Nitrogen dynamics within and between decomposing leaves, bark and branches in Eucalyptus planted forests

Soil Biology & Biochemistry 101 (2016) 55e64

Contents lists available at ScienceDirect

Soil Biology & Biochemistry journal homepage: www.elsevier.com/locate/soilbio

Nitrogen dynamics within and between decomposing leaves, bark and branches in Eucalyptus planted forests Antoine Versini a, c, d, *, Jean-Paul Laclau b, Louis Mareschal b, c, Claude Plassard e, Leki Alpiche Diamesso c, Jacques Ranger d, Bernd Zeller d a

CIRAD, UPR Recyclage et risque, Station de La Bretagne, 97743 Saint-Denis, Reunion CIRAD, UMR Eco&Sols, Ecologie Fonctionnelle & Biog
eochimie des Sols & Agro-
ecosyst emes, F34060 Montpellier, France Centre de Recherche sur la Durabilit
e et la Productivit
e des Plantations Industrielles, BP 1291 Pointe Noire, Congo d INRA, UR 1138, Biog
eochimie des Ecosyst emes Forestiers, Champenoux, France e INRA, UMR Eco&Sols, Ecologie Fonctionnelle & Biog
eochimie des Sols & Agro-
ecosyst emes, F34060 Montpellier, France b c

a r t i c l e i n f o

a b s t r a c t

Article history: Received 12 November 2015 Received in revised form 26 May 2016 Accepted 28 June 2016 Available online 15 July 2016

Nitrogen transfer between litter components is often presented as a key mechanism responsible for the synergistic effect of litter mixtures on decomposition rates. The litter cover is a heterogeneous environment stemming from the input of chemically distinct materials and the transfer of nutrients in this patchy environment is likely to fulfil the specific needs of microbial communities in each microenvironment. Our study aimed to gain insight into the factors controlling N dynamics within and between leaves and woody components in the litter cover. We used 15N labelling to discriminate endogenous and exogenous N and to measure N transfers between three types of litter components, viz. leaves and twigs (L þ T), bark and branches, in 162 litter bags for more than 2 years in two Congolese Eucalyptus forests with contrasting N status (low-N vs high-N litter). Large quantities of N were released from N-rich L þ T at the end of the study while early release of leachable N was only observed for the high-N L þ T. Exogenous N was only incorporated in N-poor litter components (bark and branches) and a net increase in N compared to the initial quantities only occurred in the low-N bark. The bi-directional N transfers observed between litter components were most likely microbially-mediated rather than driven by abiotic leaching. Nitrogen transfers were controlled by the N status of both source and sink litter components, contrary to the diffusion theory based on concentration gradient. For a given source, more N was transferred to N-rich than to N-poor sink components. Our results suggested that the microbial community might control both the quantity of N available to be transferred to other microsites and the quantity that is actually transferred, presumably because the potential for N immobilization may be limited in N-poor litter components. Interactions among micro-environments can favor chemical convergence from distinct litter components to humified organic matter along the decay continuum. © 2016 Elsevier Ltd. All rights reserved.

Keywords: Litter decomposition Nitrogen mineralization Nutrient transfer 15 N-labelling Eucalyptus plantations

1. Introduction In forest ecosystems, litter fall and subsequent litter decomposition constitute major fluxes of energy and matter in the biogeochemical cycles (Sayer, 2006; Swift et al., 1979). Although the role of climatic conditions, litter biochemical composition and decomposer communities in decomposition processes have been extensively studied (Aerts, 1997; Couteaux et al., 1995; Zhang et al.,

* Corresponding author. CIRAD, UPR Recyclage et Risque, Station de La Bretagne, 40 chemin de Grand Canal, CS 12014, 97743 Saint-Denis, Cedex 9, Reunion. E-mail address: [email protected] (A. Versini). http://dx.doi.org/10.1016/j.soilbio.2016.06.034 0038-0717/© 2016 Elsevier Ltd. All rights reserved.

2008), predicting the decomposition rates of forest litter remains a challenging task. A major reason is probably the patchy nature of litter cover and more particularly, the chemical diversity among plant organs and tree species in forest ecosystems. In most cases, litter mixtures combining several species decompose at different rates from what would be predicted from the dynamics of each species separately (Blumfield et al., 2004; Gartner and Cardon, 2004). Non-additive litter decomposition indicates that interactions can occur between micro-environments stemming from the input of chemically distinct litter components (Cuchietti et al., 2014). In this heterogeneous environment, the transfer of nutrients between decomposing litter components is likely to play a

56

A. Versini et al. / Soil Biology & Biochemistry 101 (2016) 55e64

major role in the decomposition process, fulfilling the specific needs of microbial communities in each micro-environment €ttenschwiler et al., 2005). Most of the studies dealing with (Ha nutrient dynamics within the litter cover have focused on N transfer between leaves from different species in a natural forest perspective. Some studies based on budget calculations suggested that net transfers of nutrients between leaves of the litter cover commonly occur from nutrient-rich tree species towards nutrient-poor tree species (Briones and Ineson, 1996; Salamanca et al., 1998). Over recent decades, isotopic studies based on 15N labelling have improved our understanding of the processes driving N dynamics within and between litter components. They demonstrated that the release of endogenous N throughout leaf decomposition is commonly associated with the incorporation of exogenous N in forest ecosystems (Berg, 1988; Blair et al., 1992; D’Annunzio et al., 2008; Setala et al., 1996; Zeller et al., 2000). Schwendener et al. (2005) showed 15N transfers from N-rich to N-poor leaves in mixture, but no synergistic effect on N retention within decomposing leaves. Schimel and H€ attenschwiler (2007) showed in 15Nlabelled leaf mixtures that N transfers were higher from N-rich to N-poor leaves than the reverse. They also showed that these transfers were controlled by the N status of the source leaf, rather than by the difference in N status between the leaves or by the characteristics of the sink leaves. Their results agree with N mineralization theories (Schimel and Bennett, 2004), which consider that microbes at a specific microsite control how much N is available to move to another microsite in the soil. The composition of microbial communities may therefore be a key factor controlling nutrient transfers. Lummer et al. (2012) recently demonstrated that 15N can be predominantly transferred by saprotrophic fungi in temperate forests, rather than passively by leaching. Their results suggested that N-rich bacterial-dominated litter can be a source of small quantities of N, whereas N-poor fungal-dominated litter can act as an important N source in litter mixtures. They concluded that both absolute and relative differences in initial C:N ratios of co-occurring species in the litter layer need to be considered for understanding N dynamics in decomposing litter mixtures. As far as we are aware, isotopic studies have never been carried out to experimentally quantify nutrient transfers between leaves, bark and branches throughout the decomposition processes in forest ecosystems. In Congolese Eucalyptus plantations, branch and bark components account for about 25% of the litterfall dry mass over the rotation cycle (Laclau et al., 2003a) and 45% of the organic matter left on the forest floor at harvesting (Versini et al., 2014b). Woody components with high C:N ratios are likely to immobilise N in the litter cover and reduce N leaching after forest disturbance (Vitousek and Matson, 1985), as experimentally shown by mixing litter components from Eucalyptus plantations in a lysimeter experiment (Gomez-Rey et al., 2008). Gradual release of N from the litter cover synchronized with plant uptake after disturbance may also contribute to explaining very low losses by leaching in planted forests managed in short rotations in tropical sandy soils (Mareschal et al., 2013; Versini et al., 2014a; Muller da Silva et al., 2013). The influence of the initial N concentration of decomposing leaves on net N fluxes (release or accumulation) in litter components throughout the decomposition processes is well documented (Berg and Laskowski, 2005). However, our poor understanding of the factors controlling the N dynamics between leaves and woody components in the litter cover prevent finetuning of the fertilization timing in tropical planted forests. We set out to gain new insight into N fluxes between leaves and twigs (L þ T), bark and branches throughout the decomposition of

litter components in tropical forest ecosystems. A complete factorial design using litter-bags containing 15N-labelled components was set up in two sites of contrasting N status in the Congo to estimate the dynamics of endogenous and exogenous N in each litter component, and the quantities of N transferred between the litter components. We put forward the hypotheses that i) N release is related to the initial N status of the litter component, and ii) N transfers between litter components are controlled by the N status €ttenschwiler (2007). of the source, as proposed by Schimel and Ha 2. Materials and methods 2.1. Study area The study was conducted at Kondi in the coastal area of Congo (4 340 S, 11 540 E, 100 m elevation). The climate is sub-equatorial with a rainy season from October to May and a dry season from June to September. Mean annual rainfall is about 1350 mm, and the mean annual temperature is 25  C with limited seasonal variations of about 5  C (Versini et al., 2013). The soil is classified as Ferralic Arenosols (FAO classification). Briefly, this soil is characterized by a homogeneous sandy texture down to more than ten metres, moderately acidic soil pH, and very low quantities of exchangeable base cations and organic matter. The soil mineralogy is dominated by quartz and kaolinite and nutrient bearing minerals are very scarce. A thorough description of the soil at our study area can be found in Mareschal et al. (2011). The experiment was set up in AprileMay 2009, between the previous harvest (March 2009) and re-planting Eucalyptus trees (June 2009), in two adjacent sites (200 m away) growing on the same soil type. Carbon and nutrient cycling have been intensively studied in these two sites since afforestation of the native herbaceous savannah (e.g. Mareschal et al., 2013; Nouvellon et al., 2012; Versini et al., 2013, 2014b). While the hemicellulose, cellulose and lignin contents in each litter component were of the same order of magnitude for the two sites (considering that only one composite sample per site and per component was analysed), the N concentrations were 35e46% lower in one stand (low-N plot) compared to the other (high-N plot) as a result of contrasting N accumulation patterns for the two Eucalyptus clones of each site (Table 1). Further details on Eucalyptus clone characteristics and silvicultural backgrounds of each site are given in Versini et al. (2014b). 2.2. Labelling of tree components 15

N-labelled litter material was obtained from a previous field trial initiated 20 km from our study area. During this trial, started in 2005, 20 trees of each Eucalyptus clone were watered with a 15Nammonium-nitrate solution (20% 15N atom excess, single application). The labelled trees were cut in March 2009 and three components were distinguished as follows: leaves and twigs <1 cm in diameter (L þ T), bark and branches (>1 cm in diameter). The final enrichment levels (d15N) were 259 and 232‰ for L þ T, 440 and 329‰ for bark, and 264 and 193‰ for branches of the low-N and high-N Eucalyptus clones, respectively. 2.3. Experimental design The same tree components (L þ T, bark, branches) were collected unlabelled for the two clones at Kondi in March 2009. The N contents were not significantly different for unlabelled and labelled tree components (data not shown). Litter-bags were prepared using a 1 mm nylon mesh, with internal dimensions of 22 cm  30 cm  5 cm. This mesh size allowed retention of leaf

A. Versini et al. / Soil Biology & Biochemistry 101 (2016) 55e64

15N-labelled

litter during fragmentation, thereby preventing litter loss from the bags. We established a complete factorial design in which each bag contained one 15N-labelled component and two unlabelled components. Three types of bags were therefore prepared: 15N-labelled L þ T with unlabelled branches and bark, 15N-labelled branches with unlabelled L þ T and bark, and 15N-labelled bark with unlabelled L þ T and branches (Fig. 1). The biomass of tree components was measured destructively when the two sites were harvested in March 2009 (Versini et al., 2014a). The mass of the litter components in the litter-bags was matched to the quantities of litter left on the surface of the forest floor at harvesting. In the site producing low-N litter, each bag contained approximately 150 g DM of L þ T, 95 g DM of bark and 40 g DM of branches. In the site with high-N litter, each bag was filled with 145 g DM of L þ T, 100 g DM of bark and 26 g DM of branches. The high-N litter-bags were exclusively put into the high-N site and the low-N litter-bags were exclusively put into the low-N site. The experimental design therefore comprised two sites (high-N vs low-N litter), 3 combinations of labelled vs unlabelled litter components, 9 sampling dates, and 3 replicates of litter-bags per sampling date for a total of 162 litter-bags. The bags were placed in the inter-row on the forest floor accumulated during the previous rotation cycle in three subplots (at a minimum distance of 10 m from each other) on 29 April 2009 in the low-N site and on 25 May 2009 in the high-N site.

57

litter

Unlabelled litter Leaves Branches Bark +twigs

Leaves Branches Bark +twigs

15N

15N

transfer from leaves+twigs to branches and bark

transfer from branches to leaves+twigs and bark

15N

transfer from bark to leaves+twigs and branches

Fig. 1. Simplified diagram of the experimental design, illustrating the reciprocal labelling approach used to evaluate multi-directional N transfers between litter components within each litter-bag.

2.4. Sampling and chemical analyses Three bags for each combination of labelled vs unlabelled litter components were recovered every 3 months in each site over two years. The three components in each bag were separated and roots were carefully removed by hand. Samples were oven-dried till constant weight at 65  C and weighed. Samples were then ground and sent to France for analysis. Carbon and N concentrations as well as 15N excess were determined in each sample by Elemental Analysis-Isotopic Ratio Mass Spectrometer (Delta S, Thermofinnigan, Germany) at INRA (Nancy, France). Water-extractable, hemicellulose, cellulose and lignin proportions were determined for one sample of mean C:N ratio for each component in each site, according to Van Soest et al. (1991) with modifications. The chitin assay was used to determine the glucosamine content which can be used as an indicator of fungal colonization (Joergensen and Wichern, 2008; Wallander et al., 2013). Glucosamine is assumed to be mainly derived from the chitin contained in the fungal cells. The contribution of bacterial murein to total glucosamine concentration was found to be negligible in roots and rhizosphere soil samples (Appuhn and Joergensen, 2006) and the contribution of chitin from the exoskeleton of microarthropods to the glucosamine content of soils is probably minimal, as their biomass is typically below 0.5% of the fungal biomass (Beare, 1997; Joergensen and Wichern, 2008). In our study, the contribution of non-biomass microbial residues to glucosamine content was probably very low in comparison with soils since litter was freshly colonized and

fungal chitin rapidly degraded (Fernandez et al., 2016). Samples (15 mg of dry weight) for each sampling date, site and litter component were acid-hydrolysed in 1 ml of 6N HCl for 16 h at 80  C and the released glucosamine was assayed colorimetrically as described previously (Vignon et al., 1986). The dry mass of the samples and the quantities of N transferred were expressed per unit area (g DM m2 and g N m2, respectively) since the quantities of litter components in the litter-bags were matched with the quantities of litter left on the forest floor at harvesting (g m2). The amount of 15N tracer (15N excess, mg 15 N m2) was calculated from the equation: 15

  Nexcess¼ A15 Nsample A15 Ncontrol Nsample masssample 

where A15N were the abundance units

(1)

 15 N A15 N ¼ 15 Nþ14 N; % , Nsample

was the N concentration in the sample (mg N g1 DM) and mass2 15 sample was the dry mass of the sample (g DM m ). A Ncontrol was the background value for the same unlabelled litter component measured at the outset of the study in each site. Exogenous N can be incorporated into litter components concomitantly to the release of endogenous N throughout the

Table 1 Initial N concentration, C:N ratio, water- and detergent-soluble, hemicellulose, cellulose and lignin fractions of litter components in the sites with low-N and high-N litter. Final N contents after 2 years of decomposition are also indicated. Means and standard deviations (n ¼ 9) are indicated for initial N concentrations, C:N ratios and final N quantities. Sites

Litter component

Low-N litter Leaves þ Twigs Bark Branches High-N Leaves þ Twigs litter Bark Branches

N concentration (%)

C:N ratio

Water-soluble (%)

Detergent-soluble (%)

Hemicellulose (%)

Cellulose (%)

Lignin (%)

Final N amount (g m2)

± ± ± ± ± ±

43.1 ± 1.3 228.6 ± 10.9 350.7 ± 13.2 26.8 ± 1.0 123.6 ± 5.4 219.1 ± 15.3

28 22 11 32 23 9

14 4 4 16 4 7

9 8 7 12 7 10

31 53 65 20 47 63

18 14 13 20 19 11

11.2 ± 4.4 3.7 ± 0.7 0.8 ± 0.3 12.2 ± 1.3 4.2 ± 1.0 0.7 ± 0.4

1.17 0.20 0.15 1.94 0.37 0.23

0.34 0.09 0.07 0.77 0.15 0.18

58

A. Versini et al. / Soil Biology & Biochemistry 101 (2016) 55e64

decomposition period (Zeller et al., 2001). The quantity of endogenous N within a litter component over the decomposition period was calculated from the 15N mass excess recovery in the 15Nlabelled litter component and the quantities of N. 15 N TX

EndoNTX ¼ 15

NT0

 NT0

(2)

where EndoNTX was the quantity of endogenous N still remaining on date TX (g N m2), 15NTX and 15NT0 were the quantities of excess 15 N contained in the litter component on date TX and on the date of bag positioning on the soil surface (T0), respectively (mg 15N m2), NT0 was the quantity of N contained in the same litter component on date T0 (g N m2). The proportion of exogenous N in each litter component was calculated on each date deducting the endogenous N content from the total N content:

ExoNTX ¼ NTX  EndoNTX

(3)

The gross transfer of N between two litter components inside the litter-bags between T0 and TX was calculated from the quantity of 15N initially contained in the source component transferred into the sink component until date TX: 15 N sin kTX

TransNTX ¼ 15

NsourceT0

 NsourceT0

(4)

where TransNTX was the quantity of N transferred from the source component to the sink component on date TX (g N m2), 15Nsink-TX was the quantity of excess 15N recovered in the sink component on date TX (mg 15N m2), 15Nsource-T0 and Nsource-T0 were the respective quantities of excess 15N (mg 15N m2) and total N (g N m2) initially contained in the source component. Lastly, the net transfer of N between two litter components was deduced from differential gross N-flows between these two components.

2.5. Statistics The two sites were distinguished to bring out the main trends of N dynamics within leaves and twigs, bark and branches, respectively. To determine whether net N transfers occurred between litter components, we compared gross N transfers between two litter components using t-tests on each sampling date. In addition, a 3-way mixed-effects ANOVA was performed for each of the three litter combinations (1/L þ T 4 bark, 2/L þ T 4 branches, 3/ bark 4 branches) to assess significant differences of bilateral gross transfers between two litter components over the whole study period. The fixed factors were site (low-N and high-N site), time (sampling dates), transfer direction (from N-rich to N-poor component vs from N-poor to N-rich component) and the interactions between these factors. The random factor was site*transfer direction*repetitions and residuals were modelled by a first-order autoregressive correlation model to account for the correlations between sampling dates. We compared the quantity of N transferred from a given component into the other two components with t-tests in order to investigate sink control after 1 and 2 years of study. On the other hand, the source effect was tested by comparison of N transfers from two different components to the same components. To evaluate whether 15N transfer was controlled by the N status of the source or the sink litter component, we performed a 3-way ANOVA, with time (first and second years), the component type (L þ T, bark or branch) of the source and the component type of the sink as factors. The SAS 9.2 software package (SAS Inc., Cary, NC, USA) was

used and the probability level chosen to determine significance was P < 0.05. 3. Results 3.1. Main changes during decomposition of the litter components in the litter-bags Steady decomposition of the L þ T component was observed with an average 36% of mass loss in the first year and 29% in the second year (Fig. 2ab). By contrast, the mass loss of bark and branches increased from the first year (16e20%) to the second year (43e46%). While mass loss rates were similar for the low-N and high-N sites in the first year of decomposition, they were about 13% higher for the high-N site than for the low-N site in the second year. In the L þ T component, the water-soluble proportion dropped from 30% to 12% over the first year of decomposition and then remained unchanged over the second year (Fig. 2cd). Glucosamine contents reached approx. 2.5 mg g1 for the L þ T components, after the first four months of the dry season across the two sites (Fig. 2gh). After one year of decomposition, mean glucosamine contents reached 1.3, 0.8 and 4.5 mg g1 for branches, bark and L þ T, respectively. Over the second year of decomposition, glucosamine contents was multiplied by 2 in L þ T and branch components and by a factor of 5 in the bark, irrespective of the site (Fig. 2gh). 3.2. Dynamics of endogenous and exogenous N over 2 years of decomposition In the L þ T component, N quantities steadily decreased in both sites, down to 35 ± 9% of the initial N quantity after two years of decomposition (Fig. 3). Whatever the site, exogenous N at the end of the study period accounted for only about 10% of the initial N content, but 27% of the remaining N in the L þ T component. The C:N ratio of the L þ T component decreased from 43 to 29 over the first 577 days of the study in the site with the low-N litter, and from 25 to 21 in the site with the high-N litter (Fig. 2ef). On the last two sampling dates, N contents sharply decreased in the L þ T component, which led to an increase in the C:N ratios up to the initial values in the two sites. In bark, endogenous N was released concomitantly with the incorporation of exogenous N in both sites (Fig. 3). In the site with high-N litter, 40% of the initial N content was recovered as endogenous N after two years of decomposition and 37% as exogenous N. The endogenous N was more strongly retained over time in the decomposing bark of the low-N litter, with 86% of the initial N quantity still remaining as endogenous N after 2 years of decomposition (Fig. 3). Large quantities of exogenous N incorporated in bark over the 2 years of decomposition led to a decrease in the C:N ratio and a net increase in total N in the site with the low-N litter (i.e. net N accumulation). In branches, N was slowly released with 77% of the initial quantity of N still remaining after two years of decomposition, on average, in the two sites (Fig. 3). Although 63% of the endogenous N in branches was released after 2 years of decomposition, large amounts of exogenous N were incorporated throughout the decomposition period. Exogenous N accounted for about 40% of the total N content in branches after 2 years of decomposition in the low-N litter, and 63% in the high-N litter. 3.3.

15

N transfer between the components in the litter-bags

Gradual 15N enrichment occurred in all the unlabelled litter components over the decomposition period, which led to N transfers ranging from 0.05 g N m2 (from bark to branches) to

A. Versini et al. / Soil Biology & Biochemistry 101 (2016) 55e64

Low-N litter

High-N litter

Remaining mass (%)

100

a

80 60

b

Branches

40 20

Bark L+T

0 40 Water-soluble fraction (%)

59

c

d

20

0

Glucosamine content (μg mg DM -1 )

f

e

300

50 40

200

30 20

100

10

0

0

16

h

g

12

C:N ratio

C:N ratio

400

8 4 0

0

200

400

600

800

0

200

400

600

800

Days from April 29, 2009 Fig. 2. Time series of remaining mass relative to the initial amount for the low-N litter a) and the high-N litter b), water-soluble fraction for the low-N litter c) and the high-N litter d), C:N ratio for the low-N litter e) and the high-N litter f) and glucosamine contents (an indicator of fungal colonization) for the low-N litter g) and the high-N litter h) for leaves and twigs (empty circles), bark (grey-filled triangles) and branches (black-filled squares). The C:N ratio of leaves and twigs are represented with a second y-axis (dashed empty circles). Standard errors are shown for remaining mass (n ¼ 9), C:N ratios (n ¼ 9) and glucosamine contents (n ¼ 2) but the water-soluble fraction was determined for bulk samples (n ¼ 1). Dry seasons (JuneeSeptember) are indicated by shaded areas.

1.66 g N m2 (from L þ T to bark) over 2 years, on average, for the two sites (Fig. 4). The gross N transfer from N-rich L þ T to N-poor bark was significantly greater than the reverse transfer (from bark to L þ T) on most of the sampling dates (Table 2, Fig. 4a). The gross N transfers between L þ T and branches were of the same magnitude in the two directions (Table 2, Fig. 4b). Despite higher N concentrations in bark than in branches (Table 1), N transfers were significantly greater from branches to bark than the reverse (Table 2, Fig. 4c). After two years of decomposition, gross N transfers from bark and branches to L þ T accounted for 7 ± 1% and 1 ± 0% of the N quantity in the leaves, respectively (Fig. 4 and Table 1). Gross N transfers from L þ T and branches amounted to 41 ± 14% and 3 ± 2% of the amount of N within the bark, respectively, after 2 years of decomposition. The quantities of N in the branches at the end of the study period originated from L þ T for 35 ± 12% and from bark for 6 ± 3%. The quantities of N transferred from bark and branches to other litter components were of the same magnitude in the first and the second year of decomposition (Fig. 5). By contrast, the quantities of N transferred from L þ T to bark and branches were 3.5-fold higher in the second year of decomposition than in the first year. The quantities of N transferred were about 6 times as high from L þ T to

bark as from L þ T to branches (Fig. 5). Similarly, the quantities of N transferred from bark to L þ T were 6 times as high as from bark to branches. Nevertheless, the quantities of N transferred from the branches to the other components (L þ T and bark) were of the same magnitude. The total quantity of N transferred among the three litter components in the litter-bags (sum of gross transfers) accounted for approx. 15% of the total quantity of N remaining after 2 years of decomposition, across the two sites. The 3-way ANOVA showed that N transfers between litter components after 1 and 2 years of decomposition were controlled by both source and sink components (Table 3). 4. Discussion 4.1. Nitrogen dynamics within decomposing litter components Nitrogen was retained in the three litter components throughout the 2 years of decomposition as indicated by the continuous decrease in the C:N ratio. This well-documented pattern suggests strong microbial immobilization of N after mineralization, whereas C was mainly lost by microbial respiration (Melillo et al., 1989). This result was not surprising in our tropical

A. Versini et al. / Soil Biology & Biochemistry 101 (2016) 55e64

High-N litter

120 100 80 60 40 20 0

a

b

Total-N Endo-N Exo-N

c

160

d

120

Bark

Relative N amount (% )

Low-N litter

Leaves and twigs

60

80 40 0 160

e

f

Branches

120 80 40

0 0

200

400

600

800

0

200

400

600

800

Days from April 29, 2009 Fig. 3. Time series of total N quantity (empty squares), endogenous N (black-filled circles) and exogenous N (grey-filled triangles) quantities compared to the initial N quantity (%) in leaves and twigs of the low-N litter a) and the high-N litter b), bark of the low-N litter c) and the high-N litter d) and branches of the low-N litter e) and the high-N litter f). Endogenous-N quantities were estimated from equation (2) and exogenous-N quantities deduced from total and endogenous-N quantities using equation (3). Standard errors are shown for total N (n ¼ 9) and endogenous N (n ¼ 3) amounts. Dry seasons (JuneeSeptember) are indicated by shaded areas.

forest ecosystem since N contents within litter components are commonly low relative to microbial requirements. Nutrient immobilization is the main mechanism that decomposers use to ensure their stoichiometric balance when litter is nutrientdepleted (Manzoni et al., 2010). In agreement with our first hypothesis, the time series of N mineralization were greatly influenced by the N status of the litter components. The rapid release of leachable N, defined as the leaching phase by Berg and Laskowski (2005), was only observed for the component with the highest N content (i.e. L þ T in the site with N-rich litter) while large losses of water-soluble compounds from the L þ T component occurred in both sites throughout the first year of decomposition. This pattern suggests rapid microbial immobilization of N after mineralization in the site with low-N litter, while the higher N availability in the site with high-N litter was likely to supply the N requirements of microbes, and therefore to induce a net N release. The sharp increase in fungal colonization over the same period in the L þ T component of the high-N site suggests that a fraction of the mineralized N was immobilized by fungi. Large quantities of N were released in the last stage of decomposition in the two sites, but only for the L þ T component. Surprisingly, the C:N ratios for the L þ T component rose back close to initial values at the end of the study period, while the C:N ratios of bark and branches were still decreasing. This feature suggests that a threshold was reached after 1.5 years of decomposition in the L þ T component, and that N mineralization offset immobilization beyond that threshold. Empirical breakpoint values between net immobilization and net mineralization have been shown in tropical and temperate agro-ecosystems (Seneviratne, 2000; Trinsoutrot

et al., 2000; Palm et al., 2001), but no general pattern emerged among ecosystems (Berg and Laskowski, 2005; Manzoni et al., 2010). The incorporation of exogenous N was only substantial for the bark and branch components in the litter-bags. In particular, N accumulation after 2 years of decomposition (i.e. net increase relative to initial amounts) was only observed in the bark of the site with low-N litter. The exogenous N was most likely derived from other litter components contained in the same bag and/or from the surrounding litter (Frey et al., 2003). The amounts of exogenous N derived from throughfall solutions in the litter components were probably very low due to foliar N uptake in the crown of Eucalyptus trees in the Congo (Laclau et al., 2003b) and since tree crowns did not reach the middle of the inter-row where the bags were placed in these young planted forests. At the end of the study period, about half the quantity of N in the branches was exogenous, mainly supplied by L þ T. Berglund et al. (2013) considered that large quantities of N transfers between litter components supported the theory of a litter continuum from distinct parent material to homogeneous soil organic matter (Melillo et al., 1989). In our study, the early release of N from L þ T in the site with high-N litter, N accumulation in bark of the low-N litter, as well as the final release of N from L þ T demonstrated the influence of the N status on N mineralization and illustrated how chemically-distinct litter components decomposing in mixtures converged towards common chemical properties. Nevertheless, such large quantities of N transferred between litter components in our study raise questions about the nature of these fluxes and the pathways involved.

A. Versini et al. / Soil Biology & Biochemistry 101 (2016) 55e64

Fig. 4. Reciprocal transfers of N between leaves and twigs (L þ T), bark and branches in the litter-bags estimated from 15N measurements using equation (4). Nitrogen transfer from L þ T to bark (black-filled squares) and from bark to L þ T (white-filled squares) a), from L þ T to branches (black-filled squares) and from branches to L þ T (whitefilled squares) b), from bark to branches (black-filled squares) and from branches to bark (white-filled squares) c). Standard errors are shown (n ¼ 6) and significant differences between the quantities of N transferred on each date are indicated by * (P < 0.05). Dry seasons (JuneeSeptember) are indicated by shaded areas.

4.2. Nitrogen transfer between litter components Three main pathways could be involved in N transfer between litter components: abiotic transfer via leaching of nitrate, ammonium and organic compounds in gravitational solutions (Tukey,

61

1970); mineralization and immobilization by microbial communities growing on each litter component (Schimel and H€ attenschwiler, 2007) via water-film transport; and fungal transport of N via mycelial hyphae. Leaching is known to mainly occur in the first phase of decomposition (Tukey, 1970; Ibrahima et al., 1995), which is consistent with the sharp decrease in hydrosoluble contents observed along the first year in the present study. However, the N transfer rates were constant or increased from the first to the second year even though hydrosoluble losses and precipitation were at their highest level during the first year (Versini et al., 2013). This pattern suggests a low contribution of N leaching to account for total N transfers between litter components, in agreement with previous studies (Tiunov 2009, Lummer et al., 2012). The bark and branch components were not strongly colonized by fungi throughout the first year of decomposition, consistently with the shift from bacterial- to fungal-dominated microbial communities in the second year, in the first 5 cm of the mineral soil layer (Versini et al., 2014b). Our results suggest that N immobilization within woody components as well as N transfers from these components during the first year of decomposition were mainly mediated by bacterial communities. Schimel and H€ attenschwiler (2007) showed that 15e30% of the amount of 15N initially contained in bacteria growing on leaf litter were transferred to others leaf components over 28 days of incubation. Glucosamine contents and N transfers from leaves to woody components increased concomitantly during the second year of decomposition in our study. The fungal colonization of litter components decomposing in close proximity was susceptible to increase N transfers from N-rich to N-poor litter components since transport and diffusion of elements via fungal mycelia is a well-documented process (Olsson, 1995; Tlalka et al., 2002; Tiunov, 2009; Lummer et al., 2012). Despite some limitations of the chitin assay method, the pattern observed in our study suggests that fungi might be strongly involved in N transfers from leaves to woody components during the second year of the study. In agreement with our second hypothesis, N transfers between litter components were related to the N status of the source component. The richer N source supplied much more N to a given litter component than the poorer N source in our mixtures of woody and non-woody litter components (Fig. 5). This finding confirms that microbial communities in a specific site control the quantity of N available to be transferred, as predicted by the N mineralization theory (Schimel and Bennett, 2004). Nevertheless, N transfers were also controlled by the N status of the sink component in our study. Indeed, L þ T and bark (sources) supplied more N to the richer N component than to the poorer N component in the litter-bags. This result contrasted with those of Schimel and H€ attenschwiler (2007) who reported a one-way control of the N concentration of the source over N transfers between leaves of different N status. The strong sink effect on N transfers in our study was probably the result of the large difference in N status between litter components. This two-way control of N source and N sink did not match with the diffusion theory that argues that nutrient transfer rates should be a function of a concentration gradient from the source to the sink (Hillel, 1998; Lindahl et al., 2001; Schimel and

Table 2 Effect of site (low-N and High-N sites), time (8 sampling dates), transfer direction (TD; N-poor to N-rich component or N-rich to N-poor component), and factor interactions on N transfers for each combination of litter component using a 3-way ANOVA (n ¼ 3). The quantity of N transferred was calculated from Equation (4) with the 15N recovered in unlabelled litter components on each sampling date. Significant effects (P < 0.05) are indicated in bold and with an asterisk. Combination

Site

Time

Transfer direction (TD)

Site  time

Site  TD

Time  TD

Site  time  TD

L þ T 4 Bark L þ T 4 Branches Bark 4 Branches

F1,8 ¼ 0.1 F1,8 ¼ 0.1 F1,8 ¼ 5.8*

F7,56 ¼ 16.0* F7,56 ¼ 4.6* F7,56 ¼ 13.0*

F1,8 ¼ 13.5* F1,8 ¼ 0.4 F1,8 ¼ 10.1*

F7,56 ¼ 0.8 F7,56 ¼ 1.2 F7,56 ¼ 5.4*

F1,8 ¼ 0.3 F1,8 ¼ 0.2 F1,8 ¼ 0.1

F7,56 ¼ 6.7* F7,56 ¼ 1.8 F7,56 ¼ 3.5*

F7,56 ¼ 0.7 F7,56 ¼ 1.1 F7,56 ¼ 3.3*

62

A. Versini et al. / Soil Biology & Biochemistry 101 (2016) 55e64

1st year N transferred (g N m-2)

1.5

2nd year b

a

*

Sinks L+T Ba Br

1

*

0.5

*

* 0 L+T

Bark

Branches

L+T

Bark

Branches

Sources (15N-labelled component) Fig. 5. Gross N transfers over the first year a) and the second year of decomposition b) between the source components and the sink components in the litter-bags. Nitrogen fluxes were estimated from 15N measurements using equation (4). Nitrogen fluxes over the second year were calculated deducting fluxes of the first year from those calculated for the whole study period of 2 years. Error bars indicate SEs of the mean across the two sites (n ¼ 6). * for each source component indicates significant differences between the quantities of N transferred to the 2 sink components.

Table 3 Effect of time (first and second years), source and sink litter components on N transfers using a 3-way ANOVA (n ¼ 6). The quantity of N transferred was calculated from Equation (4) with the 15N recovered in unlabelled litter components after one and two years of decomposition. The N fluxes over the second year were calculated deducting fluxes of the first year from those calculated for the whole study period. Significant effects are indicated in bold, *** indicates a significant effect at the probability level of P < 0.001, ** for P < 0.01 and * for P < 0.05. Time

Source

Sink

Time  source

Time  sink

Source  sink

Time  source  sink

F1,59 ¼ 4.5*

F2,59 ¼ 14.2***

F2,59 ¼ 10.6***

F2,59 ¼ 5.3**

F2,59 ¼ 2.9

F1,59 ¼ 6.5*

F1,59 ¼ 3.1

€ttenschwiler, 2007). Indeed, gross N transfers were much higher Ha from L þ T to bark than from L þ T to branches in our study, which suggests that the N status of the sink can limit the magnitude of the N transfers. Our results indicate that the potential for a given material to receive N is limited and positively correlated to its initial N concentration. This finding is in agreement with Berg and Cortina (1995), who showed empirically that the N increase in a litter component throughout decomposition was linearly related to its initial N concentration. Which mechanisms explain the link between the initial N status of a litter component and its potential for N retention? We hypothesise that this relationship might be regulated by the microbial communities associated to the litter component, but a physico-chemical mechanism such as lignin-N chemical binding may also be involved (Berg and Laskowski, 2005). Further research using fine scale imaging techniques such as transmission electron microscopy combined to nanoscale secondary ion mass spectrometry shall be used to address these key issues. Our study also showed that the sink effect on N transfers was not detectable for the poorest N source of the study (i.e. branches). Indeed, the branch component gave as much N to L þ T as to the bark component. Nitrogen availability, limited by microbial needs, was probably too low in N-poor branches to saturate the bark component and induce a sink effect, as compared to the L þ T sink component. Overall, these results show that the N status of the source component in the litter controls the quantity of N available to be transferred, while the N status of the sink component limits the magnitude of N transferred. This finding drove general pattern of gross N transfers but not always the net N balance between the components. Although N concentrations were lower in the branches than in the bark, much more N was transferred from the branches to the bark than the reverse. Moreover, the net N transfers were balanced between L þ T and branches despite huge differences in N status. Low gross N

transfers from L þ T and bark to branches cannot be attributable to low immobilization rate in branches since the N availability was probably low in this N-poor component as discussed above. This feature may rather indicate that the nitrogen of L þ T and bark components was not readily available to branches and may be preferentially transferred between them. Substantial net N transfer were observed from N-rich L þ T to N-poor bark supporting the assumption that fungi were involved in N transfers between these two components. Indeed, nutrient transport can apparently occur in all possible directions in fungal mycelia, but concentration gradients result in net nutrient transfer from N-rich sources to N-poor sinks (Lindahl et al., 2001). The net N transfers from L þ T and branches in favor of the bark component explained the pattern of exogenous N incorporation, demonstrating that the bark component exhibited a strong potential for N accumulation. Debarking on site should therefore be recommended to reduce N losses by deep leaching in nutrient-poor tropical planted forests. To conclude, the use of isotopically labelled litter components allowed us to demonstrate that bi-directional N transfer occurs between leaves and woody litter in a tropical planted forest. The use of microbial biomass measurements was not sufficient to explicitly demonstrate the predominance of microbial mechanisms over N transfers but provided support to the hypothesis developed in this study. Microbial communities might control both the quantity of N available to be transferred to other microsites and the quantity that is actually transferred, presumably because the potential for N immobilization may be limited in N-poor litter components. Different micro-environments can be in close proximity in the litter cover and N transfers are therefore likely to occur between microbial communities growing on chemically-distinct litter components. These interactions among micro-environments are likely to favor chemical convergence from distinct litter components to humified organic matter along the decay continuum. Fungal

A. Versini et al. / Soil Biology & Biochemistry 101 (2016) 55e64

communities may play a particular role in this process by promoting concentration gradients and net N transfers between components of contrasting N status. Consequently, temporal changes in microbial community composition might affect microenvironment interactions and thereby the decomposition rate of the whole litter cover. Determining the relative importance of physico-chemical, bacterial and fungal pathways driving nutrient transfers remains challenging in long-period in situ studies but antibiotic substances shall be sparingly used to suppress biological drivers. Our study highlights the importance of considering total litter diversity in the mixture in decomposition studies, whereas woody bark and branch components are often neglected. Further isotopic studies including both single-component and mixedcomponent litter-bags should be carried out to elucidate the link between nutrient transfers and the so-called non-additive effect. Acknowledgements We wish to acknowledge the constructive feedback of two anonymous reviewers and of editor-in-chief K. Ritz, they made a number of very insightful points and suggestions for improvement of iterations of our manuscript. We acknowledge CRDPI, the Republic of Congo and the EFC company. We thank J.-C. Mazoumbou, T. Matsoumbou and S. Ngoyi for litter-bag sampling and processing. We are grateful to Stephan H€ attenschwiler for his suggestions and to Peter Biggins for reviewing the present paper. The authors would like to thank the certified facility in Functional Ecology (PTEF OC 081) at UMR 1137 EEF and UR1138 BEF (INRA Nancy-Lorraine) for the isotopic analyses. The PTEF facility is supported by the French National Research Agency through the Laboratory of Excellence ARBRE (ANR-11-LABX-0002-01). References Aerts, R., 1997. Climate, leaf litter chemistry and leaf litter decomposition in terrestrial ecosystems: a triangular relationship. Oikos 79, 439e449. Appuhn, A., Joergensen, R.G., 2006. Microbial colonisation of roots as a function of plant species. Soil Biol. Biochem. 38, 1040e1051. Beare, M.H., 1997. Fungal and bacterial pathways of organic matter decomposition and nitrogen mineralization in arable soil. In: Brussard, L., Ferrera-Cerrato, R. (Eds.), Soil Ecology in Sustainable Agricultural Systems. CRC Press, Boca Raton, pp. 37e70. Berg, B., 1988. Dynamics of nitrogen (15N) in decomposing Scots pine (Pinus sylvestris) needle litter. Long-term decomposition in a Scots pine forest. VI. Can. J. Bot. 66, 1539e1546. Berg, B., Cortina, J., 1995. Nutrient dynamics in some decomposing leaf and needle litter types in a Pinus sylvestris forest. Scand. J. For. Res. 10, 1e11. Berg, B., Laskowski, R., 2005. Nitrogen dynamics in decomposing litter. In: Bjorn, B., Ryszard, L. (Eds.), Advances in Ecological Research. Academic Press, pp. 157e183. Berglund, S.L., Ågren, G.I., Ekblad, A., 2013. Carbon and nitrogen transfer in leaf litter mixtures. Soil Biol. Biochem. 57, 341e348. Blair, J.M., Crossley, D.A., Callaham, L.C., 1992. Effects of litter quality and microarthropods on N dynamics and retention of exogenous 15N in decomposing litter. Biol. Fertil. Soils 12, 241e252. Blumfield, T.J., Xu, Z.H., Saffigna, P.G., 2004. Carbon and nitrogen dynamics under windrowed residues during the establishment phase of a second-rotation hoop pine plantation in subtropical Australia. For. Ecol. Manag. 200, 279e291. Briones, M.J.I., Ineson, P., 1996. Decomposition of eucalyptus leaves in litter mixtures. Soil Biol. Biochem. 28, 1381e1388. Couteaux, M., Bottner, P., Berg, B., 1995. Litter decomposition, climate and litter quality. Trends Ecol. Evol. 10, 63e66.
rez Harguindeguy, N., Cuchietti, A., Marcotti, E., Gurvich, D.E., Cingolani, A.M., Pe 2014. Leaf litter mixtures and neighbor effects: low-nitrogen and high-lignin species increase decomposition rate of high-nitrogen and low-lignin neighbours. Appl. Soil Ecol. 82, 44e51. ^te, J.-F., Saint-Andre
, L., 2008. DecompoD’Annunzio, R., Zeller, B., Nicolas, M., Dho sition of European beech (Fagus sylvatica) litter: combining quality theory and 15 N labelling experiments. Soil Biol. Biochem. 40, 322e333. Fernandez, C.W., Langley, J.A., Chapman, S., McCormack, M.L., Koide, R.T., 2016. The decomposition of ectomycorrhizal fungal necromass. Soil Biol. Biochem. 93, 38e49. Frey, S.D., Six, J., Elliott, E.T., 2003. Reciprocal transfer of carbon and nitrogen by decomposer fungi at the soilelitter interface. Soil Biol. Biochem. 35, 1001e1004.

63

Gartner, T.B., Cardon, Z.G., 2004. Decomposition dynamics in mixed-species leaf litter. Oikos 104, 230e246. Gomez-Rey, M.X.G., Vasconcelos, E., Madeira, M., 2008. Effects of eucalypt residue management on nutrient leaching and soil properties. Eur. J. For. Res. 127, 379e386. €ttenschwiler, S., Tiunov, A.V., Scheu, S., 2005. Biodiversity and litter decompoHa sition in terrestrial ecosystems. Annu. Rev. Ecol. Evol. Syst. 36, 191e218. Hillel, D., 1998. Environmental Soil Physics. Academic Press, San Diego. Ibrahima, A., Joffre, R., Gillon, D., 1995. Changes in litter during the initial leaching phase: an experiment on the leaf litter of mediterranean species. Soil Biol. Biochem. 27, 931e939. Joergensen, R.G., Wichern, F., 2008. Quantitative assessment of the fungal contribution to microbial tissue in soil. Soil Biol. Biochem. 40, 2977e2991. Laclau, J.-P., Deleporte, P., Ranger, J., Bouillet, J.-P., Kazotti, G., 2003a. Nutrient dynamics throughout the rotation of Eucalyptus clonal stands in Congo. Ann. Bot. 91, 879e892. Laclau, J.-P., Ranger, J., Bouillet, J.-P., Nzila, J.D., Deleporte, P., 2003b. Nutrient cycling in a clonal stand of Eucalyptus and an adjacent savanna ecosystem in Congo. 1. Chemical composition of rainfall, throughfall and stemflow solutions. For. Ecol. Manag. 176, 105e119. Lindahl, B., Finlay, R., Olsson, S., 2001. Simultaneous, bidirectional translocation of 32 P and 33P between wood blocks connected by mycelial cords of Hypholoma fasciculare. New Phytol. 150, 189e194. Lummer, D., Scheu, S., Butenschoen, O., 2012. Connecting litter quality, microbial community and nitrogen transfer mechanisms in decomposing litter mixtures. Oikos 121, 1649e1655. Manzoni, S., Trofymow, J.A., Jackson, R.B., Porporato, A., 2010. Stoichiometric controls on carbon, nitrogen, and phosphorus dynamics in decomposing litter. Ecol. Monogr. 80, 89e106. Mareschal, L., Nzila, J.D.D., Turpault, M.P., Thongo M’Bou, A., Mazoumbou, J.C., Bouillet, J.P., Ranger, J., Laclau, J.P., 2011. Mineralogical and physico-chemical properties of Ferralic Arenosols derived from unconsolidated Plio-Pleistocenic deposits in the coastal plains of Congo. Geoderma 162, 159e170. Mareschal, L., Laclau, J.P., Nzila, J.D.D., Versini, A., Koutika, L.S., Mazoumbou, J.C., Deleporte, P., Bouillet, J.P., Ranger, J., 2013. Nutrient leaching under Eucalyptus plantations managed in short rotations after afforestation of an African savanna: a 14-year time series. For. Ecol. Manag. 307, 242e254. Melillo, J.M., Aber, J.D., Linkins, A.E., Ricca, A., Fry, B., Nadelhoffer, K.J., 1989. Carbon and nitrogen dynamics along the decay continuum: plant litter to soil organic matter. Plant soil 115, 189e198. Muller da Silva, P.H., Poggiani, F., Libardi, P.L., Gonçalves, A.N., 2013. Fertilizer management of eucalypt plantations on sandy soil in Brazil: initial growth and nutrient cycling. For. Ecol. Manag. 301, 67e78. Nouvellon, Y., Epron, D., Marsden, C., Kinana, A., Le Maire, G., Deleporte, P., Saint
, L., Bouillet, J.P., Laclau, J.P., 2012. Age-related changes in litter inputs Andre explain annual trends in soil CO2 effluxes over a full Eucalyptus rotation after afforestation of a tropical savannah. Biogeochemistry 111, 515e533. Olsson, S., 1995. Mycelial density profiles of fungi on heterogeneous media and their interpretation in terms of nutrient reallocation patterns. Mycol. Res. 99, 143e153. Palm, C.A., Gachengo, C.N., Delve, R.J., Cadisch, G., Giller, K.E., 2001. Organic inputs for soil fertility management in tropical agroecosystems: application of an organic resource database. Agriculture. Ecosyst. Environ. 83, 27e42. Salamanca, E.F., Kaneko, N., Katagiri, S., 1998. Effects of leaf litter mixtures on the decomposition of Quercus serrata and Pinus densiflora using field and laboratory microcosm methods. Ecol. Eng. 10, 53e73. Sayer, E.J., 2006. Using experimental manipulations to access the roles of leaf litter in the functioning of forest ecosystems. Biol. Rev. 81, 1e31. Schimel, J.P., Bennett, J., 2004. Nitrogen mineralization: challenges of a changing paradigm. Ecology 85, 591e602. €ttenschwiler, S., 2007. Nitrogen transfer between decomposing Schimel, J.P., Ha leaves of different N status. Soil Biol. Biochem. 39, 1428e1436. Schwendener, C.M., Lehmann, J., de Camargo, P.B., Luiz~ ao, R.C.C., Fernandes, E.C.M., 2005. Nitrogen transfer between high- and low-quality leaves on a nutrient15 poor Oxisol determined by N enrichment. Soil Biol. Biochem. 37, 787e794. Seneviratne, G., 2000. Litter quality and nitrogen release in tropical agriculture: a synthesis. Biol. Fertil. Soils 31, 60e64. Setala, H., Marshall, V.G., Trofymow, J.A., 1996. Influence of body size of soil fauna on litter decomposition and 15N uptake by polar in a pot trial. Soil Biol. Biochem. 28, 1661e1675. Swift, M.J., Heal, O.W., Anderson, J.M., 1979. Decomposition in terrestrial ecosystems. In: Studies in Ecology. Blackwell Scient. Publ., p. 372 Tiunov, A.V., 2009. Particle size alters litter diversity effects on decomposition. Soil Biol. Biochem. 41, 176e178. Tlalka, M., Watkinson, S.C., Darrah, P.R., Fricker, M.D., 2002. Continuous imaging of amino-acid translocation in intact mycelia of Phanerochaete velutina reveals rapid, pulsatile fluxes. New Phytol. 153, 173e184. Trinsoutrot, I., Recous, S., Bentz, B., Lineres, M., Cheneby, D., Nicolardot, B., 2000. Biochemical quality of crop residues and carbon and nitrogen mineralization kinetics under nonlimiting nitrogen conditions. Soil Sci. Soc. Am. J. 64, 918e926. Tukey, H.B., 1970. The leaching of substances from plants. Annu. Rev. Plant Physiol. 21, 305e324. Van Soest, P.J., Robertson, J.B., Lewis, B.A., 1991. Methods for dietary fiber, neutral detergent fiber, and nonstarch polysaccharides in relation to animal nutrition. J. Dairy Sci. 74, 3583e3597.

64

A. Versini et al. / Soil Biology & Biochemistry 101 (2016) 55e64

Versini, A., Nouvellon, Y., Laclau, J.-P., Kinana, A., Mareschal, L., Zeller, B., Ranger, J., Epron, D., 2013. The manipulation of organic residues affects tree growth and heterotrophic CO2 efflux in a tropical Eucalyptus plantation. For. Ecol. Manag. 301, 79e88. Versini, A., Mareschal, L., Matsoumbou, T., Zeller, B., Ranger, J., Laclau, J.-P., 2014a. Effects of litter manipulation in a tropical Eucalyptus plantation on leaching of mineral nutrients, dissolved organic nitrogen and dissolved organic carbon. Geoderma 232e234, 426e436.
, L., Versini, A., Zeller, B., Derrien, D., Mazoumbou, J.-C., Mareschal, L., Saint-Andre Ranger, J., Laclau, J.-P., 2014b. The role of harvest residues to sustain tree growth and soil nitrogen stocks in a tropical Eucalyptus plantation. Plant Soil 376, 245e260. Vignon, C., Plassard, C., Mousain, D., Salsac, L., 1986. Assay of fungal chitin and
ge
tale 24, 201e207. estimation of mycorrhizal infection. Physiol. Ve Vitousek, P.M., Matson, P.A., 1985. Disturbance, nitrogen availability and nitrogen

losses in an intensively managed loblolly pine plantation. Ecology 66, 1360e1376. Wallander, H., Ekblad, A., Godbold, D.L., Johnson, D., Bahr, A., Baldrian, P., Bjork, R.G., Kieliszewska-Rokicka, B., Kjoller, R., Kraigher, H., Plassard, C., Rudawska, M., 2013. Evaluation of methods to estimate production, biomass and turnover of ectomycorrhizal mycelium in forests soils e a review. Soil Biol. Biochem. 57, 1034e1047. Zeller, B., Colin-Belgrand, M., Dambrine, E., Martin, F., Bottner, P., 2000. Decomposition of 15N-labelled beech litter and fate of nitrogen derived from litter in a beech forest. Oecologia 123, 550e559. Zeller, B., Colin-Belgrand, M., Dambrine, E., Martin, F., 2001. Fate of nitrogen released from 15N-labeled litter in European beech forests. Tree Physiol. 21, 153e162. Zhang, D., Hui, D., Luo, Y., Zhou, G., 2008. Rates of litter decomposition in terrestrial ecosystems: global patterns and controlling factors. J. Plant Ecol. 1, 85e93.