Soil-induced impacts on forest structure drive coarse woody debris stocks across central Amazonia

This article was downloaded by [1862229837]On 23 April 2014 At 0655Publisher Taylor amp FrancisInforma Ltd Registered in England and Wales Registered Number 1072954 Registered office MortimerHouse 37-41 Mortimer Street London W1T 3JH UK

Plant Ecology amp DiversityPublication details including instructions for authors and subscription informationhttpwwwtandfonlinecomloitped20

Soil-induced impacts on forest structure drive coarsewoody debris stocks across central AmazoniaDemeacutetrius L Martinsa Juliana Schiettia Ted R Feldpauschb Flaacutevio J Luizatildeoc Oliver LPhillipsd Ana Andradee Carolina V Castilhof Susan G Lauranceg Aacutetila Oliveirah Ieda LAmaralh Joseacute J Toledoi Laynara F Luglij Joseacute Luiz Purri Veiga Pintod Erick M OblitasMendozaj amp Carlos A Quesadad

a Programa de Poacutes-Graduaccedilatildeo em Ecologia Instituto Nacional de Pesquisas da AmazocircniaManaus Brasilb Geography College of Life and Environmental Sciences University of Exeter Exeter UKc Coordenaccedilatildeo de Pesquisa em Dinacircmica Ambiental Instituto Nacional de Pesquisas daAmazocircnia Manaus Brasild School of Geography University of Leeds Leeds UKe Biological Dynamics of Forest Fragments Project National Institute for AmazonianResearch (INPA) and Smithsonian Tropical Research Institute Manaus Brazilf Brazilian Agricultural Research Corporation ndash EMBRAPA Centro de Pesquisa Agroflorestalde Roraima Boa Vista Brasilg School of Marine and Tropical Biology James Cook University Cairns Australiah Tropical Ecology Assessment and Monitoring Network (TEAM) Campus de RorainopolisRorainopolis Brasili Universidade Estadual de Roraima Campus de Rorainopolis Rorainopolis Brasilj Programa de Poacutes Graduaccedilatildeo em Ciecircncias Florestais Instituto Nacional de Pesquisas daAmazonia Manaus BrasilPublished online 14 Mar 2014

To cite this article Demeacutetrius L Martins Juliana Schietti Ted R Feldpausch Flaacutevio J Luizatildeo Oliver L Phillips AnaAndrade Carolina V Castilho Susan G Laurance Aacutetila Oliveira Ieda L Amaral Joseacute J Toledo Laynara F Lugli Joseacute LuizPurri Veiga Pinto Erick M Oblitas Mendoza amp Carlos A Quesada (2014) Soil-induced impacts on forest structure drivecoarse woody debris stocks across central Amazonia Plant Ecology amp Diversity DOI 101080175508742013879942

To link to this article httpdxdoiorg101080175508742013879942

PLEASE SCROLL DOWN FOR ARTICLE

Taylor amp Francis makes every effort to ensure the accuracy of all the information (the ldquoContentrdquo) containedin the publications on our platform However Taylor amp Francis our agents and our licensors make norepresentations or warranties whatsoever as to the accuracy completeness or suitability for any purpose ofthe Content Any opinions and views expressed in this publication are the opinions and views of the authorsand are not the views of or endorsed by Taylor amp Francis The accuracy of the Content should not be reliedupon and should be independently verified with primary sources of information Taylor and Francis shallnot be liable for any losses actions claims proceedings demands costs expenses damages and otherliabilities whatsoever or howsoever caused arising directly or indirectly in connection with in relation to orarising out of the use of the Content

This article may be used for research teaching and private study purposes Any substantial or systematicreproduction redistribution reselling loan sub-licensing systematic supply or distribution in any

form to anyone is expressly forbidden Terms amp Conditions of access and use can be found at httpwwwtandfonlinecompageterms-and-conditions

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014

Soil-induced impacts on forest structure drive coarse woody debris stocks across centralAmazonia

Demeacutetrius L Martinsa Juliana Schiettia Ted R Feldpauschb Flaacutevio J Luizatildeoc Oliver L Phillipsd Ana AndradeeCarolina V Castilhof Susan G Lauranceg Aacutetila Oliveirah Ieda L Amaralh Joseacute J Toledoi Laynara F LuglijJoseacute Luiz Purri Veiga Pintod Erick M Oblitas Mendozaj and Carlos A Quesadad

aPrograma de Poacutes-Graduaccedilatildeo em Ecologia Instituto Nacional de Pesquisas da Amazocircnia Manaus Brasil bGeography College of Lifeand Environmental Sciences University of Exeter Exeter UK cCoordenaccedilatildeo de Pesquisa em Dinacircmica Ambiental Instituto Nacional dePesquisas da Amazocircnia Manaus Brasil dSchool of Geography University of Leeds Leeds UK eBiological Dynamics of ForestFragments Project National Institute for Amazonian Research (INPA) and Smithsonian Tropical Research Institute Manaus BrazilfBrazilian Agricultural Research Corporation ndash EMBRAPA Centro de Pesquisa Agroflorestal de Roraima Boa Vista Brasil gSchool ofMarine and Tropical Biology James Cook University Cairns Australia hTropical Ecology Assessment and Monitoring Network(TEAM) Campus de Rorainopolis Rorainopolis Brasil iUniversidade Estadual de Roraima Campus de Rorainopolis RorainopolisBrasil jPrograma de Poacutes Graduaccedilatildeo em Ciecircncias Florestais Instituto Nacional de Pesquisas da Amazonia Manaus Brasil

(Received 20 December 2012 final version received 20 December 2013)

Background Coarse woody debris (CWD) is an essential component in tropical forest ecosystems and its quantity varieswidely with forest typesAims Relationships among CWD soil forest structure and other environmental factors were analysed to understand thedrivers of variation in CWD in forests on different soil types across central AmazoniaMethods To estimate CWD stocks and density of dead wood debris 75 permanent forest plots of 05 ha in size wereassessed along a transect that spanned ca 700 km in undisturbed forests from north of the Rio Negro to south of the RioAmazonas Soil physical properties were evaluated by digging 2-m-deep pits and by taking auger samplesResults Soil physical properties were the best predictors of CWD stocks 37 of its variation was explained by effectivesoil depth CWD stocks had a two-fold variation across a gradient of physical soil constraints (ie effective soil depthanoxia and soil structure) Average biomass per tree was related to physical soil constraints which in turn had a strongrelationship with local CWD stocksConclusions Soil physical properties appear to control average biomass per tree (and through this affect forest structure anddynamics) which in turn is correlated with CWD production and stocks

Keywords anoxia effective soil depth carbon forest dynamics line intercept sampling soil physical propertiestopographic index tropical forest vegetation structure necromass

Introduction

The interaction between different carbon stocks and flowsconstitute the carbon cycle Of the different stocks above-ground biomass is most often assessed in tropical forestshowever coarse woody debris (CWD) is also an essentialcomponent because of its role in biogeochemical cycles(Chambers et al 2000 Clark et al 2002 Wilcke et al2005 Palace et al 2008) Within tropical forests CWDaccounts for 6ndash25 of total above-ground carbon stocks(Nascimento and Laurance 2002 Rice et al 2004 Bakeret al 2007 Palace et al 2012) implying a total pan-Amazon CWD carbon stock of ca 10 Pg (Chao et al2009a) The variation in CWD stocks across the Amazonbasin is thought to be modulated by environmental factorssuch as hydrology and soils and by forest biomass itself(Rice et al 2004 Baker et al 2007 Chao et al 2009a)

Amazonia holds a great diversity of tree species (terSteege et al 2000) and its forests vary substantially in bothvegetation dynamics (Phillips et al 2004 Quesada et al

2012) and structure (Baker et al 2004 Malhi et al 2006Nogueira et al 2008 Feldpausch et al 2011) Our currentunderstanding suggests that CWD stocks generallydecrease from north-eastern to south-western Amazonia(Baker et al 2007 Chao et al 2009a) Spatial variation inCWD stocks across the landscape may respond both toshort-term climatic disturbances (eg Phillips et al 2009Negroacuten-Juaacuterez et al 2010) and to long-term differences inforest dynamics in response to environmental characteris-tics (Keller et al 2004 Malhi et al 2006 Chao et al2009a) Soils represent an important environmental gradi-ent in Amazonia with a wide variety of soil types across thebasin and with diverse chemical and physical conditions(Quesada et al 2010 2011) Variations in soil physicalproperties across the basin have been related to a largeproportion of the variation in tree turnover rates and meanforest wood density with disturbance levels and vegetationstructure of Amazonian forests being related to differentsoil types (Quesada et al 2012)

Corresponding author Email emaildemetriusgmailcom

Plant Ecology amp Diversity 2014httpdxdoiorg101080175508742013879942

copy 2014 Botanical Society of Scotland and Taylor amp Francis

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014

Very few studies have tried to understand landscape-scale drivers of CWD stocks Kissing and Powers (2010)working in secondary forests in Costa Rica showed strongpositive correlations between stand age and the amount ofCWD Chao et al (2009a) working in mature forests inAmazonia showed that there was a relationship betweenforest structure and CWD in particular with regard tobiomass wood density of living trees and mass of indivi-dual dead stems Although these studies successfully asso-ciated CWD stocks with forest structure and dynamics toour knowledge there has been no analysis of a potentialeffect of edaphic properties on CWD stocks Since edaphicfactors such as effective soil depth and structure areimportant factors controlling forest structure and dynamics(Jirka et al 2007 Quesada et al 2012) they are likely tobe related to both the production and the stocks of CWDWe hypothesise that because poor soil physical conditionsimpose constraints on tree growth and survival they resultin increased stem turnover rates and in turn limit themaximum size that trees can attain This way smallertrees yield smaller CWD stocks Therefore we may expectlandscape-scale variation in soils to be linked to variationin CWD stocks The forests south of the Rio Amazonasrepresent a vast but poorly studied region in centralAmazonia both in terms of vegetation and soil Broadlythis region (Figure 1) is characterised by hydromophicsoils (RADAMBRASIL 1978 Sombroek 2000) of poorphysical structure (Quesada et al 2011) in contrast to soilsnorth of Manaus which are dominated by well-draineddeep soils This region is also expected to have largevariation in above-ground biomass (AGB) (IBGE 1997)Central Amazonia therefore represents an ideal testing

ground for exploring edaphic and vegetation linkageswith CWD stocks

We examined stocks of CWD as a function of vari-ables related to biomass and stem density and soil proper-ties across central Amazonia in order to understand thefactors that modulate the variation of these stocksSpecifically we tested the hypothesis that CWD stockswere larger in soils with no physical constraints and smal-ler in soils with increased physical constraints

Materials and methods

Study sites

Fieldwork was conducted across a ca 700-km-long transect(Figure 1) in central Amazonia over a 1-year period (2010ndash2011) Data were collected in permanent plots of 05 halocated north and south of the Rio Amazonas in the Stateof Amazonas Brazil There were two sites north of theAmazonas river the Adolpho Ducke Forest Reserve (here-after Ducke Reserve ndash 18 plots) and the Biological Dynamicsof Forest Fragments Project site (BDFFP ndash 12 plots) Thesouthern sites were located in the Purus ndash Madeira interflu-vial zone on a ca 600 km transect established along theManausndashPorto Velho road (BR-319 ndash 45 plots)

The Ducke Reserve has 10000 ha of mature terrafirme tropical moist forest and is situated at the peripheryof the city of Manaus (02deg 95prime S 59deg 95prime W) The topo-graphy is undulating with alternating plateaux and rivuletvalleys The vegetation has a 30ndash37 m tall closed canopywith emergent trees reaching 45 m (Ribeiro et al 1999)Mean annual precipitation is 2524 mm (Coordenaccedilatildeo de

Figure 1 Spatial distribution of coarse woody debris (CWD) stocks and values of the topographyc index (TI) Size of red circles areproportional to variation in CWD stocks High values of the topographic index (light grey to white) indicate poorly drained areas

2 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014

Pesquisas em Clima e Recursos Hiacutedricos INPA unpub-lished data) In general soils are deep well-drained andhave low bulk density Ferralsols and acrisols are foundalong the slopes and plateaux which are highly weatheredand have favourable physical conditions (ie stable aggre-gate structure associated with good drainage) (Chauvelet al 1987 Quesada et al 2010) Near streams and valleybottoms wet and sandy soils (podzols) occur but thesewere not included in this study A total of 18 plots weresampled on acrisols and ferralsols for CWD and soilsPlots were at least 1 km apart and were 250 m long and20 m wide (05 ha) following the topographic contour(Magnusson et al 2005)

The BDFFP study site is located 80 km north ofManaus (2ordm 30prime S 60ordm 00prime W) Data were collected inmature terra firme tropical moist forest at least 1000 maway from fragment edges in forest fragments gt500 ha(Laurance et al 1998) The forest canopy was 30ndash37 mtall with emergent trees reaching up to 55 m Annualmean precipitation ranged from 1900ndash3500 mm(Nascimento and Laurance 2002) CWD and soil weresampled in twelve 05 ha plots (positioned independentlyof topographic features (Laurance et al 1998)) over fer-ralsols and acrisols

The plots located south of the Rio Amazonas werespaced along the BR-319 road on the interfluvial areabetween the Purus and Madeira rivers Plots located closerto Manaus had closed lowland evergreen forest vegetation(IBGE 1997) while plots located closer to Porto Velhohad a more open-type lowland evergreen forest Thisentire region is characterised by a flat topography withelevations varying between 30ndash50 m (asl) over large dis-tances Mean annual precipitation in this area varies from2155ndash2624 mm obtained from WorldClim global coverageat 25-min resolution (Hijmans et al 2005) The soils arepredominantly Plinthosols and Gleysols (Sombroek 2000)generally having varying degrees of soil water saturationand anoxic conditions Soil physical structure is generallyrestrictive to root growth with very high bulk density inthe subsoil and thus these soils have varying degrees ofhardness and effective soil depth Subsoil layers that limitroot penetration are frequent and vary from 30ndash100 cm indepth (RADAMBRASIL 1978 Sombroek 2000) In allsites we sampled a total of 45 plots deployed in ninesite clusters the clusters being at distances between 40and 60 km apart (Figure 1) Each site cluster was com-posed of a 5-km-long transect with five plots of250 times 20 m in size at intervals of 1 km following thetopographic contour

Coarse woody debris (CWD) stocks

Field sampling of dead wood was made by (1) line intersect(van Wagner 1968) for fallen dead wood and (2) belttransects for standing dead trees (Palace et al 2007 Chaoet al 2008) For line intersect sampling every piece offallen dead woody material (trees palms lianas) with adiameter gt10 cm that crossed the transect line was

measured and classified into one of three decay classesfollowing Chao et al (2008) (1) recently fallen solidwood sometimes presenting minor degradation (2) soundwood but already showing some sign of decay such as theabsence of bark (3) heavily decayed wood In partly buriedmaterial two perpendicular measures were taken and theirmean was recorded as diameter In plots that followed thetopographic contour the central line of the 250-m long plotwas used as the intersect line In square plots the intersectline was also 250 m length but followed the plot perimeterEach 250-m transect was considered as an independentindividual sample of CWD To reduce biased estimationarising from multiple crossing of CWD pieces and endpointpartial intersection (Affleck et al 2005) we counted onceonly each piece of CWD (eg Gregoire and Valentine2003) Pieces at the endpoint of an intersect line wereincluded only if at least 50 of them was touched by thetransect line As these plots along the contour did not run ina straight line and sometimes doubled up and ran at anacute angle we took provisions to avoid multiple crossingsampling bias We discounted areas where the same pieceof dead wood crossed the same transect line more thanonce To compensate for lost plot area caused by multiplecrossing an identical area was added to the plot to keep thetotal sampling area at 05 ha As we assumed that theorientation of pieces of dead wood on the forest floor wasrandom we did not see advantage in using one line intersectdesign over another (see Bell et al 1996)

The belt transects for estimating standing dead treesand broken snags were 20 m wide along the 250 m trans-ect line Standing dead stems with a diameter gt10 cm weremeasured at 13 m height or at the lowest part of the snagtrying to avoid buttress roots where possible If the snagwas shorter than 13 m the measurement was taken at thehighest point possible The height of snags taller than 2 mwas measured with a digital hypsometer (Vertex LaserVL400 Ultrasonic-Laser Hypsometer III Hagloumlf Sweden)to the point where the diameter was 10 cm The length anddiameter (gt10 cm) of attached branches in standing deadtrees were visually estimated To account for wood densityvariation following decay standing dead trees and theiroccasional branches were also classified in the same wayas wood for the line intersects

CWD wood density

Samples of dead wood (n = 726) that crossed the lineintersect in the plots were collected for measuring thedensity of CWD (dry weight per unit volume) A chainsawwas used to cut a disk sample from hard pieces Softerwood pieces were sampled by using a bush knife Thedisks were sub-sampled randomly Void spaces were takeninto account for volume estimation by visually estimatingtheir proportion (Keller et al 2004) but were not used fordensity correction which may have caused an overestima-tion of up to 10 in some decay classes (Keller et al2004 Chao et al 2008)

Soil drives coarse woody debris stocks in central Amazonia 3

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014

CWD density was determined as the ratio of oven drymass and fresh wood volume The water-displacementmethod was used to determine fresh volume (Chave2005) Before measuring the volume segments of samplesin classes 1 and 2 were pre-wetted for about 2 h to fillpores with water because dry wood absorbs more waterand leads to overestimating density values As material indecay class 3 was very friable samples in this class weresaturated with water for several minutes After volumemeasurement the segment samples were oven-dried at 60ordmC until constant mass (Keller et al 2004) The density ofeach sample segment was then calculated and used toaverage the density of each decay class at each site

Vegetation

Vegetation parameters (basal area density of trees andpalms above-ground biomass and wood density of liveindividuals) were acquired from a database of the perma-nent vegetation plots As tree-height data were unavailablefor the permanent sample plots an allometric model pre-sented in Feldpausch et al (2012) to estimate tree height(H) was applied to reduce bias in biomass estimates

H frac14 48131 1 exp 00375 D08228

(1)

where D is the tree diameter at breast heightTo estimate plot-level dry above-ground biomass

(AGB) we used the allometric model developed byFeldpausch et al (2012) The variables included in thismodel were tree diameter at breast height (D) wooddensity (ρτ) and height (H) for tree T

AGB frac14 exp 29205thorn 09894 ln D2 ρτ H

(2)

To account only for stocks of biomass that related tobranches ge10 cm diameter we reduced the estimated AGBvalues of each tree by 15 (N Higuchi unpublished datacited in Chambers et al 2000)

Wood density of living trees was obtained from adatabase (Chave et al 2009 Zanne et al 2009) Theindividuals in each plot were matched to wood density atspecies level In cases where this information was unavail-able matches were made by genus average or family (as inBaker et al 2004) When species information for a treewas missing the mean density of known trees weightedby basal area of the plot was used Species-level identifi-cations were made for 537 of stems with an additional379 identified to genus 62 to family and 22 uni-dentified At the BR-319 transect plots (south from the RioAmazonas) there were no floristic data available For thoseplots an average living wood density was estimated foreach plot by sampling wood cores in at least 20 trees perplot (trees gt30 cm diameter only with a total of 1005 treessampled in the region by J Schietti unpublished data)

Soil data

Soil sampling followed a standard protocol (httpwwwgeogleedsacukprojectsrainforpagesmanualstodownloadhtml) for a full description see Quesada et al (2010) Weused the World Reference Bases for soil resources to classifysoil types (IUSS Working group WRB 2006) Three soil pitswere dug at the Ducke Reserve and three at the BDFFP siteone soil pit was dug in six out of nine site clusters along theBR-319 To increase spatial coverage of soil propertiesauger sampling was carried out in plots without soil pits atall sites All pits were 2 m in depth even if the effective soildepth was shallower Effective soil depth is defined here asthe depth where clear impeding layers to root growth occurSoil was sampled from the pit walls to estimate bulk densityusing specially designed container rings of known volume atthe following depths 0ndash10 10ndash20 20ndash30 30ndash50 50ndash100100ndash150 150ndash200 cm

Topography and soil properties that could limit rootgrowth were assessed semi-quantitatively (Table 1Quesada et al 2010) The score for each category wasthen summed to form a general index of soil physicalquality (Π) Π1 was formed by the sum of scores for

Table 1 Score table for soil physical and chemical properties

Category Score

Effective soil depth (soil depthpresenceabsence of hardpans)

Shallow soils (lt20 cm) 4Medium shallow (20ndash50 cm) 3Hardpan or rock that allows vertical root growthother soils between 50ndash100 cm deep

2

Hardpan rocks or C horizon ge100 cm deep 1Deep soils ge150 cm 0

Soil structureVery dense very hard very compact withoutaggregation restrictive to roots

4

Dense compact little aggregationlower root restriction

3

Hard medium to high density andor with weakor block-like structure

2

Loose sand slightly dense well-aggregated insub-angular blocks discontinuous pans

1

Good aggregation friable low density 0

TopographyVery steep (gt45ordm) 4Steep (20ndash44ordm) 3Gently undulating (8ndash19ordm) 2Gently sloping (1ndash8ordm) 1Flat (0ordm) 0

Anoxic conditionsConstantly flooded patches of stagnant water 4Seasonally flooded soils with high clay content and verylow porosity andor dominated by plinthite

3

Deep saturated zone (up to 50 cm below surface) redoxfeatures

2

Deep saturated zone (maximum height ofsaturation gt100 cm deep) deep redox features

1

Unsaturated conditions 0

4 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014

effective soil depth soil structure quality topography andindicators of anoxia Π2 equalled Π1 less anoxia Thesesemi-quantitative scores were used in statistical analysesSoil fertility was similar across the entire study area(RADAMBRASIL 1978)

Additional environmental data

Mean annual precipitation and precipitation in the driestquarter were obtained from the WorldClim global databaseat 30 arc seconds (ca 1 km) resolution (Hijmans et al 2005)

The topography data were obtained using a digitalelevation model (DEM) at 90-m spatial resolution fromShuttle Radar Topography Mission (SRTM) A topo-graphic index (TI) that estimates drainage for each SRTMpixel (Moore et al 1991) was calculated using ArcMap(ESRI 2005)

TI frac14 lnα

tan β

(3)

where α is the contributing upslope drainage area and β isthe slope

Sites with higher TI values have greater drainage con-straints (eg water saturation) This topographic featuremay be important as there is a relationship between TIand tree species distribution (Feldpausch et al 2006) thatcould be related to CWD distribution across the landscape

Calculations

Volume of line intersect sampling (VLIS) (m3 haminus1) andfallen volume in each decay class were estimated usingthe following equation (van Wagner 1968)

VLIS frac14 π2 Pd2i

8 L(4)

where di is the diameter (cm) of each CWD segment i andL (m) is the length of the transect line

For the estimation of standing dead volume (VBelt)(m3 haminus1) Smalianrsquos formula was used

VBelt frac14 Hπ D1

2

2thornπ D22

22

(5)

where H (m) is the height of the tree D1 and D2 are thediameters (cm) at 13 m above the ground and at the top ofthe snag respectively To estimate D2 a taper function wasused (Chambers et al 2000)

D2 frac14 159 D1ethH0091THORN (6)

where D2 is the diameter at height H for a trunk of givenD1 This is an equation defined for central Amazoniantrees and often used in other studies (Clark et al 2002

Palace et al 2007) CWD (Mg haminus1) in each of the threedecay classes was calculated as follows

CWD frac14 Viρi (7)

where V (m3 haminus1) and ρ (Mg mminus3) correspond to volumeand density in decay class I respectively

To calculate error for each CWDi (ECWD) the followingequation was used

ECWD frac14 V Eρ thorn ρEV (8)

where Eρ and EV are the errors in density and volumerespectively Equation (8) is valid when V and the densityof the material in the respective class are not correlated(Keller et al 2004) In this study covariance between Vand density although significant (P = 00175) was verysmall (r2adj = 001965) To estimate total error in mass ineach decay class we used a conservative approach bysumming the errors of each component

Statistical analysis

For the statistical analysis plots were divided into threegroups (1) no physical restriction (NR index Π1

value lt2) (2) low levels of soil physical restriction(LRL index Π1 value between 2 and 6 and anoxiavalue lt1) and (3) high restriction levels (HRL index Π1

value gt6 and anoxia value gt1 Physically restricted soils(LRL and HRL) occurred in the interfluvial region but notnorth of Manaus

Each plot was considered as a sample unit in linearregressions (n = 75) Correlations were used to choosewhich non-collinear variables could be combined in thesame regression model In an attempt to better understandlandscape-scale CWD patterns CWD relationships withenvironmental climatic and edaphic variables wereexplored by using mixed models (nlme package in RPinhiero et al 2012) with a random intercept as thestudy had a hierarchical design Therefore the BDFFPthe Ducke Reserve and each of the nine site clustersalong the BR-319 were all considered as groups withinwhich the individual plots were nested CWD values weretransformed using the natural logarithm (ln) to improvenormality To understand the variance explained by themodels we used a method suggested by Nakagawa andShielzeth (2013) for obtaining a marginal R2 (R2

(m))which describes the proportion of variance explained bythe fixed factor and conditional R2 (R2

(c)) which describesthe proportion of variance explained by both the fixed andrandom factors To compare mean wood density of decayclasses in each forest type we also used mixed models Allanalyses were carried out in R version 300 (RDevelopment Core Team 2013)


Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014

Results

Variations in edaphic properties

Sites located north of the Amazonas river had no physicalsoil restriction (Figure 2(a)) All of these soils were deephad low subsoil bulk density had good particle aggrega-tion (good structure friable) and were well drained(Table 2) Conversely soils in the southern plots (BR-319) were generally shallow (maximum effective soildepth varying from 20ndash100 cm) with high subsoil bulkdensity (Table 2) little or no aggregation (deficient struc-ture very hard and compact) thus being generally root-restrictive and had varying levels of anoxic conditions(from seasonally flooded with patches of stagnating

water to soils showing redox features such as mottling)(Table 2) Some site clusters were severely constrained (Π1

values ranging from 6ndash11) while other plotssite clustershad lower Π1 values ranging from 2ndash6 All soils along theBR-319 had poorer physical conditions when compared tothe predominantly ferralsolsacrisols at plots north of theAmazonas river

Stocks of CWD

The volume of CWD varied significantly among soilgroups (ie NR LRL and HRL) and decay classes(Table 3) The volume of total CWD in forests growingon NR soils (695 plusmn 111 m3 haminus1) was similar to that on

Π1

(a)

Π2

(b)

Effective soil depth

(c)

Soil structure

(d)

Anoxia

(e)

Topography

(f)

Topographic index

(g)

Mean tree diameter (cm)

(h)

Mean tree height (m)

(i)

AGB (Mg haminus1)

(j)

ln C

WD

(Mg

haminus1

)ln

CW

D (M

g ha

minus1)

ln C

WD

(Mg

haminus1

)ln

CW

D (M

g ha

minus1)

ln C

WD

(Mg

haminus1

)ln

CW

D (M

g ha

minus1)

ln C

WD

(Mg

haminus1

)ln

CW

D (M

g ha

minus1)

ln C

WD

(Mg

haminus1

)ln

CW

D (M

g ha

minus1)

ln C

WD

(Mg

haminus1

)

0 2 4 6 8 10

15

25

35

45

0 2 4 6 8

15

25

35

45

0 1 2 3 4

15

25

35

45

4

15

25

35

45

0 1 2 3 0 1 2 3 4

15

25

35

45

15

25

35

45

0 1 2 3 4

7 8 9 10 11 12 13 14

15

25

35

45

18 20 22 24 26

15

25

35

45

14 15 16 17 18

15

25

35

45

100 150 200 250 300 350

15

25

35

45

02 03 04 05

15

25

35

45

AGBper tree (Mg)

(k)

High restriction levelLow restriction levelNo restriction

Figure 2 Simple relationships between coarse woody debris (CWD) and environmental variables All CWD values were lntransformed

6 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014

LRL soil (695 plusmn 116 m3 haminus1) In contrast forests onHRL had significantly less CWD (338 plusmn 20 m3 haminus1) thanforests on the other two soil groups

Densities of CWD samples were significantly differentamong decay classes decreasing considerably with degreeof decomposition (Table 3) Nevertheless there was nosignificant difference among soil groups that grew on soilswith different levels of soil physical constraints (decayclass P lt 0001 soil groups P = 076)

CWD stocks varied in a predictable way across ourstudy area (Figure 1) and also varied considerably at thesite cluster level with the northern sites showing thelargest variation For instance CWD ranged from 67ndash729 Mg haminus1 among the plots of the Ducke Reserve Incomparison CWD stocks varied little and were consis-tently lower at site clusters 1 to 5 along the BR-319 road(just south of Manaus) and also at site cluster 11 located

at the far south end of the BR-319 road Along the middle(site clusters 6 to 10) CWD was locally highly variable

Total CWD stocks followed the same pattern as totalCWD volume which is expected as CWD stock estimateis a function of site-specific CWD density values and thedensity of decay classes did not vary significantly amongsoil groups (Table 3) Forests in NR soil had a mean CWDstock of 331 plusmn 71 Mg haminus1 (Table 5) and these values didnot differ significantly from LRL soils However CWDstocks for HRL soils were significantly and substantiallylower than in both other soil types (soil groups P lt 0001decay class P lt 0001)

Standing and fallen fractions of CWD

Significant differences in fallen CWD were found amongall soil groups (P lt 0001) (Table 4) Mean stocks ofstanding CWD represented 20ndash30 of the total CWD inthe study area and this fraction did not differ significantlyamong soil groups (P = 008) with NR showing higherstocks than HRL Fallen dead wood CWD stocks weresignificantly highest in LRL intermediate in NR forestsand lowest in HRL (Table 4) The proportion of fallenstocks to total CWD did not differ among soil groupsAlso the ratio of standing to fallen dead wood was notdifferent among soil groups The CWD to AGB ratio onthe NR soils (013 plusmn 001) and LRL (017 plusmn 001) wassignificantly greater than on HRL soil (008 plusmn 001)(P lt 0001)

Vegetation

Each of the three soil groups was associated with a distinctforest structure (Table 5) Above-ground biomass washighest at the NR forests and lowest at HRL and withthe AGB at LRL sites not being significantly different

Table 3 Mean (plusmnSE) coarse woody debris (CWD) volume CWD density and CWD mass in forests on soils with no (NR) low (LRL)and high (HRL) restriction in central Amazonia

NRamx LRLamx HRLany

Wood density (cmminus3)per decay class

Class 1A 068 plusmn 002 (75) 067 plusmn 004 (20) 061 plusmn 002 (88)Class 2B 055 plusmn 002 (66) 053 plusmn 003 (43) 048 plusmn 001 (176)Class 3C 032 plusmn 001 (88) 034 plusmn 002 (24) 033 plusmn 002 (97)

CWD volume (m3 haminus1)Class 1M 123 plusmn 30 198 plusmn 38 69 plusmn 12Class 2N 261 plusmn 47 299 plusmn 34 157 plusmn 13Class 3N 311 plusmn 34 198 plusmn 44 111 plusmn 12Total 695 plusmn 111 695 plusmn 116 337 plusmn 37

CWD mass (Mg haminus1)Class 1X 84 plusmn 23 13 plusmn 33 42 plusmn 09Class 2Y 144 plusmn 31 153 plusmn 27 77 plusmn 0 8Class 3X 103 plusmn 14 68 plusmn 19 41 plusmn 07Total 331 plusmn 71 351 plusmn 72 161 plusmn 26

Number in parentheses is the number of samples (n) Uppercase superscript letters indicate statistically significant differences among decay classes(P lt 005) lowercase superscript letters are for differences among soil groups Multiple comparison tests (Tukey HSD) are labelled by the letters A B Cand a for density in different decay classes M N m and n for CWD volume X Y x and y for CWD mass

Table 2 Range of scores edphic physical conditions in the threedifferent soil groups based on their physical properties anoxialevel and topographic position in two regions of centralAmazonia Brazil

Soil parameter NR LRL HRL

Soil type Ferralsolacrisol Plinthosol GleysolplinthosolAnoxia 0 0ndash1 2ndash4Depth 0 0ndash2 1ndash4Structure 0ndash1 1ndash2 2ndash4Topography 0ndash2 0ndash1 0ndash1Bulk density

(g cmminus3)08ndash12 10ndash16 12ndash17

Π1 0ndash2 2ndash6 6ndash11Π2 0ndash2 2ndash6 4ndash8

NR no physical restriction LRL low level of physical restriction HRLhigh level of physical soil restriction bulk density was measured at areference depth of 50 cm Π1 sum of the four soil parameters Π2 sum ofsoil parameters with the exception of the anoxia parameter See Table1 forscore values


Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014

from HRL Stem density was significantly higher in HRLthan in either NR or LRL (variables associated with indi-vidual tree size were usually significantly differentamong soil groups For instance the average biomassper tree was significantly different among soil groups(NR gt LRL gt HRL Table 5) Mean tree height (estimatedfrom DBH) was also significantly different among thethree soil groups (Table 5) mean DBH was significantlylower in HRL soil than in NR and LRL (Table 5)

Determinants of CWD across landscape

CWD was significantly related to average biomass per treethe only significant forest structure parameter directly relatedto CWD (r2m = 009 r2m = 031 Figure 2(j)) This shows thattrees in HRL soil are generally smaller and store individuallyless biomass than those on LRL and NR soils with LRLshowing an intermediary behaviour Considering further therelationship between CWD stocks and parameters related toaverage maximum tree size (mean tree diameter estimated

height and AGB per tree Figure 2) we observed a clearseparation among the different soil groups with forests con-sistently showing lower CWD on HRL where trees weresmaller and high CWD on NR where trees were largerForests on LRL consistently appeared as an intermediarygroup with some superposition on NR but with a clearseparation from HRL despite these two groups occurring inthe same geographical area (HRL and LRL only occur alongthe BR-319 interfluvial area) Since plot-level variation invegetation biomass stocks could potentially influence CWDwhereby larger AGB stocksmay produce larger CWD stockswe repeated our analyses after normalising data using aCWD to above-ground biomass ratio (CWDAGBTable 6) This resulted in stem density and wood specificgravity becoming significantly related to CWD

The presence and magnitude of soil physical con-straints varied greatly across the study area and weregenerally negatively related to CWD (Figures 3(a) to (e)Table 6) Individual soil parameters were significantlyrelated to CWD with effective soil depth and anoxiabeing the best correlated variable (Table 6) Topographyincluding the continuous topographic index (TI a proxyfor hydrological gradients) had no significant relationshipwith CWD due to the characteristics of the study sitesdiscussed above

Π1 which represents the combination of all physicalparameters was strongly related to CWD (Table 6) Thisvaried from score 0 (very good physical conditions) to 11(high level of root growth restriction Figure 2(a)) with thesoils having high levels of physical constraints (Π1 gt 6)showing much lower values of CWD The index Π2 showeda similar trend to Π1 (Figure 2(b) Table 6) The only differ-ence betweenΠ1 andΠ2 was the absence of anoxia inΠ2Wefound no significant relationship between CWD and climaticvariables (mean annual precipitation and precipitation in thedriest quarter of the year (Table 6))

Discussion

Large-scale patterns in central Amazonia

CWD is a substantial fraction of forest carbon stocks Wefound that by adding CWD stocks to the above-groundbiomass pool total the above-ground wood mass stocks inforests in NR LRL and HRL soils increased by ca 1317 and 8 respectively We also found large variabilityin CWD stocks often with considerable variability at locallevel (Figure 1) However there was low variation inCWD in the first 300 km from Manaus of the BR-319road as well as at 600 km All those sites (site clusters 1 to5 and site cluster 11) had the lowest CWD stocks Theyall had in common very high levels of soil physical con-straints such as deficient soil structure shallow soil depthand anoxia suggesting that the investigated soil propertiescould indeed be the driving mechanism for low CWDin waterlogged forests The largest variability in CWDwas observed at sites where soil physical propertieswere not restrictive It is likely that at such sites sporadicand largely stochastic mortality events are the main

Table 4 Mean (plusmnSE) mass of fallen and standing coarse woodydebris (CWD) in forests growing on soils with no (NR) low(LRL) or high (HRL) physical restriction in plots in centralAmazonia north and south of the Rio Amazonas

NRax LRLaby HRLbz

StandingClass 1 38 plusmn 11 27 plusmn 09 12 plusmn 03Class 2 42 plusmn 10 27 plusmn 07 22 plusmn 05Class 3 24 plusmn 05 16 plusmn 07 10 plusmn 02

FallenClass 1 47 plusmn 14 106 plusmn 25 30 plusmn 06Class 2 102 plusmn 23 131 plusmn 14 55 plusmn 06Class 3 79 plusmn 12 51 plusmn 13 32 plusmn 04

Statistically significant differences at P lt 005 among soil groups (NRLRL HRL) are indicated by a and b for total standing CWD and by x yand z for total fallen CWD

Table 5 Average (plusmnSE) above-ground biomass (AGB) stemdensity mean tree height and DBH coarse woody debris(CWD) and CWDAGB ratio in soils with no (NR) low (LRL)and high (HRL) levels of restriction in plots in central Amazonianorth and south of the Rio Amazonas

NR LRL HRL

AGB(Mg haminus1)

2482 plusmn 61a 2188 plusmn 166b 1988 plusmn 72b

Stems (haminus1) 5979 plusmn 87a 6353 plusmn 272a 7662 plusmn 303b

AGB per tree(Mg)

042 plusmn 001a 035 plusmn 002b 026 plusmn 002c

Mean height (m) 165 plusmn 01a 160 plusmn 01b 154 plusmn 01c

DBH (cm) 231 plusmn 03a 227 plusmn 04a 203 plusmn 03b

CWD(Mg haminus1)

331 plusmn 31a 337 plusmn 27a 168 plusmn 12b

CWDAGB 013 plusmn 001a 014 plusmn 001a 008 plusmn 001b

Different letters indicate significant differences between means (P lt 005)in each row CWDAGB ratio of total CWD to AGB for trees gt10 cmDBH

8 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014

determinants of CWD stocks at any one point in time Thismay be particularly important in small plots (05 ha)

Despite large differences in soil physical conditionsno differences in the proportions of standingfallen CWDstocks were observed across our study area Standingfallen ratios in our plots (029059) were higher thanthose found by Palace et al (2007) in the Brazilian statesof Mato Grosso and Para (014017) but much lower thanvalues presented by Delaney et al (1998) in Venezuela(080) These differences among regions suggest that theratio of fallen to standing CWD varies across Amazonia asvaries the mode of death (standing vs fallen) (Chao et al2009b)

In addition to the observed relationships amongCWD soils and vegetation structure it seems likelythat variations in wood decomposition rates may be asource of variability for CWD stocks at landscape

level For instance we noted that low stocks in HRLforests were similar to the ones reported by Martius(1997) in fertile floodplain forests (vaacuterzea) in centralAmazonia and by Chao et al (2008) from a floodplainforest in Peru These studies argued that the lowerCWD stocks in these areas should be a result of higherwood decomposition rates under the cycle of wettingand drying Here we argue that differences in foreststructure such as average tree diameter (DBH per tree)may also be a source of variation in wood decomposi-tion rates (van Geffen et al 2010) Stem thickness andsurface area may exert controls on decomposition withthinner trees associated with greater decompositionrates As wood density which is commonly recognisedas a primary wood trait that affects decomposition rates(Chambers et al 2000 Chao et al 2009a Chave et al2009) did not vary significantly across our study area

Table 6 Relationships between independent variables and coarse woody debris (CWD) stocks in 75 plots across central Amazonia

Variable AIC Fixed effect Intercept r2m r2c P

CWD with soil physical constraintsDepth 92397 minus0258 3411 037 037 lt0001Anoxia 96682 minus0233 3421 033 033 lt0001Π1 98058 minus0084 352 033 033 lt0001Π2 101135 minus0123 3548 032 032 lt0001Structure 104635 minus0168 3446 029 030 lt0001Topography 109159 0112 3015 003 034 016

CWDAGB with soil physicalconstraints

Depth 9897 minus0215 minus2094 027 027 lt0001Anoxia 102362 minus0192 minus2089 024 024 lt0001Π1 10365 minus0068 minus2010 022 022 lt0001Π2 10581 minus0100 minus1992 020 020 lt0001Structure 10900 minus0130 minus2087 016 016 0013Topography 11191 0111 minus2425 003 024 018

CWD with TI 10976 minus0052 3565 003 032 0217CWDAGB with TI 11291 minus004 minus199 002 023 0341

CWD with vegetationAGB per tree 105746 1748 2498 011 032 0013Stems density 107132 minus0001 3773 008 039 005Wood specific gravity 110107 minus1137 386 001 037 0335Height 110218 0119 117 002 030 0278AGB 110419 0001 2738 001 035 0415Basal area 110742 001 2776 001 035 0557DBH 11731 004 2199 003 035 0259CWDAGB with vegetationAGB per tree 11282 0766 minus2633 002 022 0288AGB 11058 minus0003 minus1720 004 035 0077Stems density 10604 minus0001 minus1435 014 029 0004Basal area 11181 minus0025 minus1750 003 034 0166Wood specific gravity 10852 minus2581 minus0543 008 031 0026DBH 12097 0021 minus2854 001 027 0538Height 11346 0034 minus2937 001 023 0758

CWD with climateTotal precipitation 109655 0001 1258 005 037 0234Precipitation in the driest quarter 110937 0001 288 001 036 0717CWDAGB with climateTotal precipitation 11340 0000 minus3099 001 028 0612Precipitation in the driest quarter 11366 0000 minus2431 000 028 0927

AIC Akaike Information Criterion


Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014

therefore differences in tree diameter may be the pri-mary driver of decomposition

Another source of CWD variation in floodplain soilshas been suggested by Martius (1997) who argued thatflooding may redistribute CWD from higher lying areas tolower forests This cannot be applied in our study area sinceplots are not located adjacent to large rivers Out of the 45 ofour interfluvial plots only nine were located in flooding-affected areas but none of them were close to high-energyndashhigh-volume rivers that could carry wood away None of theother plots that had high values of anoxia (anoxia value gt2)had indicators of large-scale flooding Therefore we infer amechanistic role for soil physical properties whereby stag-nant soil water creates an anaerobic environment that inhi-bits deep root growth (Gale and Barfod 1999) This mayresult in a small size of trees and increased tree mortalitywhich may decrease CWD stocks

Underlying causes of variation

Soil and CWD Sites north of the Amazonas river had nosoil physical restriction In such edaphic conditions CWDproduction may be driven by stochastic patterns of treemortality mostly related to senescence and storms (Galeand Barfod 1999 Toledo et al 2012) As with the southernsites restrictive soil physical conditions appeared to beimportant predictors of CWD most likely resulting fromedaphic influences on forest dynamics at the waterloggedsites Topography in these areas was flatter than in thenorth but other soil parameters were good predictors ofCWD Physical properties such as shallow soils with highbulk density poor aggregation and severe anoxic condi-tions can restrict deep root growth In addition such soilconditions limit tree establishment and tend to increasetree mortality rates (Gale and Barfod 1999 Gale and Hall2001 Quesada et al 2012) From all edaphic propertiesstudied effective depth and anoxia seem to be the most

related to CWD in our study area (Table 6) However weobserved that instead of increasing the volume of CWDand CWD stocks severe soil physical conditions led to adecreasing in CWD stocks In our study area soil physicalrestrictions are likely to affect CWD by changing the over-all forest structure ndash reducing average tree size and therebyalso increasing decomposability ndash rather than by selectinglow-wood-density species common to more dynamic for-ests such as occurs in western Amazonia (Chao et al2009a) This was supported by the fact that average plotwood density did not vary with soil physical limitations inour study area As soil water saturation exerts controls onsoil weathering and development it may imply that soildepth and structure are correlated with soil anoxia levels(Quesada et al 2010 2011) In this case relationshipsbetween these soil variables and CWD could beinterpreted as reflecting combined soil-vegetation effects(Figure 3)

Vegetation and CWD The lack of a relationship betweenCWD and biomass found here can be compared to thosepresented by Chao et al (2009a) who found weak rela-tionships between CWD stocks and above-ground biomassacross a broader area in Amazonia Above-ground bio-mass per tree was a better predictor of CWD howeverthe relationship was relatively weak

Different levels of soil physical restrictions appearedto significantly correlated with forest structure (Table 5)implying an important influence on how and for howlong living biomass is stored in forest ecosystems(Quesada et al 2012) Soil restrictions may decreaseaverage residence time of trees (Quesada et al 2012)resulting in a population of thinner and shorter trees thatindividually store less biomass (also with more indivi-duals per ha) On the other hand we hypothesise for-ests on soils without physical limitations tend to bepopulated by larger trees simply because they can live

Figure 3 Typical Ferralsol for NR sites (BDFFP Manaus) deep soils present good particle aggregation low bulk density and nophysical impediments to root growth such as hardpans anoxic conditions (a) Typical Plinthosol occurring at the interfluve of the Purusand Madeira rivers (Site cluster 1) soil have shallow effective depth and very high bulk density which restrict root growth Soft orangecolouration in the first 50 cm and deep mottling indicate water fluctuation common in these soils (b)

10 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014

longer The death of individuals with a high biomassresults in a high mass of individual dead stem deadwood mass small trees such as those observed inHRL soil would contribute smaller amounts of deadwood even if at slightly higher mortality rates than inthe other soil groups For instance NR and LRL had 16and 13 higher AGB per tree respectively than HRLTherefore inputs of dead wood from mortality in bothof these soil groups should be greater than in HRL sitesas we found a two-fold difference of CWD stocksbetween NR or LRL and HRL sites Hence our studyreinforces an important relationship already pointed outby Chao et al (2009a) between mortality of mass inputand CWD the size of stems that die may be moreimportant for CWD stocks than the number of stemsthat die Furthermore trees with higher biomass alsohave larger diameters and therefore lower potentialdecomposition rates (van Geffen et al 2010) The bal-ance of these factors should result in higher CWDstocks in NR and LRL soil and lower in HRL

Moreover as LRL sites had an intermediate level ofedaphic restrictions we also speculate that CWD stocksthere were similar to those found in NR probably becauseof subtle differences in tree mortality rates and tree sizebetween NR and LRL While forests on LRL soil hadcertain edaphic restrictions (eg soils shallower than NRsoils) they had similar AGB to forests on NR soil but withdifferences in stem density Lower biomass per tree inLRL should have yielded lower CWD stocks than inNR but it was not the case It may have been due tothat fact that in the presence of some edaphic restrictionstree mortality in LRL increased slightly such as inQuesada et al (2012) (but not determined in this study)and this resulted in equal or higher CWD stocks than thosefound on NR

CWDAGB ratio was not constant across the studiedforests CWD contributed proportionally less in HRL for-ests (Table 5) than in NR and LRL Proportions of CWDAGB at NR and LRL sites were larger than proportions innorth-western Amazonia (0103 plusmn 0011) but similar tothose in eastern Amazonia (0132 plusmn 0013 (Chao et al2009a)) Furthermore the CWDAGB ratios in this studywere lower than those presented by Palace et al (2007) inMato Grosso and Paraacute (019ndash020) This points to theimportance of including CWD measurement in local car-bon balance studies since there is not an invariant propor-tion of AGB across Amazonia

CWD stocks have usually been expressed as the CWDAGB ratio following the rationale that variation in CWDstocks should reflect the variation of AGB We suggest thatthe use of the CWDAGB ratio may not always be infor-mative as AGB is a function of varying combinations oftree size and number of individuals (for example a similarAGB can be attained by a few large trees or by many smalltrees) and such variations in how the wood component isstored in AGB stocks may disconnect CWD stocks fromAGB Therefore we suggest that over large spatial scalesthe CWDAGB ratio should be used cautiously for the

following reasons (1) CWD stocks are a function of deadwood input from mortality and decomposition rates(fluxes) which are more influenced by tree size at themoment of death than by whole stand AGB (2) CWDAGB is not a constant proportion varying widely at largescales (3) AGB is not related to CWD in the scale of thisstudy and is only weakly related to CWD stocks in a widerscale (Chao et al 2009a) because stand AGB values areweakly related to individual tree size

Conclusions

The findings of this study fill a gap in understanding thecauses of CWD variation across central Amazonia Wefound that differences in CWD stocks across the studyarea were related to a gradient of soil physical conditionswhich affected forest structure and dynamics and in turninfluenced CWD stocks CWD was found to be positivelyrelated to biomass per tree and negatively related with soilphysical restriction We suggest that edaphic constraintmay act on vegetation structure by decreasing individualtree biomass at the time of death (earlier death) by redu-cing tree height diameter and individual biomass Suchchanges on vegetation structure may result in a reductionin the mass of individual dead stems along with increasedrates of stem mortality and decomposition This study thushighlights the importance of soil properties and its mod-ulating power over forest structure and so influencingCWD across large-scale soil gradients

AcknowledgementsThis contribution is derived from Demetrius L Martinsrsquo MScthesis with a fellowship from the Brazilian National ResearchCouncil (CNPq) Financial support for fieldwork and additionaltraining was received from the Gordon and Betty MooreFoundation through the RAINFOR project OP is supported bya European Research Council Advanced Grant and is a RoyalSociety Wolfson Research Fellow Logistical support was pro-vided by BDFFP PPBio and Large Scale Biosphere-AtmosphereProgramme in Amazonia (LBA) Part of this manuscript wasdeveloped during the 2011 RAINFOR (Gordon and BettyMoore Foundation)-UFAC workshop in Rio Branco AcreBrazil We thank Luciano A Castilho and Aires da S Lopesfor help with field work and Gabriel M Moulatlet for providingcorrected SRTM images for the interfluvial zone Thaise Emiliohelped with map preparation We also give special thanks toKuo-Jung Chao Michael Palace Michael Keller Bruce NelsonPhilip Fearnside and Laszlo Nagy for their valuable commentsThis is publication number 634 in the Biological Dynamics ofForest Fragment Project (BDFFP-INPASTRI) technical series

Supplemental dataSupplemental data for this article can be accessed here http101080175508742013879942

Notes on contributorsDemeacutetrius L Martins has a MSc from INPA and is a masterrsquosstudent investigating necromass stocks variation acrossAmazonia with support from the Gordon and Betty Moore


Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014

Foundation grant to RAINFOR His research interests includevegetation functioning dynamics and structure and whatmechanisms are responsible for their variation across landscape

Juliana Schietti is a PhD student and is interested in understand-ing how the variations in forest structure and function are relatedto hydrological processes

Ted R Feldpausch is a lecturer His research interests include theecology and effects of global change on tropical forests andsavannas land-use change and forest disturbance and recovery

Flaacutevio J Luizao is a senior researcher He has wide experience inecological issues in Amazonia especially in forest ecosystemecology His research interests include nutrient and organic mat-ter cycling soil ecology the soil-plant relationship and agrofor-estry systems

Oliver L Phillips is Professor of Tropical Ecology He leads theRAINFOR network of scientists to work together to monitorunderstand and predict the behaviour of Amazon forests

Ana Andrade manages the botanical collection of the ForestSurvey Association of Amazonas and takes part in projects ofthe Biological Dynamics of Forest Fragments Project (BDFFP)She has research interests in plant ecology and taxonomic identi-fication of the trees of central Amazonia

Carolina V Castilho is a researcher and is interested in long-termecological studies and uses permanent plots to evaluate changesin biomass structure and tree species composition in tropicalforests

Susan Laurance is a professor Her research interests are in landuse and climate change impacts in tropical forest communities

Aacutetila de Oliveira is interested in the structure and diversity oftropical forests

Ieda Leao do Amaral is interested in the structure and diversity oftropical forests

Joseacute J Toledo is a professor He has experience in forest ecologyand his research interests are the causes of tree mortality andwood decomposition in tropical forests

Laynara F Luglirsquos has a MSc from INPA and is interested inunderstanding biogeochemical cycles in tropical forests

Joseacute Luiz Purri da Veiga Pinto has a MSc from INPA and isinterested in studying soils and root productivity

Erick M Oblitas Mendoza has an MSc from INPA and has beenstudying the carbon stocks of Amazonian soils He is currently aPhD candidate working on root productivity

Carlos A Quesada is a researcher His research interests includebiogeochemical cycles Amazonian soils and their interactionwith forest structure and dynamics

ReferencesAffleck DLR Gregoire TG Valentine HT 2005 Design unbiased

estimation in line intersect sampling using segmented transectsEnvironmental and Ecological Statistics 12 139ndash154

Baker TR Honorio CE Phillips OL van der Heijden G MartinJ Garcia M Silva Espejo J 2007 Low stocks of coarsewoody debris in a south-western Amazon forest Oecologia152 495ndash504

Baker TR Phillips OL Malhi Y Almeida S Arroyo L Di FioreA Erwin T Killeen TJ Laurance SG Laurance WF et al2004 Variation in wood density determines spatial patternsin Amazonian forest biomass Global Change Biology 10545ndash562

Bell G Kerr A McNickle D Woollons R 1996 Accuracy of theline intersect method of post-logging sampling under orien-tation bias Forest Ecology and Management 84 23ndash28

Chambers JQ Higuchi N Schimel JP Ferreira LV Melack JM2000 Decomposition and carbon cycling of dead trees in tropi-cal forests of the central Amazon Oecologia 122 380ndash388

Chao KJ Phillips OL Baker TR 2008 Wood density and stocksof coarse woody debris in a northwestern Amazonian land-scape Canadian Journal of Forest Research 38 267ndash276

Chao KJ Phillips OL Baker TR Peacock J Lopez-Gonzalez GMartiacutenez RV Monteagudo A Torres-Lezama A 2009aAfter trees die quantities and determinants of necromassacross Amazonia Biogeosciences 6 1615ndash1626

Chao KJ Phillips OL Monteagudo A Torres-Lezama AVaacutesquez Martiacutenez R 2009b How do trees die Mode ofdeath in northern Amazonia Journal of Vegetation Science20 260ndash268

Chauvel A Lucas Y Boulet R 1987 On the genesis of the soilmantle of the region of Manaus Central Amazonia BrazilCell Mol Life Sci 43 234ndash241

Chave J 2005 Measuring wood density for tropical forest treesMeasuring wood density for tropical forest trees - A fieldmanual for the CTFS sites 7

Chave J Coomes DA Jansen S Lewis SL Swenson NG ZanneAE 2009 Towards a worldwide wood economics spectrumEcology letters 12 351ndash366

Clark DB Clark DA Brown S Oberbauer SF Veldkamp E2002 Stocks and flows of coarse woody debris across atropical rain forest nutrient and topography gradient ForestEcology and Management 164 237ndash248

Delaney M Brown S Lugo AE Lezama AT Quintero NBTorres-Lezama A 1998 The Quantity and Turnover ofDead Wood in Permanent Forest Plots in Six Life Zones ofVenezuela Biotropica 30 2ndash11

ESRI 2005 ESRI ArcGIS 9 Users Manual httpwwwesricomFeldpausch TR Banin L Phillips OL Baker TR Lewis SL

Quesada CA Affum-Baffoe K Arets EJMM Berry NJBird M et al 2011 Height-diameter allometry of tropicalforest trees Biogeosciences 8 1081ndash1110

Feldpausch TR Lloyd J Lewis SL Brienen RJW Gloor EMendoza AM Lopez-Gonzalez G Banin L Salim KAAbu Salim K Affum-Baffoe K et al 2012 Tree heightintegrated into pan-tropical forest biomass estimatesBiogeosciences 9 3381ndash3403

Feldpausch TR McDonald AJ Passos CAM Lehmann J RihaSJ 2006 Biomass harvestable area and forest structureestimated from commercial timber inventories and remotelysensed imagery in southern Amazonia Forest Ecology andManagement 233 121ndash132

Gale N Barfod AS 1999 Canopy tree mode of death in awestern Ecuadorian rain forest Journal of TropicalEcology 15 415ndash436

Gale N Hall P 2001 Factors determining the modes of treedeath in three Bornean rain forests Journal of VegetationScience 12 337ndash348

Gregoire TG Valentine HT 2003 Line intersect sampling Ell-shaped transects and multiple intersections Environmentaland Ecological Statistics 10 263ndash279

Hijmans RJ Cameron SE Parra JL Jones PG Jarvis A 2005Very high resolution interpolated climate surfaces for globalland areas International Journal of Climatology 25 1965ndash1978

IBGE 1997 Recursos naturais e meio ambiente uma visatildeo doBrasil Instituto Brasileiro de Geografia e Estatiacutestica 2nd ed

IUSS (International Union of Soil Science) Working Group WRBWorld reference base for soil resources 2006 A frameworkfor international classification correlation and communicationWorld Soil Resources Report 103 FAO Rome

12 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014

Jirka S McDonald AJ Johnson MS Feldpausch TR Couto EGRiha SJ 2007 Relationships between soil hydrology andforest structure and composition in the southern BrazilianAmazon Journal of Vegetation Science 18 183ndash194

Keller M Palace M Asner GP Pereira R Silva JNM 2004Coarse woody debris in undisturbed and logged forests inthe eastern Brazilian Amazon Global Change Biology 10784ndash795

Kissing LB Powers JS 2010 Coarse woody debris stocks as afunction of forest type and stand age in Costa Rican tropicaldry forest long-lasting legacies of previous land use Journalof Tropical Ecology 26 467ndash471

Laurance WF Ferreira LV Rankin-de Merona JM Laurance SG1998 Rain Forest Fragmentation and the Dynamics ofAmazonian Tree Communities Ecology 79 2032ndash2040

Magnusson WE Lima AP Luizatildeo R Luizatildeo F Costa FRCCastilho CV Kinupp VF 2005 RAPELD umamodificaccedilatildeo do meacutetodo de Gentry para inventaacuterios de biodi-versidade em siacutetios para pesquisa ecoloacutegica de longaduraccedilatildeo Biota Neotropica (Ed Portuguesa 5 2 1ndash6

Malhi Y Wood D Baker TR Wright J Phillips OL Cochrane TMeir P Chave J Almeida S Arroyo L et al 2006 Theregional variation of above-ground live biomass in old-growthAmazonian forests Global Change Biology 12 1107ndash1138

Martius C 1997 The central Amazon flooplain - ecology of apulsing system In Junk WJ editor The Central AmazonFlooplain - Ecology of a Pulsing System Springer VerlagGermany pp 267ndash276

Moore ID Grayson RB Ladson AR 1991 Digital terrain mod-eling a review of hydrological geomorphological and bio-logical applications Hidrological Processes 5 3ndash30

Nakagawa S Schielzeth H 2013 A general and simple methodfor obtaining R2 from generalized linear mixed-effects mod-els Methods in Ecology and Evolution 4 133ndash142

Nascimento HEM Laurance WF 2002 Total aboveground bio-mass in central Amazonian rainforests a landscape-scalestudy Forest Ecology and Management 168 311ndash321

Negroacuten-Juaacuterez RI Chambers JQ Guimaratildees G Zeng H RauppCFM Marra DM Ribeiro GHPM Saatchi SS Nelson BWHiguchi N 2010 Widespread Amazon forest tree mortalityfrom a single cross-basin squall line event GeophysicalResearch Letters 37 1ndash5

Nogueira EM Fearnside PM Nelson BW Franca MB OliveiraACA 2008 Tree height in Brazilrsquos lsquoarc of deforestationrsquoshorter trees in south and southwest Amazonia imply lowerbiomass Forest Ecology and Management 255 2963ndash2972

PalaceM Keller M Asner GP Silva J Passos C 2007 Necromassin undisturbed and logged forests in the Brazilian AmazonForest Ecology and Management 238 309ndash318

Palace M Keller M Frolking S Hurtt H 2012 A review ofabove ground necromass in tropical forests In Sudarshana PNageswara-Rao M and Soneji JR editors Tropical ForestsInTech Croatia 215ndash252

Palace M Keller M Silva H 2008 Necromass productionstudies in undisturbed and logged Amazon forestsEcological Applications 18 873ndash884

Phillips OL Aragatildeo LEOC Lewis SL Fisher JB Lloyd JLopez-Gonzalez G Malhi Y Monteagudo A Peacock JQuesada CA et al 2009 Drought sensivity of the AmazonRainforest Science 323 1344ndash1347

Phillips OL Baker T Arroyo L Higuchi N Killeen T LauranceWF Lewis SL Lloyd J Malhi Y Monteagudo A et al 2004Patterns and process in Amazon tree turnover 1976ndash2001

Philosophical Transactions of the Royal Society of LondonSeries B 359 437ndash462

Pinheiro J Bates D DebRoy S Sarkar D 2012 nlme linear andnonlinear mixed effects models R Packag version 3 1ndash103Vienna Austria R Found Stat Comput

Quesada CA Lloyd J Anderson LO Fyllas NM Schwarz MCzimczik CI 2011 Soils of Amazonia with particularreference to the RAINFOR sites Biogeosciences 81415ndash1440

Quesada CA Lloyd J Schwarz M Patintildeo S Baker TRCzimczik CI Fyllas NM Martinelli LA Nardoto GBSchmerler J et al 2010 Variations in chemical and physicalproperties of Amazon forest soils in relation to their genesisBiogeosciences 7 1515ndash1541

Quesada CA Phillips OL Schwarz M Czimczik CI Baker TRPatintildeo S Fyllas NM Hodnett MG Herrera R Almeida Set al 2012 Basin-wide variations in Amazon forest structureand function are mediated by both soils and climateBiogeosciences 9 2203ndash2246

R Development Core Team 2013 R A Language andEnvironment for Statistical Computing R Foundation forStatistical Computing Vienna (Austria)

RADAMBRASIL 1978 Levantamento dos RecursosNaturais Rio de Janeiro (Brazil) Ministeacuterio deMinas e Energia Departamento Nacional de ProduccedilatildeoMineral

Ribeiro JEL Hopkins MJG Vicentini A Sothers CA Costa MABrito JM Souza MAD Martins LH Lohmann LGAssunccedilatildeo PACL 1999 Flora da Reserva Ducke guia deidentificaccedilatildeo das plantas vasculares de uma floresta de terrafirme na Amazocircnia Central Flora Manaus - Brazil INPA -UFAM

Rice AH Pyle EH Saleska SR Hutyra L Palace M Keller Mde Camargo PB Portilho K Marques DF Wofsy SC 2004Carbon Balance and Vegetation Dynamics in an Old-Growth Amazonian Forest Ecological Applications 1455ndash71

Sombroek W 2000 Amazon landforms and soils realtion tobiological diversity Acta Amazonica 30 81ndash100

Ter Steege H Sabatier D Castellanos H van Andel TDuivenvoorden J Adalardo de Oliveira A Ek R LilwahR Maas P Mori S 2000 An analysis of the floristiccomposition and diversity of Amazonian forests includingthose of the Guiana Shield Journal of Tropical Ecology 16801ndash828

Toledo JJ Magnusson WE Castilho CV Nascimento HEM2012 Tree mode of death in Central Amazonia Effects ofsoil and topography on tree mortality associated withstorm disturbances Forest Ecology and Management263 253ndash261

Wilcke W Hess T Bengel C Homeier J Valarezo C Zech W2005 Coarse woody debris in a montane forest in Ecuadormass C and nutrient stock and turnover Forest Ecology andManagement 205 139ndash147

Zanne AE Lopez-Gonzalez G Coomes DA Ilic J Jansen SLewis SL Miller R Swenson NG Wiemann M Chave J2009 Data from Towards a worldwide wood economicsspectrum Dryad Digital Repository

van Geffen KG Poorter L Sass-Klaassen U van Logtestijn RSPCornelissen JHC 2010 The trait contribution to wooddecomposition rates of 15 Neotropical tree speciesEcology 91 3686ndash3697

van Wagner CE 1968 The line intersect method in forest fuelsampling Forest science 24 469ndash483


Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014

form to anyone is expressly forbidden Terms amp Conditions of access and use can be found at httpwwwtandfonlinecompageterms-and-conditions

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014







Introduction







Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014





Study sites




2 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014








CWD wood density



Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014


Vegetation


H frac14 48131 1 exp 00375 D08228

(1)




(2)



Soil data




Category Score



2



4


3


2


1




3


2


1


4 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014





TI frac14 lnα

tan β

(3)



Calculations



8 L(4)



VBelt frac14 Hπ D1

2

2thornπ D22

22

(5)


















Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014

Results





Stocks of CWD


Π1

(a)

Π2

(b)


(c)

Soil structure

(d)

Anoxia

(e)

Topography

(f)

Topographic index

(g)


(h)


(i)

AGB (Mg haminus1)

(j)

ln C

WD

(Mg

haminus1

)ln

CW

D (M

g ha

minus1)

ln C

WD

(Mg

haminus1

)ln

CW

D (M

g ha

minus1)

ln C

WD

(Mg

haminus1

)ln

CW

D (M

g ha

minus1)

ln C

WD

(Mg

haminus1

)ln

CW

D (M

g ha

minus1)

ln C

WD

(Mg

haminus1

)ln

CW

D (M

g ha

minus1)

ln C

WD

(Mg

haminus1

)

0 2 4 6 8 10

15

25

35

45

0 2 4 6 8

15

25

35

45

0 1 2 3 4

15

25

35

45

4

15

25

35

45

0 1 2 3 0 1 2 3 4

15

25

35

45

15

25

35

45

0 1 2 3 4

7 8 9 10 11 12 13 14

15

25

35

45

18 20 22 24 26

15

25

35

45

14 15 16 17 18

15

25

35

45

100 150 200 250 300 350

15

25

35

45

02 03 04 05

15

25

35

45

AGBper tree (Mg)

(k)



6 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014








Vegetation



NRamx LRLamx HRLany













Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014







Discussion




NRax LRLaby HRLbz





NR LRL HRL

AGB(Mg haminus1)



AGB per tree(Mg)




CWD(Mg haminus1)




8 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014















Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014









10 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014






Conclusions






Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014








































12 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014



































Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014







Introduction







Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014





Study sites




2 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014








CWD wood density



Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014


Vegetation


H frac14 48131 1 exp 00375 D08228

(1)




(2)



Soil data




Category Score



2



4


3


2


1




3


2


1


4 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014





TI frac14 lnα

tan β

(3)



Calculations



8 L(4)



VBelt frac14 Hπ D1

2

2thornπ D22

22

(5)


















Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014

Results





Stocks of CWD


Π1

(a)

Π2

(b)


(c)

Soil structure

(d)

Anoxia

(e)

Topography

(f)

Topographic index

(g)


(h)


(i)

AGB (Mg haminus1)

(j)

ln C

WD

(Mg

haminus1

)ln

CW

D (M

g ha

minus1)

ln C

WD

(Mg

haminus1

)ln

CW

D (M

g ha

minus1)

ln C

WD

(Mg

haminus1

)ln

CW

D (M

g ha

minus1)

ln C

WD

(Mg

haminus1

)ln

CW

D (M

g ha

minus1)

ln C

WD

(Mg

haminus1

)ln

CW

D (M

g ha

minus1)

ln C

WD

(Mg

haminus1

)

0 2 4 6 8 10

15

25

35

45

0 2 4 6 8

15

25

35

45

0 1 2 3 4

15

25

35

45

4

15

25

35

45

0 1 2 3 0 1 2 3 4

15

25

35

45

15

25

35

45

0 1 2 3 4

7 8 9 10 11 12 13 14

15

25

35

45

18 20 22 24 26

15

25

35

45

14 15 16 17 18

15

25

35

45

100 150 200 250 300 350

15

25

35

45

02 03 04 05

15

25

35

45

AGBper tree (Mg)

(k)



6 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014








Vegetation



NRamx LRLamx HRLany













Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014







Discussion




NRax LRLaby HRLbz





NR LRL HRL

AGB(Mg haminus1)



AGB per tree(Mg)




CWD(Mg haminus1)




8 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014















Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014









10 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014






Conclusions






Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014








































12 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014



































Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014





Study sites




2 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014








CWD wood density



Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014


Vegetation


H frac14 48131 1 exp 00375 D08228

(1)




(2)



Soil data




Category Score



2



4


3


2


1




3


2


1


4 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014





TI frac14 lnα

tan β

(3)



Calculations



8 L(4)



VBelt frac14 Hπ D1

2

2thornπ D22

22

(5)


















Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014

Results





Stocks of CWD


Π1

(a)

Π2

(b)


(c)

Soil structure

(d)

Anoxia

(e)

Topography

(f)

Topographic index

(g)


(h)


(i)

AGB (Mg haminus1)

(j)

ln C

WD

(Mg

haminus1

)ln

CW

D (M

g ha

minus1)

ln C

WD

(Mg

haminus1

)ln

CW

D (M

g ha

minus1)

ln C

WD

(Mg

haminus1

)ln

CW

D (M

g ha

minus1)

ln C

WD

(Mg

haminus1

)ln

CW

D (M

g ha

minus1)

ln C

WD

(Mg

haminus1

)ln

CW

D (M

g ha

minus1)

ln C

WD

(Mg

haminus1

)

0 2 4 6 8 10

15

25

35

45

0 2 4 6 8

15

25

35

45

0 1 2 3 4

15

25

35

45

4

15

25

35

45

0 1 2 3 0 1 2 3 4

15

25

35

45

15

25

35

45

0 1 2 3 4

7 8 9 10 11 12 13 14

15

25

35

45

18 20 22 24 26

15

25

35

45

14 15 16 17 18

15

25

35

45

100 150 200 250 300 350

15

25

35

45

02 03 04 05

15

25

35

45

AGBper tree (Mg)

(k)



6 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014








Vegetation



NRamx LRLamx HRLany













Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014







Discussion




NRax LRLaby HRLbz





NR LRL HRL

AGB(Mg haminus1)



AGB per tree(Mg)




CWD(Mg haminus1)




8 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014















Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014









10 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014






Conclusions






Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014








































12 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014



































Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014








CWD wood density



Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014


Vegetation


H frac14 48131 1 exp 00375 D08228

(1)




(2)



Soil data




Category Score



2



4


3


2


1




3


2


1


4 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014





TI frac14 lnα

tan β

(3)



Calculations



8 L(4)



VBelt frac14 Hπ D1

2

2thornπ D22

22

(5)


















Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014

Results





Stocks of CWD


Π1

(a)

Π2

(b)


(c)

Soil structure

(d)

Anoxia

(e)

Topography

(f)

Topographic index

(g)


(h)


(i)

AGB (Mg haminus1)

(j)

ln C

WD

(Mg

haminus1

)ln

CW

D (M

g ha

minus1)

ln C

WD

(Mg

haminus1

)ln

CW

D (M

g ha

minus1)

ln C

WD

(Mg

haminus1

)ln

CW

D (M

g ha

minus1)

ln C

WD

(Mg

haminus1

)ln

CW

D (M

g ha

minus1)

ln C

WD

(Mg

haminus1

)ln

CW

D (M

g ha

minus1)

ln C

WD

(Mg

haminus1

)

0 2 4 6 8 10

15

25

35

45

0 2 4 6 8

15

25

35

45

0 1 2 3 4

15

25

35

45

4

15

25

35

45

0 1 2 3 0 1 2 3 4

15

25

35

45

15

25

35

45

0 1 2 3 4

7 8 9 10 11 12 13 14

15

25

35

45

18 20 22 24 26

15

25

35

45

14 15 16 17 18

15

25

35

45

100 150 200 250 300 350

15

25

35

45

02 03 04 05

15

25

35

45

AGBper tree (Mg)

(k)



6 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014








Vegetation



NRamx LRLamx HRLany













Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014







Discussion




NRax LRLaby HRLbz





NR LRL HRL

AGB(Mg haminus1)



AGB per tree(Mg)




CWD(Mg haminus1)




8 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014















Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014









10 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014






Conclusions






Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014








































12 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014



































Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014


Vegetation


H frac14 48131 1 exp 00375 D08228

(1)




(2)



Soil data




Category Score



2



4


3


2


1




3


2


1


4 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014





TI frac14 lnα

tan β

(3)



Calculations



8 L(4)



VBelt frac14 Hπ D1

2

2thornπ D22

22

(5)


















Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014

Results





Stocks of CWD


Π1

(a)

Π2

(b)


(c)

Soil structure

(d)

Anoxia

(e)

Topography

(f)

Topographic index

(g)


(h)


(i)

AGB (Mg haminus1)

(j)

ln C

WD

(Mg

haminus1

)ln

CW

D (M

g ha

minus1)

ln C

WD

(Mg

haminus1

)ln

CW

D (M

g ha

minus1)

ln C

WD

(Mg

haminus1

)ln

CW

D (M

g ha

minus1)

ln C

WD

(Mg

haminus1

)ln

CW

D (M

g ha

minus1)

ln C

WD

(Mg

haminus1

)ln

CW

D (M

g ha

minus1)

ln C

WD

(Mg

haminus1

)

0 2 4 6 8 10

15

25

35

45

0 2 4 6 8

15

25

35

45

0 1 2 3 4

15

25

35

45

4

15

25

35

45

0 1 2 3 0 1 2 3 4

15

25

35

45

15

25

35

45

0 1 2 3 4

7 8 9 10 11 12 13 14

15

25

35

45

18 20 22 24 26

15

25

35

45

14 15 16 17 18

15

25

35

45

100 150 200 250 300 350

15

25

35

45

02 03 04 05

15

25

35

45

AGBper tree (Mg)

(k)



6 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014








Vegetation



NRamx LRLamx HRLany













Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014







Discussion




NRax LRLaby HRLbz





NR LRL HRL

AGB(Mg haminus1)



AGB per tree(Mg)




CWD(Mg haminus1)




8 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014















Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014









10 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014






Conclusions






Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014








































12 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014



































Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014





TI frac14 lnα

tan β

(3)



Calculations



8 L(4)



VBelt frac14 Hπ D1

2

2thornπ D22

22

(5)


















Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014

Results





Stocks of CWD


Π1

(a)

Π2

(b)


(c)

Soil structure

(d)

Anoxia

(e)

Topography

(f)

Topographic index

(g)


(h)


(i)

AGB (Mg haminus1)

(j)

ln C

WD

(Mg

haminus1

)ln

CW

D (M

g ha

minus1)

ln C

WD

(Mg

haminus1

)ln

CW

D (M

g ha

minus1)

ln C

WD

(Mg

haminus1

)ln

CW

D (M

g ha

minus1)

ln C

WD

(Mg

haminus1

)ln

CW

D (M

g ha

minus1)

ln C

WD

(Mg

haminus1

)ln

CW

D (M

g ha

minus1)

ln C

WD

(Mg

haminus1

)

0 2 4 6 8 10

15

25

35

45

0 2 4 6 8

15

25

35

45

0 1 2 3 4

15

25

35

45

4

15

25

35

45

0 1 2 3 0 1 2 3 4

15

25

35

45

15

25

35

45

0 1 2 3 4

7 8 9 10 11 12 13 14

15

25

35

45

18 20 22 24 26

15

25

35

45

14 15 16 17 18

15

25

35

45

100 150 200 250 300 350

15

25

35

45

02 03 04 05

15

25

35

45

AGBper tree (Mg)

(k)



6 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014








Vegetation



NRamx LRLamx HRLany













Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014







Discussion




NRax LRLaby HRLbz





NR LRL HRL

AGB(Mg haminus1)



AGB per tree(Mg)




CWD(Mg haminus1)




8 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014















Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014









10 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014






Conclusions






Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014








































12 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014



































Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014

Results





Stocks of CWD


Π1

(a)

Π2

(b)


(c)

Soil structure

(d)

Anoxia

(e)

Topography

(f)

Topographic index

(g)


(h)


(i)

AGB (Mg haminus1)

(j)

ln C

WD

(Mg

haminus1

)ln

CW

D (M

g ha

minus1)

ln C

WD

(Mg

haminus1

)ln

CW

D (M

g ha

minus1)

ln C

WD

(Mg

haminus1

)ln

CW

D (M

g ha

minus1)

ln C

WD

(Mg

haminus1

)ln

CW

D (M

g ha

minus1)

ln C

WD

(Mg

haminus1

)ln

CW

D (M

g ha

minus1)

ln C

WD

(Mg

haminus1

)

0 2 4 6 8 10

15

25

35

45

0 2 4 6 8

15

25

35

45

0 1 2 3 4

15

25

35

45

4

15

25

35

45

0 1 2 3 0 1 2 3 4

15

25

35

45

15

25

35

45

0 1 2 3 4

7 8 9 10 11 12 13 14

15

25

35

45

18 20 22 24 26

15

25

35

45

14 15 16 17 18

15

25

35

45

100 150 200 250 300 350

15

25

35

45

02 03 04 05

15

25

35

45

AGBper tree (Mg)

(k)



6 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014








Vegetation



NRamx LRLamx HRLany













Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014







Discussion




NRax LRLaby HRLbz





NR LRL HRL

AGB(Mg haminus1)



AGB per tree(Mg)




CWD(Mg haminus1)




8 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014















Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014









10 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014






Conclusions






Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014








































12 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014



































Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014








Vegetation



NRamx LRLamx HRLany













Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014







Discussion




NRax LRLaby HRLbz





NR LRL HRL

AGB(Mg haminus1)



AGB per tree(Mg)




CWD(Mg haminus1)




8 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014















Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014









10 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014






Conclusions






Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014








































12 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014



































Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014







Discussion




NRax LRLaby HRLbz





NR LRL HRL

AGB(Mg haminus1)



AGB per tree(Mg)




CWD(Mg haminus1)




8 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014















Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014









10 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014






Conclusions






Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014








































12 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014



































Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014















Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014









10 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014






Conclusions






Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014








































12 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014



































Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014









10 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014






Conclusions






Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014








































12 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014



































Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014






Conclusions






Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014








































12 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014



































Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014








































12 DL Martins et al

Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014



































Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014



































Dow

nloa

ded

by [

186

222

983

7] a

t 06

55 2

3 A

pril

2014

Soil-induced impacts on forest structure drive coarse woody debris stocks across central Amazonia

Documents

Transcript of Soil-induced impacts on forest structure drive coarse woody debris stocks across central Amazonia