The water footprint of wood for lumber, pulp, paper, fuel and firewood

ContentslistsavailableatScienceDirect

Advances

in

Water

Resources

journalhomepage:www.elsevier.com/locate/advwatres

The

water

footprint

of

wood

for

lumber,

pulp,

paper,

fuel

and

firewood

Joep

F.

Schyns

a,∗

_,

_Martijn

_J.

_Booij

a

_,

_Arjen

_Y.

_Hoekstra

a,b

a Twente Water Centre, University of Twente, PO Box 217, 7500 AE Enschede, The Netherlands

b Institute of Water Policy, Lee Kuan Yew School of Public Policy, National University of Singapore, 259770, Singapore

a

r

t

i

c

l

e

i

n

f

o

Article history: Received 10 January 2017 Revised 15 May 2017 Accepted 16 May 2017 Available online xxx Keywords: Water consumption Evapotranspiration Forestry Timber Bioenergy Ecosystem services

a

b

s

t

r

a

c

t

Thispaperpresentsthe firstestimateofglobal wateruse inthe forestrysectorrelatedtoroundwood productionforlumber,pulp,paper,fuelandfirewood.Fortheperiod1961–2010,weestimateforest evap-orationatahighspatialresolutionlevelandattributetotalwaterconsumptiontovariousforestproducts, includingecosystemservices.Globalwaterconsumptionforroundwoodproductionincreasedby25%over 50yearsto961×109 _m3_{/y(96%green;4%blue)in2001–2010.Thewaterfootprintperm}3 _ofwoodis

significantlysmallerin(sub)tropicalforestscomparedtotemperate/borealforests,because(sub)tropical forestshostrelativelymorevaluenexttowoodproductionintheformofotherecosystemservices.In termsofeconomicwaterproductivityandenergyyieldfrombio-ethanolperunitofwater,roundwoodis rathercomparablewithmajorfood,feedandenergycrops.Recyclingofwoodproductscouldeffectively reducethewaterfootprintoftheforestrysector,therebyleavingmorewateravailableforthegeneration ofotherecosystemservices.Intensificationofwoodproductioncanonlyreducethewaterfootprintper unitofwoodifthe additionalwoodvalueperhaoutweighsthelossofvalueofotherecosystem ser-vices,whichisoftennotthecasein(sub)tropicalforests.Theresultsofthisstudycontributetoamore completepictureofthehumanappropriationofwater,thusfeedingthedebateonwaterforfoodorfeed versusenergyandwood.

© 2017TheAuthors.PublishedbyElsevierLtd. ThisisanopenaccessarticleundertheCCBYlicense.(http://creativecommons.org/licenses/by/4.0/)

1. Introduction

Although precipitation is renewable, it is limited in time and space, andsoareits subsequentpathwaysasgreenandblue wa-ter flows (Schynsetal., 2015; Hoekstra, 2013).There are alterna-tive competinguses forthese limited flows, which makes fresh-water a scarce resource.This explains the interest inthe human appropriationofwater(Posteletal.,1996; Rockströmetal.,1999; Rockström and Gordon, 2001; Hoekstra and Mekonnen, 2012) in relation to a maximum sustainable level (Hoekstra and Wiedmann, 2014) or planetary boundary (Steffen et al., 2015; Rockstrometal.,2009).Freshwatersustainsterrestrialandaquatic ecosystemsandis usedfortheproduction ofgoods andservices. Importantwaterconsumingsectorsareagriculture,industries, mu-nicipalitiesandforestry.Multiplestudieshavequantifiedtheglobal blue and green water consumption for producing crop and live-stockproducts,andforfulfillingindustrialandmunicipaldemands (HoekstraandMekonnen,2012; Rostetal.,2008; Hanasakietal., 2010; LiuandYang,2010; Liuetal.,2009; SiebertandDöll,2010; MekonnenandHoekstra,2011; Wadaetal.,2014; Dölletal.,2012).

∗ _{Corresponding author.}

E-mail address: j.f.schyns@utwente.nl (J.F. Schyns).

As recently identified by Vanham (2016), we do not know how much water is used in the forestry sector for the production of woodproducts such aslumber, pulp andpaper,firewood or bio-fuel.

Forestevaporationaccountsfor45–58%ofthetotalvapourflow fromland toatmosphere(Rockströmetal., 1999; Rockströmand Gordon, 2001; Oki and Kanae, 2006). With the term evapora-tionwe refer to theentirevapour flux fromland toatmosphere, including evaporation through the process of plant transpiration (Savenije,2004).Determiningwhichpartoftheevaporationis ap-propriatedfortheproductionofroundwood(woodintherough)is notasstraightforwardasitisforcrops.Forcrops,allevaporation fromthecropfieldduringthegrowingseasonisusuallyattributed tocrop production.Thismakes sense,since cropfieldsare gener-allyused quiteintensively fora distinct purpose(providing food, feedorfibre).Forestsonthe other handprovidenumerousother ecosystemservicesnexttotheprovision ofwood(Costanzaetal., 1997),dependingontheintensityofforestexploitation.Therefore, forest evaporation is to be attributed to roundwood production basedontherelativevalueofroundwoodproductioncomparedto thevalueofotherecosystemservicesprovidedbytheforest.

Thereareafewstudiesthat haveattributedforestevaporation to woodproducts. Van Oel and Hoekstra (2012) madea first es-http://dx.doi.org/10.1016/j.advwatres.2017.05.013

0309-1708/© 2017 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license. ( http://creativecommons.org/licenses/by/4.0/ )

timateof the waterfootprint ofpaper inthe main pulp produc-ing countries. Chiuand Wu (2013) estimatedthe waterfootprint ofethanolfrom woodresidues fromthe southeastUnited States. TianandKe(2012)madeestimatesofthewaterfootprintof lum-ber, panels,pulp andpaper inChina. However, these studies did not account for the value of wood production relative to other forest values (Van Oel and Hoekstra, 2012; Chiu and Wu, 2013; TianandKe, 2012). Launiainenetal.(2014)arguethatoneshould not attribute forest evaporation of rain-fed managed forests to endproducts atall, based on the argumentthat the evaporation ofthese forests isnot significantly different than that of natural forests(no net difference). However, for the purpose of measur-ingtheamountofevaporationthatisappropriatedbyroundwood productionand therefore not available for other uses we should measuretotal(notnet)waterconsumption(Hoekstra,2017).

Theobjectiveofthispaperistoprovidethefirstestimateofthe globalwaterconsumptionrelatedtoroundwoodproductionandto subsequentlyattributethistovariousend-usesofwood.Our analy-sisisathighspatialresolution(30× 30)fortheperiod1961–2010 andincludesanumberofinnovations:

- Global high-resolution estimates of actual evaporation from productionforests,distinguishingthecontributionofgreen wa-ter (precipitation)andbluewater(groundwaterthrough capil-laryrise).

- Attribution of forest evaporation to roundwood production basedontherelativevalueofroundwoodproductioncompared tothevalueofotherecosystemservicesprovidedbytheforest. - Estimatesofthegreenandbluewaterfootprintsofwood prod-ucts, includingsawnwood,wood-basedpanels,woodpulp, pa-perandwood-basedenergycarriers.

2. Methodanddata

2.1.Method

We follow the method of waterfootprint assessment to esti-matethewaterconsumption associatedwithroundwood produc-tion for lumber, pulp, paper, fuel and firewood (Hoekstra et al., 2011).Firstly,weestimatethevolumeofwaterconsumedthatcan beattributedtoroundwoodproductionper30× 30gridcellper yearover the period 1961–2010 (Section 2.1.1). Secondly,we es-timatetheperiod-average waterfootprint perunit of roundwood produced(Section2.1.2).Finally,weattributethewaterfootprintof roundwoodproductiontovariousend-usesofwood(Section2.1.3). Throughoutthispaperweusethetermwaterfootprinttoreferto theconsumptivepartonly(greenplus blue)andexclude thegrey componentthatexpresseswaterpollution.

2.1.1. Waterconsumptionattributedtoroundwoodproduction

Thevolumeofwaterconsumedthatcanbeattributedto round-woodproduction(WU,inm3_{/y)ingridcellxinyeartisestimated}

as:

WU[x,t]=

(

Eact[x,t]× Arw[x,t]+Pact[x,t]× fwater[x]

)

× fvalue,rw[x,t] (1)

inwhichEact is the actualforest evaporation(m/y), Arw thearea

usedfor roundwood production(m2_{), P}

act the actual roundwood

harvested(m3_/y),f

waterthevolumetricmoisturecontentoffreshly

harvestedwood(m3 _water/m3_wood),andf

value,rwadimensionless

fractionthat represents the relative value of roundwood produc-tioncompared to thevalue of other ecosystem servicesprovided bytheforest.

Annualactualforestevaporation

Eact (m/y)isestimatedusingthemethodof Zhangetal.(2001):

Eact[x,t]=Pr[x,t]



1+wE0[x,t] Pr[x,t] 1+wE0[x,t] Pr[x,t]+ Pr[x,t] E0[x,t]



(2) in whichPr is the annualprecipitation (m/y), wa dimensionless coefficientrepresentingplantwateravailability,andE0theannual

potential forest evaporation (m/y). We apply w=2, which is the bestfit valueforforestsbased ona studythat includes56 forest catchmentsaround the world(Zhanget al., 1999). We determine

E0 basedonthemeanannualtemperature(T,in0C)usingthe

em-piricalequation derivedby Komatsu etal.(2012), whichthey de-rivedforZhang’sequationbyregressing829forestEactdatapoints:

E0[x,t]=



0.488T2[x,t]+27.5T[x,t]+412



× 10−3 ₍₃₎

Thefactor10−3istoconvertmmtom.

Distinctionbetweengreenandbluewateruse

Thedistinction betweengreenandbluewateruseis madeby applyingafractionthatrepresentsthepartofwaterusethat orig-inatesfromcapillaryrise(fblue):

WUgreen[x,t]=WU[x,t]×

(

1− fblue[x,t]

)

(4)

WUblue[x,t]=WU[x,t]× fblue[x,t] (5)

Weestimatef_bluebasedontwomainassumptions:

- Capillaryrise is atits maximumin a verydry year(Eact/Pr =

1)andmoveslinearlytozeroinanextremelywetyear(Eact/Pr

=0).Awaterpotentialgradientisrequiredtomovewater up-wardfromthegroundwatertable.Whenthesoilisdrythis gra-dientis strong. Ifthe soil issaturated this gradient is absent andtherewillbenocapillaryrise.

- The distance that needs to be bridged by capillary rise (dcap,

inm)isdefinedasthedifferencebetweenthegroundwater ta-bledepth(zg) andtherootdepth oftheforest type (zr), both

inm belowacertain referencelevel.The maximumheight of capillaryrise (dcap,max,in m) dependson the soil type. When

dcap isnon-limiting(≤0), therootstake up asharedcap,max of

zrthrough capillaryrise underverydryconditions.Thisshare

decreases linearly to zero when dcap approachesdcap,max

(be-yond,thereisnocapillaryuptakeatall).

Theseassumptionscanbecombinedintoasingleequationthat applieswhen0≤ dcap <dcap,max:

fblue[x,t]= dcap,max[x] zr[x] Eact[x,t] Pr[x,t]



1−zg[x]− zr[x] dcap,max[x]



(6)

2.1.2. Waterfootprintperunitofroundwoodproduction

Since wood production cycles are commonly multi-decadal (Bauhus etal., 2009),we calculatethe waterfootprintper unitof productionasaperiod-average.Thewaterfootprintofroundwood production(WFrw,inm3 water/m3 roundwood) forthe periodof

myearsisdefinedas:

WFrw[x]= m  t=1 WU[x] m  t=1 Pact[x] (7)

2.1.3. Waterfootprintperunitofendproduct

The waterfootprint per unit ofend product pproduced with roundwoodfromgridcellxisestimatedbymultiplyingWFrwwith

aconversionfactor(f_conversion,inm3_{roundwood/unit ofend}

prod-uct):

2.2. Data

2.2.1. Woodharvestedarea

We obtainedwood harvestedarea maps (asfraction ofa grid cell)at30× 30resolutionforeachyearintheperiod1961–2004 from Chinietal.(2014).For2005–2010,wekeepthepatternfrom 2004. Hurttetal.(2011)estimatedthewoodharvestpatternwith a globalland-usemodel thattakes,among others,nationalwood harvestdataasinput,constrainswoodharvestingbythepresence of forests,andgives preferenceto wood harvestingnear existing land-use (proximity to infrastructure or local markets). We took thesumofthefivedifferentlandtypesfromwhichwoodcanbe harvestedasdistinguishedby Hurttetal.(2011).

Weapply threerestrictionstothesemaps.Firstly,we assumed that roundwood production only takes place in those grid cells thathaveaforestcoveraccordingtotheIGBPDISCoverlandcover database (Loveland etal., 2009). Secondly,we consider grid cells withanaverage Eact overthestudyperiodoflessthan 100mm/y

to be unsuitable for forest growth that enables wood harvest-ing and hence remove those grid cellsfrom our final map. This threshold is derived from Komatsu et al. (2012), who collected 829 forest Eact data points at locations spread over the world,

of which only three (0.4%) have an Eact smaller than 100mm/y.

Thirdly,weassumedthatnowoodisharvestedfromgridcellsthat are entirelylocated within a protected area of IUCN category Ia (strictnaturereserves),Ib(wildernessareas)orII(nationalparks) from the year that these areas received this status. The data on protected areashave beenobtained fromIUCN andUNEP-WCMC (2016).

We made one exception to the above procedure. The Peo-ple’s Democratic Republic of Ethiopia (1961–1992) had a signifi-cantcontributiontoworldroundwoodproductionaccordingto na-tionalstatistics (FAOSTAT,2016a).However, thecellswherewood harvesting took place in this country according to the map by Chinietal.(2014)havenoforestcoveraccordingtotheIGBP DIS-Cover dataset.Toavoidneglectofthisroundwoodproduction,we assigned themostcommon forest type inthe regionto the cells wherewoodwasharvested:tropicalevergreenbroadleafforest. Fi-nally,wescalethewoodharvestedareamapsonthenationallevel totheareausedforroundwoodproductionestimatedbasedonthe Global Forest Resources Assessment2015 (Köhl et al., 2015) (see SI).

2.2.2. Actualroundwoodproductiononthegridlevel

Nationalannualstatisticsonactualroundwoodproductionfrom coniferous (C) andnon-coniferous (NC)forest covering the study period havebeen obtained from FAOSTAT, 2016a). We downscale these data to the grid level in two steps. Firstly, we estimate the maximum sustainable production in a grid cell by multiply-ing the wood harvested area with a long-term maximum sus-tainable wood yield (Section 2.2.3). Therein, we distinguish be-tweenCandNCproductionbyassumingthatCwoodisproduced in needleleaf forests and NC wood in broadleaf forests and that mixedforestcontributestoboth CandNCproduction(fifty–fifty). For a small number of countries, in some years, reported pro-duction concerns C and/or NC wood,while our mapscontain no grid cells yielding that type ofwood (e.g. only NC productionis reported, but all grid cells in the wood harvest map are of the needleleaf type). In these cases,we overwrite the dominant for-esttypeinallaffectedgridcellsforthatyeartomixedforest. Sec-ondly,wedistributethenationalannualstatisticsoverallgridcells usedforroundwoodproductioninthatyear,accordingtothe esti-matedmaximumsustainable productionforthat roundwoodtype (CorNC).

2.2.3. Long-termmaximumsustainablewoodyield

Therate ofwoodproduction variesoverthe ageof theforest standfollowingans-shapedcurvethatisdifferentforeachspecies, location andtype of management (Lutz, 2011). The meanannual incrementisthe averageproduction rateatanyparticular ageof the forest, calculated as the total growing stock volume divided bythe ageoftheforest stand (Lutz, 2011; Jürgensenet al.,2014; Blanchez,1997).Weobtainedminimumandmaximumforest plan-tationyields(inm3_{/ha/y)fordifferenttreespeciesinvarious}

coun-triesaroundtheworldfrom Brown(2000).Theseyieldsrepresent the mean annual increment for the likely rotation length of the foreststand.Weassumethatforestsareofamixedage,thattrees areharvestedattheirlikelyrotationlengthandthatnaturallosses areminimal.Underthesecircumstances,weconsidertheyieldsby Brown(2000)tobeagoodproxyofthelong-termmaximum sus-tainablewoodyield(Ysus).

Ultimately,weneedanestimateofYsus foreachgridcellinour

roundwoodproductionmaps.Toarrivethere,(i)wedeterminethe dominantforest type and climate zone of each grid cell; (ii) we assumecharacteristictreespeciesforeachforesttype;(iii)we de-terminethedominantclimatezoneineachcountryinthedataset by Brown(2000);(iv)from thisdatasetwe calculatethe average

Ysusofatreespeciesperclimatezoneand(v)weassignthoseYsus

estimatestothegridcells.DetailsaredescribedintheSI.The fol-lowingassumptions aremadeunder(ii), whichare looselybased ontheforesttypedescriptionsof Matthewsetal.(2000):

-Evergreenneedleleafforestyieldspine(Pinusspecies)inall cli-matezones.

-Evergreenbroadleafforestyieldseucalyptus(Eucalyptusspecies) inallclimatezones.

-Deciduousneedleleafforestyieldslarch(Larixspecies)inall cli-matezones.

-Deciduousbroadleafforestyieldsoak(Quercusspecies)inall cli-matezones.

-Mixedforestinthetropicalandsubtropicalzoneyieldsa50–50 mixofpineandeucalyptus.

-Mixedforestinthetemperatezoneyieldsa50–50mixofpine andoak.

-Mixedforestintheborealzoneyieldsa50–50mixofpineand larch.

The resulting Ysus estimates per forest type and climatezone

are presentedin Table 1.The climate zonesand forest types are mappedin Fig.1.

2.2.4. Meteorologicaldata

For each 30 _{× 30} grid cell and each year in our study pe-riod,weestimatedtheannualprecipitation(Pr) andannualmean temperature(T) basedon dailydataobtainedfrom deGraafetal. (2014).

2.2.5. Fractionofwateruseoriginatingfromcapillaryrise

Rooting depths were derived from Canadell et al. (1996) (Table2).Thegroundwater tabledepthper30× 30 gridcellhas beenestimatedby averagingoverthe30× 30map by Fanetal. (2013). The maximum height ofcapillary rise is estimated using anempirical relationbasedonthe soil’sgrain sizeandvoidratio (detailsinSI).

2.2.6. Volumetricmoisturecontentoffreshlyharvestedwood

Thefraction fwater is estimatedby multiplying a specieswood

density with the equilibrium moisture content (t water/t oven dried wood)(derivation inthe SI). The wooddensity foreach of the characteristic tree species considered in this study has been estimated from Zanne et al. (2009) (Table 3). The equilibrium moisture content is estimated per grid cell for each year with

Fig. 1. Forest types and climate zones for the grid cells where roundwood is produced. Data obtained from Loveland et al. (2009) and Van Velthuizen et al. (2007 ) as described in the SI.

Fig. 2. The relative value of ecosystems services for tropical (left) and temperate/boreal forests (right). Data source: Costanza et al. (2014 ). Descriptions of the ecosystem services can be found in Costanza et al. (1997 ).

the function of Simpson (1998) that takes temperature and rel-ativehumidity asinputs. We applied the mean annual tempera-ture (Section 2.2.4) and a climate-average relative humidity per grid cell. The latter is estimated based on the 10 × 10 grid-ded monthly mean relative humidity data for 1961–1990 from Newetal.(2002). Wetook theaverage ofallmonths and subse-quentlytheaverageofall10×10gridcellswithina30×30 grid cell.

2.2.7. Valuefractionofroundwoodproduction

We baseourestimate ofthevaluefractionofroundwood pro-duction (fvalue,rw) on Costanza et al. (2014), who estimated the

valueof17ecosystemsservices(inmonetaryunits/ha)aroundthe year 2011 for (sub)tropical forests and temperate/boreal forests, separately(Fig.2).Weassumethattheservicelabelled‘raw mate-rials’by Costanzaetal.(2014)primarilyreferstoroundwood pro-duction.Non-woodforestproductsthatarenotofinterestforthis studyareincluded underother services,e.g. foodandfood addi-tives(‘food production’) andplantand animalparts for pharma-ceuticalproducts(‘geneticresources’).

The data in Fig.2 refer to theentire forest biomes,while we areinterestedinproductionforestsspecifically.Therefore,wefirst distribute the monetary values per hectare of the services over productionandnon-productionforestsforthereferenceyear2010

(whichliesclosesttothereportingyearby Costanzaetal.(2014)). Secondly,we scalethevaluesbackintimeanddisaggregatethem spatially over the grid cells. Therein, we distinguish three cate-goriesofecosystemservicevalues:

- The value of roundwood productionthat varieswiththe vol-umeofroundwoodproduced.

- The value ofthe services pollination, biological control, habi-tat/refugia, recreation and culture that are inversely propor-tionalto the intensity offorest exploitation, which isdefined astheactual woodproductionoverthe maximumsustainable woodproduction.

- Thevalueoftheotherservicesgivenin Fig.2thatareinvariable withtheintensityofforestexploitation.

For the year 2010, and averaged per biome, the resulting ecosystem service values are consistent with those reported in Fig.2.DetailsandassumptionsaredescribedintheSI.Ultimately, wecalculatefvalue,rwpergridcellperyearusingEq.(SI.10).

2.2.8. Woodtoendproductconversionfactors

Conversion factors forsawnwood, panels,pulp, paperand en-ergywoodproductsareobtainedfromUNECE/FAO(2010)and rep-resent averages ofreportedvaluesby countriesinthe UNECE re-gion. The energy values represent higher heating values (HHV).

Ta b le 1 Long-t erm maximum sus tainable yield per fo re st type and climat e zone (in m 3 /ha/y). Dat a es timat e d base d on Br o w n (20 0 0 ). Not eac h fo re st type is pr esent in eac h climat e zone as indicat e d with a h y phen. Climat e zone Ev er gr een nee d leleaf (pine) Ev er gr een br oadleaf (eucal yp tus) Deciduous nee d leleaf (lar ch ) Deciduous br oadleaf (oak) Mix e d Tr o p ic s 13 .5 9 –9 11 .5 Subtr o pics, summer ra in fa ll 13 .5 11 .5 12 11 .5 12 .5 Subtr o pics, wint e r ra in fa ll 8 10 12 5 9 Te m per at e 7 –8 5 6 Bor e al, oceanic & sub-continent al 6 –4 –5 Bor e al, continent a l & ar ctic 3 –4 –3 .5 Table 2

Rooting depth (m). Data derived from Canadell et al. (1996 ).

Climate zone Evergreen Deciduous Mixed Tropics & subtropics, summer rainfall 7 4 5 .5 Subtropics, winter rainfall a _{5 5 5}

Temperate 4 3 3 .5

Boreal & arctic 2 2 2

a Values for sclerophyllous forest

Table 3

Wood densities of the characteristic tree species considered in this study. Data from Zanne et al.

(2009 ) as described in Chave et al. (2009 ). Data rep-

resent the average of all entries for a species. Species Wood density (t/m 3 )

Pinus (Pine) 0 .4 Eucalyptus (Eucalyptus) 0 .8 Larix (Larch) 0 .5 Quercus (Oak) 0 .7

SomeadditionaldataontheHHVofsoftwood,hardwood,ethanol andcharcoalareobtainedfrom Speight(2010).Thewaterfootprint ofan A4 (=1/16m2_{) sheet ofpaperof 80g/m}2 _{in l/sheet is}

esti-matedbymultiplyingthewaterfootprintofpaperinm3_/twitha

factor0.005(=80/16/1000). 3. Results

3.1. Waterconsumptionattributedtoroundwoodproduction

The global water consumption attributed to roundwood pro-ductionincreasedby 25%over 50years,from768 × 109 _m3_/yin

1961–1970to961× 109_m3_{/yin2001–2010(forbothdecades:96%}

green;4%blue).Thewaterconsumptionequalstheevaporated vol-umeattributed to roundwood production,since the share of the waterincorporatedinthe harvestedwoodis negligible(0.01% on average). Fig.3showsthewaterconsumptionattributedto round-woodproduction(WU) andthevalue fractionofroundwood pro-duction(fvalue,rw)inthebiomes(sub)tropicalforestsandtemperate

andborealforests,separately.WUissignificantlysmallerinthe for-mercomparedtothelattercausedbythedifferenceinf_value,rwfor thosebiomes(Fig.2andSI).For(sub)tropicalforests,anincreasing trendinWUisobserved,drivenby increasesinthearea usedfor roundwood production and the volume of roundwood produced (see Fig. SI.2). For temperate and boreal forests, a moderate in-creasingtrendinWUisvisibledueto anincreasedarea usedfor roundwoodproduction.Inter-annualvariationislargerinthiscase. VariationinWUiscausedbyvariationinf_value,rw,whichinturnis mainlydrivenbyvariationinthevolumeofroundwoodproduced (Fig. SI.2). The latter explainsthe sudden decline in fvalue,rw and

WUafter1990whenthestatistics(FAOSTAT,2016a)showadropin roundwoodproduction(intheformerUSSR).Inbothforestbiomes, varying forest evaporationrates add to the temporal variation in

WU(Fig.SI.2).

3.2.Waterfootprintperunitofroundwoodproduction

The study period average water footprint per unit of round-wood production (WFrw) is presented in Fig. 4. Besides the

dif-ferences between the (sub)tropical and temperate/boreal zones (Section3.1),spatialvariationinWFrwismostlyexplainedby

vary-ingforestevaporationrates(TableSI.2).ThedecadeaverageWFrw

increased withabout ten percent over the study period in tem-perateandborealzones,whileitvariedwithinfivepercentinthe (sub)tropics.

Fig. 3. Water consumption attributed to roundwood production. Period: 1961–2010.

Fig. 4. The water footprint per unit of roundwood production (m 3 water/m 3 roundwood) at 30 × 30  _{resolution. The table next to the legend shows the average (production-}

weighted) water footprint per climate zone: boreal, continental & arctic (Bca); boreal, oceanic & sub-continental (Bsc); temperate (Tmp); subtropics, winter rainfall (Swr); subtropics, summer rainfall (Ssr); tropics (Tro). Period: 1961–2010. Note that not all grid cells were necessarily used for roundwood production in each year.

The average capillaryrise contribution to WFrw is mapped in

Fig.5. The areas with a capillary rise contribution of morethan 50% are mostly found in Russia and Canada. Blue water consti-tutesasignificantpartofthetotalwaterconsumptionattributedto roundwoodproductionincountriesliketheBahamas(32%), Gam-bia(28%),theNetherlands (24%)andSomalia(23%).Variations in thecapillaryrisecontributionaremainlyexplainedbythe ground-waterdepth. Milleretal.(2010)foundforasemi-aridoaksavanna

intheperiod2005–2008(averageEact/Prratioof0.7),thatthe

av-eragecontributionofcapillaryrisetotheevaporationovertheyear wasabout22%.Forgridcellswithacapillaryrisecontributionto evaporationandanEact/Prratioofatleast0.7,wefoundthis

con-tributiontobe18%onaverage.

Fig.6showstheaverageWFrw foreachofthemainroundwood

producing countries. There is a clear distinction between coun-trieswithproductionforestsinmainly(sub)tropicalversus

temper-Fig. 5. The average capillary rise contribution as a fraction of the forest water consumption ( f blue ). Resolution: 30 × 30  . Period: 1961–2010.

Fig. 6. The average (production-weighted) water footprint per unit of roundwood production (m 3 water/m 3 roundwood) for the main roundwood producing countries.

Period: 1961–2010.

ate/borealzones.Amongthemainroundwoodproducingcountries, JapanhasonaveragethelargestWFrw,resultingfroma

combina-tion of a relatively highforest evaporationrate with a relatively lowwoodyield.

Although pronounced spatial variations in WFrw occur, one

shouldbecautiousinevaluatingthesedifferencesintermsof bet-terorworse.Therelevanceofthedatapresentedratherliesinthe fact thatthey canformabasis forfurtherstudyinto the

alterna-tive uses of the same water to produce more or differentgoods andservicesinthesamearea(see Section4.3).

3.3.Waterfootprintperunitofendproduct

The water footprints of various end products derived from roundwood, based on globalaverages, are given in Table 4. The valuesvarydependingonthe originoftheroundwood,since the

Table 4

The water footprint of various end products derived from roundwood (rw) in m 3 water per unit of end product. Based on global average water footprint of

roundwood weighted by production: 390 m 3 /m 3 coniferous rw; 231 m 3 /m 3 non-coniferous rw; 293 m 3 /m 3 rw on average. Conversion factors are derived from

UNECE/FAO (2010 ). Additional data sources required to determine the conversion factors for energy wood products are indicated in the Table notes.

FAOSTAT code Product name Wood type Conversion factor Water footprint

Sawnwood

1632 Sawnwood coniferous 1 .86 m 3 rw/m 3 sawnwood _{726 m} 3 /m 3 sawnwood

1633 Sawnwood non-coniferous 1 .88 m 3 rw/m 3 sawnwood _{433 m} 3 /m 3 sawnwood

Veneer and plywood

1634 Veneer sheets – 2 .21 m 3 rw/m 3 sheets _{648 m} 3 /m 3 sheets

1634 Veneer sheets coniferous 2 .08 m 3 rw/m 3 sheets _{812 m} 3 /m 3 sheets

1634 Veneer sheets non-coniferous 2 .35 m 3 rw/m 3 sheets _{542 m} 3 /m 3 sheets

1640 Plywood – 2 .07 m 3 rw/m 3 panels _{607 m} 3 /m 3 panels

1640 Plywood coniferous 2 .01 m 3 rw/m 3 panels _{785 m} 3 /m 3 panels

1640 Plywood non-coniferous 2 .13 m 3 rw/m 3 panels _{491 m} 3 /m 3 panels

Wood panels from wood particles a

1646 Particle board – 2 .76 m 3 rw/m 3 panels _{809 m} 3 /m 3 panels

1646 Particle board coniferous 2 .64 m 3 rw/m 3 panels _{1031 m} 3 /m 3 panels

1646 Particle board non-coniferous 2 .87 m 3 rw/m 3 panels _{662 m} 3 /m 3 panels

1647 Hardboard – 3 .56 m 3 rw/m 3 panels _{1044 m} 3 /m 3 panels

1647 Hardboard coniferous 3 .41 m 3 rw/m 3 panels _{1331 m} 3 /m 3 panels

1647 Hardboard non-coniferous 3 .71 m 3 rw/m 3 panels _{855 m} 3 /m 3 panels

1648 MDF – 2 .95 m 3 rw/m 3 panels _{865 m} 3 /m 3 panels

1648 MDF coniferous 2 .82 m 3 rw/m 3 panels _{1101 m} 3 /m 3 panels

1648 MDF non-coniferous 3 .07 m 3 rw/m 3 panels _{708 m} 3 /m 3 panels

1650 Insulating board – 1 .46 m 3 rw/m 3 panels _{428 m} 3 /m 3 panels

1650 Insulating board coniferous 1 .39 m 3 rw/m 3 panels _{543 m} 3 /m 3 panels

1650 Insulating board non-coniferous 1 .52 m 3 rw/m 3 panels _{350 m} 3 /m 3 panels

Wood pulp

1654 Mechanical wood pulp – 2 .50 m 3 rw/t pulp _{733 m} 3 /t pulp

1655 Semi-chemical wood pulp – 2 .67 m 3 rw/t pulp _{783 m} 3 /t pulp

1656 Chemical wood pulp – 4 .49 m 3 rw/t pulp _{1316 m} 3 /t pulp

1660 Unbleached sulphite pulp – 4 .64 m 3 rw/t pulp _{1360 m} 3 /t pulp

1661 Bleached sulphite pulp – 4 .95 m 3 rw/t pulp _{1451 m} 3 /t pulp

1662 Unbleached sulphate pulp – 4 .45 m 3 rw/t pulp _{1305 m} 3 /t pulp

1663 Bleached sulphate pulp – 4 .55 m 3 rw/t pulp _{1334 m} 3 /t pulp

1667 Dissolving wood pulp – 5 .65 m 3 rw/t pulp _{1656 m} 3 /t pulp

Paper and paperboard

1612 Uncoated mechanical – 3 .32 m 3 rw/t paper _{973 m} 3 /t paper

1616 Coated papers – 3 .70 m 3 rw/t paper _{1085 m} 3 /t paper

1617 Case materials – 3 .88 m 3 rw/t paper _{1137 m} 3 /t paper

1618 Folding boxboard – 3 .75 m 3 rw/t paper _{1099 m} 3 /t paper

1621 Wrapping papers – 3 .82 m 3 rw/t paper _{1120 m} 3 /t paper

1622 Other papers packaging – 3 .75 m 3 rw/t paper _{1099 m} 3 /t paper

1671 Newsprint – 2 .87 m 3 rw/t paper _{841 m} 3 /t paper

1674 Printing + writing paper – 3 .51 m 3 rw/t paper _{1029 m} 3 /t paper

1675 Other paper + paperboard – 3 .29 m 3 rw/t paper _{965 m} 3 /t paper

1676 Household + sanitary paper – 4 .35 m 3 rw/t paper _{1275 m} 3 /t paper

1681 Wrapg + packg paper + board – 3 .25 m 3 rw/t paper _{953 m} 3 /t paper

1683 Paper + paperboard not else specified – 3 .29 m 3 rw/t paper _{965 m} 3 /t paper

Energy wood products

– Firewood coniferous 0 .12 m 3 rw/GJ b _{47 m} 3 /GJ

– Firewood non-coniferous 0 .09 m 3 rw/GJ c _{21 m} 3 /GJ

– Pellets – 0 .14 m 3 rw/GJ _{41 m} 3 /GJ

– Pressed logs and briquettes – 0 .23 m 3 rw/GJ _{67 m} 3 /GJ

– Bark and chipped fuel – 0 .10 m 3 rw/GJ _{29 m} 3 /GJ

– Wood-based ethanol – 0 .33 m 3 rw/GJ d _{97 m} 3 /GJ

– Wood-based ethanol – 7 .71 m 3 rw/m 3 ethanol _{2260 m} 3 /m 3 ethanol

1630 Wood charcoal – 0 .20 m 3 rw/GJ e _{59 m} 3 /GJ

a For wood panels from wood particles, we assume that particles are produced from green/rough sawnwood without losses and that 1 m 3 of green sawnwood

has a solid wood equivalent of 1 m 3 ) ( _{UNECE/FAO, 2010 ).}

b Higher heating value of softwood = 20.9 GJ/t softwood ( _{Speight, 2010 ); wood basic density of coniferous fuelwood logs = 0.42 dry t/green m} 3

( UNECE/FAO, 2010 ).

c Higher heating value of hardwood = 20.0 GJ/t hardwood ( _{Speight, 2010 ); wood basic density of non-coniferous fuelwood logs = 0.54 dry t/green m} 3

( UNECE/FAO, 2010 ).

d Higher heating value of ethanol = 29.7 GJ/t ethanol ( _{Speight, 2010 ); ethanol density = 0.789 t/m} 3 . e Higher heating value of charcoal = 29.6 GJ/t charcoal ( _{Speight, 2010 ).}

water footprint per cubic metre of roundwood produced varies aroundtheglobe(Fig.4).Theglobalaveragewaterfootprintofone A4sheet80g printingandwritingpaperis5.1landrangesfrom 1.0l/sheetinthesubtropicswithsummerrainfallto12.9l/sheetin thetemperatezone.

4. Discussion

4.1. Comparisonwithpreviousestimates

Aroughcomparisoncanbemadebetweenourestimatesofthe water footprint of roundwood and those by Van Oel and Hoek-stra(2012) for themain pulpproducing countries. Ourestimates

ofactual evaporationrates areabout30%higher,whileourwood yields are about45% lower.Wespecifically estimatethe evapora-tion of forests, while Van Oel and Hoekstra (2012) used a gen-eral actual evaporationmap (which probablyunderestimates for-est evaporation). Where Van Oel and Hoekstra (2012) use rough woodyieldestimatespercountry/region,woodyieldsinourstudy arederived fromnationalproductionandarea statisticsthatwere downscaledtothegridlevel.Moreover, weusedifferent underly-ing mapsofwhichgridcellsare usedforroundwoodproduction, whichcontributesto differentspatialaverage estimatesof evapo-rationratesandwaterfootprints.Withoutapplicationofthevalue fraction of roundwoodproduction, our waterfootprint of round-woodestimatesforthemainpulpproducingcountriesare signifi-cantlyhigherthanthoseby VanOelandHoekstra(2012).After ap-plyingthevaluefractions(fvalue,rw),ourestimatesareroughly20%

and140%ofthoseby VanOelandHoekstra(2012)fortropicaland temperate/boreal zones, respectively. We used the same woodto paperconversionfactoras VanOel andHoekstra(2012),so differ-encesinthewaterfootprintofpaper(assumingarecoveryrateof zero)arealsoexplainedbytheabove.

When we compare the water footprint of seven wood prod-ucts inChina, we find that our estimatesare 5–29% of those by TianandKe(2012).Weused differentmethodsanddata,butthe largestdifferenceisprobablyexplainedbythefactthatweapplya valuefraction.

Forthesoutheastern UnitedStates, ChiuandWu(2013) found thatthegreenwaterfootprintofethanolfromforestwoodresidue is about 400–443 l/l and that the blue water footprint in the forestry stage isminimal. Ourestimated waterfootprint per unit ofroundwoodinthisregionisabout70l/l(Fig.4).Witha round-woodtobio-ethanolconversionfactorof6.8fortheUnitedStates (UNECE/FAO,2010),thistranslatesintoaquite similarwater foot-print of 476 l/l. Where we applied a value fraction to attribute forest evaporationtoroundwoodproductionfollowedbya round-wood to bio-ethanolconversion factor, Chiu andWu (2013) allo-catedforestevaporationtobio-ethanolproductionbasedonan es-timatedweightfractionofharvestedwoodresidueforbio-ethanol inthe totalabove-groundwoodmass,which alsogreatlyreduces theamountofevaporationattributedtothebio-ethanol.

4.2. Uncertaintiesregardingmethodanddata

4.2.1. Moisturerecycling

Precipitation over land relies onterrestrial evaporation (mois-turerecycling)toa varyingextentaround theglobe(Vander Ent etal.,2010)andforestsplayanimportantroleinthis(Ellisonetal., 2012). When attributing forest evaporation to forestry products, one could argue to reduce total forest evaporation by the por-tionofevaporationthatreturnsasprecipitation(inthesamearea), basedontheideathatthisreturningwatercanbeusedagainand thereforeisnotreallyconsumed(Launiainenetal.,2014).However, greenforestevaporationstemsfromtheprecipitationamountthat already includes the recycled moisture. Reducing the attributed evaporationby the recycledpartwouldwronglysuggest that the recycled water is left for use for other purposes. It is not addi-tional water that can be additionally allocated.As mentioned in theintroduction,we areinterestedinthisquestionofwater allo-cation: whichpartofthe available flowis beingappropriatedfor roundwood production? Therefore, we deliberately attribute the totalforestevaporation(thatisreducedbasedonavaluefraction) toroundwoodproduction,whateverrateofmoisturerecycling.

4.2.2. Uncertaintiesregardingdata

The estimatesofthewaterfootprint ofroundwoodproduction provided in thisstudy are subjectto a number of uncertainties. Sincethefractionofwaterintheharvestedwoodturnedouttobe

negligible(Section3.1),themainvariablesgoverningtheendresult aretheforestevaporation(Eact),theareausedforroundwood

pro-duction(Arw), the volume ofroundwood produced (Pact) andthe

valuefractionofroundwoodproduction(f_value,rw).

Out ofthese, we expect the leastuncertainty inEact andPact.

The estimate of Eact is relatively straightforward and bound by

annual precipitation and potential evaporation. Pact is based on

downscaled national statistics covering the entire study period, although the downscaling to the grid level involved coarse data onlong-termmaximumsustainablewoodyields.Thecurrentdata limitations regarding Arw (Kuemmerle et al., 2013) makes our

estimate of Arw rather uncertain, since it is based on a

mod-eled wood harvest patternthat wasscaled to an estimatedarea usedforroundwoodproductionbasedon nationalstatistics avail-able from1990 onwards. The estimated relative value of ecosys-tem services from which we derived fvalue,rw is associated with

somelimitationsaselaboratelydescribedby Costanzaetal.(1997) and Costanza et al. (2014). The estimates are based on a lim-ited number of valuation studies that reflect the state at a cer-tain point in time (Costanza et al., 2014). Besides, uncertainties are associated withwillingness-to-pay estimates and aggregation ofvaluesatspecificlocationstolargerspatialandtemporalscales (Costanzaetal.,2014).Furthermore,weneededtomakeanumber ofassumptionsfordisaggregatingthevalueofecosystem services intimeandspaceasoutlinedintheSI.

4.3.Sustainabilityofthewaterfootprint

Thisstudyhasprovided spatially-explicitestimatesofthe wa-terfootprintofroundwoodproductionandvariousforestproducts. Oneshouldbecautiousinevaluatingdifferencesinthewater foot-printsofa similarproductfromtwodifferentregions intermsof betterorworse.Therelevanceofthedatapresentedratherliesin thefactthattheycanformabasisforfurtherstudyintothe alter-nativeusesofthesamewatertoproducemoreordifferentgoods andservices.

Tojudgethesustainabilityofthewaterfootprintofroundwood production(volume/time),onewouldneedtoplacethegreenand blue water components in the context of maximum sustainable levels of green and blue water consumption and consider the competition for the limited green and blue water resources be-tween different demands. This assessment was out of the scope ofthisstudy,sincemaximum sustainablelevels arecurrentlynot knownfor greenwater (Schyns etal., 2015; Hoekstra and Wied-mann, 2014), the major component of the water footprint of roundwoodproduction.Besides,forunderstanding competing de-mandsforwaterandthepotential conflictbetween(green)water useforroundwoodproductionand(green)wateruseforother pur-poseslikecropsforfood,feedorbioenergy,abroaderstudywould berequired.Nevertheless,wecanroughlycontextualizethewater footprintofroundwoodproductionbasedonpreviouswork.

4.3.1. Additionoftheforestrysectortothewaterfootprintof

humanity

We can place the global water consumption attributed to roundwoodproductioninthecontextoftheglobalwaterfootprint fortheperiod1996–2005asestimatedby HoekstraandMekonnen (2012),whoconsideredthefollowingfivesectors:cropproduction, pasture,watersupplyinanimalraising,industrialproduction,and domestic water supply. Additionof the forestry sector raises the globalconsumptive(greenplusblue)waterfootprintofproduction fortheperiod1996–2005by12%.

4.3.2. Trade-offsbetweenwaterforfood,feed,energyandwood

Theestimatedwaterfootprintsofroundwoodrepresentthe vol-umeofwaterthatisallocated towoodproduction,albeit

implic-itlythroughland-usedecisions(RockströmandGordon,2001). Al-ternatively,this watercould be used for the generation of other terrestrialecosystemservicesorcropproduction(Rockströmetal., 1999; Rockströmand Gordon, 2001). We made a rough compar-ison between the value of water for roundwood and three ma-jor food/feed crops (Table 5) as well as the water footprint of bio-ethanolper unit ofenergy fromthesefoursources (Table6). Bothregarding economic water productivity and the water foot-printof bio-ethanol,roundwoodis comparablewithmaize, rank-ingsomewhat better comparedtowheat andworse comparedto sugarbeet.Itshould benotedthat thewaterfootprintof second-generationbio-ethanolobtainedfromcropresiduesissmallerthan thewaterfootprintoffirst-generationbio-ethanolfromthesecrops (Mathioudakisetal.,2017).

Mekonnen et al.(2015) compared the water footprint ofheat fromvarious energysources, including that from firewoodbased onVanOelandHoekstra(2012).Althoughourestimatesofthe wa-ter footprintof heatfrom wood(i.e.firewood, pellets, briquettes, bark,chips,charcoal)aredifferent(Section4.1),theyremainorders of magnitude larger than the water footprint from other energy sourcessuchascoal,lignite,oil,gasandnuclear(Mekonnenetal., 2015). From thisperspective, burningwoodforthe generation of heat and electricity still is not recommended (Mekonnen et al., 2015).

4.4.Reductionofthewaterfootprint

4.4.1. Intensificationvs.extensificationofwoodproduction

Intensification of wood production has two counteracting ef-fects on the water footprint per unit of roundwood produced (WFrw, Eq.7). Effect A is that the value of wood production

in-creases, partially at the expense of other ecosystem service val-ues(fvalue,rwincreases),suchthatthewaterconsumptionattributed

toroundwoodproductionincreases.Effect Bisthatmorewoodis produced per ha with the same amount ofwater. Intensification ofwoodproductioncan onlyreduce WFrw ifthe additionalwood

valueperha(effectB)outweighsthelossofvalueofother ecosys-temservices(effectA).

Therelationshipbetweenfvalue,rwandtheintensityofforest

ex-ploitation(seeSI)determineswhethereffectAisstrongerthan ef-fectBorviceversaandhencewhetherWFrw increases(when

ef-fectA>effectB)ordecreases(wheneffectA<effectB)with in-tensifiedproduction.Thisrelationship isdifferent in(sub)tropical forestscomparedtotemperate/borealforests,andfurthermore de-pends on the long-term maximum sustainable yield (Ysus): the

higher Ysus the larger the theoretical potential to obtain a high

valueofwoodproductionfromtheforest.

For (sub)tropical forestswe found that intensificationleadsto an increase in WFrw for Ysus <25m3/ha (which is always the

casein our study; see Table 1). For temperate/boreal forests we found that intensification results in an increase in WFrw for Ysus

<4.5m3_{/ha, buta decreasein WF}

rw forhigher Ysus.Althoughwe

recognizethatfurtherresearchisneededintothevalueofforests andtheir maximum sustainableyields – withmore spatiotempo-raldetailthanwasavailableforthisstudy– thefollowinggeneral ruleseemstoapply:inforestswitharelativelyhighYsus,

intensifi-cationcanbebeneficialintermsofwateruseefficiency,sincethe positiveeffectofintensification(effectB)canoutweighthelossof valueofotherecosystemservices(effectA).

4.4.2. Recycling

The water footprint of roundwood can effectively be reduced through recycling. The use of recycled wood nullifies the at-tributedevaporationtoroundwoodproduction,sincenonewwood is produced. In this study, recovery rates were not considered. Hence,waterfootprintestimatesrefertonewlyproducedproducts.

Ta b le 5 Economic wa te r pr oducti vity (EWP) of r o undw ood (r w) compar e d to thr e e ma jor f ood/f ee d cr ops. EWP is calculat e d as the price di vide d by the gr een plus blue wa te r fo o tp ri n t (WF). Global av er ag e WF of cr ops obt a ine d fr om Mek o nnen and Hoek st ra (20 11 ) and ra n g e s fr om Mek o nnen and Hoek st ra (20 14 ). Alt e rnati v e uses Price a WF min b (m 3 /unit of pr oduct) WF avg (m 3 /unit of pr oduct) WF max c (m 3 /unit of pr oduct) EWP max US$/m 3 ) EWP avg (US$/m 3 ) EWP min (US$/m 3 ) R o undw ood 94 $/m 3 rw 68 m 3 /m 3 rw 293 m 3 /m 3 rw 584 m 3 /m 3 rw 1 .4 0 .3 0 .2 Wheat 289 $/t 992 m 3 /t 1 620 m 3 /t 209 1 m 3 /t 0 .3 0 .2 0 .1 Maize 34 9 $/t 542 m 3 /t 10 2 8 m 3 /t 1 385 m 3 /t 0 .6 0 .3 0 .3 Sug a r be e t 81 $/t 58 m 3 /t 10 8 m 3 /t 151 m 3 /t 1 .4 0 .8 0 .5 a Price fo r ro undw ood is the a v er ag e ex p o rt unit price in UNECE countries fo r the period 20 05–20 14 obt a ine d fr om ( UNECE/F A O, 20 1 5 ). Prices fo r cr ops ar e av er ag e pr oducer prices fo r the period 20 05–20 14 obt a ine d fr om ( FA O S T A T, 20 1 6 b). b WF at 20th per centile of pr oduction. c WF at 80th per centile of pr oduction.

Ta b le 6 Wa te r fo o tp ri n t (WF) of bio-e thanol fr om r o undw ood (r w) compar e d to the WF of firs t-g ener ation bio-e thanol fr om thr e e cr ops. Global a v er ag e WF of cr ops obt a ine d fr om ( Mek o nnen and Hoek st ra , 20 11 ) and ra n g e s fr om ( Mek o nnen and Hoek st ra , 20 14 ). Wa te r fo o tp ri n ts re fe r to the wa te r use in the biomass pr oduction s tag e (cr o p and w ood gr o w th). Bio-e thanol fr om Ener gy yield a WF min b (m 3 /unit of pr oduct) WF avg (m 3 /unit of pr oduct) WF max c (m 3 /unit of pr oduct) WF min (m 3 /GJ) WF avg (m 3 /GJ) WF max (m 3 /GJ) R o undw ood 3 .0 GJ/m 3 rw 68 m 3 /m 3 rw 293 m 3 /m 3 rw 584 m 3 /m 3 rw 23 98 19 5 Wheat 10 .2 GJ/t 992 m 3 /t 1 620 m 3 /t 209 1 m 3 /t 98 15 9 20 6 Maize 10 .0 GJ/t 542 m 3 /t 10 2 8 m 3 /t 1 385 m 3 /t 54 10 3 13 9 Sug a r be e t 2.6 GJ/t 58 m 3 /t 10 8 m 3 /t 151 m 3 /t 22 41 58 a Ener gy yield of r o undw ood is the in v e rse of the con v ersion fa ct o r fo r w ood-base d e thanol in Ta b le 4 . Ener gy yield of cr ops obt a ine d fr om ( Mek o nnen and Hoek st ra , 20 11 ). b WF at 20th per centile of pr oduction. c WF at 80th per centile of pr oduction.

Van Oel and Hoekstra, (2012) already concluded that increasing paperrecovery ratesisa powerfulwayto reduce thewater foot-printofpaper.Otherwoodproductscanalsoberecycledinvarious ways.Wooden pallets or furniturecan be reused or be remanu-factured fromrecovered wood,just like particle board (Falk and McKeever, 2004). In construction, wood recovered during demo-lition is potentially suitable for reuse or remanufacture, particu-larly into flooring (Falk and McKeever, 2004). Chipped or shred-dedwoodcan beused asbasis forfuel,landscapingmulch, com-postingbulkagent,sewagesludgebulkingmedium,oranimal bed-ding (Falk and McKeever, 2004). Ideally, the cascading use prin-ciple is applied, in which wood is used, recycled and reused as longaspossiblebeforeultimatelybeingusedasanenergysource (Dammer etal., 2016). It isobviousthat reducedconsumption of end products from wood will eventually reduce the total water consumptionrelatedtoroundwoodproduction.

5. Conclusion

The global water consumption attributed to roundwood pro-ductionforlumber,pulp,paper,fuel andfirewood hasrisenfrom 768 × 109 _m3_{/y in 1961–1970to 961 × 10}9_m3_{/y in 2001–2010.}

Recyclingof woodproducts could effectivelyreduce this volume, thereby leavingmore water available for thegeneration of other ecosystemservices.Intensificationofwoodproductioncanonly re-ducethewaterfootprint perunit ofwoodiftheadditionalwood value per ha outweighsthe lossof value ofother ecosystem ser-vices,whichisoftennotthecasein(sub)tropicalforests. Alterna-tively usingthe waterfor crop productionis generallynot bene-ficial(even apartfromthenegativeeffectsofconvertingforest to cropland),sinceroundwoodisrathercomparablewithmajorfood, feedandenergycropsintermsofeconomicwaterproductivityand energyyieldfrombio-ethanolperunitofwater.Theresultsofthis studycontributetoamorecompletepictureofthehuman appro-priationofwaterandfeedintothedebateonwaterforfood,feed, energyandwood.

Acknowledgments

Thepresentwork was(partially) developedwithin the frame-workofthePantaRheiResearchInitiativeoftheInternational As-sociationofHydrologicalSciences(IAHS)andhasbeenmade pos-sible by grants from the Water Footprint Network and Deltares. We thank the reviewers of an earlier version of this manuscript fortheirconstructivecomments.

Supplementarymaterials

Supplementary material associated with this article can be found,intheonlineversion,at doi:10.1016/j.advwatres.2017.05.013. References

Bauhus, J, Puettmann, K, Messier, C, 2009. Silviculture for old-growth attributes. For- est Ecol. Manag. 258, 525–537. http://dx.doi.org/10.1016/j.foreco.2009.01.053 . Blanchez, J. , 1997. Forest resources and roundwood supply in the Asia Pacific coun-

tries: situation and outlook to the year 2010. Asia-Pacific Forestry Towards 2010. Asia-Pacific Forestry Sector Outlook Study Working Paper Series (FAO) . Brown, C. , 20 0 0. The Global Outlook for Future Wood Supply From Forest Planta-

tions. FAO, Forestry Policy and Planning Division .

Canadell, J, Jackson, RB, Ehleringer, JB, Mooney, HA, Sala, OE, Schulze, E-D, 1996. Maximum rooting depth of vegetation types at the global scale. Oecologia 108, 583–595. http://dx.doi.org/10.10 07/bf0 0329030 .

Chave, J, Coomes, D, Jansen, S, Lewis, SL, Swenson, NG, Zanne, AE, 2009. Towards a worldwide wood economics spectrum. Ecol. Lett. 12, 351–366. http://dx.doi.org/

10.1111/j.1461-0248.2009.01285.x .

Chini, LP , Hurtt, GC , Frolking, S , 2014. Harmonized Global Land Use for Years 1500 -2100, V1. ORNL Distributed Active Archive Center .

Chiu, Y-W , Wu, M , 2013. The water footprint of biofuel produced from forest wood residue via a mixed alcohol gasification process. Environ. Res. Lett. 8, 035015 .

Costanza, R , d’Arge, R , de Groot, R , Farber, S , Grasso, M , Hannon, B , et al. , 1997. The value of the world’s ecosystem services and natural capital. Nature 387, 253–260 .

Costanza, R, de Groot, R, Sutton, P, van der Ploeg, S, Anderson, SJ, Kubiszewski, I, et al., 2014. Changes in the global value of ecosystem services. Global Environ. Change 26, 152–158. http://dx.doi.org/10.1016/j.gloenvcha.2014.04.002 . Döll, P, Hoffmann-Dobrev, H, Portmann, FT, Siebert, S, Eicker, A, Rodell, M, et al.,

2012. Impact of water withdrawals from groundwater and surface water on con- tinental water storage variations. J. Geodyn. 59–60, 143–156. http://dx.doi.org/

10.1016/j.jog.2011.05.001 .

Dammer, L , Bowyer, C , Breitmayer, E , Eder, A , Nanni, S , Allen, B , et al. , 2016. Mapping Study on Cascading Use of Wood Products. WWF, Gland, Switzerland . de Graaf, IEM, van Beek, LPH, Wada, Y, Bierkens, MFP, 2014. Dynamic attribution

of global water demand to surface water and groundwater resources: effects of abstractions and return flows on river discharges. Adv. Water Resour. 64, 21–33.

http://dx.doi.org/10.1016/j.advwatres.2013.12.002 .

Ellison, D, Futter, MN, Bishop, K, 2012. On the forest cover–water yield debate: from demand- to supply-side thinking. Global Change Biol. 18, 806–820. http://dx.doi.

org/10.1111/j.1365-2486.2011.02589.x .

Falk, RF , McKeever, DB , 2004. Recovering wood for reuse and recycling: a United States perspective. Management of Recovered Wood Recycling, Bioenergy and Other Options. European COST E31 Conference. Thessaloniki .

Fan, Y, Li, H, Miguez-Macho, G, 2013. Global patterns of groundwater table depth. Science 339, 940–943. http://dx.doi.org/10.1126/science.1229881 .

FAOSTAT, Forestry Production, 2016a, Rome, Italy, http://faostat3.fao.org (accessed 29.07.16).

FAOSTAT, Producer prices - annual 2016b, Rome, Italy, http://faostat3.fao.org/

download/P/PP/E (accessed 11.10.16).

Hanasaki, N, Inuzuka, T, Kanae, S, Oki, T, 2010. An estimation of global virtual water flow and sources of water withdrawal for major crops and livestock products using a global hydrological model. J. Hydrol. 384, 232–244. http://dx.doi.org/10.

1016/j.jhydrol.2009.09.028 .

Hoekstra, AY, Mekonnen, MM, 2012. The water footprint of humanity. Proc. Natl. Acad. Sci. 109, 3232–3237. http://dx.doi.org/10.1073/pnas.1109936109 . Hoekstra, AY, Wiedmann, TO, 2014. Humanity’s unsustainable environmental foot-

print. Science 344, 1114–1117. http://dx.doi.org/10.1126/science.1248365 . Hoekstra, AY , Chapagain, AK , Aldaya, MM , Mekonnen, MM , 2011. The Water Foot-

print Assessment manual: Setting the Global Standard. Earthscan, London . Hoekstra, AY. , 2013. The Water Footprint of Modern Consumer Society. Routledge,

London, UK .

Hoekstra, AY., 2017. Water Footprint Assessment: Evolvement of a New Research Field. Water Resour. Manag. http://dx.doi.org/10.1007/s11269- 017- 1618- 5 . Hurtt, GC, Chini, LP, Frolking, S, Betts, RA, Feddema, J, Fischer, G, et al., 2011. Har-

monization of land-use scenarios for the period 1500–2100: 600 years of global gridded annual land-use transitions, wood harvest, and resulting secondary lands. Clim. Change 109, 117–161. http://dx.doi.org/10.1007/s10584- 011- 0153- 2 . IUCN and UNEP-WCMC, 2016. The World Database on Protected Areas (WDPA). UNEP-WCMC, Cambridge, UK. Available at: www.protectedplanet.net (accessed 05.16).

Jürgensen C, W Kollert, A Lebedys. Assessment of Industrial Roundwood Produc- tion from Planted Forests. Planted Forests and Trees Working Paper Series, No. FP/48/E. The Food and Agricultural Organization of the United Nations (FAO). Rome 2014.

Köhl, M, Lasco, R, Cifuentes, M, Jonsson, Ö, Korhonen, KT, Mundhenk, P, et al., 2015. Changes in forest production, biomass and carbon: Results from the 2015 UN FAO global forest resource assessment. Forest Ecol. Manag. 352, 21–34. http:

//dx.doi.org/10.1016/j.foreco.2015.05.036 .

Komatsu, H, Cho, J, Matsumoto, K, Otsuki, K, 2012. Simple modeling of the global variation in annual forest evapotranspiration. J. Hydrol. 420–421, 380–390. http:

//dx.doi.org/10.1016/j.jhydrol.2011.12.030 .

Kuemmerle, T, Erb, K, Meyfroidt, P, Müller, D, Verburg, PH, Estel, S, et al., 2013. Chal- lenges and opportunities in mapping land use intensity globally. Curr. Opin. En- viron. Sustain. 5, 4 84–4 93. http://dx.doi.org/10.1016/j.cosust.2013.06.002 . Launiainen, S, Futter, MN, Ellison, D, Clarke, N, Finér, L, Högbom, L, et al.,

2014. Is the water footprint an appropriate tool for forestry and forest prod- ucts: the fennoscandian case. Ambio 43, 244–256. http://dx.doi.org/10.1007/

s13280-013-0380-z .

Liu, J, Yang, H, 2010. Spatially explicit assessment of global consumptive water uses in cropland: green and blue water. J. Hydrol. 384, 187–197. http://dx.doi.org/10.

1016/j.jhydrol.2009.11.024 .

Liu, J, Zehnder, AJB, Yang, H, 2009. Global consumptive water use for crop produc- tion: the importance of green water and virtual water. Water Resour. Res. 45, W05428. http://dx.doi.org/10.1029/20 07wr0 06051 .

Loveland, TR , Brown, J , Ohlen, D , Reed, B , Zhu, Z , Yang, L , et al. , 2009. ISLSCP II IGBP DISCover and SiB Land Cover, 1992-1993. ORNL Distributed Active Archive Center .

Lutz, J , 2011. How Trees Grow. Forest Research Notes 8 (2) .

Mathioudakis, V, Gerbens-Leenes, PW, Van der Meer, TH, Hoekstra, AY, 2017. The water footprint of second-generation bioenergy: a comparison of biomass feed- stocks and conversion techniques. J. Clean. Prod. 148, 571–582. http://dx.doi.org/

10.1016/j.jclepro.2017.02.032 .

Matthews, E , Payne, R , Rohweder, M , Murray, S , 20 0 0. Pilot Analysis of Global Ecosystems: Forest Ecosystems. World Resources Institute .

Mekonnen, MM, Hoekstra, AY, 2011. The green, blue and grey water footprint of crops and derived crop products. Hydrol. Earth Syst. Sci. 15, 1577–1600. http://

dx.doi.org/10.5194/hess-15-1577-2011 .

Mekonnen, MM, Hoekstra, AY, 2014. Water footprint benchmarks for crop produc- tion: a first global assessment. Ecol. Indic. 46, 214–223. http://dx.doi.org/10.

1016/j.ecolind.2014.06.013 .

Mekonnen, MM, Gerbens-Leenes, PW, Hoekstra, AY, 2015. The consumptive water footprint of electricity and heat: a global assessment. Environ. Sci.: Water Res. Technol. 1, 285–297. http://dx.doi.org/10.1039/C5EW0 0 026B .

Miller, GR, Chen, X, Rubin, Y, Ma, S, Baldocchi, DD, 2010. Groundwater uptake by woody vegetation in a semiarid oak savanna. Water Resour. Res. 46, W10503.

http://dx.doi.org/10.1029/20 09WR0 08902 , n/a-n/a.

New, M , Lister, D , Hulme, M , Makin, I , 2002. A high-resolution data set of surface climate over global land areas. Clim. Res. 21, 1–25 .

Oki, T, Kanae, S, 2006. Global hydrological cycles and world water resources. Science 313, 1068–1072. http://dx.doi.org/10.1126/science.1128845 .

Postel, SL, Daily, GC, Ehrlich, PR, 1996. Human appropriation of renewable fresh wa- ter. Science 271, 785–788. http://dx.doi.org/10.1126/science.271.5250.785 . Rockström, J, Gordon, L, 2001. Assessment of green water flows to sustain major

biomes of the world: implications for future ecohydrological landscape man- agement.. Phys. Chem. Earth, Part B: Hydrol., Oceans Atmos. 26, 843–851. http:

//dx.doi.org/10.1016/S1464-1909(01)0 0 096-X .

Rockström, J , Gordon, L , Folke, C , Falkenmark, M , Engwall, M , 1999. Linkages among water vapor flows, food production, and terrestrial ecosystem services. Conserv. Ecol 3, 5 .

Rockstrom, J , Steffen, W , Noone, K , Persson, A , Chapin, FS , Lambin, EF , et al. , 2009. A safe operating space for humanity. Nature 461, 472–475 .

Rost, S, Gerten, D, Bondeau, A, Lucht, W, Rohwer, J, Schaphoff, S, 2008. Agricultural green and blue water consumption and its influence on the global water system. Water Resour. Res. 44, W09405. http://dx.doi.org/10.1029/20 07wr0 06331 . Savenije, HHG., 2004. The importance of interception and why we should delete the

term evapotranspiration from our vocabulary. Hydrol. Process. 18, 1507–1511.

http://dx.doi.org/10.1002/hyp.5563 .

Schyns, JF, Hoekstra, AY, Booij, MJ, 2015. Review and classification of indicators of green water availability and scarcity. Hydrol. Earth Syst. Sci 19, 4581–4608.

http://dx.doi.org/10.5194/hess-19-4581-2015 .

Siebert, S, Döll, P, 2010. Quantifying blue and green virtual water contents in global crop production as well as potential production losses without irrigation. J. Hy- drol. 384, 198–217. http://dx.doi.org/10.1016/j.jhydro1.2009.07.031 .

Simpson, WT. , 1998. Equilibrium Moisture Content of Wood in Outdoor Locations in the United States and Worldwide. US Dept. of Agriculture, Forest Service, Forest Products Laboratory .

Speight, JG. , 2010. The Biofuels Handbook. Royal Society of Chemistry .

Steffen, W , Richardson, K , Rockström, J , Cornell, SE , Fetzer, I , Bennett, EM , et al. , 2015. Planetary boundaries: guiding human development on a changing planet. Science 347 .

Tian, MH, Ke, SF, 2012. Virtual water content and trade analysis of primary woody products in China. Int. Forest. Rev. 14, 380–390. http://dx.doi.org/10.

1505/146554812802646639 .

UNECE/FAO, 2010. Forest Product Conversion Factors for the UNECE Region. Geneva Timber and Forest Discussion Paper 49. UNECE/FAO, Geneva .

UNECE/FAO, TIMBER database: Export and Import Unit Prices (1964-2014), 2015, http://www.unece.org/fileadmin/DAM/timber/statsdata/unit- price- 1964- 2014.

xlsx (accessed 10.10.16).

Van der Ent, RJ, Savenije, HHG, Schaefli, B, Steele-Dunne, SC, 2010. Origin and fate of atmospheric moisture over continents. Water Resour. Res. 46, W09525. http://

dx.doi.org/10.1029/2010WR009127 , n/a-n/a.

Van Oel, PR, Hoekstra, AY, 2012. Towards quantification of the water footprint of paper: a first estimate of its consumptive component. Water Resour. Manag. 26, 733–749. http://dx.doi.org/10.1007/s11269-011-9942-7 .

Van Velthuizen, H , Huddleston, B , Fischer, G , Salvatore, M , Ataman, E , Nachter- gaele, F , et al. , 2007. Mapping biophysical factors that influence agricultural pro- duction and rural vulnerability. Environment and natural resources series 11. FAO & IIASA, Rome .

Vanham, D, 2016. Does the water footprint concept provide relevant information to address the water–food–energy–ecosystem nexus? Ecosyst. Serv. 17, 298–307.

http://dx.doi.org/10.1016/j.ecoser.2015.08.003 .

Wada, Y, Wisser, D, Bierkens, MFP, 2014. Global modeling of withdrawal, allocation and consumptive use of surface water and groundwater resources. Earth Syst. Dyn. 5, 15–40. http://dx.doi.org/10.5194/esd- 5- 15- 2014 .

Zanne, AE , Lopez-Gonzalez, G , Coomes, DA , Ilic, J , Jansen, S , Lewis, SL , et al. , 2009. Data from: towards a worldwide wood economics spectrum. Dryad Data Repos. . Zhang, L , Dawes, W , Walker, G , 1999. Predicting the effect of vegetation changes on catchment average water balance. Technical Report 99/12. Cooperative Research Centre for Catchment Hydrology, CSIRO Land and Water .

Zhang, L, Dawes, WR, Walker, GR, 2001. Response of mean annual evapotranspira- tion to vegetation changes at catchment scale. Water Resour. Res. 37, 701–708.