in herbaceous plants. New Phytologist 135, 109e114.

Cornwell, W.K., Cornelissen, J.H.C., Amatangelo, K., Dorrepaal, E., Eviner, V.T.,Godoy, O., Hobbie, S.E., Hoorens, B., Kurokawa, H., Pérez-Harguindeguy, N.,Quested, H.M., Santiago, L.S., Wardle, D.A., Wright, I.J., Aerts, R., Allison, S.D., vanBodegom, P., Brovkin, V., Chatain, A., Callaghan, T.V., Díaz, S., Garnier, E.,Gurvich, D.E., Kazakou, E., Klein, J.A., Read, J., Reich, P.B., Soudzilovskaia, N.A.,Vaieretti, M.V., Westoby, M., 2008. Plant species traits are the predominantcontrol on litter decomposition rates within biomes worldwide. Ecology Letters11, 1065e1071.

Del Grosso, J., Ojima, D.S., Parton, W.J., Stehfest, E., Heistemann, M., DeAngelo, B.,Rose, S., 2009. Global scale DAYCENT model analysis of greenhouse gas emis-sions and mitigation strategies for cropped soils. Global and Planetary Change67, 44e50.

Fogel, R., Cromack Jr., K., 1977. Effect of habitat and substrate quality on Douglas firlitter decomposition in Western Oregon. Canadian Journal of Botany 55,1632e1640.

Gessner, M.O., 2005. Proximate lignin and cellulose. In: Graca, M.A.S., Bärlocher, F.,Gessner, M.O. (Eds.), Methods to Study Litter Decomposition. A Practical Guide.Springer Verlag, The Netherlands, pp. 115e120.

Gholz, H.L., Wedin, D.A., Smitherman, S.M., Harmon, M.E., Parton, W.J., 2000. Long-term dynamics of pine and hardwood litter in contrasting environments:toward a global model of decomposition. Global Change Biology 6, 751e765.

Gillon, D., Joffre, R., Ibrahima, A., 1994. Initial litter properties and decay rate:a microcosm experiment on Mediterranean species. Canadian Journal of Botany72, 946e954.

Gillon, D., Joffre, R., Ibrahima, A., 1999. Can litter decomposability be predicted bynear infrared reflectance spectroscopy? Ecology 80, 175e186.

Goebel, M., Hobbie, S.E., Bulaj, B., Zadworny, M., Archibald, D.D., Oleksyn, J.,Reich, P.B., Eissenstat, D.M., 2011. Decomposition of the finest root branchingorders: linking belowground dynamics to fine-root function and structure.Ecological Monographs 81, 89e102.

Hatakka, A., 2001. Biodegradation of lignin. In: Hofman, M., Stein, A. (Eds.),Biopolymers. Lignin, Humic Substances and Coal, vol. 1. Wiley, Weinheim,pp. 129e180.

Hättenschwiler, S., Coq, S., Barantal, S., Tanya, I., 2011. Leaf traits and decompositionin tropical rainforests: revisiting some commonly held views and towardsa new hypothesis. New Phytologist 189, 950e965.

Huang, Y., Stankiewicz, B.A., Eglinton, G., Snape, C.E., Evans, B., Latter, P.M.,Ineson, P., 1998. Monitoring biomacromolecular degradation of Calluna vulgarisin a 23-year field experiment using solid state 13C-NMR and pyrolysis-GC/MS.Soil Biology and Biochemistry 30, 1517e1528.

Hunt Jr., E.R., Piper, S.C., Nemani, R., Keeling, C.D., Otto, R.D., Running, S.W., 1996.Global net carbon exchange and intra-annual atmospheric CO2 concentrationspredicted by an ecosystem process model and three-dimensional atmospherictransport model. Global Biogeochemical Cycles 10, 431e456.

Incerti, G., Bonanomi, G., Giannino, F., Piermatteo, D., Castaldi, S., Fioretto, A.,Papa, S., De Marco, A., Fierro, A., Maggi, O., Persiani, A.M., Rutigliano, F.A.,Feoli, E., Virzo De Santo, A., Mazzoleni, S., 2011. Litter decomposition in Medi-terranean ecosystems: modelling the controlling role of climatic conditions andlitter quality. Applied Soil Ecology 49, 148e157.

Jolliffe, I.T., 2002. Principal Component Analysis. Springer, Berlin, Germany.

Kögel-Knabner, I., 2002. The macromolecular organic composition of plant andmicrobial residues as inputs to soil organic matter. Soil Biology and Biochem-istry 34, 139e162.

Legendre, P., Legendre, L., 1998. Numerical Ecology, second English ed. Elsevier,Amsterdam, the Netherlands.

Liski, J., Nissinen, A., Erhard, M., Taskinen, O., 2003. Climatic effect on litterdecomposition from arctic tundra to tropical rainforest. Global Change Biology9, 575e584.

Liski, J., Palosuo, T., Peltoniemi, M., Sievänen, R., 2005. Carbon and decompositionmodel Yasso for forest soils. Ecological Modelling 189, 168e182.

Mathers, N.J., Jalota, R.K., Dalal, R.C., Boyd, S.E., 2007. 13C-NMR analysis of decom-posing litter and fine roots in the semi-arid Mulga Lands of southern Queens-land. Soil Biology and Biochemistry 39, 993e1006.

McClaugherty, C., Berg, B., 1987. Cellulose, lignin and nitrogen concentrations as rateregulating factors in late stages of forest litter decomposition. Pedobiologia 30,101e112.

Meentemeyer, V., 1978. Macroclimate and lignin control of litter decompositionrates. Ecology 59, 465e472.

Melillo, J.M., Aber, J.D., Muratore, J.F., 1982. Nitrogen and lignin control of hardwoodleaf litter decomposition dynamics. Ecology 63, 621e626.

Olson, J.S., 1963. Energy storage and the balance of producers and decomposers inecological systems. Ecology 44, 322e331.

Ono, K., Hiradate, S., Morita, S., Ohse, K., Hirai, K., 2011. Humification processes ofneedle litters on forest floors in Japanese cedar (Cryptomeria japonica) andHinoki cypress (Chamaecyparis obtusa) plantations in Japan. Plant and Soil 338,171e181.

Pane, C., Spaccini, R., Piccolo, A., Scala, F., Bonanomi, G., 2011. Compost amendmentsenhance peat suppressiveness to Pythium ultimum, Rhizoctonia solani andSclerotinia minor. Biological Control 56, 115e124.

Parton, W.J., Ojima, D.S., Cole, C.V., Schimel, D.S., 1994. A General Model for SoilOrganic Matter Dynamics: Sensitivity to Litter Chemistry, Texture, andManagement. In: Quantitative Modelling of Soil Forming Processes, SSSASpecial Publication 39, 147e167 pp.

Pastor, J., Post, W.M., 1986. Influence of climate, soil moisture, and succession onforest carbon and nitrogen cycles. Biogeochemistry 2, 3e27.

Preston, C.M., Nault, J.R., Trofymow, J.A., 2009. Chemical changes during 6 years ofdecomposition of 11 litters in some Canadian forest sites. Part 2. 13C abundance,solid-state 13C NMR spectroscopy and the meaning of ‘‘lignin’’. Ecosystems 12,1078e1102.

Rovira, P., Vallejo, V.R., 2007. Labile, recalcitrant, and inert organic matter inMediterranean forest soils. Soil Biology and Biochemistry 39, 202e213.

Rovira, P., Kurz-Besson, C., Coûteaux, M.M., Vallejo, V.R., 2008. Changes in litterproperties during decomposition in European pine forests: a differential ther-mogravimetry and scanning calorimetry study. Soil Biology and Biochemistry40, 172e185.

Rovira, P., Rovira, R., 2010. Fitting litter decomposition datasets to mathematicalcurves: towards a generalised exponential approach. Geoderma 155, 329e343.

Spaccini, R., Piccolo, A., Haberhauer, G., Gerzabek, M.H., 2000. Transformation oforganic matter from maize residues into labile and humic fractions of threeEuropean soils as revealed by 13C distribution and CPMAS-NMR spectra.European Journal of Soil Science 51, 583e594.

Swift, M.J., Heal, O.W., Anderson, J.M., 1979. Decomposition in Terrestrial Ecosys-tems. In: Studies in Ecology, vol. 5. Blackwell Scientific Publications, Oxford.

Taylor, B.R., Parkinson, D., Parsons, W.F.J., 1989. Nitrogen and lignin content aspredictors of litter decay rates: a microcosm test. Ecology 70, 97e104.

Yang, H.S., Janssen, B.H., 2000. Amono-compartmentmodel of carbonmineralizationwith a dynamic rate constant. European Journal of Soil Science 51, 517e529.