Geoderma 232–234 (2014) 426–436

Contents lists available at ScienceDirect

Geoderma

j ourna l homepage: www.e lsev ie r .com/ locate /geoderma

Effects of littermanipulation in a tropical Eucalyptus plantation on leaching ofmineral nutrients, dissolved organic nitrogen and dissolved organic carbon

Antoine Versini a,b,c,⁎, Louis Mareschal a,b, Tiburce Matsoumbou b, Bernd Zeller c,Jacques Ranger c, Jean-Paul Laclau a,d

a CIRAD, UMR111, Ecologie Fonctionnelle & Biogéochimie des Sols & Agro-écosystèmes, F34060 Montpellier, Franceb Centre de Recherche sur la Durabilité et la Productivité des Plantations Industrielles, BP 1291 Pointe Noire, Republic of Congoc INRA, UR1138, Biogéochimie des Ecosystèmes Forestiers, Champenoux, Franced UNESP, Universidade Estadual de São Paulo, Departamento de Recursos Naturais, Botucatu, SP, Brazil

⁎ Corresponding author at: CIRAD, UMR111, Ecologie FoSols & Agro-écosystèmes, F34060 Montpellier, France. Tel

E-mail address: tonioleneo@yahoo.fr (A. Versini).

http://dx.doi.org/10.1016/j.geoderma.2014.05.0180016-7061/© 2014 Elsevier B.V. All rights reserved.

a b s t r a c t

a r t i c l e i n f o

Article history:Received 22 July 2013Received in revised form 21 May 2014Accepted 23 May 2014Available online xxxx

Keywords:Dissolved nutrientDOCDONDrainageTropical soilEucalyptus

Although many studies have shown that soil solution chemistry can be a reliable indicator of biogeochemicalcycling in forest ecosystems, the effects of littermanipulations on thefluxes of dissolved elements in gravitationalsoil solutions have rarely been investigated.We estimated the fluxes of NH4-N, NO3-N, K, Ca,Mg, Na, Cl, dissolvedorganic nitrogen (DON) and dissolved organic carbon (DOC) over the first two years after re-planting Eucalyptustrees in the coastal area of Congo. Two treatments were replicated in two blocks after clear-cutting 7-year-oldstands: in treatment R, all the litter above the mineral soil was removed before planting, and in a double slash(DS) treatment, the amount of harvest residues was doubled. The soil solutions were sampled down to adepth of 4 m and the water fluxes were estimated using the Hydrus 1Dmodel parameterized from soil moisturemeasurements in 4 plots. Isotopic and spectroscopic analytical techniques were used to assess the changes indissolved organic matter (DOM) properties throughout the transfer in the soil. The first year after planting, thefluxes of NH4-N, K, Ca, Mg, Na, Cl and DOC in the topsoil of the DS treatment were 2–5 times higher than in R,which showed that litter was a major source of dissolved nutrients. Nutrient fluxes in gravitational solutionsdecreased sharply in the second year after planting, irrespective of the soil depth, as a result of intense nutrientuptake by Eucalyptus trees. Losses of dissolved nutrients were noticeably low in these Eucalyptus plantationsdespite a low cation exchange capacity, a coarse soil texture and large amounts of harvest residues left on-siteat the clear cut in the DS treatment. All together, these results clarified the strong effect of litter manipulationobserved on eucalypt growth in Congolese sandy soils. DOM fluxes, as well as changes in δ13C, C:N andaromaticity of DOM throughout the soil profile showed that the organic compounds produced in the litterlayer were mainly consumed by microorganisms or retained in the topsoil. Below a depth of 15 cm, most ofthe DOC and the DON originated from the first 2 cm of the soil and the exchanges between soil solutions andsoil organic matter were low.

© 2014 Elsevier B.V. All rights reserved.

1. Introduction

Planted forests are expanding rapidly worldwide to meet theincreasing demand in forest products (FAO, 2010). Plantation forestsdesigned to meet social, economic, and environmental objectives arelikely to provide key ecosystem services and can therefore contributeto preserving natural forests (Paquette and Messier, 2010). However,the sustainability of tropical plantations is of concern since largeamounts of nutrients are exported with biomass harvesting in highlyweathered soils (Gonçalves et al., 2013; Nambiar, 2008). The manage-ment of organic residues at harvesting in tropical forests can strongly

nctionnelle & Biogéochimie des.: +33 6 59 66 29 88.

impact ecosystem functioningwith direct consequences on soil physicalproperties, nutrient stocks, carbon (C) cycling, plant growth, microbialcommunities and soil fauna (Sayer, 2006). Saint-André et al. (2008)showed in a network of tropical experiments that the poorest was thesoil, the highest was tree response to harvest residue manipulations.This pattern might reflect a key role of the nutrients released through-out residue decomposition on tree growth in nutrient-poor soils(Laclau et al., 2010; Versini et al., 2013). Although the chemicalcomposition of soil solutions is recognized as highly sensitive to man-agement practices, the effects of litter manipulation on dissolved nutri-ent fluxes have, surprisingly, received little attention in forestecosystems (Laclau et al., 2003a; Piirainen et al., 2009; Smethurstet al., 2001).

The factors driving dissolved organicmatter (DOM) fluxes in the soilare of immense importance since DOM has been recognized to play a

427A. Versini et al. / Geoderma 232–234 (2014) 426–436

fundamental role in soil functioning and C sequestration. However,studies dealing with DOM (and particularly dissolved organic nitrogen,DON) dynamics are still scarce in tropical forests (Chantigny, 2003; Fujiiet al., 2009, 2013; Möller et al., 2005). Decomposing litter is a potentialsource of DOM (Currie et al., 1996; Riaz et al., 2012) and litter manipu-lations at the soil surface are therefore likely to modify the inputs of or-ganic matter (OM) into the mineral soil (Kalbitz et al., 2007;Klotzbücher et al., 2012; Park and Matzner, 2003). The forest floor isan organic layer (O horizon) consisting of organic material in variousstages of decomposition, above themineral soil, and is therefore expect-ed to be a major source of DOM for the mineral soil. Some studies sug-gested that fresh litter was the major source of DOM entering into themineral soil (Currie and Aber, 1997; Michalzik and Matzner, 1999)while recent research provided evidences that the main source ofDOM can be the most humified part (Oa) of the O horizon (Fröberget al., 2005; Kalbitz et al., 2007; Klotzbücher et al., 2012). DOM can betransferred trough gravitational soil solutions in deep soil layers orcan interact with soil organic matter (SOM) along the soil profile(Fröberg et al., 2006; John et al., 2003; Sanderman et al., 2008). Sandysoils, that are characterized by low adsorption capacities and high infil-tration rates, are usually considered as large exporters of litter-derivedDOM in tropical regions (Aitkenhead-Peterson et al., 2007; Sandermanet al., 2008).

Clonal Eucalyptus plantations have been set up in the coastal plainsof Congo since the 70s for pulpwood production purpose. Even thoughseveral decades of breeding and researches in silviculture increasedthe productivity of these commercial plantations, the low amounts offertilizers applied make their sustainability highly sensitive to nutrientlosses by leaching during the post-harvest phase (Laclau et al., 2005;Mareschal et al., 2013). After afforestation of the native savanna, clonalEucalyptus stands are harvested every 6–9 years in Congo, before asharp decrease in annual increment. Mean productivity is about20 m3 ha−1 year−1, which is at the lower range of commercial planta-tions worldwide (Gonçalves et al., 2013). Most of the plots are re-planted after harvesting and the silvicultural practices are similar toother tropical commercial plantations managed to produce highamounts of biomass.

The objective of this studywas to investigate the effects of litter ma-nipulations at the harvest of tropical Eucalyptus plantations on thechemistry of soil solutions and the losses of nutrients, dissolved organiccarbon (DOC) and DON by leaching. Soil solutions were acidic at thisstudy site and a minor contribution of inorganic carbon to total carbonfluxes was therefore expected. Two contrasting treatments (removalvs addition of harvest residues at the soil surface) were compared toestimate the contribution of the O horizon to the input of DOM and dis-solved nutrients into themineral soil. Isotopic and spectroscopic analyt-ical techniqueswere used to assess changes inDOMchemical propertiesthroughout their transfer from the topsoil to deep soil layers.

The hypothesis of the study were that i— the O horizon was a majorsource of DOMand nutrients, highlighting the role of littermanagementon Eucalyptus growth in Congolese plantations and ii—harvesting led to

Table 1Soil properties and stocks of total C, total N, exchangeable nutrients, Fe and Al oxy-hydrox(Laclau et al., 2003b; Mareschal et al., 2011).

Soil layer Particle sizedistribution (%)

δ13C (‰)a Concentrations (g kg−1)

cm Clay Silt Sand Plot A Plot B C N K Ca Mg

0–5 7.7 2.1 87.7 −23.7 −19.2 8.90 0.53 0.01 0.02 0.015–15 7.1 2.1 88.6 −19.9 −16.7 4.87 0.29 0.01 0.01 0.0115–50 7.0 2.1 89.1 −15.0 −13.7 2.71 0.17 0.01 0.01 0.0050–100 9.9 2.2 86.6 −19.0 −19.0 1.79 0.11 0.01 0.01 0.00100–200 10.3 2.4 86.1 −24.0 −24.0 1.27 0.08 0.01 0.01 0.00200–400 11.3 2.7 85.1 −25.0 −25.0 0.90 0.06 0.00 0.01 0.00

a δ13C wasmeasured in the 0–5 and 5–15 cm soil layers just before clear-cutting in stands Asame plots in 2006 (Epron et al., 2009) and values below a depth of 50 cm were measured by

large losses of nutrients and DOMby leachingwhen fresh residueswereadded at the soil surface.

2. Material and methods

2.1. Study site description

The study site of Kondi was located on the coastal area of Congo(4° 34′ S, 11° 54′ E, 100 m elevation). The climate is sub-equatorialwith a rainy season from October to May and a dry season from Juneto September.Mean annual rainfall is about 1220mm, andmean annualtemperature is 25.5 °C with seasonal variations of about 5.0 °C. The soilis classified as Ferralic Arenosols (FAO). Briefly, this soil is characterizedby a homogeneous sandy texture down tomore than 10m, amoderate-ly acidic soil pH, and very low amounts of exchangeable ‘base’ cationsand organic matter (Table 1). The soil mineralogy is dominated byquartz and kaolinite and nutrient bearing minerals are very scarce(Mareschal et al., 2011). The vertical distribution of SOM in these soilswas characterized by high variations of δ13C along the soil profile as aconsequence of C3–C4 type vegetal successions. Before the afforestationof savanna with Eucalyptus trees, the upper 40 cm of soil was entirelycomposed of C4-derived OM (δ13C of −13‰) from the savanna(Trouvé et al., 1994), the proportion of this OMdecreased progressivelydown to the depth of 3 mwhere the SOMwas entirely composed of oldand stable C3-derived OM (δ13C of−25‰) originating from a forest re-placed 3000 years ago by the herbaceous savanna (Schwartz et al.,1992). The afforestation with Eucalyptus (δ13C of−30.5‰) led to a pro-gressive substitution of C4-SOM by C3-SOM in the upper 50 cm of thesoil profile and to an exponential δ13C decrease in SOM relative totime since afforestation (Epron et al., 2009; Trouvé et al., 1994)(Table 1).

The present study was carried out in two adjacent stands (A and Bconsidered as two blocks in this study) planted on native savannawith the same clone of a Eucalyptus hybrid (PF1, unknown species,clone 1-41). The A stand was afforested in 1992 and the B stand in2001, at a stocking density of 532 trees per ha. The A standwas harvest-ed in 2001, the stumps were killed by glyphosate application and thestand was re-planted with a more productive clone (18-52) from thehybrid Eucalyptus urophylla (ST Blake) × Eucalyptus grandis (Hill exMaid.), at a stocking density of 800 trees per ha. The soil propertiesand the chemistry of soil solutions were studied in these two standsover this period (Laclau et al., 2003b, 2005; Mareschal et al., 2013).The two stands were harvested in March 2009 and two treatmentswere immediately replicated in each stand (i.e. 4 plots):

R (Removed): the O horizon was removed from the plot (litter fromthe previous rotation as well as harvest residues),

DS (Double Slash): only debarked commercial-sized boles (top-endover-bark diameter N 2 cm) were removed. Harvest residues of the Rplots were added in the DS and uniformly distributed on the ground.This treatment left on the soil surface 6.4 kg m−2 of dry matter instand A and 4.5 kgm−2 in stand B, of which about 35%was fresh leaves.

ide down to a depth of 4 m calculated from concentrations measured at the same site

Stocks of element (g m−2)

Na Fe Al C N K Ca Mg Na Fe Al

0.01 13.00 25.00 627 37 1 2 1 1 917 17630.01 13.15 26.09 682 40 1 2 1 1 1842 36520.00 13.31 27.72 1328 81 3 3 1 1 6521 13,5840.00 14.52 30.59 1386 87 5 4 1 4 11,254 23,7050.01 15.53 34.94 1964 125 10 9 3 11 24,066 54,1520.01 16.53 40.94 2801 182 14 16 7 23 51,580 127,723

and B (Derrien, unpublished data), values in the 15–50 cm soil layer weremeasured in theTrouvé et al. (1994) in the same soil type under a nearby savanna.

428 A. Versini et al. / Geoderma 232–234 (2014) 426–436

Although land use history was different in the two blocks, thetreatments were similar (R and DS for litter and harvest residues ofthe previous Eucalyptus stand). The two stands were replanted withthe same clone (18-52) in June 2009 at a stocking density of 800 treesper ha, after glyphosate application to kill the stumps and the understo-ry. Ammonium nitrate fertilizer was applied in September 2009 at adose of 4.3 g N m−2 in all the plots.

2.2. Soil water sampling

Eight sets of nine zero-tension lysimeters (40 cm × 2.5 cm) werecarefully installed in 1997 at 2 cm soil depth (two sets per treatmentand per stand) to sample gravitational solutions below the O horizonand the surface soil. Soil solutions were also collected using ceramiccup lysimeters connected to a vacuum pump and maintained manually(daily checking) at a constant suction of about— 70 kPa. Two replicatesof tension lysimeters (TL) were set up horizontally at depths of 15 and50 cm, 1, 2 and 4 m in 1997 in each plot (a total of four TL at eachdepth per treatment in the two stands) as well as at the depth of 6 min stand A. The replicates of lysimeters were set up at different distancesfrom trees and the distance between them was more than 10 m. Soilsolutions were collected in these plots from 1998 to 2003 to assess nu-trient leaching after savanna afforestation (Mareschal et al., 2013). Thesoil solutions were collected weekly in glass bottles placed in closedpits where they were protected from light and sharp variations intemperature.

Once a week from February 2009 (1 month before harvesting andtreatment establishment) to June 2011, the solutions were carried tothe laboratory where they were kept at +4 °C. Samples were filtered(0.45 μm) and pooled proportionally to the volumes collected eachweek for a single monthly sample per lysimeter. The pH and the electri-cal conductivity were measured with a S47 SevenMulti™ (MettlerToledo, Switzerland). Two subsamples were sent to the BEF-INRA labo-ratory (Nancy, France) and to the CIRAD laboratory (Montpellier,France) for chemical analyses. The remaining soil solutions werefreeze-dried (freeze dryer Alpha 1-2 with manifold, Christ, Germany),and sent to the BEF-INRA laboratory (Nancy, France) for δ13C analyses.

2.3. Drymatter and nutrient contents in the O horizon and tree components

The aboveground litter that accumulated at the soil surface beforeharvesting was sampled at 10 random positions in each stand using a2 m2 frame. The amounts of aboveground harvest residues were esti-mated from allometric relationships established in the same site thanthe present study (Saint-André et al., 2005) and census made on 1358trees just before the harvest in the two stands. The amount of below-ground residues was measured in stand A on 12 trees covering the dis-tribution of basal areas, just before harvesting. Fine root biomass(diameter b 2 mm) was quantified using a root auger (inner diameter

Table 2Dry matter and nutrient stocks in aboveground and belowground harvest residues and litter fr

Dry matter Stocks (g m−2)

(g m−2) N K

Leaves in DS 1895 ±284 29.8 ±5.6 9.1 ±1.0Barks in DS 1245 ±37 3.4 ±0.6 3.8 ±0.3Branches in DS 490 ±412 0.8 ±0.2 0.4 ±0.1Litter in DS 1805 ±596 18.7 ±3.1 1.2 ±0.4Fine roots 248 ±62 0.8 ±0.1 0.2 ±0.0Medium-sized roots 375 ±134 0.9 ±0.2 0.3 ±0.1Coarse roots 712 ±706 1.5 ±0.2 0.3 ±0.1Total R 1335 ±1224 3.3 ±0.5 0.9 ±0.2Total DS 6770 ±1351 56.0 ±3.3 15.4 ±1.2

Aboveground harvest residueswere deposited at the soil surface only in the DS treatment. Stan(n = 12 trees), medium-sized and coarse root drymatter (n = 6 trees) and for nutrient stocksaboveground tree components and only in stand A for roots.

of 8 cm) down to a depth of 3 m, at 13 positions close to each sampledtree (12 × 13 samples per soil layer). Medium-sized (2–10mm in diam-eter) and coarse roots (diameter N 10 mm) were excavated in sixtrenches down to a depth of 3 m (3.3 m × 3.7 m × 3.0 m) (Levillainet al., 2011; Saint-André et al., 2005). The concentrations of N, K, Ca,Mg and Na were determined after mineralization by ICP-AES emissionspectroscopy (Vista, Varian, Palo Alto, California, USA) and the concen-trations of Cl by potentiometric titration in the CIRAD laboratory.

At the end of the study period (June 2011, two years after planting),eight treeswere destructively sampled in each treatment of standA. In asubsample of each aboveground tree component, N was determinedusing the Kjeldahl method after digestion in sulfuric acid (TE036/1,Tecnal, Brazil). Calcium and Mg were determined by atomic absorptionspectrophotometry (AAnalyst 100, Perkin Elmer, USA) and K by flameemission spectrophotometry (B462, Micronal, Brazil). Allometricrelationships were established using Proc NLP (SAS 9.2) to predict thebiomass and nutrient contents of aboveground tree components (leaf,branch, bark and trunk) from the diameter at breast height and theheight of the trees. Allometric equations were applied to the censusmade two years after planting.

2.4. Chemistry of soil solutions

Total K, Ca, Mg, and Na in soil solutions were determined by ICPemission spectroscopy (JY 180 Ultrace, Jobin Yvon, Longjumeau,France) and total N and C concentrations were measured using a TOC/TN analyzer (TOC-VCSN, TNM-1, Shimadzu, Kyoto, Japan). Nitrate,NH4-N and Cl were measured by colorimetry (Evolution II, Alliance in-struments). DON concentrations were estimated by subtracting inor-ganic N concentrations (NO3 + NH4) from total N concentrations.

The DOMaromaticity was assessed for each soil solution sample (fora total of 960 samples) by ultraviolet (UV) absorption spectrophotome-try. Specific absorption at 254 nm (i.e., measured absorbance divided bythe DOC concentration), referred to specific ultraviolet absorbance(SUVA), has been used for many years to assess the aromaticity ofDOM (Dilling and Kaiser, 2002; Jaffrain et al., 2007; Sanderman et al.,2008). This method is rapid, requires little sample preparation, is non-invasive, and only a small sample is necessary (Chen et al., 2002). Ultra-violet absorption spectra were measured at a constant temperature(25 °C)with a UV–visible spectrophotometer (UV-1700, Shimadzu Cor-poration, Kyoto, Japan). Spectra were acquired between 200 and500 nm at a scan rate of 150 nmmin−1 and with a uniform data pointinterval of 0.5 nm. Samples were placed in 1.5 ml UV-cuvettes(Plastibrand®, Wertheim, Germany).

The δ13C composition of DOM was measured in order to assess theorigin and the modifications of DOM throughout their transfer insoil profiles. Taking advantage that the C4-type organic matter of savan-na grass is less depleted in 13C than the C3-type organic matter ofEucalyptus (Epron et al., 2009; Trouvé et al., 1994), it was possible to

om the previous rotation just after clear cutting.

Ca Mg Na Cl

5.1 ±1.4 4.5 ±0.7 5.1 ±0.2 3.0 ±1.44.6 ±0.4 2.4 ±1.1 3.8 ±0.5 4.3 ±2.70.5 ±0.2 0.3 ±0.0 0.3 ±0.1 0.4 ±0.12.6 ±1.1 1.4 ±0.7 0.6 ±0.2 0.3 ±0.10.3 ±0.0 0.3 ±0.1 0.4 ±0.1 0.3 ±0.10.5 ±0.1 0.4 ±0.1 0.5 ±0.2 0.5 ±0.10.6 ±0.1 0.3 ±0.1 0.6 ±0.2 0.5 ±0.11.3 ±0.2 1.0 ±0.3 1.6 ±0.4 1.2 ±0.2

14.1 ±1.8 9.6 ±1.0 11.4 ±1.0 9.2 ±3.9

dard deviations are given for aboveground drymatter (n = 2 stands), fine root dry matter(n = 4 samples for nutrient concentrations). Samples were collected in the two stands for

Table3

Biom

ass,nu

trient

conc

entrations

andnu

trient

stocks

intree

compa

rtmen

tsin

thetw

otrea

tmen

tstw

oye

arsafterplan

ting

.

Biom

ass

Conc

entrations

(gkg

−1 )

Stoc

ks(g

m−2)

(gm

−2)

NK

CaMg

Na

NK

CaMg

Na

Leav

esR

156

±6

*18

.4±0.1

*4.2

±0.3

*3.3

±1.0

*2.4

±0.2

*3.2

±0.4

2.9

±0.1

*0.7

±0.0

*0.5

±0.0

*0.4

±0.0

*0.5

±0.0

*DS

289

±18

16.9

±0.9

4.9

±0.6

5.8

±0.6

2.2

±0.2

2.9

±0.4

4.9

±0.3

1.4

±0.1

1.7

±0.1

0.6

±0.0

0.8

±0.1

Barks

R13

5±4

*3.7

±0.4

1.7

±0.3

*3.2

±1.4

*1.3

±0.4

*4.4

±0.5

*0.5

±0.0

*0.2

±0.0

*0.4

±0.0

*0.2

±0.0

*0.6

±0.0

*DS

224

±12

3.8

±0.3

2.4

±0.4

8.3

±0.5

2.0

±0.3

3.4

±0.3

0.8

±0.0

0.5

±0.0

1.9

±0.1

0.4

±0.0

0.8

±0.0

Bran

ches

R30

4±12

*3.4

±0.5

*0.6

±0.2

*0.9

±0.3

*0.4

±0.1

*1.4

±0.3

*1.0

±0.0

*0.2

±0.0

*0.3

±0.0

*0.1

±0.0

*0.4

±0.0

*DS

569

±36

2.8

±0.3

1.0

±0.4

2.1

±0.6

0.5

±0.1

1.2

±0.1

1.6

±0.1

0.6

±0.0

1.2

±0.1

0.3

±0.0

0.7

±0.0

Woo

dR

650

±21

*1.8

±0.3

0.5

±0.1

0.2

±0.1

*0.1

±0.0

*0.7

±0.1

*1.2

±0.0

*0.3

±0.0

*0.2

±0.0

*0.1

±0.0

*0.4

±0.0

*DS

1136

±66

1.8

±0.3

0.6

±0.1

0.4

±0.1

0.1

±0.0

0.6

±0.1

2.0

±0.1

0.7

±0.0

0.4

±0.0

0.1

±0.0

0.7

±0.0

Total

R12

45±43

*4.6

±0.2

1.2

±0.2

*1.2

±0.4

*0.6

±0.1

*1.6

±0.1

*5.7

±0.2

*1.5

±0.1

*1.5

±0.1

*0.7

±0.0

*2.0

±0.1

*DS

2216

±13

24.3

±0.2

1.5

±0.1

2.4

±0.2

0.7

±0.0

1.3

±0.0

9.5

±0.6

3.3

±0.2

5.3

±0.3

1.6

±0.1

2.9

±0.2

Stan

dard

deviations

aregive

nforco

ncen

trations

(n=

8tree

s),b

iomassan

dstoc

ks(n

=2stan

ds).Sign

ificant

diffe

renc

esbe

twee

nthetw

otreatm

ents

(Pb0.05

)areindicatedby

*.

429A. Versini et al. / Geoderma 232–234 (2014) 426–436

discriminate old savanna-derived DOM and Eucalyptus-derived DOM.Samples collected at different depths over the whole study periodwere analyzed for δ13C with an Elemental Analyzer–Isotope RatioMass Spectrometer (EA–IRMS, Delta S, Thermo Finnigan, Germany) forthe two stands and the two treatments (139 samples). Additionally toSOM properties showed in Table 1, the C:N ratio and δ13C values ofthe 0–2 cm soil layer in the two stands were measured at harvesting.

2.5. Water balance modeling

Soil water content was monitored with two–three TDR probes(Trase, Soil moisture, USA) buried at various distances from trees atdepths of 15 and 50 cm, 1, 2, 3 and 4 m in four plots (treatments Rand DS in stands A and B). Amodel based on Richard's equation for sim-ulating one dimensional water flow (Hydrus 1D) was calibrated toquantify water flow at the depths where lysimeters were installed(Simunek et al., 2008). An internal drainage field experiment was car-ried out to calibrate thewater retention curves and hydraulic conductiv-ity parameters of the different soil layers. In each stand, 3 m3 of waterwas poured into an area of about 8 m2 inside two concentric steelrings and equipped with TDR probes down to a depth of 3 m (Laclauet al., 2005). A plastic cover prevented evaporation. Soil water contentsand pressure heads were measured hourly the first day then daily forseveral weeks to estimate retention curves and soil hydraulic conduc-tivity in each stand from the inverse method of the Hydrus model. Soilhydraulic properties were not re-measured because they aremainly de-pendent on soil physical properties that are not expected to change in afew years.

Fine root densitieswere assessed down to a depth of 3m before har-vesting and at 6, 12, 18 and 24 months after planting in treatments Rand DS. Simulations were run at a daily time step for 6 month periodsafter planting using fine root distributions measured in each treatment.Penman evapotranspiration (PET) was computed daily distinguishingbetween evaporation (10% of PET before harvesting) and tree transpira-tion (90% of PET). After harvesting, soil evaporation was considered asinversely proportional to stand LAI in each treatment, decreasing from50% of PET in R and 30% of PET in DS before planting, to 10% at canopyclosure in both treatments (i.e. two years after planting). The contribu-tions of soil evaporation and tree transpiration to the actual standevapotranspiration were roughly estimated to fit the predictions ofsoil water content to themeasurements at the depths of the TDR probesthroughout the study period.

2.6. Data analyses

The fluxes of dissolved elements were calculated monthly for eachdepth, treatment and stand, multiplying mean concentrations in the soilsolutions by the flux of gravitational water estimated from the Hydrusmodel. Throughfall and stemflow fluxes were estimated from the dailyrainfall using the equations developed in Eucalyptus plantations at thesame site (Laclau et al., 2005). Tension lysimeters did not sample soil so-lutions at a few dates when a water flux was predicted by the Hydrusmodel. Missing concentrations were then estimated as the mean of thevalues measured in samples collected in the same lysimeter at the datesimmediately before and after. Chemical analyses were performedindividually for the two lysimeters at each depth in each plot. Below adepth of 2 m, only mineral nutrients were analyzed at each date.Selected properties of DOM (δ13C and SUVA) were studied at a fewdates down to a depth of 6 m. The nutrient contents in the O horizonat harvesting and in tree components at two years after planting werecalculated multiplying the nutrient concentrations by the dry matterof each compartment.

Two-way ANOVAs were used to assess the effect of both treatmentsand stands on nutrient concentrations in solutions and stocks in treebiomass compartments two years after planting. Mixed models withsoil depth as a random factor were used to assess the effects of the

0 200 400 600 800

0 200 400 600 8000 200 400-200 600 800

0 200 400 600 800

0 200 400-200 600 800

0 200 400 600 800

0 200 400 600 800

20

30

10

2

1

4

5

2

3

1

3

4

5

1

2

20

40

60

-200

-200

-200

-200

-200

0

20

30

40

10

4

6

0

2

0 200 400-200 600 8000

20

30

40

10

20

30

40

10

00 200 400-200 600 800

0

Day since harvesting (March 15, 2009)

DON DOC

KC

on

cen

trat

ion

s (m

g l

-1)

NaCl

8

NH4-NCaMg

NO3-N

* ** * ****** * * *

**

* * * * *

**

*

0

0

0

0

Fig. 1. Time-course of element concentrations at a depth of 15 cmover the study period. The vertical solid line indicates the date of harvesting and the vertical dotted line indicates the dateof planting for treatments R (open symbol) and DS (filled symbol) in stand A (circles) and stand B (triangles). Significant differences (P b 0.05) in element concentrations between R andDS are indicated by *.

430 A. Versini et al. / Geoderma 232–234 (2014) 426–436

year after planting, the treatment and the stand on element fluxes. Datawere processed using the SAS 9.2 software package (SAS Inc., Cary, NC,USA). The probability level used to determine significance was P b 0.05.

3. Results

3.1. Nutrients stocks within soil layers and tree components

Large amounts of nutrients were contained in aboveground residuesandwere deposited at the soil surface at harvesting in DS (Table 2). Thestocks of N, K, Ca,Mg, Na and Cl in the O horizonwere, respectively, 2.8-,

Fig. 2.Dynamics of soil water contentsmeasured by 2-3 TDR probes per depth (full circles) andin treatments R (a) and DS (b), as well as at a depth of 4 m in treatments R (c) and DS (d). Thedate of planting.

12.4-, 5.0-, 6.2-, 17.8- and 23.5-fold larger after deposition of the harvestresidues than before harvesting. While a high proportion of N wascontained in the litter that accumulated before harvesting, the other nu-trients were mainly contained within the leaves and the bark of harvestresidues. The amounts of nutrients in litter left on-siteweremuch lowerin R than in DS: N, K, Ca, Mg, Na and Cl contents within roots at harvestin R represented only 6, 6, 9, 11, 14 and 13% of the total amounts withinbelow- and aboveground litter in DS, respectively (Table 2).

The aboveground biomasswas 178%higher inDS (2216 gm−2) thanin R (1245 g m−2) at two years after planting (Table 3). The concentra-tions of N and Na in most of tree compartments were significantly

simulated over the study periodwith theHydrus 1Dmodel (solid lines) at a depth of 15 cmvertical solid line indicates the date of harvesting and the vertical dotted line indicates the

Fig. 3. Fluxes of gravitational water estimated down to a depth of 5 m the first year afterplanting in treatment R (open circle) and treatment DS (filled circle) and the secondyear after planting in treatment R (open triangle) and treatment DS (filled triangle).

431A. Versini et al. / Geoderma 232–234 (2014) 426–436

higher in R than in DS at two years of age while the concentrations of Kand Ca were lower in R than in DS. Nutrient contents in two-year-oldtrees were significantly lower in R than in DS as a result of the lowbiomass of tree compartments in R.

3.2. Chemical composition of gravitational soil solutions in the top soil

Harvesting increased the concentrations of NH4-N, NO3-N, K, Ca,Mg,Na, Cl, DON andDOC at a depth of 15 cmcompared to the baseline at theend of the previous rotation (Fig. 1). The highest concentrations wereobserved at the beginning of the first rainy season following harvesting.The concentrations dropped down to pre-harvest values in the secondyear after planting, except for DOC and Na in stand A. Concentrationsof Na, K and Cl at a depth of 15 cm were about 3-fold as high in DS asin R over the first year after planting. The concentrations of Ca, Mg,NH4-N, NO3-N and DON at a depth of 15 cm did not differ significantlybetween the two treatments. Nitrate was the dominant form of N insoil solutions sampled at a depth of 15 cm in both treatments with anaverage concentrations of 3.71 mg N l−1, while mean concentrationswere 0.86 and 0.42mg l−1 for DON and NH4-N, respectively. The chem-ical composition of soil solutions was less influenced by the treatmentsin deep soil layers than in the topsoil (data not shown).While DOC con-centrations in soil solutions sampled at 2 cm soil depthwere on averagemore than twice as high in DS as in R over the first two years after plant-ing, the differences between the two treatments decreased sharply inthe mineral soil to become equivalent from 15 cm depth (Fig. 1).

Table 4P-values of stand, treatment and year effects on the annual fluxes of water and dissolved elem

Water N-NH4 N-NO3 K

Stand 0.98 0.95 0.13 0.97Treatment 0.13 0.10 0.87 b .0001Year after harvest b .0001 b .0001 b .0001 0.01Treatment × year 0.13 0.22 0.29 0.02

The sampling depth was used as a random factor in the mixed model. Significant effects at P b

3.3. Fluxes of water and dissolved elements throughout soil profiles

The Hydrus model parameterized at the same study site than in thepresent study made it possible to predict satisfactorily soil water con-tents at the depths of the TDR probes (Fig. 2). Rainfall amounted to1210 mm in the first year after planting and 1540 mm in the secondyear (Fig. 3). Fluxes of gravitational waters in deep soil layers were86 mm year−1 lower in R than in DS in the first year after planting, asa result of higher evaporation rates in the bare soils of the R plots thanunder the O horizon in the DS plots (Fig. 3). About 180 days after har-vesting, soil moisture in the 0–6 cm soil layer was no longer differentbetween the two treatments (Versini et al., 2013). Therefore, differencesin deep drainage between the treatments disappeared the second yearafter planting. This pattern was a consequence of decreasing evapora-tion rates in R (resulting from litter accumulation at the soil surface inR), as well as lower transpiration rates in R than in DS (due to a signifi-cantly lower tree growth in R than in DS). Indeed, Versini et al. (2013)estimated in stand A that belowground and aboveground biomasses inR amounted to 2.8 and 11.2 t ha−1 at age 2 years, respectively, whilethey amounted to 3.3 and 16.5 t ha−1 in DS.

Litter manipulations and time since harvesting significantly influ-enced the fluxes ofmost of the dissolved elements in gravitational solu-tions (Table 4). By contrast, those fluxes did not differ significantlybetween the two stands, except for the Na and DON fluxes that weresignificantly higher in stand A than in stand B. At a depth of 2 cm, thefluxes of K, Na and Cl were 4.7, 3.2 and 2.3-fold higher in DS than in Rplots in the first year after planting. At a depth of 1 m, the fluxes of Cl,Na, K and DOC in gravitational solutions amounted to 1.4, 1.2, 0.5 and1.6 g m−2 in R and 6.4, 5.8, 2.3 and 3.8 g m−2 in DS, respectively, inthe first year after planting (Fig. 4). The DOC fluxes at 2 cm deepamounted to 13.4 g m−2 year−1 in R and 38.0 g m−2 year−1 in DSover the study period and they dropped down to 1.9 g m−2 year−1 inR and 2.6 g m−2 year−1 in DS at a depth of 1 m. The fluxes of Ca andMg were only significantly higher in DS than in R at 2 cm soil depth(Fig. 4). The DON, NO3-N and NH4-N fluxes were not significantly influ-enced by treatments, at all soil depths. The fluxes of dissolved nutrientssharply decreased at the onset of the second year after planting with 4-fold lower nutrient fluxes over the soil profile for most of nutrients andby 44-fold lower NO3-N fluxes in the second than in the first year(Table 4). At 15 cmdepth, NO3-N, NH4-N, Ca andMg fluxes dropped, re-spectively, from about 11.4, 0.5, 0.8 and 1.3 g m−2 in the first year afterplanting to 0.5, 0.3, 0.3 and 0.2 g m−2 in the second year (Fig. 4). Thefluxes were very low in the sub-soil, whatever the element and thetreatment, with mean values of 2.0, 1.0, 0.1, 0.04, 0.17, 2.9, 0.2, 0.2 and0.04 g m−2 year−1 for Cl, Na, K, Mg, Ca, DOC, DON, NH4-N and NO3-Nat a depth of 4 m over the first two years after planting, respectively(Fig. 4).

3.4. Changes in DOM quality

DOMcharacteristics were stronglymodified throughout the transferof gravitational waters in the mineral soil in both treatments (Fig. 5).While mean SUVA values at a depth of 2 cm were 2.6 l mg C−1 m−1

in R and 3.4 l mg C−1 m−1 in DS over the first two years after planting,

ents.

Ca Mg Na Cl DON DOC

0.16 0.36 0.01 0.98 0.02 0.270.23 0.11 b.0001 b.0001 0.33 b .00010.07 b.0001 b.0001 b.0001 b .0001 0.110.31 0.18 b.0001 b.0001 0.53 0.14

0.05 are indicated in bold.

Fig. 4. Fluxes of dissolved elements in gravitational solutions down to a depth of 4m thefirst year after planting in treatment R (open circle) and treatment DS (filled circle) and the secondyear after planting in treatment R (open triangle) and treatment DS (filled triangle).

432 A. Versini et al. / Geoderma 232–234 (2014) 426–436

C:N ratio δ13CSUVA

Depth 15 cm

Depth 50 cm

Depth 1 m

Depth 2 m

Depth 3 m

Depth 4 m

Depth 6 m

Depth 2 cm

1 2 3 -30 -24 -18 -15 -120 10 20 30 0 -21-27

SOMRainLitter DOM DS DOM R

a b c

* * *

Fig. 5.Mean and standard errors of C:N ratios in soil organicmatter (SOM) and DOC:DON ratios in dissolved organicmatter (DOM) (a), SUVA of DOM (b), and δ13C in SOM andDOM (c) intreatments R and DS over the first 2 years after planting. Significant differences (P b 0.05) in DOM characteristics between DS and R are indicated by *. SOM characteristics are given inTable 1.

433A. Versini et al. / Geoderma 232–234 (2014) 426–436

they decreased down to 0.4 l mg C−1 m−1 at a depth of 6 m in bothtreatments (Fig. 5). The aromatic content (i.e. SUVA) and the DOC:DON ratio of DOMwere significantly lower in R than in DS at 2 cm soildepth. However, differences between treatments were nomore signifi-cant in the underlyingmineral soil layers. The δ13C values of DOM in thetwo treatments were 5–11‰ lower than those of SOM in the upper 1 m(Fig. 5c). The δ13C values for DOM in DS were intermediate betweenthose in R and of the litter. While δ13C values of SOMwere strongly de-pendant on soil depth, δ13C values of DOMwere littlemodified through-out the transfer of gravitational solutions. Mean δ13C values of DOMacross the treatments ranged from −28.5‰ in the top soil to−25.5‰at a depth of 2 m. δ13C values of DOM in stand B peaked after plantingat −14‰ and −19‰ at the depths of 15 cm and 1 m, respectively(Fig. 6). This pattern in C4-free plots suggested a transient release ofsavanna-derived DOM about nine years after afforestation.

4. Discussion

4.1. Effects of litter manipulation on nutrient leaching

The O horizon was a major source of nutrients dissolved in gravita-tional solutions. The fluxes of K, Na and Cl in gravitational solutionssampled in the first year after planting at a depth of 2 cm in DS

Fig. 6. Changes of DOC δ13C at soil depths of 2 cm (empty circles), 15 cm (gray squares)and 1 m (black triangles), over the first 2 years after planting in the stand B. The solidand the dashed lines indicate the δ13C of soil organic matter at the depths of 15 and 1 m,respectively.

represented 62–114% of the total amounts of those elements containedin the O horizon at planting. The potassium is easily leached in the can-opy and fromdecomposing litter (Chang et al., 2007;Maisto et al., 2011;Rees et al., 2006). The amount of Na and Cl derived from the O horizonoffset the reduction of dry depositions after harvesting, as observed forCl in other forests after disturbance (Kauffman et al., 2003; Lovett et al.,2005). Contrarily to the other elements, NO3-N fluxes at 15 cm depthwere similar for the two treatments and higher than the fluxes at2 cmdepth. This patternwas consistentwith a lack of effect of litterma-nipulation on the production of mineral N measured in situ from 7 to24 months after planting Eucalyptus trees at the same site (Nzila et al.,2002). The large production of NO3-N in the first year after harvestingin the upper soil layer also resulted from the application of ammoniumnitrate fertilizer in all the plots at the end of the first dry season afterplanting. Together, the rapid solubilization of 4.3 g N m−2 applied inthe ammonium nitrate fertilizer plus the net mineralization of 5.5 g Nm−2 year−1 estimated by Nzila et al. (2002) in an adjacent experimentof the same agematchedwith the estimations of NO3-N leached from the0–15 cm soil layer in the first year after planting (9.9 g N m−2 year−1).Nutrient leaching in the upper soil layers reflected contrasting dynamicsof release from the O horizon as well as contrasting affinities of the ionsto the exchange surfaces in the soil and large differences in tree uptakerates. The fluxes of Ca and NH4-N sharply decreased in the 0–15 cmsoil layer, probably as a result of the high Ca affinity for OM (DeSutteret al., 2006; Gomez-Rey et al., 2008), the use of NH4-N by microbes andpreferential uptake by trees. A weak adsorption of mobile anions onoxy-hydroxides might have contributed to delay their transfer in verydeep soil layers (Duwig et al., 2003).

The sharp decrease in dissolved nutrient fluxes from the first to thesecond year after planting might be accounted for by a concomitantdrop in nutrient release from leave decomposition and a rise in nutrientdemand by Eucalyptus trees. The rapid growth of Eucalyptus fine rootsled to a dense colonization of the soil volume down to a depth of 3 min the second year after planting (Bouillet et al., 2002). This featureplays amajor role to avoid nutrient losses by deep drainage in Eucalyptusplantations established on deep tropical soils (Laclau et al., 2013). Nutri-ent contents in aboveground tree components at age two years weregenerally lower than the fluxes in gravitational solutions at a depth of15 cm cumulated over the first year after harvesting. In the R treatment,the fluxes of N, K, Ca,Mg andNa in soil solutions at a depth of 15 cmoverthe first year after planting represented, respectively, 172, 128, 86, 163and 95% of the contents of those nutrients in above-ground tree

434 A. Versini et al. / Geoderma 232–234 (2014) 426–436

components at two years of age. This pattern points out the quantitativeimportance of the fluxes of dissolved nutrients to fulfill the nutritionaldemand. Nutrient losses by deep drainage were very low over thestudy period. Most of the nutrients leached in the topsoil that were nottaken up by trees were therefore retained on soil exchange surfaces inthe upper 4 m of soil or accumulated in the O horizon through litterfallin the second year after planting. In the DS treatment, the fluxes of N,K, Ca, Mg and Na in gravitational solutions at a depth of 15 cm overthe first year after planting represented, 169, 289, 68, 146 and 303% ofthe stocks in aboveground tree components at two years of age, respec-tively. Large amounts of dissolved nutrients that are not been taken uptwo years after planting in the DS treatment were therefore stored inthe soil and were likely to be taken up by trees after canopy closure.This pattern might contribute to explaining the prolonged effect of littermanipulations on tree growth up to the harvest age in the Congo (Laclauet al., 2010; Saint-André et al., 2008).

4.2. Sources and dynamics of dissolved organic matter

The upper soil layer was a major source of DOM since the highestfluxes were found at depth of 2 cm in both treatments. The lack of litterin the R treatment in the first year after planting suggested that the DOCwas probably produced from the coarsest fraction of the topsoil OM,which is generally constituted of particulate OM less 13C enriched thanthe bulk SOM (Epron et al., 2009). Indeed, the δ13C values of the DOMin the R treatment (−25.0‰ for stand B and −27.2‰ for stand A)were intermediate between Eucalyptus litter δ13C (−30.5‰) andSOM δ13C in the 0–5 cm soil layer (−19.2‰ for stand B and −23.7‰for stand A) while the O horizon has been removed and root litter ex-cluded from the zero-tension lysimeters. DOMproperties in themineralsoil usually resemble partly humified SOM processed by microbes(Currie et al., 1996). Sanderman et al. (2008) found similitudes betweenDOC properties in the mineral soil and humified silt- and clay-associated OM fractions, and hypothesized that DOC came from a spa-tially limited C pool that was smaller but more dynamic than the bulkSOC. DOM source in the coarse OM in the present study, rather than inthe fine fractions of the soil, might reflect differences of C pool dynamicsamong ecosystems and in particular themicrobial activity associated tothese OM fractions. The large contribution of the O horizon to the pro-duction of DOC was shown by the significantly higher DOC fluxes,DOC:DON ratio and aromaticity at 2 cm depth collected in DS than inR. The DOC:DON ratio of 28 at 2 cm in DS was intermediate betweenthe C:N ratio of the Oa layer of about 36 and the C:N ratio of about 20in the 0–2 cm soil layer. The DOC:DON ratio at 2 cm in R was close tothe C:N ratio of the soil, as expected since the litter was entirely re-moved in this treatment. The differences between the DOC fluxes inthe two treatments suggest that about 73 and 49% of the amounts ofDOC at a depth of 2 cm in the DS treatment derived from the O horizonin the first and the second year after planting, respectively. Assumingthat DOC compounds in DS originated both from coarse OM, as mea-sured in R, and from Eucalyptus litter, a rough calculation from δ13Cvalues confirmed that about 73 and 54% of the amount of DOC was de-rived from Eucalyptus litter in the DS treatment. Contrarily to DOCfluxes, DON fluxes were not significantly influenced by litter manipula-tion over the study period. The results of the present study showed thatthe O horizon was a major source of DOC in the DS treatment, in agree-ment with other studies in forest ecosystems (Kalbitz et al., 2007;Qualls, 2000; Riaz et al., 2012). However, the two first centimeters ofthe mineral soil were a non-negligible source of DOC and providedmost of the DON production. Yano et al. (2004) also found that themain source of DOC and DON was the mineral topsoil (0–10 cm) andnot the O horizon in an American coniferous old-growth stand. The pro-duction of DOM then resulted mainly from root exudation and biologi-cal turnover in the rhizosphere. In the present study, the contribution ofroot exudates should be low in the first year after harvesting becausefine root densities were low. Most of the DOM originating from the

0–2 cm layer came probably from the decomposition of the coarsestfraction of SOM. Furthermore, a recent study at the same site pointedout the role of particulate OM to sustain topsoil N stocks and to pro-vide mineral N (Versini et al., 2014).

Unlike what was expected, DOM losses by deep drainage were verylow despite considerable addition of fresh residues at harvesting. Only2.1 g DOCm−2 year−1 and 0.2 g DONm−2 year−1where leached on av-erage at a depth of 2 m, across the treatments and the years after plant-ing. DOC losses were at the lower range of values reported in temperateand boreal forest ecosystems (Neff and Asner, 2001), and close to DOCfluxes in Indonesian tropical forests (Fujii et al., 2011). DON fluxeswere within the range (from 0.0 to 0.9 g N m−2 year−1) reportedunder the rooting zone in temperate forest ecosystems (Michalziket al., 2001) and close to DON fluxes in Indonesian tropical forest onsandy Ultisol (0.4 g N m−2 year−1, Fujii et al., 2013). The results of thepresent study did not support the widely accepted idea of large exportof DOC in tropical sandy soils (Aitkenhead-Peterson et al., 2007; Fujiiet al., 2011; Sanderman et al., 2008). Low losses resulted from a sharpdecrease in DOC and DON fluxes from 2 cm to 15 cm, combined witha slow and progressive decrease between the depths of 15 cm and2 m. Most of DOM derived from the O horizon in the DS treatmentwas retained or mineralized in the upper 15 cm of the soil since DOMcharacteristics were no longer different between treatments at depths≥ 15 cm. The DOM decrease along the soil profile probably resultedfrom sorption on mineral surfaces (Guggenberger and Kaiser, 2003;Kaiser and Guggenberger, 2000; Sanderman and Amundson, 2008)and/or mineralization of the most labile compounds contained inDOM (Cleveland et al., 2004; Kalbitz et al., 2003). The decline in DOC:DON ratio and in aromaticity of DOM with soil depth suggested thatthe decrease in DOC and DON fluxes in the mineral soil resulted morefrom a preferential retention of high C:N hydrophobic compounds onmineral surfaces than from a mineralization of low C:N hydrophiliccompounds. Hydrophobic organic compounds have in general ahigher affinity with the soil matrix than hydrophilic compounds(Guggenberger and Zech, 1993; Kaiser et al., 2001). The clay fractionof the soil was dominated by kaolinite in the study site (Mareschalet al., 2011), which is known to have low affinity with DOC (Benkeet al., 1999). However, the amounts of Al and Fe oxy-hydroxide werelarge (Mareschal et al., 2011) and they are usually the largest sorbentsfor DOM in soils (Kaiser et al., 1996). The relationship between organicC and iron plus aluminum oxy-hydroxide proposed by Kindler et al.(2011) suggested that about 80–100% of the amount of DOC in gravita-tional solutions could be retained by sorbtion in themineral soil since Feand Al concentrations largely exceed organic C concentrations in thissite. In contrast, δ13C values of DOM showed that exchanges of C be-tween gravitational waters and SOM were low, probably because reac-tive C pools in the soil, such as organo-mineral associations are scarcein sandy soils (Sanderman et al., 2008). Nevertheless, the potential forsequestration of DOC appeared to be limited compared to carbon stocksor to other sources of carbon in this forest ecosystem. Indeed, the mea-surements made by Versini et al. (2013) in the same plots showed thataboveground litterfall, belowground litter and root turnover annuallybrought into the soil 201, 72 and 92 g C·m−2 in DS and 5, 67 and 70g C·m−2 in R over the two first years after re-planting. The low DOCfluxes in the subsoil explained why most of the SOM was still derivedfrom savanna below the depth of 50 cm, 15 years after afforestationwith Eucalyptus trees (Trouvé et al., 1994).

5. Conclusions

In agreement with the first hypothesis, the nutrition of Eucalyptustrees in this nutrient-poor sandy soil relied on the rapid release of nutri-ents from the organic layer over the first year after planting, as well ason themineralization of organicmatter in the upper soil layer. However,despite the coarse texture of the soil and large inputs of harvest resi-dues, Eucalyptus plantations were very conservative of nutrient pools.

435A. Versini et al. / Geoderma 232–234 (2014) 426–436

The organic layer was a major source of DOC, which was partially con-sumed by microorganisms and/or retained in the upper soil layer,while DON was mainly produced in the first 2 cm of the soil profile.Most of the DOC fluxes in the mineral soil probably derived from themost humified coarse fraction of the twofirst centimeters of themineralsoil. Exchanges between the aqueous phase and SOMwere low over thefirst two years after planting Eucalyptus trees, and DOM propertiesremained unchanged below a depth of 15 cm. Further studies focusingon themineral topsoil are needed to assess the long term consequencesof litter management on the dynamic of organic matter in tropicalforest.

Acknowledgments

The authors acknowledge CRDPI, the Republic of Congo and EFC fortheir support. The authors are particularly grateful to J.C. Mazoumbou,S. Ngoyi, A. Diamesso and A. Kinana from the CRDPI for field measure-ments. The authors thank CIRAD researchers such as L. Saint-André,P. Vigneron, P. Bouvet and D. Epron for sample transports from Congoto French laboratories. The authors also thank K. Alary and N. Bouarfafrom the CIRAD-AMI laboratory of Montpellier, G. Nourisson andS. Bienaimé from the BEF-INRA laboratory of Nancy for carrying outthe analyses. Finally, the authorswould like to thank the certified facilityin Functional Ecology (PTEF OC 081) from UMR 1137 EEF and UR1138BEF in research center INRA Nancy-Lorraine for its contribution to theisotopic analysis. The PTEF facility is supported by the French NationalResearch Agency through the Laboratory of Excellence ARBRE (ANR-12-LABXARBRE-01).

References

Aitkenhead-Peterson, J.A., Smart, R.P., Aitkenhead, M.J., Cresser, M.S., McDowell, W.H.,2007. Spatial and temporal variation of dissolved organic carbon export from gaugedand ungaugedwatersheds of Dee Valley, Scotland. Effect of land cover and C:N.WaterResour. Res. 43, W05442.

Benke, M.B., Mermut, A.R., Shariatmadari, H., 1999. Retention of dissolved organic carbonfrom vinasse by a tropical soil, kaolinite, and Fe-oxides. Geoderma 91, 47–63.

Bouillet, J.-P., Laclau, J.-P., Arnaud, M., M'Bou, A.T., Saint-André, L., Jourdan, C., 2002.Changes with age in the spatial distribution of roots of Eucalyptus clone in Congo: im-pact on water and nutrient uptake. For. Ecol. Manage. 171, 43–57.

Chang, S.-C., Wang, C.-P., Feng, C.-M., Rees, R., Hell, U., Matzner, E., 2007. Soil fluxes ofmineral elements and dissolved organic matter following manipulation of leaf litterinput in a Taiwan Chamaecyparis forest. For. Ecol. Manage. 242, 133–141.

Chantigny, M.H., 2003. Dissolved and water-extractable organic matter in soils: a reviewon the influence of land use and management practices. Geoderma 113, 357–380.

Chen, J., Gu, B., LeBoeuf, E.J., Pan, H., Dai, S., 2002. Spectroscopic characterization of thestructural and functional properties of natural organic matter fractions. Chemosphere48, 59–68.

Cleveland, C.C., Neff, J.C., Townsend, A.R., Hood, E., 2004. Composition, dynamics, and fateof leached dissolved organic matter in terrestrial ecosystems: results from a decom-position experiment. Ecosystems 7, 175–285.

Currie, W.S., Aber, J.D., 1997. Modeling leaching as a decomposition process in humidmontane forests. Ecology 78, 1844–1860.

Currie, W.S., Aber, J., McDowell, W.H., Boone, R.D., Magill, A.H., 1996. Vertical transport ofdissolved organic C and N under long-term N amendments in pine and hardwoodforests. Biogeochemistry 35, 471–505.

DeSutter, T.M., Pierzynski, G.M., Baker, L.R., 2006. Flow-through and batch methods fordetermining calcium–magnesium and magnesium–calcium selectivity. Soil Sci. Soc.Am. J. 70, 1280–1285.

Dilling, J., Kaiser, K., 2002. Estimation of the hydrophobic fraction of dissolved organicmatter in water samples using UV photometry. Water Res. 36, 5037–5044.

Duwig, C., Becquer, T., Charlet, L., Clothier, B.E., 2003. Estimation of nitrate retention in aFerralsol by a transient-flow method. Eur. J. Soil Sci. 54, 505–515.

Epron, D., Marsden, C., Thongo M'Bou, A., Saint-André, L., d'Annunzio, R., Nouvellon, Y.,2009. Soil carbon dynamics following afforestation of a tropical savannah with Euca-lyptus in Congo. Plant Soil 323, 309–322.

FAO, 2010. Planted Forests in Sustainable Forest Management. A Statement of Principles,.Fröberg, M., Kleja, D.B., Bergkvist, B., Tipping, E., Mulder, J., 2005. Dissolved organic carbon

leaching from a coniferous forest floor — a field manipulation experiment. Biogeo-chemistry 75, 271–287.

Fröberg, M., Berggren, D., Bergkvist, B., Bryant, C., Mulder, J., 2006. Concentration andfluxes of dissolved organic carbon (DOC) in three Norway spruce stands along a cli-matic gradient in Sweden. Biogeochemistry 77, 1–23.

Fujii, K., Uemura, M., Hayakawa, C., Funakawa, S., Sukartiningsih, Kosaki, T., Ohta, S., 2009.Fluxes of dissolved organic carbon in two tropical forest ecosystems of East Kaliman-tan, Indonesia. Geoderma 152, 127–136.

Fujii, K., Hartono, A., Funakawa, S., Uemura, M., Kosaki, T., 2011. Fluxes of dissolved organ-ic carbon in three tropical secondary forests developed on serpentine and mudstone.Geoderma 163, 119–126.

Fujii, K., Hayakawa, C., Funakawa, S., Sukartiningsih, Kosaki, T., 2013. Fluxes of dissolvedorganic carbon and nitrogen in cropland and adjacent forest in a clay-rich Ultisol ofThailand and a sandy Ultisol of Indonesia. Soil Tillage Res. 126, 267–275.

Gomez-Rey, M.X., Vasconcelos, E., Madeira, M., 2008. Effects of eucalypt residue manage-ment on nutrient leaching and soil properties. Eur. J. For. Res. 127, 379–386.

Gonçalves, J.L.d.M., Alvarez, C.A., Higa, A.R., Silva, L.D., Alfenas, A.C., Stahl, J., Frosini deBarros Ferraz, S., De Pauma Lima, W., Brancalion, P.H.S., Hubner, A., Bouillet, J.-P.,Laclau, J.-P., Nouvellon, Y., Epron, D., 2013. Integrating genetic and silvicultural strat-egies to minimize abiotic and biotic constraints in Brazilian eucalypt plantations. For.Ecol. Manage. 301, 6–27.

Guggenberger, G., Kaiser, K., 2003. Dissolved organic matter in soil: challenging the para-digm of sorptive preservation. Geoderma 113, 293–310.

Guggenberger, G., Zech, W., 1993. Dissolved organic carbon control in acid forest soils ofthe Fichtelgebirge (Germany) as revealed by distribution patterns and structuralcomposition analyses. Geoderma 59, 109–129.

Jaffrain, J., Gérard, F., Meyer, M., Ranger, J., 2007. Assessing the quality of dissolved organicmatter in forest soils using ultraviolet absorption spectrophotometry. Soil Sci. Soc.Am. J. 71, 1851–1858.

John, B., Ludwig, B., Flessa, H., 2003. Carbon dynamics determined by natural 13C abun-dance in microcosm experiments with soils from long-term maize and rye monocul-tures. Soil Biol. Biochem. 35, 1193–1202.

Kaiser, K., Guggenberger, G., 2000. The role of DOM sorption to mineral surfaces in thepreservation of organic matter in soils. Org. Geochem. 31, 711–725.

Kaiser, K., Guggenberger, G., Zech, W., 1996. Sorption of DOM and DOM fractions to forestsoils. Geoderma 74, 281–303.

Kaiser, K., Guggenberger, G., Zech, W., 2001. Isotopic fractionation of dissolved organiccarbon in shallow forest soils as affected by sorption. Eur. J. Soil Sci. 52, 585–597.

Kalbitz, K., Schwesig, D., Schmerwitz, J., Kaiser, K., Haumaier, L., Glaser, B., Ellerbrock, R.,Leinweber, P., 2003. Changes in properties of soil-derived dissolved organic matterinduced by biodegradation. Soil Biol. Biochem. 35, 1129–1142.

Kalbitz, K., Meyer, A., Yang, R., Gerstberger, P., 2007. Response of dissolved organic matterin the forest floor to long-term manipulation of litter and throughfall inputs. Biogeo-chemistry 86, 301–318.

Kauffman, S.J., Royer, D.L., Chang, S.B., Berner, R.A., 2003. Export of chloride afterclearcutting in the Hubbard Brook sandbox experiment. Biogeochemistry 63, 23–33.

Kindler, R., Siemens, J.A.N., Kaiser, K., Walmsley, D.C., Bernhofer, C., Buchmann, N., Cellier,P., Eugster, W., Gleixner, G., GrŨNwald, T., Heim, A., Ibrom, A., Jones, S.K., Jones, M.,Klumpp, K., Kutsch, W., Larsen, K.S., Lehuger, S., Loubet, B., McKenzie, R., Moors, E.,Osborne, B., Pilegaard, K.I.M., Rebmann, C., Saunders, M., Schmidt, M.W.I., Schrumpf,M., Seyfferth, J., Skiba, U.T.E., Soussana, J.-F., Sutton, M.A., Tefs, C., Vowinckel, B., Zee-man, M.J., Kaupenjohann, M., 2011. Dissolved carbon leaching from soil is a crucialcomponent of the net ecosystem carbon balance. Glob. Change Biol. 17, 1167–1185.

Klotzbücher, T., Kaiser, K., Stepper, C., van Loon, E., Gerstberger, P., Kalbitz, K., 2012. Long-term litter input manipulation effects on production and properties of dissolved or-ganic matter in the forest floor of a Norway spruce stand. Plant Soil 355, 407–416.

Laclau, J.-P., Ranger, J., Nzila, J.d.D., Bouillet, J.-P., Deleporte, P., 2003a. Nutrient cycling in aclonal stand of Eucalyptus and an adjacent savanna ecosystem in Congo: 2. Chemicalcomposition of soil solutions. For. Ecol. Manage. 180, 527–544.

Laclau, J.P., Deleporte, P., Ranger, J., Bouillet, J.P., Kazotti, G., 2003b. Nutrient dynamicsthroughout the rotation of Eucalyptus clonal stands in Congo. Ann. Bot. 91, 879–892.

Laclau, J.-P., Ranger, J., Deleporte, P., Nouvellon, Y., Saint-André, L., Marlet, S., Bouillet, J.-P.,2005. Nutrient cycling in a clonal stand of Eucalyptus and an adjacent savanna ecosys-tem in Congo: 3. Input–output budgets and consequences for the sustainability of theplantations. For. Ecol. Manage. 210, 375–391.

Laclau, J.-P., Levillain, J., Deleporte, P., Nzila, J.d.D., Bouillet, J.-P., Saint André, L., Versini, A.,Mareschal, L., Nouvellon, Y., Thongo M'Bou, A., Ranger, J., 2010. Organic residue massat planting is an excellent predictor of tree growth in Eucalyptus plantationsestablished on a sandy tropical soil. For. Ecol. Manage. 260, 2148–2159.

Laclau, J.-P., Da Silva, E.A., Rodrigues Lambais, G., Bernoux, M., Le Maire, G., Stape, J.L.,Bouillet, J.-P., Gonçalves, J.L.M., Jourdan, C., Nouvellon, Y., 2013. Dynamics of soil ex-ploration by fine roots down to a depth of 10m throughout the entire rotation in Eu-calyptus grandis plantations. Front. Plant Sci. 4, 243.

Levillain, J., ThongoM'bou, A., Deleporte, P., Saint-Andre, L., Jourdan, C., 2011. Is the simpleauger coringmethod reliable for below-ground standing biomass estimation in Euca-lyptus forest plantations? Ann. Bot. 108, 221–230.

Lovett, G.M., Likens, G.E., Buso, D.C., Driscoll, C.T., Bailey, S.W., 2005. The biogeochemistryof chlorine at Hubbard Brook, New Hampshire, USA. Biogeochemistry 72, 191–232.

Maisto, G., DeMarco, A., Meola, A., Sessa, L., Virzo De Santo, A., 2011. Nutrient dynamics inlitter mixtures of four Mediterranean maquis species decomposing in situ. Soil Biol.Biochem. 43, 520–530.

Mareschal, L., Nzila, J.D.D., Turpault, M.P., ThongoM'Bou, A., Mazoumbou, J.C., Bouillet, J.P.,Ranger, J., Laclau, J.P., 2011. Mineralogical and physico-chemical properties of FerralicArenosols derived from unconsolidated Plio-Pleistocenic deposits in the coastalplains of Congo. Geoderma 162, 159–170.

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 man-aged in short rotations after afforestation of an African savanna: a 14-year time series.For. Ecol. Manage. 307, 242–254.

Michalzik, B., Matzner, E., 1999. Dynamics of dissolved organic nitrogen and carbon in aCentral European Norway spruce ecosystem. Eur. J. Soil Sci. 50, 579–590.

Michalzik, B., Kalbitz, K., Park, J.H., Solinger, S., Matzner, E., 2001. Fluxes and concentra-tions of dissolved organic carbon and nitrogen — a synthesis for temperate forests.Biogeochemistry 52, 173–205.

436 A. Versini et al. / Geoderma 232–234 (2014) 426–436

Möller, A., Kaiser, K., Guggenberger, G., 2005. Dissolved organic carbon and nitrogen inprecipitation, throughfall, soil solution, and stream water of the tropical highlandsin northern Thailand. J. Plant Nutr. Soil Sci. 168, 649–659.

Nambiar, E.K.S., 2008. Introduction: sustained productivity of plantation forests in the tro-pics: a decade of research partnership. In: Nambiar, E.K.S. (Ed.), Site Management andProductivity in Tropical Plantation Forests. Proceedings of Workshops in Piracicaba(Brazil) 22–26 November 2004 and Bogor (Indonesia) 6–9 November 2006. CIFOR,Bogor, Indonesia, pp. 1–3.

Neff, J.C., Asner, G.P., 2001. Dissolved organic carbon in terrestrial ecosystems: synthesisand a model. Ecosystems 4, 29–48.

Nzila, J.d.D., Bouillet, J.-P., Laclau, J.-P., Ranger, J., 2002. The effects of slashmanagement onnutrient cycling and tree growth in Eucalyptus plantations in the Congo. For. Ecol.Manage. 171, 209–221.

Paquette, A., Messier, C., 2010. The role of plantations in managing the world's forests inthe Anthropocene. Front. Ecol. Environ. 8, 27–34.

Park, J.-H., Matzner, E., 2003. Controls on the release of dissolved organic carbon and ni-trogen from a deciduous forest floor investigated by manipulations of abovegroundlitter inputs and water flux. Biogeochemistry 66, 265–286.

Piirainen, S., Finer, L., Mannerkoski, H., Starr, M., 2009. Leaching of cations and sulphate aftermechanical site preparation at a boreal forest clear-cut area. Geoderma 149, 386–392.

Qualls, R.G., 2000. Comparison of the behavior of soluble organic and inorganic nutrientsin forest soils. For. Ecol. Manage. 138, 29–50.

Rees, R., Chang, S.C., Wang, C.P., Matzner, E., 2006. Release of nutrients and dissolved or-ganic carbon during decomposition of Chamaecyparis obtusa var. formosana leaves ina mountain forest in Taiwan. J. Plant Nutr. Soil Sci. 169, 792–798.

Riaz, M., Mian, I., Bhatti, A., Cresser, M., 2012. An exploration of how litter controls drain-agewater DIN, DON and DOC dynamics in freely draining acid grassland soils. Biogeo-chemistry 107, 165–185.

Saint-André, L., M'Bou, A.T., Mabiala, A., Mouvondy, W., Jourdan, C., Roupsard, O.,Deleporte, P., Hamel, O., Nouvellon, Y., 2005. Age-related equations for above- andbelow-ground biomass of a Eucalyptus hybrid in Congo. For. Ecol. Manage. 205,199–214.

Saint-André, L., Laclau, J.-P., Deleporte, P., Gava, P., Gonçalves, J., Nzila, J.d.D., Smith, C., DuToit, B., Xu, D., Sankaran, K., Marien, J., Nouvellon, Y., Bouillet, J.-P., Ranger, J., 2008.

Slash and litter management effects on eucalyptus productivity: a synthesis using agrowth and yield modelling approach. In: Nambiar, E., Ranger, J., Tiarks, A., Toma, T.(Eds.), Proceedings of Workshop in Brazil 2004 and in Indonesia 2006. CIFOR.

Sanderman, J., Amundson, R., 2008. A comparative study of dissolved organic carbontransport and stabilization in California forest and grassland soils. Biogeochemistry89, 309–327.

Sanderman, J., Baldock, J., Amundson, R., 2008. Dissolved organic carbon chemistry anddynamics in contrasting forest and grassland soils. Biogeochemistry 89, 181–198.

Sayer, E.J., 2006. Using experimental manipulations to access the roles of leaf litter in thefunctioning of forest ecosystems. Biol. Rev. 81, 1–31.

Schwartz, D., Mariotti, A., Trouvé, C., Van Den Borg, K., Guillet, B., 1992. Etude des profilsisotopiques 13C et 14C d'un sol ferralitique sableux du littoral congolais. Implicationssur la dynamique de la matière organique et l'histoire de la végétation. C. R. Acad. Sci.315, 1411–1417.

Simunek, J., Sejna, M., Saito, H., Sakai, H., Van Genuchten, T., 2008. The Hydrus-1D Soft-ware Package for Simulating the One-Dimensional Movement of Water, Heat andMultiple Solutes in Variably-Saturated Media. Version 4 Department of Environmen-tal Sciences, University of California Riverside, (281 pp).

Smethurst, P.J., Herbert, A.M., Ballard, L.M., 2001. Fertilization effects on soil solutionchemistry in three eucalypt plantations. Soil Sci. Soc. Am. J. 65, 795–804.

Trouvé, C., Mariotti, A., Schwartz, D., Guillet, B., 1994. Soil organic carbon dynamics underEucalyptus and Pinus planted on savannas in the Congo. Soil Biol. Biochem. 26,287–295.

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 heterotrophicCO2 efflux in a tropical Eucalyptus plantation. For. Ecol. Manage. 301, 79–88.

Versini, A., Zeller, B., Derrien, D., Mazoumbou, J.-C.,Mareschal, L., Saint-André, L., Ranger, J.,Laclau, J.-P., 2014. The role of harvest residues to sustain tree growth and soil nitrogenin a tropical Eucalyptus plantation. Plant Soil 376, 245–260.

Yano, Y., Lajtha, K., Sollins, P., Caldwell, B.A., 2004. Chemical and seasonal controls on thedynamics of dissolved organic matter in a coniferous old-growth stand in the PacificNorthwest, USA. Biogeochemistry 71, 197–223.