New LPJmL4 – a dynamic global vegetation model with managed … · 2020. 6. 23. · LPJmL4 – a dynamic global vegetation model with managed land ... DGVMs were further developed

Geosci Model Dev 11 1343ndash1375 2018httpsdoiorg105194gmd-11-1343-2018copy Author(s) 2018 This work is distributed underthe Creative Commons Attribution 40 License

LPJmL4 ndash a dynamic global vegetation model with managed land ndashPart 1 Model descriptionSibyll Schaphoff1 Werner von Bloh1 Anja Rammig2 Kirsten Thonicke1 Hester Biemans3 Matthias Forkel4Dieter Gerten15 Jens Heinke1 Jonas Jaumlgermeyr1 Juumlrgen Knauer6 Fanny Langerwisch1 Wolfgang Lucht15Christoph Muumlller1 Susanne Rolinski1 and Katharina Waha17

1Potsdam Institute for Climate Impact Research PO Box 60 12 03 14412 Potsdam Germany2Technical University of Munich School of Life Sciences Weihenstephan 85354 Freising Germany3Alterra Wageningen University amp Research PO Box 47 6700 AA Wageningen the Netherlands4TU Wien Climate and Environmental Remote Sensing Group Department of Geodesy and GeoinformationGusshausstraszlige 25ndash29 1040 Vienna Austria5Humboldt Universitaumlt zu Berlin Department of Geography Unter den Linden 6 10099 Berlin Germany6Max Planck Institute for Biogeochemistry Hans-Knoumlll-Str 10 07745 Jena Germany7CSIRO Agriculture amp Food 306 Carmody Rd St Lucia QLD 4067 Australia

Correspondence Sibyll Schaphoff (sibyllschaphoffpik-potsdamde)

Received 21 June 2017 ndash Discussion started 27 July 2017Revised 26 February 2018 ndash Accepted 5 March 2018 ndash Published 12 April 2018

Abstract This paper provides a comprehensive descriptionof the newest version of the Dynamic Global VegetationModel with managed Land LPJmL4 This model simulatesndash internally consistently ndash the growth and productivity ofboth natural and agricultural vegetation as coherently linkedthrough their water carbon and energy fluxes These fea-tures render LPJmL4 suitable for assessing a broad rangeof feedbacks within and impacts upon the terrestrial bio-sphere as increasingly shaped by human activities such as cli-mate change and land use change Here we describe the coremodel structure including recently developed modules nowunified in LPJmL4 Thereby we also review LPJmL modeldevelopments and evaluations in the field of permafrost hu-man and ecological water demand and improved representa-tion of crop types We summarize and discuss LPJmL modelapplications dealing with the impacts of historical and futureenvironmental change on the terrestrial biosphere at regionaland global scale and provide a comprehensive overview ofLPJmL publications since the first model description in 2007To demonstrate the main features of the LPJmL4 modelwe display reference simulation results for key processessuch as the current global distribution of natural and man-aged ecosystems their productivities and associated waterfluxes A thorough evaluation of the model is provided in a

companion paper By making the model source code freelyavailable at httpsgitlabpik-potsdamdelpjmlLPJmL wehope to stimulate the application and further development ofLPJmL4 across scientific communities in support of majoractivities such as the IPCC and SDG process

1 Introduction

The terrestrial biosphere a highly dynamic key componentof the Earth system is undergoing significant and widespreadtransformations induced by human activities such as climateand land use change Humans have by now transformedabout 40 of the terrestrial ice-free land surface into landused for agriculture and urban settlements (Ellis et al 2010)thus pushing the planetary dynamics beyond the boundariesthat have been characteristic for the past ca 12 000 years(Rockstroumlm et al 2009) These interventions put at risk im-portant functions of the biosphere such as the provisioning offloral and faunal biodiversity (Voumlroumlsmarty et al 2010) theterrestrial carbon sink (Le Queacutereacute et al 2015) and the provi-sioning of accessible fresh water (Voumlroumlsmarty et al 2010)Understanding and modelling the current and potential fu-ture dynamics of the Earth system thus renders it necessary

Published by Copernicus Publications on behalf of the European Geosciences Union

1344 S Schaphoff et al LPJmL4 ndash Part 1 Model description

to consider human activities as an integral part while rep-resenting the major dynamics of the biosphere in a spatio-temporally explicit and process-based manner accountingfor the feedbacks between vegetation global carbon and wa-ter cycling and the atmosphere This would also allow forthe numerical evaluation of potential implementation path-ways for the United Nations Sustainable Development Goals(SDGs httpssustainabledevelopmentunorg) and their im-pacts on the terrestrial environment complementing the im-portant role that dynamic biosphere models have played inthe United Nations scientific assessment reports on climatechange published by the United Nations IntergovernmentalPanel on Climate Change (IPCC 2014) By combining corefeatures of global biogeographical and biogeochemical mod-els developed in the 1990s dynamic global vegetation mod-els (DGVMs) emerged as the main tool to simulate the pro-cesses underlying the dynamics of natural vegetation types(growth mortality resource competition and disturbancessuch as wildfires) and the associated carbon and water fluxes(Cramer et al 2001 Prentice et al 2007 Sitch et al 2008Friend et al 2014) In light of strengthening human inter-ferences DGVMs were further developed to integrate ad-ditional processes that are relevant to the original researchquest of studying biogeography and biogeochemical cyclesunder climate change (Canadell et al 2007) This includesthe incorporation of human land use and the simulation ofagricultural production systems (Bondeau et al 2007 Lin-deskog et al 2013) nutrient limitation (Zaehle et al 2010Smith et al 2014) hydrological modules and river-routingschemes (Gerten et al 2004 Rost et al 2008) Knowledgederived from models that are designed to cover aspects ofthe Earth system other than terrestrial vegetation and the car-bon cycle such as models of the global water balance couldevidentially improve the DGVMsrsquo ability to also evaluatemodel performance for processes (eg river discharge) thatare closely connected to simulated vegetation and carbon cy-cle dynamics (Bondeau et al 2007 Smith et al 2014) Thedevelopment towards more comprehensive models of Earthrsquosland surface offers new possibilities for cross-disciplinary re-search DGVMs as land components of Earth system mod-els still show large uncertainties about the terrestrial carbon(C) balance under future climate change (Friedlingstein et al2013) This uncertainty partly results from differences in thesimulation of soil and vegetation C residence times (Carval-hais et al 2014 Friend et al 2014) The time that C residesin an ecosystem is thereby strongly affected by simulatedprocesses of vegetation dynamics (Ahlstroumlm et al 2015)These examples highlight the need to continuously improveprocess representations in DGVMs in order to reduce the un-certainty in projected ecosystem functioning and services un-der future climate change This requires however that modeldevelopments in specific fields or improvements for certainprocesses are synthesized and integrated into a unified in-ternally consistent model version The LundndashPotsdamndashJenaDGVM with managed Land (LPJmL Bondeau et al 2007)

originates from a former version of the model described bySitch et al (2003) and simulates the growth and geographi-cal distribution of natural plant functional types (PFTs) cropfunctional types (CFTs) and the associated biogeochemicalprocesses (mainly carbon cycling) Recent developments fo-cused on an improved energy balance model able to estimatepermafrost dynamics based on a vertical soil carbon distribu-tion scheme and a new soil hydrological scheme (Schaphoffet al 2013) Also a new process-based fire module (SPIT-FIRE) was implemented that allows for detailed simulationof fire ignition spread and effects to estimate fire impactsand emissions (Thonicke et al 2010) An updated phenol-ogy scheme was developed which now takes phenology lim-itations arising from low temperatures limited light anddrought into account (Forkel et al 2014) Further modeldevelopments encompass the parallelization of the modelto efficiently simulate river routing (Von Bloh et al 2010)and the implementation of an irrigation scheme (Rost et al2008) recently updated with a mechanistic representation ofthe three major irrigation systems (Jaumlgermeyr et al 2015)Biemans et al (2011) implemented reservoir operations andirrigation extraction and evaluated the impact on river dis-charge Other developments focused on a newly formulatedimplementation of different cropping systems in sub-SaharanAfrica (Waha et al 2013) Mediterranean agricultural planttypes (Fader et al 2015) and bioenergy crops such as sugar-cane (Lapola et al 2009) fast-growing grasses and bioen-ergy trees (Beringer et al 2011) With these implementa-tions the potential of bioenergy production under future landuse population and climate development could be exten-sively investigated (Haberl et al 2011 Popp et al 2011Humpenoumlder et al 2014) All developments the core modelstructure and recently developed modules of DGVM LPJmLversion 40 (in the following referred to as LPJmL4) will bedescribed in Sect 2 in more detail We show that the model inits present form allows for a consistent and joint quantifica-tion of climate and land use change impacts on the terrestrialbiosphere the water cycle the carbon cycle and on agricul-tural production (a systematic evaluation can be found in PartII of this paper) To give an overview of recent developmentsand applications of LPJmL4 we present the following

1 A comprehensive description of the full model with allcontributing developments since its original publicationby Sitch et al (2003) and Bondeau et al (2007) Weaim at consistently uniting all developments includ-ing undocumented and already published developmentsthus providing a comprehensive description of the fullLPJmL4 model

2 An overview of published LPJmL applications to reviewthe improvement of process understanding

3 A discussion of the presented standard LPJmL4 resultsthat give an overview of simulated biogeochemical hy-drological and agricultural patterns at the global scale

Geosci Model Dev 11 1343ndash1375 2018 wwwgeosci-model-devnet1113432018

S Schaphoff et al LPJmL4 ndash Part 1 Model description 1345

Figure 1 LPJmL4 scheme for carbon water and energy fluxes represented by the model C ndash carbon W ndash water S ndash sensible heatconduction H ndash latent heat convection c ndash energy conduction Rn ndash net downward radiation (input) PAR ndash photosynthetic active radiationEI ndash interceptionET ndash transpirationES ndash evaporation Infil ndash infiltration Perc ndash percolation Pr ndash precipitation (input) GPP ndash gross primaryproduction NPP ndash net primary production Ra ndash autotrophic respiration Rh ndash heterotrophic respiration Hc ndash carbon harvested Fc ndash carbonemitted by fire SOM ndash soil organic matter R ndash run-off Q ndash discharge

2 Model description

The original LundndashPotsdamndashJena (LPJ) DGVM was de-scribed in detail by Sitch et al (2003) This description andthe associated model evaluation focused on modelling thegrowth and geographical distribution of natural plant func-tional types (PFTs) and the associated biogeochemical pro-cesses (mainly carbon cycling) by building on the improvedrepresentation of the water balance (Gerten et al 2004)Bondeau et al (2007) introduced the representation of cropfunctional types (CFTs) and evaluated the role of agriculturefor the terrestrial carbon balance in particular This model hassince then been referred to as LPJmL (LundndashPotsdamndashJenawith managed Land) and provides the foundation for explic-itly simulating agricultural production in a changing climateand for quantifying the impacts of agricultural activities onthe terrestrial carbon and water cycle

A number of further specific model developments and ap-plications have been published but a comprehensive modeldescription of all developments and amendments is missingThe parts of LPJmL4 building on Bondeau et al (2007) notonly allow for quantifying changes in vegetation composi-tion the water cycle the carbon cycle and agricultural pro-duction but also for explicitly simulating the dynamics and

constraints within and among the modules thereby provid-ing a consistent and comprehensive representation of Earthrsquosland surface processes To demonstrate the interplay of allthese model features in the new LPJmL4 version the presentpaper documents the core model structure including equa-tions and parameters from Sitch et al (2003) and Bondeauet al (2007) and all more recent code developments Fig-ure S1 in the Supplement provides a schematic overview ofthe model structure and Fig 1 of the simulated carbon wa-ter and energy fluxes The following sections describe themodel components energy balance model and permafrost(Sect 21) plant physiology (Sect 22) plant functional(Sect 23) and crop functional types (Sect 24) soil litterand carbon pools (Sect 25) water balance (Sect 26) andland use (Sect 27)

21 Energy balance model and permafrost

The energy balance model includes the calculation of photo-synthetic active radiation daylength potential evapotranspi-ration (Sect 211) and albedo (Sect 212) The permafrostmodule is based on a new calculation of the soil energy bal-ance (Sect 213)

wwwgeosci-model-devnet1113432018 Geosci Model Dev 11 1343ndash1375 2018

211 Photosynthetic active radiation daylength andpotential evapotranspiration

Photosynthetic active radiation (PAR) is the primary energysource for photosynthesis (Sect 221) and thus for the wholecarbon cycle Total daily PAR in mol mminus2 dayminus1 is calculatedas

PAR= 05 middot cq middotRsday (1)

where cq = 46times10minus6 is the conversion factor from J to molfor solar radiation at 550 nm Half of the daily incoming so-lar irradiance Rsday is assumed to be PAR and atmosphericabsorption to be the same for PAR and Rsday (Prentice et al1993 Haxeltine and Prentice 1996)

Similar to the role of PAR for the carbon cycle potentialevapotranspiration (PET) is the primary driver of the watercycle The calculation of both PAR and PET follows the ap-proach of Prentice et al (1993) in which the calculation ofPET (mm dayminus1) is based on the theory of equilibrium evap-otranspiration Eeq (Jarvis and McNaughton 1986) given by

Eeq =s

s+ γmiddotRnday

λ (2)

where Rnday is daily surface net radiation (in J mminus2 dayminus1)and λ is the latent heat of vaporization (in J kgminus1) with aweak dependence on air temperature (Tair in C) derivedfrom Monteith and Unsworth (1990 p 376 Table A3)

λ= 2495times 106minus 2380 middot Tair (3)

where s is the slope of the saturation vapour pressure curve(in Pa Kminus1) given by

s = 2502times 106middot

exp[17269 middot Tair(2373+ Tair)]

(2373+ Tair)2 (4)

and γ is the psychrometric constant (in Pa Kminus1) given by

γ = 6505+ 0064 middot Tair (5)

Following Priestley and Taylor (1972) PET (mm dayminus1)is subsequently calculated from Eeq as

PET= PT middotEeq (6)

where PT is the empirically derived PriestleyndashTaylor coeffi-cient (PT= 132)

The terrestrial radiation balance is written as

Rn = (1minusβ) middotRs+Rl (7)

where Rn is net surface radiation Rs is incoming solar ir-radiance (downward) at the surface and Rl the outgoing netlongwave radiation flux at the surface (all in W mminus2) β is theshortwave reflection coefficient of the surface (albedo) The

calculation of albedo depending on land surface conditionsis described in Sect 212

If not supplied directly as input variables to the model theradiation terms Rs and Rl can be computed for any day andlatitude at given cloudiness levels (input) following Prenticeet al (1993) Rl can be approximated by a linear function oftemperature and the clear-sky fraction

Rl = (b+ (1minus b) middot ni) middot (Aminus Tair) (8)

where b = 02 and A= 107 are empirical constants Tair isthe mean daily air temperature in C ie any effects of di-urnal temperature variations are ignored The proportion ofbright sky (ni) is defined by ni= 1minuscloudiness The net out-going daytime longwave flux Rlnday

is obtained by multiply-ing with the length of the day in seconds

Rlnday= Rl middot daylength middot 3600 (9)

Instantaneous solar irradiance at the surface is computedfrom the solar constant accounting for ni and the angulardistance between the sunrsquos rays and the local vertical (z)

Rs = (c+ d middot ni) middotQ0 middot cos(z) (10)

where c = 025 and d = 05 are empirical constants that to-gether represent the clear-sky transmittivity (075) Q0 isthe solar irradiance at day i accounting for the variation inEarthrsquos distance to the sun

Q0 =Q00 middot (1+ 2 middot 001675 middot cos(2 middotπ middot i365)) (11)

where Q00 is the solar constant with 1360 W mminus2 The solarzenith angle (z) correction of Rs is computed from the solardeclination (δ ie the angle between the orbital plane andthe Earthrsquos equatorial plane) which varies between +234

in the Northern Hemisphere midsummer and minus234 in theNorthern Hemisphere midwinter the latitude (lat in radians)and the hour angle h ie the fraction of 2 middotπ (in radians)which the Earth has turned since the local solar noon

cos(z)= sin(lat) middot sin(δ)+ cos(lat) middot cos(δ) middot cos(h) (12)

δ =minus234 middotπ180 middot cos(2 middotπ middot (i+ 10)365) (13)

To obtain the Rsday Eq (10) needs to be integrated from sun-rise to sunset ie from minush12 to h12 where h12 is the half-day length in angular units computed as

h12 = arccos(minus

sin(lat) middot sin(δ)cos(lat) middot cos(δ)

) (14)

and thus

Rsday = (c+ d middot ni) middotQ0 middot (sin(lat) middot sin(δ) middoth12 (15)+ cos(lat) middot cos(δ) middoth12)

The duration of sunshine in a single day (daylength inhours) is computed as

daylength= 24 middoth12

π (16)

212 Albedo

Albedo (β) the average reflectivity of the grid cell was firstimplemented by Strengers et al (2010) and later improvedby considering several drivers of phenology as in Forkel et al(2014)

nPFTsumPFT=1

βPFT middotFPCPFT+Fbare (17)

middot (Fsnow middotβsnow+ (1minusFsnow) middotβsoil)

β depends on the land surface condition and is based on acombination of defined albedo values for bare soil (βsoil =

03) snow (βsnow = 07 average value taken from Lianget al 2005 Malik et al 2012) and plant-compartment-specific albedo values in which vegetation albedo (βPFT) issimulated as the albedo of each existing PFT (βPFT) FPCPFTis the foliage projective cover of the respective PFT (seeEq 57) Parameters (βleafPFT) were taken as suggested byStrugnell et al (2001) (see Table S5) Parameters βstemPFTand βlitterPFT were obtained from Forkel et al (2014) whooptimized these parameters by using MODIS albedo time se-ries Fsnow and Fbare are the snow coverage and the fractionof bare soil respectively (Strengers et al 2010)

213 Soil energy balance

The newly implemented calculation of the soil energy bal-ance as described in Schaphoff et al (2013) marks a newdevelopment and differs markedly from previous implemen-tations of permafrost modules in LPJ (Beer et al 2007)Soil water dynamics are computed daily (see Sect 26) Thesoil column is divided into five hydrological active layers of02 03 05 1 and 1 m of depth (1z) summing to 3 m (seeSect 261 and Fig 1) Soil temperatures (Tsoil in C) forthese layers are computed with an energy balance model in-cluding one-dimensional heat conduction and convection oflatent heat Freezing and thawing has been added to betteraccount for soil ice dynamics For a thermal buffer we as-sume an additional layer of 10 m thickness which is onlythermally and not hydrologically active Soil parameters forthermal diffusivity (m2 sminus1) at wilting point at 15 of wa-ter holding capacity and at field capacity and for thermal con-ductivity (W mminus1 Kminus1) at wilting point and at saturation (forwater and ice) are derived for each grid cell using soil texturefrom the Harmonized World Soil Database (HWSD) version1 (Nachtergaele et al 2009) Relationships between textureand thermal properties are taken from Lawrence and Slater(2008) The one-dimensional heat conduction equation is

partTsoil

partt= α middot

part2Tsoil

partz2 (18)

where α = λc is thermal diffusivity λ thermal conductiv-ity and c heat capacity (in J mminus3 Kminus1) Tsoil at position z

and time t is solved in its finite-difference form followingBayazıtoglu and Oumlzisik (1988)

Tsoil(t+1l) minus Tsoil(tl)

1t= (19)

α middotTsoil(tlminus1) + Tsoil(tl+1) minus 2Tsoil(tl)

for soil layers l including a snow layer and time step t withthe following boundary conditions

Tsoil(t = 1l = 1) = Tair (20)Tsoil(tl = nsoil+1) = Tsoil(tl = nsoil)

where nsoil = 6 is the number of soil layers We assume aheat flux of zero below the lowest soil layer ie below 13 mof depth The largest possible but still numerically stable timestep 1t is calculated depending on 1z and soil thermal dif-fusivity α (Bayazıtoglu and Oumlzisik 1988) which gives thestability criterion (r) for the finite-difference solution

r =α1t

(1z)2 (22)

For numerical stability (1minus 2r) needs to be gt 0 so thatr le 05 as 1z is given from soil depth and α can be calcu-lated from soil properties The maximum stable 1t can becalculated as

1t le(1z)2

2 middotα (23)

and therefore Eq (19) becomes

Tsoil(t+1l) = (24)r middot(Tsoil(tlminus1) + Tsoil(tl+1) + (1minus 2r) middot Tsoil(tl)

For the diurnal temperature range after Parton and Lo-gan (1981) at least 4 time steps per day are calculated andthe maximum number of time steps is set to 40 per dayHeat capacity (c) of the soil is calculated as the sum ofthe volumetric-specific heat capacities (in J mminus3 Kminus1) of soilminerals (cmin) soil water content (cwater) soil ice content(cice) and their corresponding shares (m in m3) of the soilbucket

c = cmin middotmmin+ cwater middotmwater+ cice middotmice (25)

The heat capacity of air is neglected because of its com-paratively low contribution to overall heat capacity Ther-mal conductivity (λ) is calculated following Johansen (1977)Sensible and latent heat fluxes are calculated explicitly forthe snow layer by assuming a constant snow density of03 t mminus3 and the resulting thermal diffusivity of 317times10minus7 m2 sminus1 Sublimation is assumed to be 01 mm dayminus1which corresponds to the lower end suggested by Gelfanet al (2004) The active layer thickness represents the depth

of maximum thawing of the year Freezing depth is calcu-lated by assuming that the fraction of frozen water is congru-ent with the frozen soil bucket The 0 C isotherm within alayer is estimated by assuming a linear temperature gradientwithin the layer and this fraction of heat is assumed to beused for the thawing or freezing process Temperature repre-sents the amount of thermal energy available whereas heattransport represents the movement of thermal energy into thesoil by rainwater and meltwater Precipitation and percola-tion energy and the amount of energy which arises from thetemperature difference between the temperature of the abovelayer (or the air temperature for the upper layer) and the tem-perature of the below layer are assumed to be used for con-verting latent heat fluxes first The residual energy is used toincrease soil temperature Tsoil is initialized at the beginningof the spin-up simulation by the mean annual air temperature

22 Plant physiology

221 Photosynthesis

The LPJmL4 photosynthesis model is a ldquobig leafrdquo representa-tion of the leaf-level photosynthesis model developed by Far-quhar et al (1980) and Farquhar and von Caemmerer (1982)These assumptions have been generalized by Collatz et al(1991 1992) for global modelling applications and for thestomatal response The ldquostrong optimalityrdquo hypothesis (Hax-eltine and Prentice 1996 Prentice et al 2000) is applied byassuming that Rubisco activity and the nitrogen content ofleaves vary with canopy position and seasonally so as to max-imize net assimilation at the leaf level Most details are as inSitch et al (2003) but a summary is provided in the follow-ing In LPJmL4 photosynthesis is simulated as a function ofabsorbed photosynthetically active radiation (APAR) tem-perature daylength and canopy conductance for each PFTor CFT present in a grid cell and at a daily time step APARis calculated as the fraction of incoming net photosyntheti-cally active radiation (PAR see Eq 1) that is absorbed bygreen vegetation (FAPAR)

APARPFT = PAR middotFAPARPFT middotαaPFT (26)

where αaPFT is a scaling factor to scale leaf-level photosyn-thesis in LPJmL4 to stand level The PFT-specific FAPARPFTis calculated as follows

FAPARPFT = FPCPFT middot((phenPFTminusFSnowGC) (27)

middot (1minusβleafPFT)minus (1minus phenPFT)

middot cfstem middotβstemPFT

where phenPFT is the daily phenological status (ranging be-tween 0 and 1) representing the fraction of full leaf cover-age currently attained by the PFT reduced by the green-leafalbedo βleafPFT and the stem albedo βstemPFT (for trees) andFSnowGC is the fraction of snow in the green canopy cfstem =

07 is the masking of the ground by stems and branches with-out leaves (Strengers et al 2010) Based on this the grossphotosynthesis rate Agd is computed as the minimum of twofunctions (details in Haxeltine and Prentice 1996)

1 The light-limited photosynthesis rate JE(mol C mminus2 hminus1)

JE = C1 middotAPAR

daylength (28)

where for C3 photosynthesis

C1 = αC3 middot Tstress middot

(piminus0lowast

pi+ 2 middot0lowast

and for C4 photosynthesis

λmaxC4

) (30)

pi is the leaf internal partial pressure of CO2 given bypi = λmiddotpa where λ reflects the soilndashplant water interac-tion (see Eq 40) and gives the actual ratio of the inter-cellular to ambient CO2 concentration and pa (in Pa) isthe ambient partial pressure of CO2 Tstress is the PFT-specific temperature inhibition function which limitsphotosynthesis at high and low temperatures αC3 andαC4 are the intrinsic quantum efficiencies for CO2 up-take in C3 and C4 plants respectively and 0lowast is thephotorespiratory CO2 compensation point

0lowast =[O2]

2 middot τ (31)

where τ = τ25 middotq(Tairminus25) middot 0110τ is the specificity factor and

it reflects the ability of Rubisco to discriminate betweenCO2 and O2 [O2] is the partial pressure of O2 (Pa) τ25is the τ value at 25 C and q10τ is the temperature sen-sitivity parameter

2 The Rubisco-limited photosynthesis rate JC(mol C mminus2 hminus1)

JC = C2 middotVm (32)

where Vm is the maximum Rubisco capacity (seeEq 35) and

C2 =piminus0lowast

(1+ [O2]

) (33)

KC and KO represent the MichaelisndashMenten constants forCO2 and O2 respectively Daily gross photosynthesis Agd isthen given by

(JE + JC minus

radic(JE + JC)2minus 4 middot θ middot JE middot JC

)2 middot θ middot daylength

The shape parameter θ describes the co-limitation of lightand Rubisco activity (Haxeltine and Prentice 1996) Sub-tracting leaf respiration (Rleaf given in Eq 46) gives thedaily net photosynthesis (And) and thus Vm is included inJC and Rleaf To calculate optimal And the zero point of thefirst derivative is calculated (ie partAndpartVm equiv 0) The thusderived maximum Rubisco capacity Vm is

Vm =1bmiddotC1

C2((2 middot θ minus 1) middot sminus (2 middot θ middot sminusC2) middot σ) middotAPAR (35)

radic1minus

C2minus 2C2minus θs

and s = 24daylength middot b (36)

and b denotes the proportion of leaf respiration in Vm forC3 and C4 plants of 0015 and 0035 respectively For thedetermination of Vm pi is calculated differently by using themaximum λ value for C3 (λmaxC3

) and C4 plants (λmaxC4

see Table S6) The daily net daytime photosynthesis (Adt) isgiven by subtracting dark respiration

Rd = (1minus daylength24) middotRleaf (37)

See Eq (46) for Rleaf and Adt is given by

Adt = AndminusRd (38)

The photosynthesis rate can be related to canopy conduc-tance (gc in mm sminus1) through the CO2 diffusion gradient be-tween the intercellular air spaces and the atmosphere

gc =16Adt

pa middot (1minus λ)+ gmin (39)

where gmin (mm sminus1) is a PFT-specific minimum canopyconductance scaled by FPC that occurs due to processesother than photosynthesis Combining both methods deter-mining Adt (Eqs 38 39) gives

0= AdtminusAdt = And+ (1minus daylength24) middotRleaf (40)minuspa middot (gcminus gmin) middot (1minus λ)16

This equation has to be solved for λ which is not possi-ble analytically because of the occurrence of λ in And in thesecond term of the equation Therefore a numerical bisec-tion algorithm is used to solve the equation and to obtain λThe actual canopy conductance is calculated as a function ofwater stress depending on the soil moisture status (Sect 26)and thus the photosynthesis rate is related to actual canopyconductance All parameter values are given in Table S6

222 Phenology

The phenology module of tree and grass PFTs is based onthe growing season index (GSI) approach (Jolly et al 2005)Thereby the continuous development of canopy greennessis modelled based on empirical relations to temperaturedaylength and drought conditions The GSI approach wasmodified for its use in LPJmL (Forkel et al 2014) so thatit accounts for the limiting effects of cold temperature lightwater availability and heat stress on the daily phenology sta-tus phenPFT

phenPFT = fcold middot flight middot fwater middot fheat (41)

Each limiting function can range between 0 (full limita-tion of leaf development) and 1 (no limitation of leaf devel-opment) The limiting functions are defined as logistic func-tions and depend also on the previous dayrsquos value

f (x)t = (42)f (x)tminus1+ (1(1+ exp(slx middot (xminus bx))minus f (x)tminus1) middot τx

where x is daily air temperature for the cold and heat stress-limiting functions fcold and fheat respectively and standsfor shortwave downward radiation in the light-limiting func-tion flight and water availability for the water-limiting func-tion fwater The parameters bx and slx are the inflectionpoint and slope of the respective logistic function τx is achange rate parameter that introduces a time-lagged responseof the canopy development to the daily meteorological con-ditions The empirical parameters were estimated by opti-mizing LPJmL simulations of FAPAR against 30 years ofsatellite-derived time series of FAPAR (Forkel et al 2014)

223 Productivity

Autotrophic respiration

Autotrophic respiration is separated into carbon costs formaintenance and growth and is calculated as in Sitchet al (2003) Maintenance respiration (Rx in g C mminus2 dayminus1)depends on tissue-specific C N ratios (for abovegroundCNsapwood and belowground tissues CNroot) It further de-pends on temperature (T ) either air temperature (Tair) foraboveground tissues or soil temperatures (Tsoil) for below-ground tissues on tissue biomass (Csapwoodind or Crootind)and phenology (phenPFT see Eq 41) and is calculated at adaily time step as follows

Rsapwood = P middot rPFT middot k middotCsapwoodind

CNsapwoodmiddot g(Tair) (43)

Rroot = P middot rPFT middot k middotCrootind

CNrootmiddot g(Tsoil) middot phenPFT (44)

The respiration rate (rPFT in g C g Nminus1 dayminus1) is a PFT-specific parameter on a 10 C base to represent the acclima-tion of respiration rates to average conditions (Ryan 1991)

k refers to the value proposed by Sprugel et al (1995) and Pis the mean number of individuals per unit area

The temperature function g(T ) describing the influenceof temperature on maintenance respiration is defined as

g(T )= exp[

30856 middot(

minus1

(T + 4602)

)] (45)

Equation (45) is a modified Arrhenius equation (Lloyd andTaylor 1994) where T is either air or soil temperature (C)This relationship is described by Tjoelker et al (1999) fora consistent decline in autotrophic respiration with temper-ature While leaf respiration (Rleaf) depends on Vm (seeEq 35) with a static parameter b depending on photosyn-thetic pathway

Rleaf = Vm middot b (46)

gross primary production (GPP calculated by Eq 34 andconverted to g C mminus2 dayminus1) is reduced by maintenance res-piration Growth respiration the carbon costs for producingnew tissue is assumed to be 25 of the remainder Theresidual is the annual net primary production (NPP)

NPP= (1minus rgr) middot (GPPminusRleafminusRsapwoodminusRroot) (47)

where rgr = 025 is the share of growth respiration (Thorn-ley 1970)

Reproduction cost

As in Sitch et al (2003) a fixed fraction of 10 of annualNPP is assumed to be carbon costs for producing reproduc-tive organs and propagules in LPJmL4 Since only a verysmall part of the carbon allocated to reproduction finally en-ters the next generation the reproductive carbon allocationis added to the aboveground litter pool to preserve a closedcarbon balance in the model

Tissue turnover

As in Sitch et al (2003) a PFT-specific tissue turnover rateis assigned to the living tissue pools (Table S8 and Fig 1)Leaves and fine roots are transferred to litter and living sap-wood to heartwood Root turnover rates are calculated on amonthly basis and the conversion of sapwood to heartwoodannually Leaf turnover rates depend on the phenology of thePFT it is calculated at leaf fall for deciduous and daily forevergreen PFTs

23 Plant functional types

Vegetation composition is determined by the fractional cov-erage of populations of different plant functional types(PFTs) PFTs are defined to account for the variety of struc-ture and function among plants (Smith et al 1993) InLPJmL4 11 PFTs are defined of which 8 are woody (2

tropical 3 temperate 3 boreal) and 3 are herbaceous (Ap-pendix A) PFTs are simulated in LPJmL4 as average in-dividuals Woody PFTs are characterized by the populationdensity and the state variables crown area (CA) and thesize of four tissue compartments leaf mass (Cleaf) fine rootmass (Croot) sapwood mass (Csapwood) and heartwood mass(Cheartwood) The size of all state variables is averaged acrossthe modelled area The state variables of grasses are repre-sented only by the leaf and root compartments The physi-ological attributes and bioclimatic limits control the dynam-ics of the PFT (see Table S4) PFTs are located in one standper grid cell and as such compete for light and soil waterThat means their crown area and leaf area index determinestheir capacity to absorb photosynthetic active radiation forphotosynthesis (see Sect 221) and their rooting profiles de-termine the access to soil water influencing their productiv-ity (see Sect 26) In the following we describe how carbonis allocated to the different tissue compartments of a PFT(Sect 231) and vegetation dynamics (Sect 232) ie howthe different PFTs interact The vegetation dynamics compo-nent of LPJmL4 includes the simulation of establishment anddifferent mortality processes

231 Allocation

The allocation of carbon is simulated as described in Sitchet al (2003) and all parameter values are given in Table S6The assimilated amount of carbon (the remaining NPP) con-stitutes the annual woody carbon increment which is allo-cated to leaves fine roots and sapwood such that four basicallometric relationships (Eqs 48ndash51) are satisfied The pipemodel from Shinozaki et al (1964) and Waring et al (1982)prescribes that each unit of leaf area must be accompanied bya corresponding area of transport tissue (described by the pa-rameter klasa) and the sapwood cross-sectional area (SAind)

LAind = klasa middotSAind (48)

where LAind is the average individual leaf area and ind givesthe index for the average individual

A functional balance exists between investment in fine rootbiomass and investment in leaf biomass Carbon allocationto Cleafind is determined by the maximum leaf to root massratio lrmax (Table S8) which is a constant and by a waterstress index ω (Sitch et al 2003) which indicates that un-der water-limited conditions plants are modelled to allocaterelatively more carbon to fine root biomass which ensuresthe allocation of relatively more carbon to fine roots underwater-limited conditions

Cleafind = lrmax middotω middotCrootind (49)

The relation between tree height (H ) and stem diameter(D) is given as in Huang et al (1992)

H = kallom2 middotDkallom3 (50)

The crown area (CAind) to stem diameter (D) relation isbased on inverting Reinekersquos rule (Zeide 1993) with krp asthe Reineke parameter

CAind =min(kallom1 middotDkrp CAmax) (51)

which relates tree density to stem diameter under self-thinning conditions CAmax is the maximum crown area al-lowed The reversal used in LPJmL4 gives the expected rela-tion between stem diameter and crown area The assumptionhere is a closed canopy but no crown overlap

By combining the allometric relations of Eqs (48)ndash(51) itfollows that the relative contribution of sapwood respirationincreases with height which restricts the possible height oftrees Assuming cylindrical stems and constant wood density(WD) H can be computed and is inversely related to SAind

SAind =Cleafind middotSLA

klasa (52)

From this follows

H =Csapwoodind middot klasa

WD middotCleafind middotSLA (53)

Stem diameter can then be calculated by invertingEq (50) Leaf area is related to leaf biomass Cleafind by PFT-specific SLA and thus the individual leaf area index (LAIind)is given by

LAIind =Cleafind middotSLA

CAind (54)

SLA is related to leaf longevity (αleaf) in a month (see Ta-ble S8) which determines whether deciduous or evergreenphenology suits a given climate suggested by Reich et al(1997) The equation is based on the form suggested bySmith et al (2014) for needle-leaved and broadleaved PFTsas follows

SLA=2times 10minus4

DMCmiddot 10β0minusβ1middotlog(αleaf) log(10) (55)

The parameter β0 is adapted for broadleaved (β0 = 22)and needle-leaved trees (β0 = 208) and for grass (β0 =

225) and β1 is set to 04 Both parameters were derivedfrom data given in Kattge et al (2011) The dry matter carboncontent of leaves (DMC) is set to 04763 as obtained fromKattge et al (2011) LAIind can be converted into foliar pro-jective cover (FPCind which is the proportion of ground areacovered by leaves) using the canopy light-absorption model(LambertndashBeer law Monsi 1953)

FPCind = 1minus exp(minusk middotLAIind) (56)

where k is the PFT-specific light extinction coefficient (seeTable S5) The overall FPC of a PFT in a grid cell is ob-tained by the product FPCind mean individual CAind and

mean number of individuals per unit area (P ) which is de-termined by the vegetation dynamics (see Sect 232)

FPCPFT = CAind middotP middotFPCind (57)

FPCPFT directly measures the ability of the canopy to inter-cept radiation (Haxeltine and Prentice 1996)

232 Vegetation dynamics

Establishment

For PFTs within their bioclimatic limits (Tcmin see Ta-ble S4) each year new woody PFT individuals and herba-ceous PFTs can establish depending on available spaceWoody PFTs have a maximum establishment rate kest of 012(saplings mminus2 aminus1) which is a medium value of tree densityfor all biomes (Luyssaert et al 2007) New saplings can es-tablish on bare ground in the grid cell that is not occupiedby woody PFTs The establishment rate of tree individuals iscalculated

ESTTREE = (58)

kest middot (1minus exp(minus5 middot (1minusFPCTREE))) middot(1minusFPCTREE)

nestTREE

The number of new saplings per unit area (ESTTREE inind mminus2 aminus1) is proportional to kest and to the FPC of eachPFT present in the grid cell (FPCTREE and FPCGRASS) Itdeclines in proportion to canopy light attenuation when thesum of woody FPCs exceeds 095 thus simulating a declinein establishment success with canopy closure (Prentice et al1993) nestTREE gives the number of tree PFTs present in thegrid cell Establishment increases the population density P Herbaceous PFTs can establish if the sum of all FPCs isless than 1 If the accumulated growing degree days (GDDs)reach a PFT-specific threshold GDDmin the respective PFTis established (Table S5)

Background mortality

Mortality is modelled by a fractional reduction of P Mor-tality always leads to a reduction in biomass per unitarea Similar to Sitch et al 2003 a background mortal-ity rate (mortgreff in ind mminus2 aminus1) the inverse of meanPFT longevity is applied from the yearly growth efficiency(greff= bminc(Cleafind middotSLA)) (Waring 1983) expressed asthe ratio of net biomass increment (bminc) to leaf area

mortgreff = P middotkmort1

1+ kmort2 middot greff (59)

where kmort1 is an asymptotic maximum mortality rate andkmort2 is a parameter governing the slope of the relationshipbetween growth efficiency and mortality (Table S6)

Stress mortality

Mortality from competition occurs when tree growth leadsto too-high tree densities (FPC of all trees exceeds gt 95 )In this case all tree PFTs are reduced proportionally to theirexpansion Herbaceous PFTs are outcompeted by expandingtrees until these reach their maximum FPC of 95 or bylight competition between herbaceous PFTs Dead biomassis transferred to the litter pools

Boreal trees can die from heat stress (mortheat inind mminus2 aminus1) (Allen et al 2010) It occurs in LPJmL4 whena temperature threshold (Tmortmin in C Table S4) is ex-ceeded but only for boreal trees (Sitch et al 2003) Tem-peratures above this threshold are accumulated over the year(gddtw) and this is related to a parameter value of the heatdamage function (twPFT) which is set to 400

mortheat = P middotmin(

twPFT1)

P is reduced for both mortheat and mortgreff

Fire disturbance and mortality

Two different fire modules can be applied in the LPJmL4model the simple Glob-FIRM model (Thonicke et al 2001)and the process-based SPITFIRE model (Thonicke et al2010) In Glob-FIRM fire disturbances are calculated as anexponential probability function dependent on soil moisturein the top 50 cm and a fuel load threshold The sum of thedaily probability determines the length of the fire seasonBurnt area is assumed to increase non-linearly with increas-ing length of fire season The fraction of trees killed withinthe burnt area depends on a PFT-specific fire resistance pa-rameter for woody plants while all litter and live grassesare consumed by fire Glob-FIRM does not specify fire igni-tion sources and assumes a constant relationship between fireseason length and resulting burnt area The PFT-specific fireresistance parameter implies that fire severity is always thesame an approach suitable for model applications to multi-century timescales or paleoclimate conditions In SPITFIREfire disturbances are simulated as the fire processes risk igni-tion spread and effects separately The climatic fire dangeris based on the Nesterov index NI(Nd) which describes at-mospheric conditions critical to fire risk for day Nd

NI(Nd)=

Ndsumif Pr(d)le 3 mm

Tmax(d) middot (Tmax(d)minus Tdew(d)) (61)

where Tmax and Tdew are the daily maximum and dew-pointtemperature and d is a positive temperature day with a pre-cipitation of less than 3 mm The probability of fire spreadPspread decreases linearly as litter moisture ω0 increases to-wards its moisture of extinction me

Pspread =

1minusω0me ω0 leme

0 ω0 gtme(62)

Combining NI and Pspread we can calculate the fire dangerindex FDI

FDI=max

01minus

1memiddot exp

(minusNI middot

nsump=1

) (63)

to interpret the qualitative fire risk in quantitative terms Thevalue of αp defines the slope of the probability risk functiongiven as the average PFT parameter (see Table S9) for allexisting PFTs (n) SPITFIRE considers human-caused andlightning-caused fires as sources for fire ignition Lightning-caused ignition rates are prescribed from the OTDLIS satel-lite product (Christian et al 2003) Since it quantifies to-tal flash rate we assume that 20 of these are cloud-to-ground flashes (Latham and Williams 2001) and that un-der favourable burning conditions their effectiveness to startfires is 004 (Latham and Williams 2001 Latham and Schli-eter 1989) Human-caused ignitions are modelled as a func-tion of human population density assuming that ignition ratesare higher in remote regions and declines with an increasinglevel of urbanization and the associated effects of landscapefragmentation infrastructure and improved fire monitoringThe function is

nhig = PD middot k(PD) middot a(ND)100 (64)

k(PD)= 300 middot exp(minus05 middotradicPD) (65)

PD is the human population density (individuals kmminus2) anda(ND) (ignitions individualminus1 dayminus1) is a parameter describ-ing the inclination of humans to use fire and cause fire ig-nitions In the absence of further information a(ND) can becalculated from fire statistics using the following approach

a(ND)=Nhobs

tobs middotLFS middotPD (66)

where Nhobs is the average number of human-caused firesobserved during the observation years tobs in a region withan average length of fire season (LFS) and the mean humanpopulation density Assuming that all fires ignited in 1 dayhave the same burning conditions in a 05 grid cell with thegrid cell size A we combine fire danger potential ignitionsand the mean fire area Af to obtain daily total burnt area with

Ab =min(E(nig) middotFDI middotAfA) (67)

We calculate E(nig) with the sum of independent esti-mates of numbers of lightning (nlig) and human-caused ig-nition events (nhig) disregarding stochastic variations Af iscalculated from the forward and backward rate of spreadwhich depends on the dead fuel characteristics fuel load inthe respective dead fuel classes and wind speed Dead plantmaterial entering the litter pool is subdivided into 1 10 100and 1000 h fuel classes describing the amount of time to dry

a fuel particle of a specific size (1 h fuel refers to leaves andtwigs and 1000 h fuel to tree boles) As described by Thon-icke et al (2010) the forward rate of spread ROSfsurface(m minminus1) is given by

ROSfsurface =IR middot ζ middot (1+8w)

ρb middot ε middotQig (68)

where IR is the reaction intensity ie the energy release rateper unit area of fire front (kJ mminus2 minminus1) ζ is the propagat-ing flux ratio ie the proportion of IR that heats adjacent fuelparticles to ignition 8w is a multiplier that accounts for theeffect of wind in increasing the effective value of ζ ρb is thefuel bulk density (kg mminus3) assigned by PFT and weightedover the 1 10 and 100 h dead fuel classes ε is the effectiveheating number ie the proportion of a fuel particle that isheated to ignition temperature at the time flaming combus-tion starts andQig is the heat of pre-ignition ie the amountof heat required to ignite a given mass of fuel (kJ kgminus1) Withfuel bulk density ρb defined as a PFT parameter surface areato volume ratios change with fuel load Assuming that firesburn longer under high fire danger we define fire duration(tfire) (min) as

tfire =241

1+ 240 middot exp(minus1106 middotFDI) (69)

In the absence of topographic influence and changing winddirections during one fire event or discontinuities of the fuelbed fires burn an elliptical shape Thus the mean fire area(in ha) is defined as follows

Af =π

4 middotLBmiddotD2

T middot 10minus4 (70)

with LB as the length to breadth ratio of elliptical fire andDT as the length of major axis with

DT = ROSfsurface middot tfire+ROSbsurface middot tfire (71)

where ROSbsurface is the backward rate of spread LB forgrass and trees respectively is weighted depending on thefoliage projective cover of grasses relative to woody PFTs ineach grid cell SPITFIRE differentiates fire effects dependingon burning conditions (intra- and inter-annual) If fires havedeveloped insufficient surface fire intensity (lt 50 kW mminus1)ignitions are extinguished (and not counted in the model out-put) If the surface fire intensity has supported a high-enoughscorch height of the flames the resulting scorching of thecrown is simulated Here the tree architecture through thecrown length and the height of the tree determine fire effectsand describe an important feedback between vegetation andfire in the model PFT-specific parameters describe the treesensitivity to or influence on scorch height and crown scorchSurface fires consume dead fuel and live grass as a func-tion of their fuel moisture content The amount of biomassburnt results from crown scorch and surface fuel consump-tion Post-fire mortality is modelled as a result of two fire

mortality causes crown and cambial damage The latter oc-curs when insufficient bark thickness allows the heat of thefire to damage the cambium It is defined as the ratio of theresidence time of the fire to the critical time for cambial dam-age The probability of mortality due to crown damage (CK)is

Pm(CK)= rCK middotCKp (72)

where rCK is a resistance factor between 0 and 1 and p isin the range of 3 to 4 (see Table S9) The biomass of treeswhich die from either mortality cause is added to the respec-tive dead fuel classes In summary the PFT composition andproductivity strongly influences fire risk through the mois-ture of extinction fire spread through composition of fuelclasses (fine vs coarse fuel) openness of the canopy and fuelmoisture fire effects through stem diameter crown lengthand bark thickness of the average tree individual The higherthe proportion of grasses in a grid cell the faster fires canspread the smaller the trees andor the thinner their bark thehigher the proportion of the crown scorched and the highertheir mortality

24 Crop functional types

In LPJmL4 12 different annual crop functional types (CFTs)are simulated (Table S10) similar to Bondeau et al (2007)with the addition of sugar cane The basic idea of CFTsis that these are parameterized as one specific representa-tive crop (eg wheat Triticum aestivum L) to represent abroader group of similar crops (eg temperate cereals) In ad-dition to the crops represented by the 12 CFTs other annualand perennial crops (other crops) are typically representedas managed grassland Bioenergy crops are simulated to ac-count for woody (willow trees in temperate regions eucalyp-tus for tropical regions) and herbaceous types (Miscanthus)(Beringer et al 2011) The physical cropping area (ie pro-portion per grid cell) of each CFT the group of other cropsmanaged grasslands and bioenergy crops can be prescribedfor each year and grid cell by using gridded land use data de-scribed in Fader et al (2010) and Jaumlgermeyr et al (2015) seeSect 27 In principle any land use dataset (including futurescenarios) can be implemented in LPJmL4 at any resolution

Crop varieties and phenology

The phenological development of crops in LPJmL4 is drivenby temperature through the accumulation of growing degreedays and can be modified by vernalization requirements andsensitivity to daylength (photoperiod) for some CFTs andsome varieties Phenology is represented as a single phasefrom planting to physiological maturity Different varietiesof a single crop species are represented by different phe-nological heat unit requirements to reach maturity (PHU)but also different harvest indices (HIopt) ie the fractionof the aboveground biomass that is harvested is typically

CFT specific but can be specified to represent specific vari-eties (Asseng et al 2013 Bassu et al 2014 Kollas et al2015 Fader et al 2010) Heat units (HUt growing de-gree days) are accumulated (HUsum) daily (see Eq 73) Thedaily heat unit increment (HUt ) is the difference betweenthe daily mean temperature of day t and the CFT-specificbase temperature (see Table S10) The increment HUt can-not be less than zero at any given day The phenologicalstage of the crop development (fPHU) is expressed as the ra-tio of accumulated (HUsum) and required phenological heatunits (PHUs see Eq 74) Physiological maturity is reachedas soon as the sufficient growing degree days have been ac-cumulated (fPHU= 10) Both unfulfilled vernalization re-quirements and unsuitable photoperiod affect the phenologi-cal development of the CFTs (see Eqs 78 and 79) Thereforethe daily increment HUt at day t is scaled by reduction fac-tors vrf for vernalization and prf for photoperiod

HUsum =

tsumt prime= sdate

HUt prime middot vrf middotprf (73)

fPHU= HUsumPHU (74)

Wheat and rapeseed are implemented as spring and wintervarieties The model endogenously determines which varietyto grow based on the average climate of past decades If inter-nally computed sowing dates for winter varieties (see belowSect 271) indicate that the winter is too long to allow forgrowing winter varieties which is prior to day 258 (90) forwheat and 241 (61) for rapeseed in the Northern Hemisphere(Southern Hemisphere) spring varieties are grown insteadThese are computed on the basis of the sowing dates (sdate)as an indication for the length of the cropping season con-strained by crop-specific limits For winter varieties of wheatand rapeseed PHU is computed as

PHU= minus 01081 middot (sdateminus keyday)2+ 31633 (75)middot (sdateminus keyday)+PHUwhigh

PHUle PHUwlow

where PHUwlow and PHUwhigh are minimum and maximumPHU requirements for winter varieties respectively Thesowing date sdate can either be internally computed (seeSect 271) or prescribed for a crop and pixel The parameterkey day is day 365 in the Northern Hemisphere and day 181in the Southern Hemisphere For spring varieties of wheatand rapeseed as well as for all other crops PHU is computedas

PHU= max(Tbaselow atemp20) middot pfCFT (76)PHUshigh ge PHUge PHUslow

where PHUslow and PHUshigh are minimum and maximumPHU requirements for spring varieties respectively Tbaselow

is the minimum base temperature for the accumulation ofheat units atemp20 is the 20-year moving average annualtemperature and pfCFT is a CFT-specific scaling factor

Vernalization requirements (PVDs) are zero for spring va-rieties and are computed for winter varieties

PVD= verndate20minus sdateminus pPVDCFT 0le PVDle 60 (77)

with pPVDCFT as a CFT-specific vernalization factor sdateas the Julian day of the year of sowing and verndate20 asthe multi-annual average of the first day of the year whentemperatures rise above a CFT-specific vernalization thresh-old (Tvern see Table S10) The effectiveness of vernaliza-tion is dependent on the daily mean temperature being in-effective below minus4 C and above 17 C and fully effectivebetween 3 and 10 C and the effectiveness scales linearlybetween minus4 and 3 and between 10 and 17 C The effec-tive number of vernalizing days vdsum is accumulated untilthe requirements (PVD) as computed in Eq (77) are met oruntil phenology has progressed over 20 of its phenologi-cal development (ie fPHUge 02) Crop varieties can be pa-rameterized as sensitive to photoperiod (ie daylength) buthere are assumed to be insensitive Parameter settings canbe adjusted for specific applications such as in model inter-comparisons (Asseng et al 2013 Bassu et al 2014 Kollaset al 2015) Photoperiod restrictions are active until the cropreaches senescence

The reduction factors are computed as

vrf = (vdsumminus 100)(PVDminus 100) (78)

forcing vrf to be between 0 and 1 and

prf = (1minuspsens) middotmin(1max(0 (daylengthminuspb) (79)

(psminuspb)))+psens

where psens is the parameterized sensitivity to photoperiod(0 1) daylength is the duration of daylight (sunrise to sun-set) in hours (see Sect 211) pb is the base photoperiod inhours and ps is the saturation photoperiod in hours

Crop growth and allocation

Photosynthesis and autotrophic respiration of crops are com-puted as for the herbaceous natural PFTs (see Sect 221 and22) Light absorption for photosynthesis is computed basedon the LambertndashBeer law (Monsi 1953) except for maizeFor maize LPJmL4 employs a linear FAPAR model (Zhouet al 2002) and a maximum leaf area index (LAImax) of5 instead of 7 as for all other CFTs (Fader et al 2010)Daily NPP accumulates to total biomass and is allocateddaily to crop organs in a hierarchical order roots leavesstorage organ mobile reserves and stem (pool) The fractionof biomass that is allocated to each compartment dependson the phenological development stage (fPHU) The fractionof total biomass that is allocated to the roots (froot) ranges

between 40 at planting and 10 at maturity modified bywater stress

froot =04minus (03 middot fPHU) middotwdf

wdf+ exp(613minus 00883 middotwdf) (80)

where the water deficit (wdf) is defined as the ratio betweenaccumulated daily transpiration and accumulated daily waterdemand since planting representing a measure of the aver-age water stress After allocation to the roots biomass is al-located to the leaves Leaf area development follows a CFT-specific shape that is controlled by phenological development(fPHU) the onset of senescence (ssn) and the shape of greenLAI decline after the onset of senescence The ideal CFT-specific development of the canopy (Eq 81) is thus describedas a function of the maximum LAI (LAImax) and the pheno-logical development (fPHU) with two turning points in thephenological development (fPHUc and fPHUk) and the cor-responding fraction of the maximum green LAI reached atthese stages (fLAImaxc and fLAImaxk )

fLAImax =fPHU

fPHU+ c middot (ck)fPHUcminusfPHU

fPHUkminusfPHUc

c =fPHUc

fLAImaxc minus fPHUc (82)

k =fPHUk

fLAImaxk minus fPHUk (83)

The onset of senescence is defined as a point in the pheno-logical development fPHUsen After the onset of senescenceie fPHUle fPHUsen no more biomass is allocated to theleaves and the maximum green LAI is computed as

fLAImax =

(1minus fPHU

1minus fPHUsen

middot (1minus fLAImaxh)+ fLAImaxh

with fLAImaxh as the green LAI fraction at which harvest oc-curs This optimal development of LAI is modified by acutewater stress For this the daily increment LAIinc which isoptimal for day t is computed as

LAIinct = (fLAImaxt minus fLAImaxtminus1) middotLAImax (85)

with fLAImaxt as the maximum green LAI of day t andfLAImaxtminus1 as the maximum green LAI of the previous dayThe daily increment LAIinc is additionally scaled with thedaily water stress (ω) which is calculated as the ratio of ac-tual transpiration and demand (see Sect 26) on that day Thecalculation of LAIinc applies to daily LAI increments whichare independent of each other The LAI on day t is accumu-lated from daily LAI increments

LAIt =tsum

t prime= sdate

LAIinct prime middotω (86)

and implies that the LAI development cannot recover fromwater-limitation-induced reductions in LAI Until the onsetof senescence the daily LAI determines the biomass allo-cated to the leaves by dividing LAI by specific leaf area(SLA) SLA is computed as in Eq (55) using the β0 value forgrasses (225) and CFT-specific αleaf values (Table S11) Itscalculation was adjusted for SLA values as given in Xu et al(2010) Biomass in the storage organ is computed by pheno-logical stage and the harvest index (HI) which describes thefraction of the aboveground biomass that is allocated to thestorage organ

fHIopt middotHIopt if HIopt ge 1

fHIopt middot (HIoptminus 1)+ 1 otherwise(87)

fHIopt = 100 middotfPHU(100 middotfPHU+exp(111minus100 middotfPHU))(88)

As the HI is defined relative to aboveground biomass rootsand tubers have HI values larger than 10 which needs to beaccounted for in the allocation of biomass to the storage or-gan (see Eq 87) If biomass is limiting (low NPP) biomassis allocated in hierarchical order starting with roots (whichcan always be satisfied as it is 40 of total biomass max-imum) followed by leaves (Cleaf where eventually the LAIis temporarily reduced impacting APAR and thus NPP) andthe storage organ (Cso) If biomass is not limiting the alloca-tion to the storage organ is computed from the harvest index(HI) and total aboveground biomass

Cso = HI middot (Cleaf+Cso+Cpool) (89)

Excess biomass after allocating to roots leaves and stor-age organ is allocated to a pool (Cpool) that represents mo-bile reserves and the stem At harvest storage organs arecollected from the field and crop residues can be left on thefield or removed (for scenario setting see eg Bondeau et al2007) If removed a fraction of 10 of the abovegroundbiomass (leaves and pool) is assumed to remain on the fieldas stubbles Stubbles and root biomass enter the litter poolsafter harvest

25 Soil and litter carbon pools

The biogeochemical processes in soil and litter are impor-tant for the global carbon balance The LPJmL4 litter poolconsists of CFT- and PFT-dependent pools for leaf root andwood The soil consists of a fast and a slow organic mat-ter pool Decomposition fluxes transferring litter carbon intosoil carbon and losses for heterotrophic respiration (Rh) aredescribed in the following section

251 Decomposition

The decomposition of organic matter pools is represented byfirst-order kinetics (Sitch et al 2003)

dC(l)dt=minusk(l) middotC(l) (90)

where C(l) is the carbon pool size of soil or litter and k(l) isthe annual decomposition rate per layer (l) in dayminus1 Inte-grating for a time interval 1t (here 1 day) yields

C(t+1l) = C(tl) middot exp(minusk(l) middot1t) (91)

where C(tl) and C(t+1l) are the carbon pool sizes at the be-ginning and the end of the day The amount of carbon de-composed per layer is

C(tl) middot (1minus exp(minusk(l) middot1t)) (92)

at which 70 of decomposed litter goes directly into the at-mosphere Rhlitter and the remaining is transferred to the soilcarbon pools 985 to the fast soil carbon pool and 15 tothe slow carbon pool (Sitch et al 2003)

Rh = Rhlitter+RhfastSoil+RhslowSoil (93)

The decomposition rates for root litter and soil (k(lPFT))are a function of soil temperature and soil moisture

k(lPFT) =1

τ10PFT

middot g(Tsoil(l)

)middot f(θ(l)) (94)

which is reciprocal to the mean residence time (τ10PFT ) Rootlitter decomposition is defined for all PFTs (03 aminus1) and forfast and slow soil carbon (003 and 0001 aminus1 respectively)as in Sitch et al (2003) p represents the different poolsThe decomposition rate of leaf and wood litter is defined asPFT-specific decomposition rates at 10 C for leaf wood androot which has been analysed and proposed by Brovkin et al(2012) for leaf and wood The temperature dependence func-tion for the fast and slow soil carbon and the leaf and rootlitter pool g(Tsoil) was already described in Eq (45) Woodlitter decomposition is calculated as follows

kwoodPFT =(Q10woodlitter

) (Tsoilminus10)100 (95)

Table S7 presents the (1τ10PFT ) used for leaves and woodand the Q10woodlitter parameter for temperature-dependentwood decomposition in the litter pool The soil moisturefunction follows Schaphoff et al (2013)

f (θ(l))= 00402minus 5005 middot θ3(l)+ 4269 middot θ2

(l) (96)

+ 0719 middot θ(l)

where θ(l) is the soil volume fraction of the layer l Parame-ters are chosen based on the assumption that rates are maxi-mal at field capacity and decline for higher θ(l) to 02 f

(θ(l))

is very small (00402) when θ(l) equals 1 due to oxygen lim-itation and when θ(l) is 0

To account for different decomposition rates in the dif-ferent soil layers a vertical soil carbon distribution is nowimplemented in LPJmL4 following Schaphoff et al (2013)Jobbagy and Jackson (2000) suggested a cumulative logndashlogdistribution of the fraction of soil organic carbon (Cfl) as afunction of depth with

Cf(l) = 10ksocmiddotlog10(d(l)) (97)

where d(l) is the relative share of the layer l in the entire soilbucket and the parameter ksoc was adjusted for the soil layerdepth now used in LPJmL4 (see Table S7) The total amountof soil carbon Cstotal is estimated from the mean annual de-composition rate kmean(l) and the mean litter input into thesoil as in (Sitch et al 2003) but is distributed to all rootlayers separately (Eq 98) The envisaged vertical soil distri-bution C(l)

C(l) =

nPFTsumPFT= 1

dksocPFT(l) middotCstotal (98)

is estimated after a carbon equilibrium phase of 2310 yearsThe mean decomposition rate for each PFT kmeanPFT can bederived from the mean annual decomposition rate kmean(l)of the spin-up years as a layer-weighted value derived fromEq (97)

kmeanPFT =

nsoilsuml=1

kmean(l) middotCf(lPFT) (99)

The annual carbon shift rates Cshift(lp) describe the organicmatter input from the different PFTs into the respective layerdue to cryoturbation and bioturbation and are designed forglobal applications

Cshift(lPFT) =Cf(lPFT) middot kmean(l)

kmeanPFT

26 Water balance

The terrestrial water balance is a pivotal element in LPJmL4as water and vegetation are linked in multiple ways

1 the coupling of plant transpiration and carbon uptakefrom the atmosphere through stomatal conductance inthe process of photosynthesis

2 the down-regulation of photosynthesis plant growthand productivity in response to soil water limitation(relative to atmospheric moisture demand) in the casethat the actual canopy conductance is below potentialcanopy conductance (in the demand function that de-scribes transpiration)

3 the effect of changes in vegetation type distributionphenology and production on evaporation transpira-tion interception run-off and soil moisture and

4 the anthropogenic stimulation of crop growth throughirrigation with water taken from rivers dams lakes andassumed renewable groundwater

These couplings of water and vegetation dynamics enablesimulations of the interacting mutual feedbacks betweenfreshwater cycling in and above the Earthrsquos surface and ter-restrial vegetation dynamics

261 Soil water balance

Advancing the former two-layer approach (Sitch et al2003) LPJmL4 divides the soil column into five hydrologicalactive layers of 02 03 05 1 and 1 m thickness (Schaphoffet al 2013) Water holding capacity (water content at per-manent wilting point at field capacity and at saturation)and hydraulic conductivity are derived for each grid cell us-ing soil texture from the Harmonized World Soil Database(HWSD) version 1 (Nachtergaele et al 2009) and relation-ships between texture and hydraulic properties from Cosbyet al (1984) see also Sect 213 Water content in soil layersis altered by infiltrating rainfall and the vertical movement ofgravitational water (percolation) Since the accuracy neededfor a global model does not justify the computational costsof an exact solution to the governing differential equationa simplified storage approach is implemented in LPJmL4Rather than calculating the infiltration and percolation of pre-cipitation at once total precipitation is divided in portionsof 4 mm that are routed through the soil one after anotherThis effectively emulates a time discretization which leadsto a higher proportion of run-off being generated for higheramounts precipitation

Infiltration

The infiltration rate of rain and irrigation water into the soil(infil in mm) depends on the current soil water content of thefirst layer as follows

infil= Pr middot

radic1minus

SW(1)minusWpwp(1)

Wsat(1) minusWpwp(1) (101)

where Wsat(1) is the soil water content at saturation Wpwp(1)the soil water content at wilting point and SW(1) the totalactual soil water content of the first layer all in millimetresPr is the amount of water in the current portion of daily pre-cipitation or applied irrigation water (maximum 4 mm) Thesurplus water that does not infiltrate is assumed to generatesurface run-off

Percolation

Subsequent percolation through the soil layers is calculatedby the storage routine technique (Arnold et al 1990) as usedin regional hydrological models such as SWIM (Krysanova

et al 1998)

FW(t+1 l) = FW(t l) middot exp(minus1t

) (102)

where FW(tl) and FW(t+1l) are the soil water content be-tween field capacity and saturation at the beginning and theend of the day for all soil layers l respectively 1t is thetime interval (here 24 h) and TTl determines the travel timethrough the soil layer in hours

TT(l) =FW(l)

HC(l)(103)

HC(l) is the hydraulic conductivity of the layer in mm hminus1

HC(l) =Ks(l) middot

(SW(l)

Wsat(l)

)β(l) (104)

whereWsat(l) is the soil water content at saturationKs(l) is thesaturated conductivity (in mm hminus1) and SW(l) the total soilwater content of the layer (in mm) Thus percolation can becalculated by subtracting FW(tl) from FW(t+1l) for all soillayers

percprime(l) = FW(tl) middot

[1minus exp

(minus1t

)](105)

The percolation perc(l) (in mm dayminus1) is limited by thesoil moisture of the lower layer similar to the infiltration ap-proach

perc(l+1) = percprime(l+1) middot

radic1minus

SW(l+1)minusWpwp(l+1)

Wsat(l+1) minusWpwp(l+1)

Excess water over the saturation levels forms lateral run-off in each layer and contributes to subsurface run-off Theformation of groundwater which is the seepage from the bot-tom soil layer has been recently introduced into LPJmL4(Schaphoff et al 2013) Both surface and subsurface run-off are simulated to accumulate to river discharge (seeSect 263)

262 Evapotranspiration

Similar to Gerten et al (2004) evapotranspiration (ET) isthe sum of vapour flow from the Earthrsquos surface to the atmo-sphere It consists of three major components evaporationfrom bare soils evaporation of intercepted rainfall from thecanopy and plant transpiration through leaf stomata The cal-culation of these different components in LPJmL4 is basedon equilibrium evapotranspiration (Eeq) and the PET as de-scribed in Sect 211 and Eqs (2) and (6)

Canopy evaporation

Canopy evaporation is the evaporation of rainfall that hasbeen intercepted by the canopy limited either by PET or theamount of intercepted rainfall I (both in mm dayminus1)

Ecanopy =min(PETI ) (107)

The amount of intercepted rainfall is given as

nPFTsumPFT=1

IPFT middotLAIPFT middotPr (108)

where IPFT is the interception storage parameter for eachPFT (Gerten et al 2004) LAIPFT the PFT-specific leaf areaper unit of grid cell area and Pr is daily precipitation inmm dayminus1

Soil evaporation

Soil evaporation (Es in mm dayminus1) only occurs from baresoil in which the vegetation cover (fv) is less than 100 The fv is the sum of all present PFT FPCs (see Eq 57)taking daily phenology into account The evaporation fluxdepends on available energy for the vaporization of water(see Eq 6) and the available water in the soil LPJmL4 as-sumes that water for evaporation is available from the up-per 03 m of the soil implicitly accounting for some cap-illary rise Evaporation-available soil water (wevap) is thusall water above the wilting point of the upper layer (02 m)and one-third of the second layer (03 m) Actual evapora-tion is then computed according to Eq (110) with w beingthe evaporation-available water relative to the water holdingcapacity in that layer whcevap

w =min(1wevapwhcevap) (109)

and thus

Es = PET middotw2middot (1minus fv) (110)

This potential evaporation flux is reduced if a portion of thewater is frozen or if the energy for the vaporization has al-ready been used to vaporize water that was intercepted bythe canopy or for plant transpiration (see Sect 26)

Plant transpiration coupled with photosynthesis

Plant transpiration (ET in mm dayminus1) is modelled as thelesser of plant-available soil water supply function (S) andatmospheric demand function (D) following Federer (1982)

ET =min(SD) middot fv (111)

S depends on a PFT-specific maximum water transport ca-pacity (Emax in mm dayminus1) and the relative water content(wr) and phenology (phenPFT)

S = Emax middotwr middot phenPFT (112)

The water accessible for plants (wr) is computed from therelative water content at field capacity (wl) and the fractionof roots (rootdistl) within each soil layer (l) as

nsoilminus1suml= 1

wl middot rootdistl (113)

rootdistl can be calculated from the proportion of roots fromsurface to soil depth z rootdistz as in Jackson et al (1996)

rootdistz =

int z0 (βroot)

zprimedzprimeint zbottom0 (βroot)z

primedzprime=

1minus (βroot)z

1minus (βroot)zbottom (114)

where βroot represents a numerical index for root distribution(for parameter values see Table S8) and rootdistl is given bythe difference rootdistz(l)minus rootdistz(lminus1) If the soil depth oflayer l is greater than the thawing depth then rootdistl is set tozero The non-zero rootdistl values are rescaled in such a waythat their sum is normalized to 1 considering the reallocationof the root distribution under freezing conditions

Plants in natural vegetation compete for water resourcesand thus only have access to the fraction of water that corre-sponds to their foliage projected cover (FPCPFT)

SPFT = S middotFPCPFT (115)

For agricultural crops water supply is also dependent ontheir root biomass bmroot

S = Emax middotwr middot (1minus exp(minus00411 middot bmroot)) (116)

Atmospheric demand (D) is a hyperbolic function of gc(see Sect 221 and Eq 39) following Monteith (1995)and employs a maximum PriestleyndashTaylor coefficient αm =

1391 describing the asymptotic transpiration rate and a con-ductance scaling factor gm = 326

D = (1minuswet) middotEeq middotαm(1+ gmgc) (117)

where ldquowetrdquo is the fraction of Eeq that was used to vaporizeintercepted water from the canopy (see Sect 26) and gc isthe potential canopy conductance If S is not sufficient to ful-fil transpiration demand gc is recalculated for D = S and thephotosynthesis rate might be adjusted (see Sect 221)

263 River routing

Description of the river-routing module

The river-routing module computes the lateral exchangeof discharge (see Sect 312 for input) between grid cellsthrough the river network (Rost et al 2008) The transportof water in the river channel is approximated by a cascade oflinear reservoirs River sections are divided into n homoge-neous segments of lengthL each behaving like a linear reser-voir Following the unit hydrograph method (Nash 1957)

the outflow Qout(t) of a linear reservoir cascade for an in-stantaneous inflow Qin is given as

Qout(t)=Qin middot1

K middot0(n)

)nminus1

middot exp(minustK) (118)

where 0(n) is the gamma function that replaces (nminus 1) toallow for non-integer values of n K is the storage parame-ter defined as the hydraulic retention time of a single linearreservoir segment of length L It can be calculated as the av-erage travel time of water through a single river segment

v (119)

where v is the average flow velocityThe river routing in LPJmL4 is calculated at a time step

of 1t = 3 h We assume a globally constant flow velocity vof 1 m sminus1 and a segment length L of 10 km to calculate theparameters n and K for each route between grid cell mid-points At the start of the simulation for each route the unithydrograph for a rectangular input impulse of length 1t isdetermined from Eq (118) Because Eq (118) assumes aninstantaneous input impulse we numerically determine theresponse to a rectangular input impulse by adding up theresponses of a series of 100 consecutive instantaneous in-put impulses From the obtained unit hydrograph the sumof outflow during each subsequent time step 1t is recordeduntil 99 of the total input impulse has been released (max-imum 24 time steps) During simulation the thus determinedresponse function is then used to calculate the convolutionintegral for the flow packages routed through the networkAn efficient parallelization of the river-routing scheme usingglobal communicators of the MPI message-passing library isdescribed in Von Bloh et al (2010)

264 Irrigation and dams

LPJmL4 explicitly accounts for human influences on the hy-drological cycle by accounting for irrigation water abstrac-tion consumption and return flows and non-agriculturalwater consumption from households industry and livestock(HIL) as well as an implementation of reservoirs and dams

Irrigation

LPJmL4 features a mechanistic representation of the worldrsquosmost important irrigation systems (surface sprinkler drip)which is key to refined global simulations of agricultural wa-ter use as constrained by biophysical processes and watertrade-offs along the river network HIL water use in each gridcell is based on Floumlrke et al (2013) (accounting for 201 km3

in the year 2000) We assume HIL water to be withdrawnprior to irrigation water LPJmL4 comes with the first inputdataset that details the global distribution of irrigation typesfor each cell and crop type (Jaumlgermeyr et al 2015) Irriga-tion water partitioning is dynamically calculated in coupling

to the modelled water balance and climate soil and vege-tation properties The spatial pattern of improved irrigationefficiencies is presented in Jaumlgermeyr et al (2015)

Irrigation water demand is withdrawn from available sur-face water ie river discharge (see Sect 26) lakes andreservoirs (Sect 265) and if not sufficient in the respec-tive grid cell requested from neighbouring upstream cellsThe amount of daily irrigation water requirements is basedon the soil water deficit resulting crop water demand (netirrigation requirements NIR) and irrigation-system-specificapplication requirements (specified below) If soil moisturegoes below the CFT-specific irrigation threshold (it) the to-tal amount (daily gross irrigation requirements) is requestedfor abstraction NIR is defined as the water needs of the top50 cm soil layer to avoid water limitation to the crop It iscalculated to meet field capacity (Wfc) if the water supply(root-available soil water) falls below the atmospheric de-mand (potential evapotranspiration see Sect 26) as

NIR=Wfcminuswaminuswice NIRge 0 (120)

where wa is actually available soil water and wice the frozensoil content in millimetres Due to inefficiencies in any irri-gation system excess water is required to meet the water de-mand of the crop To this end we calculate system-specificconveyance efficiencies (Ec) and application requirements(AR) which lead to gross irrigation requirements (GIRs inmm)

GIR=NIR+ARminusStore

Ec (121)

where ldquoStorerdquo stands in as a storage buffer (see also Fig S2for a conceptual description) For pressurized systems (sprin-kler and drip) Ec is set to 095 For surface irrigation welink Ec to soil saturated hydraulic conductivity (Ks seeSect 26) adopting Ec estimates from Brouwer et al (1989)Half of conveyance losses are assumed to occur due toevaporation from open water bodies and the remainder isadded to the return flow as drainage Indicative of applica-tion losses AR represents the excess water needed to uni-formly distribute irrigation across the field AR is calculatedas a system-specific scalar of the free water capacity

AR= (WsatminusWfc) middot duminusFW ARge 0 (122)

where du is the water distribution uniformity scalar as a func-tion of the irrigation system and FW represents the availablefree water (see Sect 26 and 262 see Jaumlgermeyr et al 2015for details)

Irrigation scheduling is controlled by Pr and the irrigationthreshold (IT) that defines tolerable soil water depletion priorto irrigation (see Table S14) Accessible irrigation water issubtracted by the precipitation amount Irrigation water vol-umes that are not released (if S gt IT) are added to Store andare available for the next irrigation event Withdrawn irriga-tion volumes are subsequently reduced by conveyance losses

Irrigation water application is assumed to occur below thecanopy for surface and drip systems and above the canopy forsprinkler systems which leads to interception losses (calcu-lated as described above) Drip systems are assumed to applyirrigation water localized to the plant root zone below the sur-face and thereby reduce soil evaporation by 60 (Sect 26)Note that drip systems are parameterized to represent a mod-est form of deficit irrigation ie to save water and not tomaximize yields For detailed parameterization of the threeirrigation systems implemented see Table 14 and Jaumlgermeyret al (2015)

265 Dams lakes and reservoirs

The operation of large reservoirs affects the seasonal dis-charge patterns downstream of the dam and the amount ofwater that is locally available for irrigation In LPJmL4reservoirs are considered starting from the prescribed yearthey were built (Biemans et al 2011) The reservoir is filleddaily with discharge from upstream locations and with localprecipitation At the beginning of an operational year whichis defined as the first month when mean monthly inflow islower than mean annual inflow the actual storage in the reser-voir is compared with the maximum storage capacity of thereservoir The reservoir outflow factor of the following yearis adjusted accordingly to compensate for inter-annual flowfluctuations Subsequently a target release is defined basedon the main purpose of the reservoir Dams built primarily forirrigation are assumed to release their water proportionally togross irrigation water demand downstream Dams built pri-marily for other purposes (hydropower flood control etc)are assumed to be designed for releasing a constant watervolume throughout the year The actual release from a reser-voir is simulated to depend on its storage capacity relative toits inflow If an irrigation purpose is defined for the reservoirpart of the outflow is diverted to irrigated lands downstreamCells receive water from the reservoirs when the followingconditions are met the cells have a lower altitude than thecell containing the reservoir and they are situated along themain river downstream or at maximum five cells upstreamThus a cell can receive water from multiple reservoirs Asirrigation demands vary daily water released from reservoirscan be stored in the conveyance system for up to 5 days Ifthe total irrigation water demand to a reservoir cannot be ful-filled all requesting fields are supplied with the same fractionof their demand (see Biemans et al 2011 for details)

27 Land use

Human land use is represented in LPJmL4 by dividing gridcells that have the same climate and soil texture input intoseparate subunits referred to as stands Stands are driven bythe same input data but changes in soil water and soil car-bon are computed separately When new stands are createdtheir soils are direct copies of the stand from which they

are generated If stands are merged soil properties are av-eraged according to the two standsrsquo size to maintain massand energy balance Natural vegetation (ie PFTs) agricul-tural crops (ie CFTs) managed grasslands and bioenergyplantations are represented on separate stands that can partlyor fully cover any grid cell The size of each stand is deter-mined by the extent of land use defined by the input dataprescribing fractions of each land use type (crops managedgrassland bioenergy plantations all as rain-fed andor irri-gated cultivation) All natural vegetation grows on a singlenatural vegetation stand on which all present PFTs competefor water and light Agricultural crops are implemented asmonocultures in which only one single crop is cultivated andthere is no competition for resources with other stands (fieldsor natural vegetation) within that cell For each crop and irri-gation system (irrigated or rain-fed) there can always only beone stand within one grid cell For irrigated crops only oneirrigation system (sprinkler surface or drip see Sect 26)can be selected At the beginning of each simulation yearpresent total agricultural land (all crops all managed grass-lands) is compared with land requirements for the year ac-cording to the input data If there is too little agriculturalland available the needed fraction is cleared (liberated) fromthe natural vegetation stand if there is too much the excessagricultural land is abandoned and merged into the naturalvegetation stand leaving new space there for the establish-ment of natural vegetation Uncultivated cropland (ie out-side the cropping period) is merged in set-aside stands sep-arated into irrigated and rain-fed to prevent irrigation waterfrom irrigated stands from transferring to rain-fed stands inoff seasons Set-aside land from irrigated agriculture is not ir-rigated during fallow periods but is kept separate because ofthe soil water content that is enhanced through irrigation dur-ing the growing period Land can be transferred between thetwo set-aside stands if the ratio between irrigated and rain-fed cropland changes Depending on the scenario setting ifintercropping is assumed a simple intercrop (grass) can begrown on the set-aside stand during the fallow period Oncea sowing date for a crop is reached (see Sect 271) the pre-scribed fraction of that crop and irrigation system is removedfrom the set-aside stand by copying the soil properties of theset-aside stand to the newly created stand and reducing theset-aside stand size accordingly The crop is then cultivatedon that newly created stand and returned to the set-aside uponharvest of the crop

271 Sowing dates

Sowing dates are simulated based on a set of rules depend-ing on climate- and crop-specific thresholds as described inWaha et al (2012) The start of the growing period is as-sumed to depend either on the onset of the wet season in trop-ical and subtropical regions or on the exceeding of a crop-specific temperature threshold for emergence in temperateregions We describe the intra-annual variability of precipi-

tation and temperature in each location using variation coef-ficients for temperature (CVtemp) and precipitation (CVprec)calculated from past monthly climate data We assume tem-perature seasonality if CVtemp exceeds 001 and precipitationseasonality if CVprec exceeds 04 Hence four seasonalitytypes can be differentiated (Fig S3)

1 temperature seasonality

2 precipitation seasonality

3 temperature and precipitation seasonality and

4 no temperature and no precipitation seasonality

For locations with a combined temperature and precipita-tion seasonality we additionally consider the mean temper-ature of the coldest month If it exceeds 10 C we assumethe absence of a cold season ie the risk of frost occurrenceis negligible assuming temperatures are high enough to sowall year-round Accordingly precipitation seasonality definesthe timing of sowing If the mean temperature of the coldestmonth is equal to or below 10 C temperature seasonalitydetermines the timing of sowing In regions with precipita-tion seasonality only the sowing date is at the onset of themain wet season The precipitation to potential evapotranspi-ration ratio is used to find moist and dry months in a year assuggested by Thornthwaite (1948) The main wet season isidentified by the largest sum of monthly precipitation to po-tential evapotranspiration ratios of 4 consecutive months be-cause the length of that period aligns well with the length ofthe growing period of the majority of the simulated crops Inregions with bimodal rainfall patterns the wet season startswith the first month of the longest wet season Crops aresown at the first wet day in the main wet season of the yearie when daily precipitation exceeds 01 mm The onset ofthe growing period depends on temperature if temperatureseasonality is detectable Accordingly crop emergence is re-lated to temperature and thus sowing starts when daily aver-age temperature exceeds a certain threshold (Tfall or TspringTable S10) Locations without any temperature or precipita-tion seasonality eg in the wet tropics crops are sown on1 January These assumptions lead to a possible adaptationof projected sowing dates

272 Management and cropping intensity

Agricultural management is represented as a distinct set ofoptions and a calibration of cropping intensity Explicit man-agement options include

1 cultivar choices (Sect 24)

2 sowing dates (Sect 271)

3 irrigation shares and type (Sect 26)

4 residue removal (Sect 24) and

5 intercrops (Sect 27)

Different irrigation systems can be represented as followsFor drip irrigation systems assuming localized subsurfacewater application soil evaporation is reduced so that only40 of the applied irrigation water is available for evapo-ration Also for rainwater management (see Sect 26) soilevaporation can be reduced mimicking agricultural man-agement systems like mulching techniques or conservationtillage (Jaumlgermeyr et al 2016) Secondly to simulate im-proved rainwater management (see Sect 26) the infiltrationcapacity can be increased mimicking agricultural manage-ment practices such as different tillage systems or organicmulching (Jaumlgermeyr et al 2016)

Other than that management options are not treated ex-plicitly in the LPJmL4 model that is it assumes no nutri-ent limitation to crop growth Current management patternswhich are desirable for eg studies of the carbon or water cy-cle can be represented by calibrating national cropping in-tensity to FAO statistics as described in Fader et al (2010)For this the maximum leaf area index (LAImax) the harvestindex parameter (HIopt) and a scaling factor for scaling leaf-level photosynthesis to stand level (αa Haxeltine and Pren-tice 1996) are scaled in combination LAImax can range be-tween 1 (lowest intensity) and 5 for maize or 7 for all othercrops (highest intensity) and αa ranges from 04 to 10 Theparameter HIopt is crop specific (see Table 11) and can be re-duced by up to 20 assuming that there are more robust butless productive varieties (Gosme et al 2010)

273 Managed grassland

On managed grassland stands only herbaceous PFTs (TrHTeH PoH see Appendix A for definition) can establish con-ditional to their bioclimatic limits (see Table S4) If morethan one herbaceous PFT establishes these compete for lightand water resources but do not interact with other stands inthat grid cell In contrast to annual C allocation as describedabove (see Sect 231) LPJmL4 simulates managed grass-lands and herbaceous biomass plantations (see Sect 274)with a daily allocation and turnover scheme in which it dy-namically computes leaf biomass per day as described belowIt therefore enables us to better represent the current pheno-logical state and suitable times for harvest

Daily allocation of managed grasslands

The allocation scheme is designed to distribute the dailybiomass increment to leaf and root biomass in a way that bestfulfils the predetermined ratio of leaf to root mass lr for thewhole plant This allows for short-term deviations from theallowed leaf to root mass ratio lr after harvest events whenmuch of the leaf biomass is removed After a harvest eventNPP is first allocated to leaves until lr is restored If moreCO2 is assimilated than needed for maintenance respiration(ie NPP is positive) assimilated carbon BI is allocated to

the root (R) and the leaf carbon pool (L) by calculating therespective increments (LIRI) (Eqs 123 and 124)

LI =min(BI max

(BI+RminusLlr

1+ 1lr 0))

RI = BIminusLI (124)

In the case of negative NPP (ie maintenance and growthrespiration are larger than the GPP of that day) both com-partments (leaves and roots) are reduced proportionally lris scaled with a measure of average growing-season waterstress (mean of daily ratios of plant water supply to atmo-spheric demand see Sect 262) to account for the functionalrelationship that plants allocate more carbon to roots underdry conditions

lr= lrp middotWsupplyWdemand (125)

Grassland harvest routine

LPJmL4 employs a default harvest scheme that attemptsto approximate the actual global grassland production of23 Gt DM (Herrero et al 2013) while avoiding degradationA harvest event of grass biomass occurs when leaf biomassincreases over the previous month Prior to the harvest eventgrass leaf biomass (Cleaf) and the biomass after the last har-vest event (MCleaf) is summed up for all grass species at themanaged grassland stand

Cleaf =sumPFT

BmlPFT

MCleaf =sumPFT

MmlPFT (126)

On the last day of each month harvest occurs and the har-vest indexHfrac is determined depending on the leaf biomassCleaf

If day (Cleaf gtMCleaf) Hfrac = 1minus1000

1000+Cleaf(127)

Harvested biomass is taken from the leaf biomass of eachherbaceous PFT Depending on the amount of carbon in theleaves the harvested fraction increases (Fig S4) and biomassharvested depends on the present leaf carbon In the absenceof any detailed information about actual grassland manage-ment systems this generic harvest routine does not representspecific management systems but allows for the simulationof regular harvest events (be it grazing or mowing) duringproductive periods of the year and the harvest amount auto-matically adjusts to productivity

274 Biomass plantations

Three biomass functional types (BFTs) were implementedin LPJmL4 (one fast-growing C4 grass and a temperate and atropical tree) to allow for the simulation of dedicated biomass

plantations (Beringer et al 2011) These BFTs are genericrepresentations of some of the most promising types of cropsfor the production of second-generation biofuels biomateri-als or energy (possibly in combination with carbon captureand storage mechanisms) Their parameterization is partlyidentical to their natural PFT equivalents tropical C4 peren-nial grass temperate broadleaved summergreen tree andtropical broadleaved raingreen tree yet with some importantmodifications to characterize the enhanced growth charac-teristics of these managed vegetation types (see Table S12)Woody energy crops are represented as short-rotation cop-pice systems (SRCs) In short intervals young tree stems arecut down to near ground stumps implemented as regularlycycles (see Table S13) A grown root system and nutrientstorage in roots and stumps enables high-yielding varieties ofpoplar willow and eucalyptus used for SRC to regrow force-fully in renewal years Until plantations need to replantedafter 40 years several harvest cycles are possible Under ef-fective pest and fire control on modern biomass plantationsmortality and fire occurrence are reduced (to zero emissions)compared to natural vegetation

3 Modelling protocol

The objective of this publication is to provide a comprehen-sive description of the LPJmL4 model Here we also providesome outputs from a standard simulation of the historic pe-riod 1901 to 2011 which is also the basis for the actual andmore comprehensive model evaluation described in the com-panion paper (Schaphoff et al 2018c)

31 Model set-up and inputs

For this simulation all carbon and water pools in the modelare initialized to zero and a spin-up simulation for 5000 yearsis conducted in which plants dynamically establish growand die following the model dynamics described above Afterthe soil carbon equilibrium phase of 2310 simulation years(see Sect 25) equilibrium soil carbon pool sizes are esti-mated and corrected depending on organic matter input andthe mean decomposition rate in each grid cell after another2690 spin-up simulation years all carbon pools have reacheda dynamic equilibrium In this phase LPJmL4 simulates onlynatural vegetation For the spin-up simulation we cyclicallyrepeat the first 30 years of climate data input and prescribeatmospheric carbon dioxide concentrations at 278 ppm Dur-ing the second phase of the spin-up simulation land use isintroduced in the year 1700 from which it is updated an-nually according to the historic land use dataset (see Faderet al 2010 and Sect 312)

311 Climate river routing and soil inputs

We use monthly climate data inputs on precipitation providedby the Global Precipitation Climatology Centre (GPCC Full

Data Reanalysis version 70 Becker et al 2013) dailymean temperature from the Climatic Research Unit (CRUTS version 323 University of East Anglia Climatic Re-search Unit 2015 Harris et al 2014) shortwave downwardradiation and net downward longwave radiation reanalysisdata from ERA-Interim (Dee et al 2011) and number ofwet days per month derived synthetically as suggested byNew et al (2000) which is used to allocate monthly pre-cipitation to individual days Precipitation is stochasticallydisaggregated while preserving monthly sum and temper-ature is linearly interpolated (Gerten et al 2004) Besidesclimate information the model is forced with invariant infor-mation on the soil texture (FAOIIASAISRICISSCASJRC2012 Nachtergaele et al 2009) and annual information onland use from Fader et al (2010) but now also explicitlydescribes sugar cane areas (see Sect 312) For the SPIT-FIRE module LPJmL4 uses additional input Dew-pointtemperature is approximated from daily minimum temper-ature (Thonicke et al 2010) Monthly average wind speedsare based on NCEP reanalysis data which were regriddedto CRU (NOAA-CIRES Climate Diagnostics Center Boul-der Colorado USA Kalnay et al 1996) Atmospheric CO2concentrations are used from the Mauna Loa station (Tansand Keeling 2015) For the transport directions we use theglobal (05times 05) simulated Topological Network (STN-30) drainage direction map (Vorosmarty and Fekete 2011)STN-30 organizes the Earthrsquos land area into drainage basinsand provides the river network topology under the assump-tion that each grid cell can drain into one of the eight next-neighbour cells detailed information on water reservoirs isobtained from the GRanD database (Lehner et al 2011) in-cluding information on storage capacity total area and mainpurpose Natural lakes are obtained from Lehner and Doumlll(2004) A complete overview of all inputs used here is givenin Table S2

312 Land use input

In principle LPJmL4 can be driven by any land use data in-formation As the default land use input file the cropping ar-eas for each of the CFTs are taken from MIRCA2000 (Port-mann et al 2010) which is a combination of crop-specificareas from Monfreda et al (2008) and areas equipped forirrigation from 1900ndash2005 from Siebert et al (2015) Mon-freda et al (2008) defined 175 crops and this number wasreduced to 26 in MIRCA2000 (therefore the group pulsesconsist of 12 individual crops) These land use patterns thathave been derived from maximum monthly growing areasper crop and grid cell have been combined and if these areasadd up to more than 1 ie when sequential cropping systemsare present the total cropland fraction was reduced to not ex-ceed the physical land area in each pixel A more detailed de-scription of the procedure can be found in Fader et al (2010)After the implementation of sugar cane as a 12th annualcrop that is explicitly represented in LPJmL4 (Lapola et al

2009) the standard land use input dataset was amended bysubtracting the sugar cane areas from the ldquoothersrdquo band andimplementing it as a separate input data band All 16 inputdata bands (CFTs 1ndash12 others managed grassland bioen-ergy grass and bioenergy trees) are included four times inthe dataset with the first 16 bands representing purely rain-fed agricultural areas and the second third and fourth setrepresenting irrigated areas of these land use types for sur-face sprinkler and drip irrigation respectively

32 Standard outputs

The multiple aspects of the terrestrial biosphere and hydro-sphere that are implemented in LPJmL4 allow for an assess-ment of multiple processes from natural and managed landwhich span ecological hydrological and agricultural com-ponents The consistent single modelling framework allowsfor an analysis of interactions among these multiple sectorsfrom local to global scale spanning seasons to centuries Alist of the key parameters calculated by LPJmL4 is shown inTable S3

Being driven by climate and land use data LPJmL4 canbe applied to quantify both climatic and anthropogenic im-pacts on the terrestrial biosphere Computed dynamics ofbiogeochemical and hydrological processes thus arise fromvegetation dynamics in natural ecosystems under climatechange and elevated atmospheric CO2 concentrations landuse change and climate- and management-driven changes inmanaged ecosystems Each grid cell can be dominated bymanaged land (croplands and pastures) but still contain frac-tions of natural vegetation and vice versa We here apply theanthromes concept (see also Ellis et al 2010) to illustrate theglobal distribution of natural vegetation and managed landas simulated by LPJmL4 (Fig 2) We use simulated vege-tation carbon potential evapotranspiration foliar projectivecover for each PFT managed grassland and CFT and com-bine these with climate input data to map natural biomes andanthromes at the global scale (see Boit et al 2016 for thealgorithm description) The composition of natural ecosys-tems is dynamically computed by LPJmL4 as the differentPFTs compete with each other Bounded by the bioclimaticlimits the modelled global distribution of forests shrublandand natural grasslands and the spatial extent of polar andalpine ecosystems and deserts are in broad agreement withthe biomes identified by Olson et al (2001) The integratedmapping of biomes and anthromes underlines the extent ofanthropogenic impacts on the terrestrial biosphere and howmuch of the potential vegetation coverage is left By apply-ing the land use input (see Sect 312) LPJmL4 simulatescropland in 27 and pasture in 16 of the ice-free globalland area Simulating biophysical and biochemical processesin densely populated or urban areas could be considered infuture model developments of LPJmL4 given their large spa-tial extent (Ellis et al 2010)

AnthromesClosed moist forestOpen moist forestShrublandSavannagrasslandTemperate forestDesertCroplandPastureBoreal forestTundraArtic desert

Figure 2 Global anthrome classes in the potential natural vegetation and agricultural areas Anthrome classes are defined using coveragewith types of dominating natural vegetation (eg tropical forest) dominant agricultural usage (eg cropland) and external drivers (egtemperature)

Carbon and water fluxes productivity and the harvest ofcrops and managed grasslands have also been quantifiedResults show that soils are the largest carbon pool of theterrestrial biosphere with maximum amounts of more than60 kg C mminus2 in the boreal zone most notably in permafrostsoils (see Fig 3a and b) Vegetation carbon pools are largestin the tropics with almost 20 kg C mminus2 and in the temper-ate zone with about 8 kg C mminus2 The large vegetation car-bon pools are a result of high net primary productivity (NPP)in tropical and subtropical ecosystems which process about1000 to 1200 g C mminus2 aminus1 respectively (see Fig 4b) Firecarbon emissions are highest in the tropics as a result ofhigh ignition probability and high biomass values simulated(Fig 4c) Crop productivity is determined by climatic condi-tions and management strategies and is currently highest inthe temperate zones of North America and Europe but alsoin regions in eastern China the irrigated Ganges Valley inIndia and temperate South America (see Fig 3c)

Over the 20th century changes in climate land use andatmospheric CO2 concentrations had distinct effects on theterrestrial carbon stocks and fluxes Global vegetation car-bon declined by 20 Pg C after 1940 and has been rising againsince 2005 whereas carbon stored in soil and litter increasedconstantly over the simulation period (see Fig S5) GPPNPP and heterotrophic respiration also follow this trend andshow considerable inter-annual variability while fire-relatedcarbon emissions declined in the 1970s and remained rel-atively stable thereafter Interception and run-off also showa positive trend while evaporation from bare soil decreased(see Fig S5)

4 Discussion

Previous versions of LPJmL4 were used in a large numberof applications to evaluate vegetation water and carbon dy-namics under current and future climate and land use changeIn total almost 100 papers were published since 2007 whichcover a wide range of model developments and process anal-yses (see references in Table S1) including 18 studies thatdescribe significant model developments The majority ofthe studies deal with modelling human land use with a fo-cus on different crop types (N = 54 studies) managed grass-lands (N = 21) and agricultural water use (N = 18) Mostwere conducted at global scale (N = 58) but also at regionalscales mainly for Europe Africa and Amazonia Many stud-ies (N = 43) investigate the potential future impacts of cli-mate change while the remaining studies evaluate the ef-fects of current or historic climates In the following wewill highlight some of the most important previous publica-tions using LPJmL by describing the most prominent fieldsof model application An important field of LPJmL modelapplication testing and subsequent development is the anal-ysis of historic events LPJmL simulation results contributedto the analysis of extreme event impacts on the biosphereglobally (Zscheischler et al 2014a b) and at pan-Europeanscale (Rammig et al 2015 Rolinski et al 2015) In com-bination with remote sensing and eddy flux data LPJmLhas been applied to estimate ecosystem respiration (Jaumlger-meyr et al 2014) and productivity (eg Jung et al 2008)The increasing trend in atmospheric CO2 amplitude couldbe explained by increasing productivity in subarctic and bo-real forest ecosystems and less so by increases in agricul-tural land and productivity (Forkel et al 2016) as simulated

04 6 11 18 48 70

LPJmL4 soil carbon [kgC mminus2]

03 1 45 168

LPJmL4 vegetation carbon [kgC mminus2]

2e+09 7e+10 5e+11 4e+12

LPJmL4 crop production [ kcal yrminus1]

Figure 3 Soil (a) and vegetation (b) carbon pool and cumulativecrop production (c) computed by LPJmL4 as an average of the timeperiod 1996ndash2005 Note values are plotted on a logarithmic scale

0 500 1000 1500 2000

LPJmL4 gross primary production [gC mminus2 yrminus1]

0 200 400 600 800 1000

LPJmL4 net primary production [gC mminus2 yrminus1]

2 8 25 80 107

LPJmL4 fire carbon emissions [gC mminus2 yrminus1]

Figure 4 Annual GPP (a) NPP (b) and fire carbon emissions (c)computed by LPJmL4 as an average of the time period 1996ndash2005Note values for fire carbon emissions are plotted on a logarithmicscale

by building on the improved phenology scheme by Forkelet al (2014) The coupling of LPJmL to the climate modelSPEEDY allowed for an investigation of climatendashvegetationfeedbacks from land use change (eg Strengers et al 2010Boisier et al 2012)

Studies on agricultural water consumption (Rost et al2008) and virtual water contents and water footprints forcrops (Fader et al 2010 2011) have contributed to illumi-nating the role of agriculture in human water consumption

The majority of LPJmL applications address the impactsof climate and land use change on various biogeochemicaland ecosystem properties of the terrestrial biosphere To eval-uate the impacts of climate change on ecosystem processesand the carbon cycle at the global scale Heyder et al (2011)performed a risk analysis of terrestrial ecosystems based onan integrated metric that considered joint changes in car-bon and water fluxes carbon stocks and vegetation struc-ture Applying CMIP3 and CMIP5 climate change scenariosto LPJmL severe impacts for the terrestrial biosphere werefound when global warming levels exceed 3 K local temper-ature in cold and tropical biomes and 4 K in the temperatebiome (Heyder et al 2011 Ostberg et al 2013) The cou-pling of LPJmL to the integrated assessment model IMAGEallowed researchers to evaluate feedbacks between land usechange the carbon balance and climate change so that thedynamics of a potential reversal of the terrestrial carbon bal-ance could be assessed (Muumlller et al 2016) Also economicfeedbacks of agricultural production were evaluated by cou-pling LPJmL to the agro-economic model MAgPIE (Lotze-Campen et al 2008) for example measures of land use pro-tection for climate change mitigation (Popp et al 2014)

Regional climate change applications investigated the roleof CO2 fertilization in Amazon rainforest stability (Rammiget al 2010) and analysed additional threats arising fromtropical deforestation (Gumpenberger et al 2010 Poulteret al 2010) Boit et al (2016) applied the anthromes conceptto LPJmL simulation results to differentiate the relative im-portance of future climate vs land use change in Latin Amer-ica In that study land use change was identified as the maindriver of biome shifts and biome degradation early in the 21stcentury while climate change impacts will dominate the sec-ond half LPJmL simulations also showed that in boreal for-est and Arctic ecosystems 100 years of future climate changemight destabilize the carbon stored in permafrost soils overseveral centuries due to feedbacks between permafrost anddynamic vegetation (Schaphoff et al 2013)

Several studies used LPJmL to investigate water limita-tions in natural ecosystems (Gerten et al 2013) and foodproduction in particular (eg Gerten et al 2008 Biemanset al 2011 2013) Gerten et al (2011) applied LPJmL toquantify ldquogreenrdquo and ldquobluerdquo water requirements for futurefood security finding that water scarcity will increase inmany countries as confirmed by other studies prepared inthe context of multi-model intercomparisons such as ISIMIP(eg Schewe et al 2014) In the context of planetary bound-

aries Gerten et al (2013) and Steffen et al (2015) haveproposed sub-global modifications to the planetary bound-ary for freshwater use and Jaumlgermeyr et al (2017) quanti-fied the therefore needed environmental flow requirementsin view of the Sustainable Development Goals Future cli-mate change effects on irrigation requirements were inves-tigated by Konzmann et al (2013) Jaumlgermeyr et al (2015)and Fader et al (2016) LPJmL simulations have also shown(Asseng et al 2015 Muumlller and Robertson 2014 Rosen-zweig et al 2014) that climate change constitutes a majorthreat to agricultural productivity especially in the tropicsbut with large uncertainties regarding the benefits from ele-vated atmospheric CO2 on crop water use and productivity(Muumlller et al 2015 Deryng et al 2016) It was also shownthat the potential for major cereal crop production may de-cline in a future climate (Pugh et al 2016b) Building on thedevelopment of bioenergy plantations (Beringer et al 2011)LPJmL was applied to analyse synergies and trade-offs ofbiomass plantations finding that demands for future bioen-ergy potentials implicit in climate mitigation and climate en-gineering portfolios could only be met at substantial envi-ronmental costs (Boysen et al 2016 Heck et al 2016) Thepresent publication describing the LPJmL4 model code andthe companion paper (Schaphoff et al 2018c) provide a thor-ough model evaluation and are intended to serve as a com-prehensive description of the current LPJmL4 model Themodel code will be published under an open source licenceat httpsgitlabpik-potsdamdelpjmlLPJmL We hope thatthis will help to promote further development and improve-ment of LPJmL4 and foster high-profile research in areassuch as multi-sectoral climate change impacts Earth systemdynamics planetary boundaries and SDGs Besides the on-going implementation of the dynamics of major nutrientssuch as nitrogen there are new plant physiological insightsthat have not yet found their way into LPJmL4 or relatedmodels (Pugh et al 2016a Rogers et al 2017) The con-sistent coverage of natural and managed ecosystems as wellas the full carbon and water dynamics linked by vegetationdynamics and land use management is central to the furtherdevelopment of LPJmL4

Code and data availability The model code of LPJmL4 is publiclyavailable through PIKrsquos gitlab server at httpsgitlabpik-potsdamdelpjmlLPJmL and an exact version of the code described hereis archived under httpsdoiorg105880pik2018002 and shouldbe referenced as Schaphoff et al (2018b) The output data fromthe model simulations described here are available at the GFZ DataServices under httpsdoiorg105880pik2017009 and can be ref-erenced as Schaphoff et al (2018a)

Appendix A Abbreviation of PFTs BFTs and CFTs

Tropical broadleaved evergreen tree TrBETropical broadleaved raingreen tree TrBRTemperate needle-leaved evergreen tree TeNETemperate broadleaved evergreen tree TeBETemperate broadleaved summergreen tree TeBSBoreal needle-leaved evergreen tree BoNEBoreal broadleaved summergreen tree BoBSBoreal needle-leaved summergreen tree BoNSTropical herbaceous TrHTemperate herbaceous TeHPolar herbaceous PoHBioenergy tropical tree BTrTBioenergy temperate tree BTeTBioenergy C4 grass BGrC4Temperate cereals TeCerRice RiceMaize MaizeTropical cereals TrCerPulses PulTemperate roots TeRoTropical roots TrRoSunflower SunFlSoybean SoyGroundnut GrNuRapeseed RapeSugar cane SuCa

The Supplement related to this article is availableonline at httpsdoiorg105194gmd-11-1343-2018-supplement

Competing interests The authors declare that they have no conflictof interest

Acknowledgements This study was supported by the GermanFederal Ministry of Education and Research (BMBF) projectldquoPalMod 23 Methankreislauf Teilprojekt 2 Modellierung derMethanemissionen von Feucht- und Permafrostgebieten mit Hilfevon LPJmLrdquo (FKZ 01LP1507C) Anja Rammig and Jens Heinkeacknowledge funding from the Helmholtz Alliance ldquoRemoteSensing and Earth System Dynamicsrdquo We thank the ClimaticResearch Unit for providing global gridded temperature input theGlobal Precipitation Climatology Centre for providing precipitationinput and the coordinators of ERA-Interim for providing short-wave downward radiation and net downward longwave radiationFurthermore we thank the authors of MIRCA2000 for providingland use input We thank Jannes Breier and Nele Steinmetz foreditorial help and Kirsten Elger for her great support in archivingdata and the LPJmL4 code We thank two anonymous reviewers fortheir helpful comments on earlier versions of the paper Finally themany people involved in the development testing and applicationof LPJmL4 and its predecessors are gratefully acknowledged

Edited by Julia HargreavesReviewed by two anonymous referees

References

Ahlstroumlm A Xia J Arneth A Luo Y and Smith B Impor-tance of vegetation dynamics for future terrestrial carbon cyclingEnviron Res Lett 10 054019 httpsdoiorg1010881748-9326105054019 2015

Allen C D Macalady A K Chenchouni H Bachelet D Mc-Dowell N Vennetier M Kitzberger T Rigling A Bres-hears D D Hogg E T Gonzalez P Fensham R ZhangZ Castro J Demidova N Lim J-H Allard G Run-ning S W Semerci A and Cobb N A global overview ofdrought and heat-induced tree mortality reveals emerging cli-mate change risks for forests Forest Ecol Manag 259 660ndash684 httpsdoiorg101016jforeco200909001 2010

Arnold J G Williams J R Nicks A D and Sammons N BSWRRB A Basin Scale Simulation Model for Soil and WaterResources Management Texas AampM University Press CollegeStation Texas 1990

Asseng S Brisson N Basso B Martre P Aggarwal P K An-gulo C Bertuzzi P Biernath C Challinor A J Doltra JGayler S Goldberg R Grant R Heng L Hooker J HuntL A Ingwersen J Izaurralde R C Kersebaum K C MuumlllerC Kumar S N Nendel C Leary G O Olesen J E Os-borne T M Palosuo T Priesack E Ripoche D SemenovM A Shcherbak I Steduto P Stoumlckle C Stratonovitch PStreck T Supit I Tao F Travasso M Waha K Wallach D

Williams J R and Wolf J Uncertainty in simulating wheatyields under climate change Nat Clim Change 3 827ndash832httpsdoiorg101038NCLIMATE1916 2013

Asseng S Ewert F Martre P Rotter R P Lobell D B Cam-marano D Kimball B A Ottman M J Wall G W WhiteJ W Reynolds M P Alderman P D Prasad P V V Aggar-wal P K Anothai J Basso B Biernath C Challinor A JDe Sanctis G Doltra J Fereres E Garcia-Vila M GaylerS Hoogenboom G Hunt L A Izaurralde R C JablounM Jones C D Kersebaum K C Koehler A-K Muller CNaresh Kumar S Nendel C OrsquoLeary G Olesen J E Palo-suo T Priesack E Eyshi Rezaei E Ruane A C SemenovM A Shcherbak I Stockle C Stratonovitch P Streck TSupit I Tao F Thorburn P J Waha K Wang E Wal-lach D Wolf J Zhao Z and Zhu Y Rising temperaturesreduce global wheat production Nat Clim Change 5 143ndash147httpsdoiorg101038nclimate2470 2015

Bassu S Brisson N Durand J-L Boote K Lizaso J JonesJ W Rosenzweig C Ruane A C Adam M Baron CBasso B Biernath C Boogaard H Conijn S Corbeels MDeryng D De Sanctis G Gayler S Grassini P HatfieldJ Hoek S Izaurralde C Jongschaap R Kemanian A RKersebaum K C Kim S-H Kumar N S Makowski DMuumlller C Nendel C Priesack E Pravia M V Sau FShcherbak I Tao F Teixeira E Timlin D and Waha KHow do various maize crop models vary in their responses toclimate change factors Glob Change Biol 20 2301ndash2320httpsdoiorg101111gcb12520 2014

Bayazıtoglu Y and Oumlzisik M N Elements of heat transferMcGraw-Hill New York 550 pp ISBN 9781439878910 1988

Becker A Finger P Meyer-Christoffer A Rudolf B SchammK Schneider U and Ziese M A description of the globalland-surface precipitation data products of the Global Precipita-tion Climatology Centre with sample applications including cen-tennial (trend) analysis from 1901ndashpresent Earth Syst Sci Data5 71ndash99 httpsdoiorg105194essd-5-71-2013 2013

Beer C Lucht W Gerten D Thonicke K and Schmul-lius C Effects of soil freezing and thawing on veg-etation carbon density in Siberia A modeling analysiswith the Lund-Potsdam-Jena Dynamic Global VegetationModel (LPJ-DGVM) Global Biogeochem Cy 21 GB1012httpsdoiorg1010292006GB002760 2007

Beringer T Lucht W and Schaphoff S Bioenergy produc-tion potential of global biomass plantations under environmen-tal and agricultural constraints GCB Bioenergy 3 299ndash312httpsdoiorg101111j1757-1707201001088x 2011

Biemans H Haddeland I Kabat P Ludwig F Hutjes RW a Heinke J von Bloh W and Gerten D Impactof reservoirs on river discharge and irrigation water supplyduring the 20th century Water Resour Res 47 W03509httpsdoiorg1010292009WR008929 2011

Biemans H Speelman L Ludwig F Moors E WiltshireA Kumar P Gerten D and Kabat P Future water re-sources for food production in five South Asian river basinsand potential for adaptation ndash A modeling study Changing wa-ter resources availability in Northern India with respect to Hi-malayan glacier retreat and changing monsoon patterns con-sequences and adaptation Sci Total Environ 468ndash469 S117ndashS131 httpsdoiorg101016jscitotenv201305092 2013

Boisier J de Noblet-Ducoudreacute N Pitman A Cruz F Delire Cvan den Hurk B van der Molen M Muumlller C and VoldoireA Attributing the biogeophysical impacts of Land-Use inducedLand-Cover Changes on surface climate to specific causes Re-sults from the first LUCID set of simulations J Geophys Res117 D12116 httpsdoiorg1010292011JD017106 2012

Boit A Sakschewski B Boysen L Cano-Crespo A ClementJ Garcia-alaniz N Kok K Kolb M Langerwisch FRammig A Sachse R van Eupen M von Bloh WClara Zemp D and Thonicke K Large-scale impact ofclimate change vs land-use change on future biome shiftsin Latin America Glob Change Biol 22 3689ndash3701httpsdoiorg101111gcb13355 2016

Bondeau A Smith P Zaehle S Schaphoff S Lucht WCramer W Gerten D Lotze-Campen Hermann Muumlller CReichstein M and Smith B Modelling the role of agri-culture for the 20th century global terrestrial carbon balanceGlob Change Biol 13 679ndash706 httpsdoiorg101111j1365-2486200601305x 2007

Boysen L R Lucht W Gerten D and Heck V Impactsdevalue the potential of large-scale terrestrial CO2 removalthrough biomass plantations Environ Res Lett 11 095010httpsdoiorg1010881748-9326119095010 2016

Brouwer C Prins K and Heibloem M Irrigation Water Manage-ment Irrigation Scheduling Training manual no 4 Tech Rep 4FAO Land and Water Development Division Rome Italy avail-able at httpwwwfaoorgdocrept7202et7202e00htm 1989

Brovkin V van Bodegom P M Kleinen T Wirth C Corn-well W K Cornelissen J H C and Kattge J Plant-driven variation in decomposition rates improves projectionsof global litter stock distribution Biogeosciences 9 565ndash576httpsdoiorg105194bg-9-565-2012 2012

Canadell J G Le Queacutereacute C Raupach M R Field C B Buiten-huis E T Ciais P Conway T J Gillett N P HoughtonR A and Marland G Contributions to accelerating atmo-spheric CO2 growth from economic activity carbon intensityand efficiency of natural sinks P Natl Acad Sci USA 10418866ndash18870 httpsdoiorg101073pnas0702737104 2007

Carvalhais N Forkel M Khomik M Bellarby J Jung MMigliavacca M Mu M Saatchi S Santoro M Thurner MWeber U Ahrens B Beer C Cescatti A Randerson J Tand Reichstein M Global covariation of carbon turnover timeswith climate in terrestrial ecosystems Nature 514 213ndash217httpsdoiorg101038nature13731 2014

Christian H J Blakeslee R J Boccippio D J Boeck W LBuechler D E Driscoll K T Goodman S J Hall J MKoshak W J Mach D M and Stewart M F Global fre-quency and distribution of lightning as observed from space bythe Optical Transient Detector J Geophys Res-Atmos 1084005 httpsdoiorg1010292002JD002347 2003

Collatz G Ball J Grivet C and Berry J A Physiologi-cal and environmental regulation of stomatal conductance pho-tosynthesis and transpiration a model that includes a lam-inar boundary layer Agr Forest Meteorol 54 107ndash136httpsdoiorg1010160168-1923(91)90002-8 1991

Collatz G J Ribas-Carbo M and Berry J A Cou-pled Photosynthesis-Stomatal Conductance Model forLeaves of C4 Plants Funct Plant Biol 19 519ndash538httpsdoiorg101071pp9920519 1992

Cosby B J Hornberger G M Clapp R B and Ginn T RA Statistical Exploration of the Relationships of Soil MoistureCharacteristics to the Physical Properties of Soils Water ResourRes 20 682ndash690 httpsdoiorg101029WR020i006p006821984

Cramer W Bondeau A Woodward F I Prentice I C BettsR A Brovkin V Cox P M Fisher V Foley J A FriendA D Kucharik C Lomas M R Ramankutty N Sitch SSmith B White A and Young-Molling C Global responseof terrestrial ecosystem structure and function to CO2 and cli-mate change results from six dynamic global vegetation modelsGlob Change Biol 7 357ndash373 httpsdoiorg101046j1365-2486200100383x 2001

Dee D P Uppala S M Simmons A J Berrisford P PoliP Kobayashi S Andrae U Balmaseda M A Balsamo GBauer P Bechtold P Beljaars A C M van de Berg I BiblotJ Bormann N Delsol C Dragani R Fuentes M Greer AJ Haimberger L Healy S B Hersbach H Holm E V Isak-sen L Kallberg P Kohler M Matricardi M McNally A PMong-Sanz B M Morcette J-J Park B-K Peubey C deRosnay P Tavolato C Thepaut J N and Vitart F The ERA-Interim reanalysis Configuration and performance of the dataassimilation system Q J Roy Meteorol Soc 137 553ndash597httpsdoiorg101002qj828 2011

Deryng D Elliott J Folberth C Muller C Pugh TA M Boote K J Conway D Ruane A C GertenD Jones J W Khabarov N Olin S Schaphoff SSchmid E Yang H and Rosenzweig C Regional dispar-ities in the beneficial effects of rising CO2 concentrationson crop water productivity Nat Clim Change 6 786ndash790httpsdoiorg101038nclimate2995 2016

Ellis E C Klein Goldewijk K Siebert S Lightman Dand Ramankutty N Anthropogenic transformation of thebiomes 1700 to 2000 Global Ecol Biogeogr 19 589ndash606httpsdoiorg101111j1466-8238201000540x 2010

Fader M Rost S Muumlller C Bondeau A and GertenD Virtual water content of temperate cereals and maizePresent and potential future patterns J Hydrol 384 218ndash231httpsdoiorg101016jjhydrol200912011 2010

Fader M Gerten D Thammer M Heinke J Lotze-Campen HLucht W and Cramer W Internal and external green-blue agri-cultural water footprints of nations and related water and landsavings through trade Hydrol Earth Syst Sci 15 1641ndash1660httpsdoiorg105194hess-15-1641-2011 2011

Fader M von Bloh W Shi S Bondeau A and Cramer WModelling Mediterranean agro-ecosystems by including agricul-tural trees in the LPJmL model Geosci Model Dev 8 3545ndash3561 httpsdoiorg105194gmd-8-3545-2015 2015

Fader M Shi S von Bloh W Bondeau A and CramerW Mediterranean irrigation under climate change more effi-cient irrigation needed to compensate for increases in irriga-tion water requirements Hydrol Earth Syst Sci 20 953ndash973httpsdoiorg105194hess-20-953-2016 2016

FAOIIASAISRICISSCASJRC Harmonized World SoilDatabase (version 12) available at httpwwwiiasaacatResearchLUCExternal-World-soil-databaseHTML 2012

Farquhar G D and von Caemmerer S Modelling of Photosyn-thetic Response to Environmental Conditions in Physiologi-cal Plant Ecology II Water Relations and Carbon Assimilation

edited by Lange O L Nobel P S Osmond C B and ZieglerH Springer Berlin Heidelberg Berlin Heidelberg 549ndash587httpsdoiorg101007978-3-642-68150-9_17 1982

Farquhar G D Caemmerer S V and Berry J A A Biochemi-cal Model of Photosynthetic CO2 Assimilation in Leaves of C3Species Planta 90 78ndash90 httpsdoiorg101007BF003862311980

Federer C A Transpirational Supply and Demand - Plant Soiland Atmospheric Effects Evaluated by Simulation Water ResourRes 18 355ndash362 httpsdoiorg101029WR018i002p003551982

Floumlrke M Kynast E Baumlrlund I Eisner S Wimmer Fand Alcamo J Domestic and industrial water uses of thepast 60 years as a mirror of socio-economic development Aglobal simulation study Global Environ Chang 23 144ndash156httpsdoiorg101016jgloenvcha201210018 2013

Forkel M Carvalhais N Schaphoff S v Bloh W Migli-avacca M Thurner M and Thonicke K Identifyingenvironmental controls on vegetation greenness phenologythrough modelndashdata integration Biogeosciences 11 7025-7050httpsdoiorg105194bg-11-7025-2014 2014

Forkel M Carvalhais N Roumldenbeck C Keeling R HeimannM Thonicke K Zaehle S and Reichstein M En-hanced seasonal CO2 exchange caused by amplified plantproductivity in northern ecosystems Science 351 696ndash699httpsdoiorg101126scienceaac4971 2016

Friedlingstein P Meinshausen M Arora V K Jones C DAnav A Liddicoat S K and Knutti R Uncertainties inCMIP5 Climate Projections due to Carbon Cycle FeedbacksJ Climate 27 511ndash526 httpsdoiorg101175JCLI-D-12-005791 2013

Friend A D Lucht W Rademacher T T Keribin R BettsR Cadule P Ciais P Clark D B Dankers R Fal-loon P D Ito A Kahana R Kleidon A Lomas M RNishina K Ostberg S Pavlick R Peylin P SchaphoffS Vuichard N Warszawski L Wiltshire A and Wood-ward F I Carbon residence time dominates uncertainty interrestrial vegetation responses to future climate and atmo-spheric CO2 P Natl Acad Sci USA 111 3280ndash3285httpsdoiorg101073pnas1222477110 2014

Gelfan A N Pomeroy J W and Kuchment L SModeling Forest Cover Influences on Snow Ac-cumulation Sublimation and Melt J Hydrom-eteor 5 785ndash803 httpsdoiorg1011751525-7541(2004)005lt0785MFCIOSgt20CO2 2004

Gerten D Schaphoff S Haberlandt U Lucht W and Sitch STerrestrial vegetation and water balancendashhydrological evaluationof a dynamic global vegetation model J Hydrol 286 249ndash270httpsdoiorg101016jjhydrol200309029 2004

Gerten D Rost S von Bloh W and Lucht W Causes of changein 20th century global river discharge Geophys Res Lett 351ndash5 httpsdoiorg1010292008GL035258 2008

Gerten D Heinke J Hoff H Biemans H Fader Mand Waha K Global water availability and requirementsfor future food production J Hydrometeor 12 885ndash899httpsdoiorg1011752011JHM13281 2011

Gerten D Hoff H Rockstroumlm J Jaumlgermeyr J KummuM and Pastor A V Towards a revised planetary bound-ary for consumptive freshwater use role of environmen-

tal flow requirements Curr Opin Env Sust 5 551ndash558httpsdoiorg101016jcosust201311001 2013

Gosme M Suffert F and Jeuffroy M-H Intensive ver-sus low-input cropping systems What is the optimal parti-tioning of agricultural area in order to reduce pesticide usewhile maintaining productivity Agr Syst 103 110ndash116httpsdoiorg101016jagsy200911002 2010

Gumpenberger M Vohland K Heyder U Poulter B MaceyK Anja Rammig Popp A and Cramer W Predictingpan-tropical climate change induced forest stock gains andlossesndashimplications for REDD Environ Res Lett 5 014013httpsdoiorg1010881748-932651014013 2010

Haberl H Erb K-H Krausmann F Bondeau A Lauk CMuumlller C Plutzar C and Steinberger J K Global bioenergypotentials from agricultural land in 2050 Sensitivity to climatechange diets and yields Biomass Bioenerg 35 4753ndash4769httpsdoiorg101016jbiombioe201104035 2011

Harris I Jones P Osborn T and Lister D Updatedhigh-resolution grids of monthly climatic observations ndashthe CRU TS310 Dataset Int J Climatol 34 623ndash642httpsdoiorg101002joc3711 2014

Haxeltine A and Prentice I C A General Model for the Light-Use Efficiency of Primary Production Funct Ecol 10 551ndash561httpsdoiorg1023072390165 1996

Heck V Gerten D Lucht W and Boysen L R Is extensiveterrestrial carbon dioxide removal a ldquogreenrdquo form of geoengi-neering A global modelling study Global Planet Change 137123ndash130 httpsdoiorg101016jgloplacha201512008 2016

Herrero M Havliacutek P Valin H Notenbaert A RufinoM C Thornton P K Bluumlmmel M Weiss F GraceD and Obersteiner M Biomass use production feed ef-ficiencies and greenhouse gas emissions from global live-stock systems P Natl Acad Sci USA 110 20888ndash20893httpsdoiorg101073pnas1308149110 2013

Heyder U Schaphoff S Gerten D and Lucht W Risk of severeclimate change impact on the terrestrial biosphere Environ ResLett 6 034036 httpsdoiorg1010881748-9326630340362011

Huang S Titus S J and Wiens D P Comparison of nonlinearheightndashdiameter functions for major Alberta tree species CanJ Forest Res 22 1297ndash1304 httpsdoiorg101139x92-1721992

Humpenoumlder F Popp A Dietrich J P Klein D Lotze-CampenH Bonsch M Bodirsky B L Weindl I Stevanovic M andMuumlller C Investigating afforestation and bioenergy CCS as cli-mate change mitigation strategies Environ Res Lett 9 064029httpsdoiorg1010881748-932696064029 2014

IPCC Summary for Policymakers in Climate Change 2014 Mit-igation of Climate Change Contribution of Working Group IIIto the Fifth Assessment Report of the Intergovernmental Panelon Climate Change edited by Edenhofer O Pichs-MadrugaR Sokona Y Farahani E Kadner S Seyboth K Adler ABaum I Brunner S Eickemeier P Kriemann B SavolainenJ Schloumlmer S von Stechow C Zwickel T and Minx J CCambridge University Press Cambridge UK and New York NYUSA 1ndash32 2014

Jackson R Canadell J Ehleringer J Mooney H SalaO and Schulze E A global analysis of root distri-

butions for terrestrial biomes Oecologia 108 389ndash411httpsdoiorg101007BF00333714 1996

Jaumlgermeyr J Gerten D Lucht W Hostert P MigliavaccaM and Nemani R A high-resolution approach to esti-mating ecosystem respiration at continental scales using op-erational satellite data Glob Change Biol 20 1191ndash1210httpsdoiorg101111gcb12443 2014

Jaumlgermeyr J Gerten D Heinke J Schaphoff S Kummu Mand Lucht W Water savings potentials of irrigation systemsglobal simulation of processes and linkages Hydrol Earth SystSci 19 3073ndash3091 httpsdoiorg105194hess-19-3073-20152015

Jaumlgermeyr J Gerten D Schaphoff S Heinke J Lucht W andRockstroumlm J Integrated crop water management might sustain-ably halve the global food gap Environ Res Lett 11 025002httpsdoiorg1010881748-9326112025002 2016

Jaumlgermeyr J Pastor A Biemans H and Gerten D Reconcil-ing irrigated food production with environmental flows for Sus-tainable Development Goals implementation Nat Commun 815900 httpsdoiorg101038ncomms15900 2017

Jarvis P G and McNaughton K Stomatal control of transpira-tion scaling up from leaf to region Adv Ecol Res 15 1ndash49httpsdoiorg101016S0065-2504(08)60119-1 1986

Jobbagy E G and Jackson R B The Vertical Distribution of SoilOrganic Carbon and Its Relation to Climate and Vegetation EcolAppl 10 423ndash436 httpsdoiorg1023072641104 2000

Johansen O Thermal conductivity of soils PhD thesisUniversity of Trondheim Trondheim Norway availableat httpoaidticmiloaioaiverb=getRecordampmetadataPrefix=htmlampidentifier=ADA044002 1977

Jolly W M Nemani R and Running S W A generalized bio-climatic index to predict foliar phenology in response to climateGlob Change Biol 11 619ndash632 httpsdoiorg101111j1365-2486200500930x 2005

Jung M Verstraete M Gobron N Reichstein M PapaleD Bondeau A Robustelli M and Pinty B Diagnos-tic assessment of European gross primary production GlobChange Biol 14 2349ndash2364 httpsdoiorg101111j1365-2486200801647x 2008

Kalnay E Kanamitsu M Kistler R Collins W DeavenD Gandin L Iredell M Saha S White G WoollenJ Zhu Y Leetmaa A Reynolds R Chelliah MEbisuzaki W Higgins W Janowiak J Mo K CRopelewski C Wang J Jenne R and Joseph DThe NCEPNCAR 40-Year Reanalysis Project B AmMeteorol Soc 77 437ndash471 httpsdoiorg1011751520-0477(1996)077lt0437TNYRPgt20CO2 1996

Kattge J Diacuteaz S Lavorel S Prentice I C Leadley P BoumlnischG Garnier E Westoby M Reich P B Wright I J Cor-nelissen J H C Violle C Harrison S P Van BODEGOMP M Reichstein M Enquist B J Soudzilovskaia N AAckerly D D Anand M Atkin O Bahn M Baker T RBaldocchi D Bekker R Blanco C C Blonder B BondW J Bradstock R Bunker D E Casanoves F Cavender-Bares J Chambers J Q Chapin Iii F S Chave J CoomesD Cornwell W K Craine J M Dobrin B H Duarte LDurka W Elser J Esser G Estiarte M Fagan W F FangJ Fernaacutendez-Meacutendez F Fidelis A Finegan B Flores OFord H Frank D Freschet G T Fyllas N M Gallagher

R V Green W A Gutierrez A G Hickler T Higgins S IHodgson J G Jalili A Jansen S Joly C A Kerkhoff A JKirkup D Kitajima K Kleyer M Klotz S Knops J M HKramer K Kuumlhn I Kurokawa H Laughlin D Lee T DLeishman M Lens F Lenz T Lewis S L Lloyd J LlusiagraveJ Louault F Ma S Mahecha M D Manning P Mas-sad T Medlyn B E Messier J Moles A T Muumlller S CNadrowski K Naeem S Niinemets Uuml Noumlllert S NuumlskeA Ogaya R Oleksyn J Onipchenko V G Onoda Y Or-dontildeez J Overbeck G Ozinga W A Patintildeo S Paula SPausas J G Pentildeuelas J Phillips O L Pillar V Poorter HPoorter L Poschlod P Prinzing A Proulx R Rammig AReinsch S Reu B Sack L Salgado-Negret B Sardans JShiodera S Shipley B Siefert A Sosinski E Soussana J-F Swaine E Swenson N Thompson K Thornton P Wal-dram M Weiher E White M White S Wright S J YguelB Zaehle S Zanne A E and Wirth C TRY ndash a globaldatabase of plant traits Global Change Biology 17 2905ndash2935httpsdoiorg101111j1365-2486201102451x 2011

Kollas C Kersebaum K C Nendel C Manevski K MuumlllerC Palosuo T Armas-Herrera C M Beaudoin N BindiM Charfeddine M Conradt T Constantin J Eitzinger JEwert F Ferrise R Gaiser T Cortazar-Atauri I G dGiglio L Hlavinka P Hoffmann H Hoffmann M P Lau-nay M Manderscheid R Mary B Mirschel W MoriondoM Olesen J E Oumlztuumlrk I Pacholski A Ripoche-WachterD Roggero P P Roncossek S Roumltter R P Ruget FSharif B Trnka M Ventrella D Waha K WegehenkelM Weigel H-J and Wu L Crop rotation modellingndashAEuropean model intercomparison Eur J Agron 70 98ndash111httpsdoiorg101016jeja201506007 2015

Konzmann M Gerten D and Heinke J Climate impacts onglobal irrigation requirements under 19 GCMs simulated witha vegetation and hydrology model Hydrol Sci J 58 88ndash105httpsdoiorg101080026266672013746495 2013

Krysanova V Muumlller-Wohlfeil D-I and Becker A Develop-ment and test of a spatially distributed hydrologicalwater qualitymodel for mesoscale watersheds Ecol Model 106 261ndash289httpsdoiorg101016S0304-3800(97)00204-4 1998

Lapola D M Oyama M D and Nobre C A Explor-ing the range of climate biome projections for tropi-cal South America The role of CO2 fertilization andseasonality Global Biogeochem Cy 23 GB3003httpsdoiorg1010292008GB003357 2009

Latham D and Williams E Lightning and forest fires chap11 in Forest fires Behavior and ecological effects 375ndash418httpsdoiorg101016B978-012386660-850013-1 2001

Latham D J and Schlieter J A Ignition probabilities of wildlandfuels based on simulated lightning discharges US Departmentof Agriculture Forest Service Intermountain Research Station24 pp 1989

Lawrence D and Slater A Incorporating organic soil intoa global climate model Clim Dynam 30 145ndash160httpsdoiorg101007s00382-007-0278-1 2008

Le Queacutereacute C Moriarty R Andrew R M Canadell J G Sitch SKorsbakken J I Friedlingstein P Peters G P Andres R JBoden T A Houghton R A House J I Keeling R F TansP Arneth A Bakker D C E Barbero L Bopp L ChangJ Chevallier F Chini L P Ciais P Fader M Feely R A

Gkritzalis T Harris I Hauck J Ilyina T Jain A K KatoE Kitidis V Klein Goldewijk K Koven C LandschuumltzerP Lauvset S K Lefegravevre N Lenton A Lima I D MetzlN Millero F Munro D R Murata A Nabel J E M SNakaoka S Nojiri Y OrsquoBrien K Olsen A Ono T PeacuterezF F Pfeil B Pierrot D Poulter B Rehder G RoumldenbeckC Saito S Schuster U Schwinger J Seacutefeacuterian R SteinhoffT Stocker B D Sutton A J Takahashi T Tilbrook B vander Laan-Luijkx I T van der Werf G R van Heuven S Van-demark D Viovy N Wiltshire A Zaehle S and Zeng NGlobal Carbon Budget 2015 Earth Syst Sci Data 7 349ndash396httpsdoiorg105194essd-7-349-2015 2015

Lehner B and Doumlll P Development and validation of a globaldatabase of lakes reservoirs and wetlands J Hydrol 296 1ndash22httpsdoiorg101016jjhydrol200403028 2004

Lehner B Liermann C R Revenga C Voumlroumlsmarty C FeketeB Crouzet P Doumlll P Endejan M Frenken K and MagomeJ High-resolution mapping of the worldrsquos reservoirs and damsfor sustainable river-flow management Front Ecol Environ 9494ndash502 httpsdoiorg101890100125 2011

Liang S Stroeve J and Box J E Mapping daily snowiceshortwave broadband albedo from Moderate ResolutionImaging Spectroradiometer (MODIS) The improved di-rect retrieval algorithm and validation with Greenland insitu measurement J Geophys Res-Atmos 110 D10109httpsdoiorg1010292004JD005493 2005

Lindeskog M Arneth A Bondeau A Waha K Seaquist JOlin S and Smith B Implications of accounting for landuse in simulations of ecosystem carbon cycling in Africa EarthSyst Dynam 4 385ndash407 httpsdoiorg105194esd-4-385-2013 2013

Lloyd J and Taylor J A On the Temperature Depen-dence of Soil Respiration Funct Ecol 8 315ndash323httpsdoiorg1023072389824 1994

Lotze-Campen H Muumlller C Bondeau A Rost S Popp Aand Lucht W Global food demand productivity growth andthe scarcity of land and water resources a spatially explicitmathematical programming approach Agr Econ 39 325ndash338httpsdoiorg101111j1574-0862200800336x 2008

Luyssaert S Inglima I Jung M Richardson A D Reich-stein M Papale D Piao S L Schulze E D Wingate LMatteucci G Aragao L Aubinet M Beer C BernhoferC Black K G Bonal D Bonnefond J M Chambers JCiais P Cook B Davis K J Dolman A J Gielen BGoulden M Grace J Granier A Grelle A Griffis T Gruumln-wald T Guidolotti G Hanson P J Harding R HollingerD Y Hutyra L R Kolari P Kruijt B Kutsch W Lager-gren F Laurila T Law B E Le Maire G Lindroth ALoustau D Malhi Y Mateus J Migliavacca M MissonL Montagnani L Moncrieff J Moors E Munger J WNikinmaa E Ollinger S V Pita G Rebmann C Roup-sard O Saigusa N Sanz M J Seufert G Sierra C SmithM L Tang J Valentini R Vesala T and Janssens I ACO2 balance of boreal temperate and tropical forests derivedfrom a global database Glob Change Biol 13 2509ndash2537httpsdoiorg101111j1365-2486200701439x 2007

Malik M J van der Velde R Vekerdy Z and SuZ Assimilation of Satellite-Observed Snow Albedo in

a Land Surface Model J Hydrometeor 13 1119ndash1130httpsdoiorg101175JHM-D-11-01251 2012

Monfreda C Ramankutty N and Foley J A Farm-ing the planet 2 Geographic distribution of crop ar-eas yields physiological types and net primary produc-tion in the year 2000 Global Biogeochem Cy 22 1ndash19httpsdoiorg1010292007GB002947 2008

Monsi M Uumlber den Lichtfaktor in den Pflanzengesellschaften undseine Bedeutung fur die Stoffproduktion Jpn J Bot 14 22ndash52httpsdoiorg101093aobmci052 1953

Monteith J and Unsworth M Principles of EnvironmentalPhysics Edward Arnold New York 2nd edn 1990

Monteith J L Accommodation between transpiring vegetationand the convective boundary layer J Hydrol 166 251ndash263httpsdoiorg1010160022-1694(94)05086-D 1995

Muumlller C and Robertson R D Projecting future crop produc-tivity for global economic modeling Agr Econ 45 37ndash50httpsdoiorg101111agec12088 2014

Muumlller C Elliott J Chryssanthacopoulos J Deryng D Fol-berth C Pugh T A and Schmid E Implications of climatemitigation for future agricultural production Environ Res Lett10 125004 httpsdoiorg1010881748-932610121250042015

Muumlller C Stehfest E Minnen J G v Strengers B BlohW v Beusen A H W Schaphoff S Kram T and Lucht WDrivers and patterns of land biosphere carbon balance reversalEnviron Res Lett 11 044002 httpsdoiorg1010881748-9326114044002 2016

Nachtergaele F van Velthuizen H Verelst L Batjes NDijkshoorn K van Engelen V Fischer G Jones AMontanarella L and Petri M Harmonized world soildatabase Food and Agriculture Organization of the United Na-tions available at httpwwwfaoorgsoils-portalsoil-surveysoil-maps-and-databasesharmonized-world-soil-database-v12en 2009

Nash J E The form of the instantaneous unit hydrograph inComptes Rendus et Rapports Assemblee Generale de Torontovol III Toronto 114ndash121 available at httpnoranercacuk508550 1957

New M Hulme M and Jones P Representing Twentieth-Century SpacendashTime Climate Variability Part II Develop-ment of 1901ndash96 Monthly Grids of Terrestrial Surface Cli-mate J Climate 13 2217ndash2238 httpsdoiorg1011751520-0442(2000)013lt2217RTCSTCgt20CO2 2000

Olson D M Dinerstein E Wikramanayake E D BurgessN D Powell G V N Underwood E C Drsquoamico J AItoua I Strand H E Morrison J C Loucks C JAllnutt T F Ricketts T H Kura Y Lamoreux J FWettengel W W Hedao P and Kassem K R Terres-trial Ecoregions of the World A New Map of Life onEarth BioScience 51 933ndash938 httpsdoiorg1016410006-3568(2001)051[0933TEOTWA]20CO2 2001

Ostberg S Lucht W Schaphoff S and Gerten D Critical im-pacts of global warming on land ecosystems Earth Syst Dynam4 347ndash357 httpsdoiorg105194esd-4-347-2013 2013

Parton W J and Logan J A A model for diurnal varia-tion in soil and air temperature Agr Meteorol 23 205ndash216httpsdoiorg1010160002-1571(81)90105-9 1981

Popp A Dietrich J Lotze-Campen H Klein D Bauer NKrause M Beringer T Gerten D and Edenhofer O Theeconomic potential of bioenergy for climate change mitigationwith special attention given to implications for the land systemEnviron Res Lett 6 034017 httpsdoiorg1010881748-932663034017 2011

Popp A Humpenoumlder F Weindl I Bodirsky B L BonschM Lotze-Campen H Muumlller C Biewald A Rolinski SStevanovic M and Dietrich J P Land-use protection forclimate change mitigation Nat Clim Change 4 1095ndash1098httpsdoiorg101038nclimate2444 2014

Portmann F T Siebert S and Doumlll P MIRCA2000 ndash Globalmonthly irrigated and rainfed crop areas around the year2000 A new high-resolution data set for agricultural andhydrological modeling Global Biogeochem Cy 24 1ndash24httpsdoiorg1010292008GB003435 2010

Poulter B Aragatildeo L Heyder U Gumpenberger M Heinke JLangerwisch F Rammig A Thonicke K and Cramer W Netbiome production of the Amazon Basin in the 21st century GlobChange Biol 16 2062ndash2075 httpsdoiorg101111j1365-2486200902064x 2010

Prentice C I Sykes M T and Cramer W A simulationmodel for the transient effects of climate change on forest land-scapes Ecol Model 65 51ndash70 httpsdoiorg1010160304-3800(93)90126-D 1993

Prentice I C Heimann M and Sitch S The Car-bon Balance of the Terrestrial Biosphere Ecosys-tem Models and Atmospheric Observations EcolAppl 10 1553ndash1573 httpsdoiorg1018901051-0761(2000)010[1553TCBOTT]20CO2 2000

Prentice I Bondeau A Cramer W Harrison S Hickler TLucht W Sitch S Smith B and Sykes M DynamicGlobal Vegetation Modeling Quantifying Terrestrial Ecosys-tem Responses to Large-Scale Environmental Change in Ter-restrial Ecosystems in a Changing World Global Changendash The IGBP Series Springer Berlin Heidelberg 175ndash192httpsdoiorg101007978-3-540-32730-1_15 2007

Priestley C H B and Taylor R J On the Assessment of Sur-face Heat Flux and Evaporation Using Large-Scale ParametersMon Weather Rev 100 81ndash92 httpsdoiorg1011751520-0493(1972)100lt0081OTAOSHgt23CO2 1972

Pugh T Muumlller C Arneth A Haverd V and SmithB Key knowledge and data gaps in modelling the influ-ence of CO2 concentration on the terrestrial carbon sinkPlants facing Changing Climate J Plant Physiol 203 3ndash15httpsdoiorg101016jjplph201605001 2016a

Pugh T Muumlller C Elliott J Deryng D Folberth C OlinS Schmid E and Arneth A Climate analogues suggestlimited potential for intensification of production on currentcroplands under climate change Nat Commun 7 12608httpsdoiorg101038ncomms12608 2016b

Rammig A Jupp T Thonicke K Tietjen B Heinke J Os-tberg S Lucht W Cramer W and Cox P Estimating therisk of Amazonian forest dieback New Phytol 187 694ndash706httpsdoiorg101111j1469-8137201003318x 2010

Rammig A Wiedermann M Donges J F Babst F von BlohW Frank D Thonicke K and Mahecha M D Coinci-dences of climate extremes and anomalous vegetation responsescomparing tree ring patterns to simulated productivity Biogeo-

sciences 12 373ndash385 httpsdoiorg105194bg-12-373-20152015

Reich P B Walters M B and Ellsworth D S Fromtropics to tundra Global convergence in plant func-tioning P Natl Acad Sci USA 94 13730ndash13734httpsdoiorg101073pnas942513730 1997

Rockstroumlm J Steffen W Noone K Persson A Chapin F SLambin E F Lenton T M Scheffer M Folke C Schellnhu-ber H J Nykvist B de Wit C A Hughes T van der LeeuwS Rodhe H Sorlin S Snyder P K Costanza R SvedinU Falkenmark M Karlberg L Corell R W Fabry V JHansen J Walker B Liverman D Richardson K CrutzenP and Foley J A A safe operating space for humanity Nature461 472ndash475 httpsdoiorg101038461472a 2009

Rogers A Medlyn B E Dukes J S Bonan G von CaemmererS Dietze M C Kattge J Leakey A D B Mercado L MNiinemets Uuml Prentice I C Serbin S P Sitch S Way D Aand Zaehle S A roadmap for improving the representation ofphotosynthesis in Earth system models New Phytol 213 22ndash42 httpsdoiorg101111nph14283 2017

Rolinski S Rammig A Walz A von Bloh W van Oi-jen M and Thonicke K A probabilistic risk assessment forthe vulnerability of the European carbon cycle to weather ex-tremes the ecosystem perspective Biogeosciences 12 1813ndash1831 httpsdoiorg105194bg-12-1813-2015 2015

Rosenzweig C Elliott J Deryng D Ruane A C MuumlllerC Arneth A Boote K J Folberth C Glotter M andKhabarov N Assessing agricultural risks of climate changein the 21st century in a global gridded crop model in-tercomparison P Natl Acad Sci USA 111 3268ndash3273httpsdoiorg101073pnas1222463110 2014

Rost S Gerten D Bondeau A Lucht W Rohwer J andSchaphoff S Agricultural green and blue water consumptionand its influence on the global water system Water Resour Res44 W09405 httpsdoiorg1010292007WR006331 2008

Ryan M Effects of climate change on plant respiration EcolAppl 1 157ndash167 httpsdoiorg1023071941808 1991

Schaphoff S Heyder U Ostberg S Gerten D HeinkeJ and Lucht W Contribution of permafrost soils tothe global carbon budget Environ Res Lett 8 014026httpsdoiorg1010881748-932681014026 2013

Schaphoff S von Bloh W Rammig A Thonicke K ForkelM Biemans H Gerten D Heinke J Jaumlgermyer JKnauer J Lucht W Muumlller C Rolinski S and Waha KLPJmL4 model output for the publications in GMD LPJmL4ndash a dynamic global vegetation model with managed landPart I ndash Model description and Part II ndash Model evaluationhttpsdoiorg105880pik2017009 2018a

Schaphoff S von Bloh W Thonicke K Biemans H ForkelM Heinke J Jaumlgermyer J Muumlller C Rolinski S Waha KStehfest E de Waal L Heyder U Gumpenberger M andBeringer T LPJmL4 Model Code V 40 GFZ Data Serviceshttpsdoiorg105880pik2018002 2018b

Schaphoff S Forkel M Muumlller C Knauer J von Bloh WGerten D Jaumlgermyer J Lucht W Rammig A ThonickeK and Waha K LPJmL4 ndash a dynamic global vegetationmodel with managed land ndash Part 2 Model evaluation GeosciModel Dev 11 1377ndash1403 httpsdoiorg105194gmd-11-1377-2018 2018c

Schewe J Heinke J Gerten D Haddeland I Arnell N WClark D B Dankers R Eisner S Fekete B M Coloacuten-Gonzaacutelez F J Gosling S N Kim H Liu X Masaki YPortmann F T Satoh Y Stacke T Tang Q Wada YWisser D Albrecht T Frieler K Piontek F WarszawskiL and Kabat P Multimodel assessment of water scarcity un-der climate change P Natl Acad Sci USA 111 3245ndash3250httpsdoiorg101073pnas1222460110 2014

Shinozaki K Yoda K Hozumi K and Kira T Aquantitative analysis of plant form-the pipe modeltheory I Basic analyses Jpn J Ecol 14 97ndash105httpsdoiorg1018960seitai143_97 1964

Siebert S Kummu M Porkka M Doumlll P Ramankutty N andScanlon B R A global data set of the extent of irrigated landfrom 1900 to 2005 Hydrol Earth Syst Sci 19 1521ndash1545httpsdoiorg105194hess-19-1521-2015 2015

Sitch S Smith B Prentice I C Arneth A Bondeau ACramer W Kaplan J O Levis S Lucht W Sykes M TThonicke K and Venevsky S Evaluation of ecosystem dynam-ics plant geography and terrestrial carbon cycling in the LPJ dy-namic global vegetation model Glob Change Biol 9 161ndash185httpsdoiorg101046j1365-2486200300569x 2003

Sitch S Huntingford C Gedney N Levy P Lomas M RPiao S L Betts R A Ciais P Cox P M FriedlingsteinP Johns C D Prentice I C and Woodward F I Evalua-tion of the terrestrial carbon cycle future plant geography andclimate-carbon cycle feedbacks using five Dynamic Global Veg-etation Models (DGVMs) Glob Change Biol 14 2015ndash2039httpsdoiorg101111j1365-2486200801626x 2008

Smith B Waringrlind D Arneth A Hickler T Leadley P Silt-berg J and Zaehle S Implications of incorporating N cy-cling and N limitations on primary production in an individual-based dynamic vegetation model Biogeosciences 11 2027ndash2054 httpsdoiorg105194bg-11-2027-2014 2014

Smith T Shugart H Woodward F and Burton P Plantfunctional types in Vegetation Dynamics amp Global ChangeSpringer 272ndash292 httpsdoiorg101007978-1-4615-2816-61993

Sprugel D G Ryan M G Brooks J R Vogt K A and Mar-tin T A Respiration from the organ level to the stand chap 8in Resource physiology of conifers Academic Press 255ndash299httpsdoiorg101016B978-0-08-092591-250013-3 1995

Steffen W Richardson K Rockstroumlm J Cornell S EFetzer I Bennett E M Biggs R Carpenter S Rde Vries W de Wit C A Folke C Gerten D HeinkeJ Mace G M Persson L M Ramanathan V Rey-ers B and Soumlrlin S Planetary boundaries Guiding humandevelopment on a changing planet Science 347 1259855httpsdoiorg101126science1259855 2015

Strengers B J Muumlller C Schaeffer M Haarsma R J Severi-jns C Gerten D Schaphoff S van den Houdt R and Oosten-rijk R Assessing 20th century climate-vegetation feedbacks ofland-use change and natural vegetation dynamics in a fully cou-pled vegetation-climate model Int J Climatol 30 2055ndash2065httpsdoiorg101002joc2132 2010

Strugnell N C Lucht W and Schaaf C A globalalbedo data set derived from AVHRR data for use inclimate simulations Geophys Res Lett 28 191ndash194httpsdoiorg1010292000GL011580 2001

Tans P and Keeling R Trends in Atmospheric Carbon DioxideNational Oceanic amp Atmospheric Administration Earth SystemResearch Laboratory (NOAAESRL) available at httpwwwesrlnoaagovgmdccggtrends 2015

Thonicke K Venevsky S Sitch S and Cramer W Therole of fire disturbance for global vegetation dynamics cou-pling fire into a Dynamic Global Vegetation Model GlobalEcol Biogeogr 10 661ndash677 httpsdoiorg101046j1466-822X200100175x 2001

Thonicke K Spessa A Prentice I C Harrison S P DongL and Carmona-Moreno C The influence of vegetation firespread and fire behaviour on biomass burning and trace gas emis-sions results from a process-based model Biogeosciences 71991ndash2011 httpsdoiorg105194bg-7-1991-2010 2010

Thornley J H M Respiration Growth and Maintenance in PlantsNature 227 304ndash305 httpsdoiorg101038227304b0 1970

Thornthwaite C W An Approach toward a Ratio-nal Classification of Climate Geogr Rev 38 55ndash94httpsdoiorg102307210739 1948

Tjoelker M Oleksyn J and Reich P B Acclimation of respira-tion to temperature and CO2 in seedlings of boreal tree species inrelation to plant size and relative growth rate Glob Change Biol5 679ndash691 httpsdoiorg101046j1365-2486199900257x1999

University of East Anglia Climatic Research Unit Harris IC and Jones P CRU TS323 Climatic Research Unit(CRU) Time-Series (TS) Version 323 of High ResolutionGridded Data of Month-by-month Variation in Climate(Jan 1901ndashDec 2014) Centre for Environmental DataAnalysis httpsdoiorg1052854c7fdfa6-f176-4c58-acee-683d5e9d2ed5 2015

Von Bloh W Rost S Gerten D and Lucht W Effi-cient parallelization of a dynamic global vegetation modelwith river routing Environ Model Softw 25 685ndash690httpsdoiorg101016jenvsoft200911012 2010

Vorosmarty C and Fekete B ISLSCP II River Routing Data(STN-30p) in ISLSCP Initiative II Collection Data set editedby Hall F G Collatz G Meeson B Los S Brown de Col-stoun E and Landis D ORNL Distributed Active ArchiveCenter httpsdoiorg103334ORNLDAAC1005 2011

Voumlroumlsmarty C J McIntyre P B Gessner M O Dudgeon DPrusevich A Green P Glidden S Bunn S E SullivanC A Liermann C R and Davies P M Global threats to hu-man water security and river biodiversity Nature 467 555ndash561httpsdoiorg101038nature09440 2010

Waha K van Bussel L G J Muumlller C and Bondeau AClimate-driven simulation of global crop sowing dates GlobalEcol Biogeogr 21 247ndash259 httpsdoiorg101111j1466-8238201100678x 2012

Waha K Muumlller C Bondeau A Dietrich J KurukulasuriyaP Heinke J and Lotze-Campen H Adaptation to climatechange through the choice of cropping system and sowing datein sub-Saharan Africa Global Environ Chang 23 130ndash143httpsdoiorg101016jgloenvcha201211001 2013

Waring R Schroeder P and Oren R Application of the pipemodel theory to predict canopy leaf area Can J Forest Res 12556ndash560 httpsdoiorg101139x82-086 1982

Waring R H Estimating Forest Growth and Efficiency in Re-lation to Canopy Leaf Area Adv Ecol Res 13 327ndash354httpsdoiorg101016S0065-2504(08)60111-7 1983

Xu R Dai J Luo W Yin X Li Y Tai X Han L ChenY Lin L and Li G A photothermal model of leaf area in-dex for greenhouse crops Agr Forest Meteorol 150 541ndash552httpsdoiorg101016jagrformet201001019 2010

Zaehle S Sitch S Prentice I C Liski J Cramer W ErhardM Hickler T and Smith B The importance of age-relateddecline in forest NPP for modeling regional carbon balancesEcol Appl 16 1555ndash1574 httpsdoiorg1018901051-0761(2006)016[1555TIOADI]20CO2 2010

Zeide B Primary Unit of the Tree Crown Ecology 74 1598ndash1602 httpsdoiorg1023071940088 1993

Zhou X Zhu Q Tang S Chen X and Wu M Intercep-tion of PAR and relationship between FPAR and LAI in sum-mer maize canopy IGARSS rsquo02 2002 IEEE International Geo-science and Remote Sensing Symposium 2002 6 3252ndash3254httpsdoiorg101109IGARSS20021027146 2002

Zscheischler J Mahecha M Von Buttlar J Harmeling SJung M Rammig A Randerson J T Schoumlllkopf B Senevi-ratne S I Tomelleri E Zaehle S and Reichstein MFew extreme events dominate global interannual variabilityin gross primary production Environ Res Lett 9 035001httpsdoiorg1010881748-932693035001 2014a

Zscheischler J Reichstein M Harmeling S Rammig ATomelleri E and Mahecha M D Extreme events in gross pri-mary production a characterization across continents Biogeo-sciences 11 2909ndash2924 httpsdoiorg105194bg-11-2909-2014 2014b

Abstract

Introduction

Model description

Energy balance model and permafrost

Photosynthetic active radiation daylength and potential evapotranspiration

Albedo

Soil energy balance

Plant physiology

Photosynthesis

Phenology

Productivity

Plant functional types

Allocation

Vegetation dynamics

Crop functional types

Soil and litter carbon pools

Decomposition

Water balance

Soil water balance

Evapotranspiration

River routing

Irrigation and dams

Dams lakes and reservoirs

Land use

Sowing dates

Management and cropping intensity

Managed grassland

Biomass plantations

Modelling protocol

Model set-up and inputs

Climate river routing and soil inputs

Land use input

Standard outputs

Discussion

Code and data availability


Competing interests

Acknowledgements

References

to consider human activities as an integral part while rep-resenting the major dynamics of the biosphere in a spatio-temporally explicit and process-based manner accountingfor the feedbacks between vegetation global carbon and wa-ter cycling and the atmosphere This would also allow forthe numerical evaluation of potential implementation path-ways for the United Nations Sustainable Development Goals(SDGs httpssustainabledevelopmentunorg) and their im-pacts on the terrestrial environment complementing the im-portant role that dynamic biosphere models have played inthe United Nations scientific assessment reports on climatechange published by the United Nations IntergovernmentalPanel on Climate Change (IPCC 2014) By combining corefeatures of global biogeographical and biogeochemical mod-els developed in the 1990s dynamic global vegetation mod-els (DGVMs) emerged as the main tool to simulate the pro-cesses underlying the dynamics of natural vegetation types(growth mortality resource competition and disturbancessuch as wildfires) and the associated carbon and water fluxes(Cramer et al 2001 Prentice et al 2007 Sitch et al 2008Friend et al 2014) In light of strengthening human inter-ferences DGVMs were further developed to integrate ad-ditional processes that are relevant to the original researchquest of studying biogeography and biogeochemical cyclesunder climate change (Canadell et al 2007) This includesthe incorporation of human land use and the simulation ofagricultural production systems (Bondeau et al 2007 Lin-deskog et al 2013) nutrient limitation (Zaehle et al 2010Smith et al 2014) hydrological modules and river-routingschemes (Gerten et al 2004 Rost et al 2008) Knowledgederived from models that are designed to cover aspects ofthe Earth system other than terrestrial vegetation and the car-bon cycle such as models of the global water balance couldevidentially improve the DGVMsrsquo ability to also evaluatemodel performance for processes (eg river discharge) thatare closely connected to simulated vegetation and carbon cy-cle dynamics (Bondeau et al 2007 Smith et al 2014) Thedevelopment towards more comprehensive models of Earthrsquosland surface offers new possibilities for cross-disciplinary re-search DGVMs as land components of Earth system mod-els still show large uncertainties about the terrestrial carbon(C) balance under future climate change (Friedlingstein et al2013) This uncertainty partly results from differences in thesimulation of soil and vegetation C residence times (Carval-hais et al 2014 Friend et al 2014) The time that C residesin an ecosystem is thereby strongly affected by simulatedprocesses of vegetation dynamics (Ahlstroumlm et al 2015)These examples highlight the need to continuously improveprocess representations in DGVMs in order to reduce the un-certainty in projected ecosystem functioning and services un-der future climate change This requires however that modeldevelopments in specific fields or improvements for certainprocesses are synthesized and integrated into a unified in-ternally consistent model version The LundndashPotsdamndashJenaDGVM with managed Land (LPJmL Bondeau et al 2007)

originates from a former version of the model described bySitch et al (2003) and simulates the growth and geographi-cal distribution of natural plant functional types (PFTs) cropfunctional types (CFTs) and the associated biogeochemicalprocesses (mainly carbon cycling) Recent developments fo-cused on an improved energy balance model able to estimatepermafrost dynamics based on a vertical soil carbon distribu-tion scheme and a new soil hydrological scheme (Schaphoffet al 2013) Also a new process-based fire module (SPIT-FIRE) was implemented that allows for detailed simulationof fire ignition spread and effects to estimate fire impactsand emissions (Thonicke et al 2010) An updated phenol-ogy scheme was developed which now takes phenology lim-itations arising from low temperatures limited light anddrought into account (Forkel et al 2014) Further modeldevelopments encompass the parallelization of the modelto efficiently simulate river routing (Von Bloh et al 2010)and the implementation of an irrigation scheme (Rost et al2008) recently updated with a mechanistic representation ofthe three major irrigation systems (Jaumlgermeyr et al 2015)Biemans et al (2011) implemented reservoir operations andirrigation extraction and evaluated the impact on river dis-charge Other developments focused on a newly formulatedimplementation of different cropping systems in sub-SaharanAfrica (Waha et al 2013) Mediterranean agricultural planttypes (Fader et al 2015) and bioenergy crops such as sugar-cane (Lapola et al 2009) fast-growing grasses and bioen-ergy trees (Beringer et al 2011) With these implementa-tions the potential of bioenergy production under future landuse population and climate development could be exten-sively investigated (Haberl et al 2011 Popp et al 2011Humpenoumlder et al 2014) All developments the core modelstructure and recently developed modules of DGVM LPJmLversion 40 (in the following referred to as LPJmL4) will bedescribed in Sect 2 in more detail We show that the model inits present form allows for a consistent and joint quantifica-tion of climate and land use change impacts on the terrestrialbiosphere the water cycle the carbon cycle and on agricul-tural production (a systematic evaluation can be found in PartII of this paper) To give an overview of recent developmentsand applications of LPJmL4 we present the following

1 A comprehensive description of the full model with allcontributing developments since its original publicationby Sitch et al (2003) and Bondeau et al (2007) Weaim at consistently uniting all developments includ-ing undocumented and already published developmentsthus providing a comprehensive description of the fullLPJmL4 model

2 An overview of published LPJmL applications to reviewthe improvement of process understanding

3 A discussion of the presented standard LPJmL4 resultsthat give an overview of simulated biogeochemical hy-drological and agricultural patterns at the global scale

2 Model description

Eeq =s

s+ γmiddotRnday

λ (2)

(2373+ Tair)2 (4)

γ = 6505+ 0064 middot Tair (5)

h12 = arccos(minus

) (14)

and thus

π (16)

212 Albedo

nPFTsumPFT=1

partTsoil

partt= α middot

part2Tsoil

partz2 (18)

1t= (19)

r =α1t

(1z)2 (22)

1t le(1z)2

2 middotα (23)

22 Plant physiology

221 Photosynthesis

JE = C1 middotAPAR

daylength (28)

(piminus0lowast

pi+ 2 middot0lowast

λmaxC4

) (30)

0lowast =[O2]

2 middot τ (31)

C2 =piminus0lowast

(1+ [O2]

) (33)

(JE + JC minus

Vm =1bmiddotC1

radic1minus

gc =16Adt

222 Phenology

223 Productivity

g(T )= exp[

30856 middot(

minus1

(T + 4602)

)] (45)

Reproduction cost

Tissue turnover

231 Allocation

klasa (52)

From this follows

CAind (54)

SLA=2times 10minus4

Establishment

ESTTREE = (58)

nestTREE

Stress mortality

twPFT1)

NI(Nd)=

Pspread =

0 ω0 gtme(62)

FDI=max

01minus

1memiddot exp

(minusNI middot

nsump=1

) (63)

a(ND)=Nhobs

tfire =241

Af =π

4 middotLBmiddotD2

HUsum =

tsumt prime= sdate

fPHU= HUsumPHU (74)

PHUle PHUwlow

(psminuspb)))+psens

fLAImax =fPHU

fPHUkminusfPHUc

c =fPHUc

k =fPHUk

fLAImax =

(1minus fPHU

1minus fPHUsen

LAIt =tsum

t prime= sdate

251 Decomposition

k(lPFT) =1

τ10PFT

middot g(Tsoil(l)

(l) (96)

+ 0719 middot θ(l)

(θ(l))

C(l) =

nPFTsumPFT= 1

kmeanPFT =

nsoilsuml=1

kmeanPFT

26 Water balance

Infiltration

infil= Pr middot

radic1minus

SW(1)minusWpwp(1)

Percolation

et al 1998)

) (102)

TT(l) =FW(l)

HC(l)(103)

HC(l) =Ks(l) middot

(SW(l)

Wsat(l)

)β(l) (104)

[1minus exp

(minus1t

)](105)

radic1minus

Canopy evaporation

nPFTsumPFT=1

Soil evaporation

and thus

nsoilminus1suml= 1

rootdistz =

int z0 (βroot)

primedzprime=

1minus (βroot)z

263 River routing

Qout(t)=Qin middot1

K middot0(n)

)nminus1

v (119)

Irrigation

Ec (121)

27 Land use

271 Sowing dates

LI =min(BI max

(BI+RminusLlr

1+ 1lr 0))

Cleaf =sumPFT

BmlPFT

MCleaf =sumPFT

MmlPFT (126)

1000+Cleaf(127)

312 Land use input

32 Standard outputs

4 Discussion

04 6 11 18 48 70

03 1 45 168

2e+09 7e+10 5e+11 4e+12

0 500 1000 1500 2000

0 200 400 600 800 1000

2 8 25 80 107

References

Abstract

Introduction

Model description



Albedo

Soil energy balance

Plant physiology

Photosynthesis

Phenology

Productivity


Allocation

Vegetation dynamics



Decomposition

Water balance

Soil water balance

Evapotranspiration

River routing

Irrigation and dams


Land use

Sowing dates


Managed grassland

Biomass plantations

Modelling protocol



Land use input

Standard outputs

Discussion



Competing interests

Acknowledgements

References

2 Model description

Eeq =s

s+ γmiddotRnday

λ (2)

(2373+ Tair)2 (4)

γ = 6505+ 0064 middot Tair (5)

h12 = arccos(minus

) (14)

and thus

π (16)

212 Albedo

nPFTsumPFT=1

partTsoil

partt= α middot

part2Tsoil

partz2 (18)

1t= (19)

r =α1t

(1z)2 (22)

1t le(1z)2

2 middotα (23)

22 Plant physiology

221 Photosynthesis

JE = C1 middotAPAR

daylength (28)

(piminus0lowast

pi+ 2 middot0lowast

λmaxC4

) (30)

0lowast =[O2]

2 middot τ (31)

C2 =piminus0lowast

(1+ [O2]

) (33)

(JE + JC minus

Vm =1bmiddotC1

radic1minus

gc =16Adt

222 Phenology

223 Productivity

g(T )= exp[

30856 middot(

minus1

(T + 4602)

)] (45)

Reproduction cost

Tissue turnover

231 Allocation

klasa (52)

From this follows

CAind (54)

SLA=2times 10minus4

Establishment

ESTTREE = (58)

nestTREE

Stress mortality

twPFT1)

NI(Nd)=

Pspread =

0 ω0 gtme(62)

FDI=max

01minus

1memiddot exp

(minusNI middot

nsump=1

) (63)

a(ND)=Nhobs

tfire =241

Af =π

4 middotLBmiddotD2

HUsum =

tsumt prime= sdate

fPHU= HUsumPHU (74)

PHUle PHUwlow

(psminuspb)))+psens

fLAImax =fPHU

fPHUkminusfPHUc

c =fPHUc

k =fPHUk

fLAImax =

(1minus fPHU

1minus fPHUsen

LAIt =tsum

t prime= sdate

251 Decomposition

k(lPFT) =1

τ10PFT

middot g(Tsoil(l)

(l) (96)

+ 0719 middot θ(l)

(θ(l))

C(l) =

nPFTsumPFT= 1

kmeanPFT =

nsoilsuml=1

kmeanPFT

26 Water balance

Infiltration

infil= Pr middot

radic1minus

SW(1)minusWpwp(1)

Percolation

et al 1998)

) (102)

TT(l) =FW(l)

HC(l)(103)

HC(l) =Ks(l) middot

(SW(l)

Wsat(l)

)β(l) (104)

[1minus exp

(minus1t

)](105)

radic1minus

Canopy evaporation

nPFTsumPFT=1

Soil evaporation

and thus

nsoilminus1suml= 1

rootdistz =

int z0 (βroot)

primedzprime=

1minus (βroot)z

263 River routing

Qout(t)=Qin middot1

K middot0(n)

)nminus1

v (119)

Irrigation

Ec (121)

27 Land use

271 Sowing dates

LI =min(BI max

(BI+RminusLlr

1+ 1lr 0))

Cleaf =sumPFT

BmlPFT

MCleaf =sumPFT

MmlPFT (126)

1000+Cleaf(127)

312 Land use input

32 Standard outputs

4 Discussion

04 6 11 18 48 70

03 1 45 168

2e+09 7e+10 5e+11 4e+12

0 500 1000 1500 2000

0 200 400 600 800 1000

2 8 25 80 107

References

Abstract

Introduction

Model description



Albedo

Soil energy balance

Plant physiology

Photosynthesis

Phenology

Productivity


Allocation

Vegetation dynamics



Decomposition

Water balance

Soil water balance

Evapotranspiration

River routing

Irrigation and dams


Land use

Sowing dates


Managed grassland

Biomass plantations

Modelling protocol



Land use input

Standard outputs

Discussion



Competing interests

Acknowledgements

References

Eeq =s

s+ γmiddotRnday

λ (2)

(2373+ Tair)2 (4)

γ = 6505+ 0064 middot Tair (5)

h12 = arccos(minus

) (14)

and thus

π (16)

212 Albedo

nPFTsumPFT=1

partTsoil

partt= α middot

part2Tsoil

partz2 (18)

1t= (19)

r =α1t

(1z)2 (22)

1t le(1z)2

2 middotα (23)

22 Plant physiology

221 Photosynthesis

JE = C1 middotAPAR

daylength (28)

(piminus0lowast

pi+ 2 middot0lowast

λmaxC4

) (30)

0lowast =[O2]

2 middot τ (31)

C2 =piminus0lowast

(1+ [O2]

) (33)

(JE + JC minus

Vm =1bmiddotC1

radic1minus

gc =16Adt

222 Phenology

223 Productivity

g(T )= exp[

30856 middot(

minus1

(T + 4602)

)] (45)

Reproduction cost

Tissue turnover

231 Allocation

klasa (52)

From this follows

CAind (54)

SLA=2times 10minus4

Establishment

ESTTREE = (58)

nestTREE

Stress mortality

twPFT1)

NI(Nd)=

Pspread =

0 ω0 gtme(62)

FDI=max

01minus

1memiddot exp

(minusNI middot

nsump=1

) (63)

a(ND)=Nhobs

tfire =241

Af =π

4 middotLBmiddotD2

HUsum =

tsumt prime= sdate

fPHU= HUsumPHU (74)

PHUle PHUwlow

(psminuspb)))+psens

fLAImax =fPHU

fPHUkminusfPHUc

c =fPHUc

k =fPHUk

fLAImax =

(1minus fPHU

1minus fPHUsen

LAIt =tsum

t prime= sdate

251 Decomposition

k(lPFT) =1

τ10PFT

middot g(Tsoil(l)

(l) (96)

+ 0719 middot θ(l)

(θ(l))

C(l) =

nPFTsumPFT= 1

kmeanPFT =

nsoilsuml=1

kmeanPFT

26 Water balance

Infiltration

infil= Pr middot

radic1minus

SW(1)minusWpwp(1)

Percolation

et al 1998)

) (102)

TT(l) =FW(l)

HC(l)(103)

HC(l) =Ks(l) middot

(SW(l)

Wsat(l)

)β(l) (104)

[1minus exp

(minus1t

)](105)

radic1minus

Canopy evaporation

nPFTsumPFT=1

Soil evaporation

and thus

nsoilminus1suml= 1

rootdistz =

int z0 (βroot)

primedzprime=

1minus (βroot)z

263 River routing

Qout(t)=Qin middot1

K middot0(n)

)nminus1

v (119)

Irrigation

Ec (121)

27 Land use

271 Sowing dates

LI =min(BI max

(BI+RminusLlr

1+ 1lr 0))

Cleaf =sumPFT

BmlPFT

MCleaf =sumPFT

MmlPFT (126)

1000+Cleaf(127)

312 Land use input

32 Standard outputs

4 Discussion

04 6 11 18 48 70

03 1 45 168

2e+09 7e+10 5e+11 4e+12

0 500 1000 1500 2000

0 200 400 600 800 1000

2 8 25 80 107

References

Abstract

Introduction

Model description



Albedo

Soil energy balance

Plant physiology

Photosynthesis

Phenology

Productivity


Allocation

Vegetation dynamics



Decomposition

Water balance

Soil water balance

Evapotranspiration

River routing

Irrigation and dams


Land use

Sowing dates


Managed grassland

Biomass plantations

Modelling protocol



Land use input

Standard outputs

Discussion



Competing interests

Acknowledgements

References

212 Albedo

nPFTsumPFT=1

partTsoil

partt= α middot

part2Tsoil

partz2 (18)

1t= (19)

r =α1t

(1z)2 (22)

1t le(1z)2

2 middotα (23)

22 Plant physiology

221 Photosynthesis

JE = C1 middotAPAR

daylength (28)

(piminus0lowast

pi+ 2 middot0lowast

λmaxC4

) (30)

0lowast =[O2]

2 middot τ (31)

C2 =piminus0lowast

(1+ [O2]

) (33)

(JE + JC minus

Vm =1bmiddotC1

radic1minus

gc =16Adt

222 Phenology

223 Productivity

g(T )= exp[

30856 middot(

minus1

(T + 4602)

)] (45)

Reproduction cost

Tissue turnover

231 Allocation

klasa (52)

From this follows

CAind (54)

SLA=2times 10minus4

Establishment

ESTTREE = (58)

nestTREE

Stress mortality

twPFT1)

NI(Nd)=

Pspread =

0 ω0 gtme(62)

FDI=max

01minus

1memiddot exp

(minusNI middot

nsump=1

) (63)

a(ND)=Nhobs

tfire =241

Af =π

4 middotLBmiddotD2

HUsum =

tsumt prime= sdate

fPHU= HUsumPHU (74)

PHUle PHUwlow

(psminuspb)))+psens

fLAImax =fPHU

fPHUkminusfPHUc

c =fPHUc

k =fPHUk

fLAImax =

(1minus fPHU

1minus fPHUsen

LAIt =tsum

t prime= sdate

251 Decomposition

k(lPFT) =1

τ10PFT

middot g(Tsoil(l)

(l) (96)

+ 0719 middot θ(l)

(θ(l))

C(l) =

nPFTsumPFT= 1

kmeanPFT =

nsoilsuml=1

kmeanPFT

26 Water balance

Infiltration

infil= Pr middot

radic1minus

SW(1)minusWpwp(1)

Percolation

et al 1998)

) (102)

TT(l) =FW(l)

HC(l)(103)

HC(l) =Ks(l) middot

(SW(l)

Wsat(l)

)β(l) (104)

[1minus exp

(minus1t

)](105)

radic1minus

Canopy evaporation

nPFTsumPFT=1

Soil evaporation

and thus

nsoilminus1suml= 1

rootdistz =

int z0 (βroot)

primedzprime=

1minus (βroot)z

263 River routing

Qout(t)=Qin middot1

K middot0(n)

)nminus1

v (119)

Irrigation

Ec (121)

27 Land use

271 Sowing dates

LI =min(BI max

(BI+RminusLlr

1+ 1lr 0))

Cleaf =sumPFT

BmlPFT

MCleaf =sumPFT

MmlPFT (126)

1000+Cleaf(127)

312 Land use input

32 Standard outputs

4 Discussion

04 6 11 18 48 70

03 1 45 168

2e+09 7e+10 5e+11 4e+12

0 500 1000 1500 2000

0 200 400 600 800 1000

2 8 25 80 107

References

Abstract

Introduction

Model description



Albedo

Soil energy balance

Plant physiology

Photosynthesis

Phenology

Productivity


Allocation

Vegetation dynamics



Decomposition

Water balance

Soil water balance

Evapotranspiration

River routing

Irrigation and dams


Land use

Sowing dates


Managed grassland

Biomass plantations

Modelling protocol



Land use input

Standard outputs

Discussion



Competing interests

Acknowledgements

References

22 Plant physiology

221 Photosynthesis

JE = C1 middotAPAR

daylength (28)

(piminus0lowast

pi+ 2 middot0lowast

λmaxC4

) (30)

0lowast =[O2]

2 middot τ (31)

C2 =piminus0lowast

(1+ [O2]

) (33)

(JE + JC minus

Vm =1bmiddotC1

radic1minus

gc =16Adt

222 Phenology

223 Productivity

g(T )= exp[

30856 middot(

minus1

(T + 4602)

)] (45)

Reproduction cost

Tissue turnover

231 Allocation

klasa (52)

From this follows

CAind (54)

SLA=2times 10minus4

Establishment

ESTTREE = (58)

nestTREE

Stress mortality

twPFT1)

NI(Nd)=

Pspread =

0 ω0 gtme(62)

FDI=max

01minus

1memiddot exp

(minusNI middot

nsump=1

) (63)

a(ND)=Nhobs

tfire =241

Af =π

4 middotLBmiddotD2

HUsum =

tsumt prime= sdate

fPHU= HUsumPHU (74)

PHUle PHUwlow

(psminuspb)))+psens

fLAImax =fPHU

fPHUkminusfPHUc

c =fPHUc

k =fPHUk

fLAImax =

(1minus fPHU

1minus fPHUsen

LAIt =tsum

t prime= sdate

251 Decomposition

k(lPFT) =1

τ10PFT

middot g(Tsoil(l)

(l) (96)

+ 0719 middot θ(l)

(θ(l))

C(l) =

nPFTsumPFT= 1

kmeanPFT =

nsoilsuml=1

kmeanPFT

26 Water balance

Infiltration

infil= Pr middot

radic1minus

SW(1)minusWpwp(1)

Percolation

et al 1998)

) (102)

TT(l) =FW(l)

HC(l)(103)

HC(l) =Ks(l) middot

(SW(l)

Wsat(l)

)β(l) (104)

[1minus exp

(minus1t

)](105)

radic1minus

Canopy evaporation

nPFTsumPFT=1

Soil evaporation

and thus

nsoilminus1suml= 1

rootdistz =

int z0 (βroot)

primedzprime=

1minus (βroot)z

263 River routing

Qout(t)=Qin middot1

K middot0(n)

)nminus1

v (119)

Irrigation

Ec (121)

27 Land use

271 Sowing dates

LI =min(BI max

(BI+RminusLlr

1+ 1lr 0))

Cleaf =sumPFT

BmlPFT

MCleaf =sumPFT

MmlPFT (126)

1000+Cleaf(127)

312 Land use input

32 Standard outputs

4 Discussion

04 6 11 18 48 70

03 1 45 168

2e+09 7e+10 5e+11 4e+12

0 500 1000 1500 2000

0 200 400 600 800 1000

2 8 25 80 107

References

Abstract

Introduction

Model description



Albedo

Soil energy balance

Plant physiology

Photosynthesis

Phenology

Productivity


Allocation

Vegetation dynamics



Decomposition

Water balance

Soil water balance

Evapotranspiration

River routing

Irrigation and dams


Land use

Sowing dates


Managed grassland

Biomass plantations

Modelling protocol



Land use input

Standard outputs

Discussion



Competing interests

Acknowledgements

References

(JE + JC minus

Vm =1bmiddotC1

radic1minus

gc =16Adt

222 Phenology

223 Productivity

g(T )= exp[

30856 middot(

minus1

(T + 4602)

)] (45)

Reproduction cost

Tissue turnover

231 Allocation

klasa (52)

From this follows

CAind (54)

SLA=2times 10minus4

Establishment

ESTTREE = (58)

nestTREE

Stress mortality

twPFT1)

NI(Nd)=

Pspread =

0 ω0 gtme(62)

FDI=max

01minus

1memiddot exp

(minusNI middot

nsump=1

) (63)

a(ND)=Nhobs

tfire =241

Af =π

4 middotLBmiddotD2

HUsum =

tsumt prime= sdate

fPHU= HUsumPHU (74)

PHUle PHUwlow

(psminuspb)))+psens

fLAImax =fPHU

fPHUkminusfPHUc

c =fPHUc

k =fPHUk

fLAImax =

(1minus fPHU

1minus fPHUsen

LAIt =tsum

t prime= sdate

251 Decomposition

k(lPFT) =1

τ10PFT

middot g(Tsoil(l)

(l) (96)

+ 0719 middot θ(l)

(θ(l))

C(l) =

nPFTsumPFT= 1

kmeanPFT =

nsoilsuml=1

kmeanPFT

26 Water balance

Infiltration

infil= Pr middot

radic1minus

SW(1)minusWpwp(1)

Percolation

et al 1998)

) (102)

TT(l) =FW(l)

HC(l)(103)

HC(l) =Ks(l) middot

(SW(l)

Wsat(l)

)β(l) (104)

[1minus exp

(minus1t

)](105)

radic1minus

Canopy evaporation

nPFTsumPFT=1

Soil evaporation

and thus

nsoilminus1suml= 1

rootdistz =

int z0 (βroot)

primedzprime=

1minus (βroot)z

263 River routing

Qout(t)=Qin middot1

K middot0(n)

)nminus1

v (119)

Irrigation

Ec (121)

27 Land use

271 Sowing dates

LI =min(BI max

(BI+RminusLlr

1+ 1lr 0))

Cleaf =sumPFT

BmlPFT

MCleaf =sumPFT

MmlPFT (126)

1000+Cleaf(127)

312 Land use input

32 Standard outputs

4 Discussion

04 6 11 18 48 70

03 1 45 168

2e+09 7e+10 5e+11 4e+12

0 500 1000 1500 2000

0 200 400 600 800 1000

2 8 25 80 107

References

Abstract

Introduction

Model description



Albedo

Soil energy balance

Plant physiology

Photosynthesis

Phenology

Productivity


Allocation

Vegetation dynamics



Decomposition

Water balance

Soil water balance

Evapotranspiration

River routing

Irrigation and dams


Land use

Sowing dates


Managed grassland

Biomass plantations

Modelling protocol



Land use input

Standard outputs

Discussion



Competing interests

Acknowledgements

References

g(T )= exp[

30856 middot(

minus1

(T + 4602)

)] (45)

Reproduction cost

Tissue turnover

231 Allocation

klasa (52)

From this follows

CAind (54)

SLA=2times 10minus4

Establishment

ESTTREE = (58)

nestTREE

Stress mortality

twPFT1)

NI(Nd)=

Pspread =

0 ω0 gtme(62)

FDI=max

01minus

1memiddot exp

(minusNI middot

nsump=1

) (63)

a(ND)=Nhobs

tfire =241

Af =π

4 middotLBmiddotD2

HUsum =

tsumt prime= sdate

fPHU= HUsumPHU (74)

PHUle PHUwlow

(psminuspb)))+psens

fLAImax =fPHU

fPHUkminusfPHUc

c =fPHUc

k =fPHUk

fLAImax =

(1minus fPHU

1minus fPHUsen

LAIt =tsum

t prime= sdate

251 Decomposition

k(lPFT) =1

τ10PFT

middot g(Tsoil(l)

(l) (96)

+ 0719 middot θ(l)

(θ(l))

C(l) =

nPFTsumPFT= 1

kmeanPFT =

nsoilsuml=1

kmeanPFT

26 Water balance

Infiltration

infil= Pr middot

radic1minus

SW(1)minusWpwp(1)

Percolation

et al 1998)

) (102)

TT(l) =FW(l)

HC(l)(103)

HC(l) =Ks(l) middot

(SW(l)

Wsat(l)

)β(l) (104)

[1minus exp

(minus1t

)](105)

radic1minus

Canopy evaporation

nPFTsumPFT=1

Soil evaporation

and thus

nsoilminus1suml= 1

rootdistz =

int z0 (βroot)

primedzprime=

1minus (βroot)z

263 River routing

Qout(t)=Qin middot1

K middot0(n)

)nminus1

v (119)

Irrigation

Ec (121)

27 Land use

271 Sowing dates

LI =min(BI max

(BI+RminusLlr

1+ 1lr 0))

Cleaf =sumPFT

BmlPFT

MCleaf =sumPFT

MmlPFT (126)

1000+Cleaf(127)

312 Land use input

32 Standard outputs

4 Discussion

04 6 11 18 48 70

03 1 45 168

2e+09 7e+10 5e+11 4e+12

0 500 1000 1500 2000

0 200 400 600 800 1000

2 8 25 80 107

References

Abstract

Introduction

Model description



Albedo

Soil energy balance

Plant physiology

Photosynthesis

Phenology

Productivity


Allocation

Vegetation dynamics



Decomposition

Water balance

Soil water balance

Evapotranspiration

River routing

Irrigation and dams


Land use

Sowing dates


Managed grassland

Biomass plantations

Modelling protocol



Land use input

Standard outputs

Discussion



Competing interests

Acknowledgements

References

klasa (52)

From this follows

CAind (54)

SLA=2times 10minus4

Establishment

ESTTREE = (58)

nestTREE

Stress mortality

twPFT1)

NI(Nd)=

Pspread =

0 ω0 gtme(62)

FDI=max

01minus

1memiddot exp

(minusNI middot

nsump=1

) (63)

a(ND)=Nhobs

tfire =241

Af =π

4 middotLBmiddotD2

HUsum =

tsumt prime= sdate

fPHU= HUsumPHU (74)

PHUle PHUwlow

(psminuspb)))+psens

fLAImax =fPHU

fPHUkminusfPHUc

c =fPHUc

k =fPHUk

fLAImax =

(1minus fPHU

1minus fPHUsen

LAIt =tsum

t prime= sdate

251 Decomposition

k(lPFT) =1

τ10PFT

middot g(Tsoil(l)

(l) (96)

+ 0719 middot θ(l)

(θ(l))

C(l) =

nPFTsumPFT= 1

kmeanPFT =

nsoilsuml=1

kmeanPFT

26 Water balance

Infiltration

infil= Pr middot

radic1minus

SW(1)minusWpwp(1)

Percolation

et al 1998)

) (102)

TT(l) =FW(l)

HC(l)(103)

HC(l) =Ks(l) middot

(SW(l)

Wsat(l)

)β(l) (104)

[1minus exp

(minus1t

)](105)

radic1minus

Canopy evaporation

nPFTsumPFT=1

Soil evaporation

and thus

nsoilminus1suml= 1

rootdistz =

int z0 (βroot)

primedzprime=

1minus (βroot)z

263 River routing

Qout(t)=Qin middot1

K middot0(n)

)nminus1

v (119)

Irrigation

Ec (121)

27 Land use

271 Sowing dates

LI =min(BI max

(BI+RminusLlr

1+ 1lr 0))

Cleaf =sumPFT

BmlPFT

MCleaf =sumPFT

MmlPFT (126)

1000+Cleaf(127)

312 Land use input

32 Standard outputs

4 Discussion

04 6 11 18 48 70

03 1 45 168

2e+09 7e+10 5e+11 4e+12

0 500 1000 1500 2000

0 200 400 600 800 1000

2 8 25 80 107

References

Abstract

Introduction

Model description



Albedo

Soil energy balance

Plant physiology

Photosynthesis

Phenology

Productivity


Allocation

Vegetation dynamics



Decomposition

Water balance

Soil water balance

Evapotranspiration

River routing

Irrigation and dams


Land use

Sowing dates


Managed grassland

Biomass plantations

Modelling protocol



Land use input

Standard outputs

Discussion



Competing interests

Acknowledgements

References

Stress mortality

twPFT1)

NI(Nd)=

Pspread =

0 ω0 gtme(62)

FDI=max

01minus

1memiddot exp

(minusNI middot

nsump=1

) (63)

a(ND)=Nhobs

tfire =241

Af =π

4 middotLBmiddotD2

HUsum =

tsumt prime= sdate

fPHU= HUsumPHU (74)

PHUle PHUwlow

(psminuspb)))+psens

fLAImax =fPHU

fPHUkminusfPHUc

c =fPHUc

k =fPHUk

fLAImax =

(1minus fPHU

1minus fPHUsen

LAIt =tsum

t prime= sdate

251 Decomposition

k(lPFT) =1

τ10PFT

middot g(Tsoil(l)

(l) (96)

+ 0719 middot θ(l)

(θ(l))

C(l) =

nPFTsumPFT= 1

kmeanPFT =

nsoilsuml=1

kmeanPFT

26 Water balance

Infiltration

infil= Pr middot

radic1minus

SW(1)minusWpwp(1)

Percolation

et al 1998)

) (102)

TT(l) =FW(l)

HC(l)(103)

HC(l) =Ks(l) middot

(SW(l)

Wsat(l)

)β(l) (104)

[1minus exp

(minus1t

)](105)

radic1minus

Canopy evaporation

nPFTsumPFT=1

Soil evaporation

and thus

nsoilminus1suml= 1

rootdistz =

int z0 (βroot)

primedzprime=

1minus (βroot)z

263 River routing

Qout(t)=Qin middot1

K middot0(n)

)nminus1

v (119)

Irrigation

Ec (121)

27 Land use

271 Sowing dates

LI =min(BI max

(BI+RminusLlr

1+ 1lr 0))

Cleaf =sumPFT

BmlPFT

MCleaf =sumPFT

MmlPFT (126)

1000+Cleaf(127)

312 Land use input

32 Standard outputs

4 Discussion

04 6 11 18 48 70

03 1 45 168

2e+09 7e+10 5e+11 4e+12

0 500 1000 1500 2000

0 200 400 600 800 1000

2 8 25 80 107

References

Abstract

Introduction

Model description



Albedo

Soil energy balance

Plant physiology

Photosynthesis

Phenology

Productivity


Allocation

Vegetation dynamics



Decomposition

Water balance

Soil water balance

Evapotranspiration

River routing

Irrigation and dams


Land use

Sowing dates


Managed grassland

Biomass plantations

Modelling protocol



Land use input

Standard outputs

Discussion



Competing interests

Acknowledgements

References

tfire =241

Af =π

4 middotLBmiddotD2

HUsum =

tsumt prime= sdate

fPHU= HUsumPHU (74)

PHUle PHUwlow

(psminuspb)))+psens

fLAImax =fPHU

fPHUkminusfPHUc

c =fPHUc

k =fPHUk

fLAImax =

(1minus fPHU

1minus fPHUsen

LAIt =tsum

t prime= sdate

251 Decomposition

k(lPFT) =1

τ10PFT

middot g(Tsoil(l)

(l) (96)

+ 0719 middot θ(l)

(θ(l))

C(l) =

nPFTsumPFT= 1

kmeanPFT =

nsoilsuml=1

kmeanPFT

26 Water balance

Infiltration

infil= Pr middot

radic1minus

SW(1)minusWpwp(1)

Percolation

et al 1998)

) (102)

TT(l) =FW(l)

HC(l)(103)

HC(l) =Ks(l) middot

(SW(l)

Wsat(l)

)β(l) (104)

[1minus exp

(minus1t

)](105)

radic1minus

Canopy evaporation

nPFTsumPFT=1

Soil evaporation

and thus

nsoilminus1suml= 1

rootdistz =

int z0 (βroot)

primedzprime=

1minus (βroot)z

263 River routing

Qout(t)=Qin middot1

K middot0(n)

)nminus1

v (119)

Irrigation

Ec (121)

27 Land use

271 Sowing dates

LI =min(BI max

(BI+RminusLlr

1+ 1lr 0))

Cleaf =sumPFT

BmlPFT

MCleaf =sumPFT

MmlPFT (126)

1000+Cleaf(127)

312 Land use input

32 Standard outputs

4 Discussion

04 6 11 18 48 70

03 1 45 168

2e+09 7e+10 5e+11 4e+12

0 500 1000 1500 2000

0 200 400 600 800 1000

2 8 25 80 107

References

Abstract

Introduction

Model description



Albedo

Soil energy balance

Plant physiology

Photosynthesis

Phenology

Productivity


Allocation

Vegetation dynamics



Decomposition

Water balance

Soil water balance

Evapotranspiration

River routing

Irrigation and dams


Land use

Sowing dates


Managed grassland

Biomass plantations

Modelling protocol



Land use input

Standard outputs

Discussion



Competing interests

Acknowledgements

References

HUsum =

tsumt prime= sdate

fPHU= HUsumPHU (74)

PHUle PHUwlow

(psminuspb)))+psens

fLAImax =fPHU

fPHUkminusfPHUc

c =fPHUc

k =fPHUk

fLAImax =

(1minus fPHU

1minus fPHUsen

LAIt =tsum

t prime= sdate

251 Decomposition

k(lPFT) =1

τ10PFT

middot g(Tsoil(l)

(l) (96)

+ 0719 middot θ(l)

(θ(l))

C(l) =

nPFTsumPFT= 1

kmeanPFT =

nsoilsuml=1

kmeanPFT

26 Water balance

Infiltration

infil= Pr middot

radic1minus

SW(1)minusWpwp(1)

Percolation

et al 1998)

) (102)

TT(l) =FW(l)

HC(l)(103)

HC(l) =Ks(l) middot

(SW(l)

Wsat(l)

)β(l) (104)

[1minus exp

(minus1t

)](105)

radic1minus

Canopy evaporation

nPFTsumPFT=1

Soil evaporation

and thus

nsoilminus1suml= 1

rootdistz =

int z0 (βroot)

primedzprime=

1minus (βroot)z

263 River routing

Qout(t)=Qin middot1

K middot0(n)

)nminus1

v (119)

Irrigation

Ec (121)

27 Land use

271 Sowing dates

LI =min(BI max

(BI+RminusLlr

1+ 1lr 0))

Cleaf =sumPFT

BmlPFT

MCleaf =sumPFT

MmlPFT (126)

1000+Cleaf(127)

312 Land use input

32 Standard outputs

4 Discussion

04 6 11 18 48 70

03 1 45 168

2e+09 7e+10 5e+11 4e+12

0 500 1000 1500 2000

0 200 400 600 800 1000

2 8 25 80 107

References

Abstract

Introduction

Model description



Albedo

Soil energy balance

Plant physiology

Photosynthesis

Phenology

Productivity


Allocation

Vegetation dynamics



Decomposition

Water balance

Soil water balance

Evapotranspiration

River routing

Irrigation and dams


Land use

Sowing dates


Managed grassland

Biomass plantations

Modelling protocol



Land use input

Standard outputs

Discussion



Competing interests

Acknowledgements

References

fLAImax =fPHU

fPHUkminusfPHUc

c =fPHUc

k =fPHUk

fLAImax =

(1minus fPHU

1minus fPHUsen

LAIt =tsum

t prime= sdate

251 Decomposition

k(lPFT) =1

τ10PFT

middot g(Tsoil(l)

(l) (96)

+ 0719 middot θ(l)

(θ(l))

C(l) =

nPFTsumPFT= 1

kmeanPFT =

nsoilsuml=1

kmeanPFT

26 Water balance

Infiltration

infil= Pr middot

radic1minus

SW(1)minusWpwp(1)

Percolation

et al 1998)

) (102)

TT(l) =FW(l)

HC(l)(103)

HC(l) =Ks(l) middot

(SW(l)

Wsat(l)

)β(l) (104)

[1minus exp

(minus1t

)](105)

radic1minus

Canopy evaporation

nPFTsumPFT=1

Soil evaporation

and thus

nsoilminus1suml= 1

rootdistz =

int z0 (βroot)

primedzprime=

1minus (βroot)z

263 River routing

Qout(t)=Qin middot1

K middot0(n)

)nminus1

v (119)

Irrigation

Ec (121)

27 Land use

271 Sowing dates

LI =min(BI max

(BI+RminusLlr

1+ 1lr 0))

Cleaf =sumPFT

BmlPFT

MCleaf =sumPFT

MmlPFT (126)

1000+Cleaf(127)

312 Land use input

32 Standard outputs

4 Discussion

04 6 11 18 48 70

03 1 45 168

2e+09 7e+10 5e+11 4e+12

0 500 1000 1500 2000

0 200 400 600 800 1000

2 8 25 80 107

References

Abstract

Introduction

Model description



Albedo

Soil energy balance

Plant physiology

Photosynthesis

Phenology

Productivity


Allocation

Vegetation dynamics



Decomposition

Water balance

Soil water balance

Evapotranspiration

River routing

Irrigation and dams


Land use

Sowing dates


Managed grassland

Biomass plantations

Modelling protocol



Land use input

Standard outputs

Discussion



Competing interests

Acknowledgements

References

251 Decomposition

k(lPFT) =1

τ10PFT

middot g(Tsoil(l)

(l) (96)

+ 0719 middot θ(l)

(θ(l))

C(l) =

nPFTsumPFT= 1

kmeanPFT =

nsoilsuml=1

kmeanPFT

26 Water balance

Infiltration

infil= Pr middot

radic1minus

SW(1)minusWpwp(1)

Percolation

et al 1998)

) (102)

TT(l) =FW(l)

HC(l)(103)

HC(l) =Ks(l) middot

(SW(l)

Wsat(l)

)β(l) (104)

[1minus exp

(minus1t

)](105)

radic1minus

Canopy evaporation

nPFTsumPFT=1

Soil evaporation

and thus

nsoilminus1suml= 1

rootdistz =

int z0 (βroot)

primedzprime=

1minus (βroot)z

263 River routing

Qout(t)=Qin middot1

K middot0(n)

)nminus1

v (119)

Irrigation

Ec (121)

27 Land use

271 Sowing dates

LI =min(BI max

(BI+RminusLlr

1+ 1lr 0))

Cleaf =sumPFT

BmlPFT

MCleaf =sumPFT

MmlPFT (126)

1000+Cleaf(127)

312 Land use input

32 Standard outputs

4 Discussion

04 6 11 18 48 70

03 1 45 168

2e+09 7e+10 5e+11 4e+12

0 500 1000 1500 2000

0 200 400 600 800 1000

2 8 25 80 107

References

Abstract

Introduction

Model description



Albedo

Soil energy balance

Plant physiology

Photosynthesis

Phenology

Productivity


Allocation

Vegetation dynamics



Decomposition

Water balance

Soil water balance

Evapotranspiration

River routing

Irrigation and dams


Land use

Sowing dates


Managed grassland

Biomass plantations

Modelling protocol



Land use input

Standard outputs

Discussion



Competing interests

Acknowledgements

References

Infiltration

infil= Pr middot

radic1minus

SW(1)minusWpwp(1)

Percolation

et al 1998)

) (102)

TT(l) =FW(l)

HC(l)(103)

HC(l) =Ks(l) middot

(SW(l)

Wsat(l)

)β(l) (104)

[1minus exp

(minus1t

)](105)

radic1minus

Canopy evaporation

nPFTsumPFT=1

Soil evaporation

and thus

nsoilminus1suml= 1

rootdistz =

int z0 (βroot)

primedzprime=

1minus (βroot)z

263 River routing

Qout(t)=Qin middot1

K middot0(n)

)nminus1

v (119)

Irrigation

Ec (121)

27 Land use

271 Sowing dates

LI =min(BI max

(BI+RminusLlr

1+ 1lr 0))

Cleaf =sumPFT

BmlPFT

MCleaf =sumPFT

MmlPFT (126)

1000+Cleaf(127)

312 Land use input

32 Standard outputs

4 Discussion

04 6 11 18 48 70

03 1 45 168

2e+09 7e+10 5e+11 4e+12

0 500 1000 1500 2000

0 200 400 600 800 1000

2 8 25 80 107

References

Abstract

Introduction

Model description



Albedo

Soil energy balance

Plant physiology

Photosynthesis

Phenology

Productivity


Allocation

Vegetation dynamics



Decomposition

Water balance

Soil water balance

Evapotranspiration

River routing

Irrigation and dams


Land use

Sowing dates


Managed grassland

Biomass plantations

Modelling protocol



Land use input

Standard outputs

Discussion



Competing interests

Acknowledgements

References

Canopy evaporation

nPFTsumPFT=1

Soil evaporation

and thus

nsoilminus1suml= 1

rootdistz =

int z0 (βroot)

primedzprime=

1minus (βroot)z

263 River routing

Qout(t)=Qin middot1

K middot0(n)

)nminus1

v (119)

Irrigation

Ec (121)

27 Land use

271 Sowing dates

LI =min(BI max

(BI+RminusLlr

1+ 1lr 0))

Cleaf =sumPFT

BmlPFT

MCleaf =sumPFT

MmlPFT (126)

1000+Cleaf(127)

312 Land use input

32 Standard outputs

4 Discussion

04 6 11 18 48 70

03 1 45 168

2e+09 7e+10 5e+11 4e+12

0 500 1000 1500 2000

0 200 400 600 800 1000

2 8 25 80 107

References

Abstract

Introduction

Model description



Albedo

Soil energy balance

Plant physiology

Photosynthesis

Phenology

Productivity


Allocation

Vegetation dynamics



Decomposition

Water balance

Soil water balance

Evapotranspiration

River routing

Irrigation and dams


Land use

Sowing dates


Managed grassland

Biomass plantations

Modelling protocol



Land use input

Standard outputs

Discussion



Competing interests

Acknowledgements

References

Qout(t)=Qin middot1

K middot0(n)

)nminus1

v (119)

Irrigation

Ec (121)

27 Land use

271 Sowing dates

LI =min(BI max

(BI+RminusLlr

1+ 1lr 0))

Cleaf =sumPFT

BmlPFT

MCleaf =sumPFT

MmlPFT (126)

1000+Cleaf(127)

312 Land use input

32 Standard outputs

4 Discussion

04 6 11 18 48 70

03 1 45 168

2e+09 7e+10 5e+11 4e+12

0 500 1000 1500 2000

0 200 400 600 800 1000

2 8 25 80 107

References

Abstract

Introduction

Model description



Albedo

Soil energy balance

Plant physiology

Photosynthesis

Phenology

Productivity


Allocation

Vegetation dynamics



Decomposition

Water balance

Soil water balance

Evapotranspiration

River routing

Irrigation and dams


Land use

Sowing dates


Managed grassland

Biomass plantations

Modelling protocol



Land use input

Standard outputs

Discussion



Competing interests

Acknowledgements

References

27 Land use

271 Sowing dates

LI =min(BI max

(BI+RminusLlr

1+ 1lr 0))

Cleaf =sumPFT

BmlPFT

MCleaf =sumPFT

MmlPFT (126)

1000+Cleaf(127)

312 Land use input

32 Standard outputs

4 Discussion

04 6 11 18 48 70

03 1 45 168

2e+09 7e+10 5e+11 4e+12

0 500 1000 1500 2000

0 200 400 600 800 1000

2 8 25 80 107

References

Abstract

Introduction

Model description



Albedo

Soil energy balance

Plant physiology

Photosynthesis

Phenology

Productivity


Allocation

Vegetation dynamics



Decomposition

Water balance

Soil water balance

Evapotranspiration

River routing

Irrigation and dams


Land use

Sowing dates


Managed grassland

Biomass plantations

Modelling protocol



Land use input

Standard outputs

Discussion



Competing interests

Acknowledgements

References

LI =min(BI max

(BI+RminusLlr

1+ 1lr 0))

Cleaf =sumPFT

BmlPFT

MCleaf =sumPFT

MmlPFT (126)

1000+Cleaf(127)

312 Land use input

32 Standard outputs

4 Discussion

04 6 11 18 48 70

03 1 45 168

2e+09 7e+10 5e+11 4e+12

0 500 1000 1500 2000

0 200 400 600 800 1000

2 8 25 80 107

References

Abstract

Introduction

Model description



Albedo

Soil energy balance

Plant physiology

Photosynthesis

Phenology

Productivity


Allocation

Vegetation dynamics



Decomposition

Water balance

Soil water balance

Evapotranspiration

River routing

Irrigation and dams


Land use

Sowing dates


Managed grassland

Biomass plantations

Modelling protocol



Land use input

Standard outputs

Discussion



Competing interests

Acknowledgements

References

LI =min(BI max

(BI+RminusLlr

1+ 1lr 0))

Cleaf =sumPFT

BmlPFT

MCleaf =sumPFT

MmlPFT (126)

1000+Cleaf(127)

312 Land use input

32 Standard outputs

4 Discussion

04 6 11 18 48 70

03 1 45 168

2e+09 7e+10 5e+11 4e+12

0 500 1000 1500 2000

0 200 400 600 800 1000

2 8 25 80 107

References

Abstract

Introduction

Model description



Albedo

Soil energy balance

Plant physiology

Photosynthesis

Phenology

Productivity


Allocation

Vegetation dynamics



Decomposition

Water balance

Soil water balance

Evapotranspiration

River routing

Irrigation and dams


Land use

Sowing dates


Managed grassland

Biomass plantations

Modelling protocol



Land use input

Standard outputs

Discussion



Competing interests

Acknowledgements

References

312 Land use input

32 Standard outputs

4 Discussion

04 6 11 18 48 70

03 1 45 168

2e+09 7e+10 5e+11 4e+12

0 500 1000 1500 2000

0 200 400 600 800 1000

2 8 25 80 107

References

Abstract

Introduction

Model description



Albedo

Soil energy balance

Plant physiology

Photosynthesis

Phenology

Productivity


Allocation

Vegetation dynamics



Decomposition

Water balance

Soil water balance

Evapotranspiration

River routing

Irrigation and dams


Land use

Sowing dates


Managed grassland

Biomass plantations

Modelling protocol



Land use input

Standard outputs

Discussion



Competing interests

Acknowledgements

References

4 Discussion

04 6 11 18 48 70

03 1 45 168

2e+09 7e+10 5e+11 4e+12

0 500 1000 1500 2000

0 200 400 600 800 1000

2 8 25 80 107

References

Abstract

Introduction

Model description



Albedo

Soil energy balance

Plant physiology

Photosynthesis

Phenology

Productivity


Allocation

Vegetation dynamics



Decomposition

Water balance

Soil water balance

Evapotranspiration

River routing

Irrigation and dams


Land use

Sowing dates


Managed grassland

Biomass plantations

Modelling protocol



Land use input

Standard outputs

Discussion



Competing interests

Acknowledgements

References

04 6 11 18 48 70

03 1 45 168

2e+09 7e+10 5e+11 4e+12

0 500 1000 1500 2000

0 200 400 600 800 1000

2 8 25 80 107

References

Abstract

Introduction

Model description



Albedo

Soil energy balance

Plant physiology

Photosynthesis

Phenology

Productivity


Allocation

Vegetation dynamics



Decomposition

Water balance

Soil water balance

Evapotranspiration

River routing

Irrigation and dams


Land use

Sowing dates


Managed grassland

Biomass plantations

Modelling protocol



Land use input

Standard outputs

Discussion



Competing interests

Acknowledgements

References

References

Abstract

Introduction

Model description



Albedo

Soil energy balance

Plant physiology

Photosynthesis

Phenology

Productivity


Allocation

Vegetation dynamics



Decomposition

Water balance

Soil water balance

Evapotranspiration

River routing

Irrigation and dams


Land use

Sowing dates


Managed grassland

Biomass plantations

Modelling protocol



Land use input

Standard outputs

Discussion



Competing interests

Acknowledgements

References

References

Abstract

Introduction

Model description



Albedo

Soil energy balance

Plant physiology

Photosynthesis

Phenology

Productivity


Allocation

Vegetation dynamics



Decomposition

Water balance

Soil water balance

Evapotranspiration

River routing

Irrigation and dams


Land use

Sowing dates


Managed grassland

Biomass plantations

Modelling protocol



Land use input

Standard outputs

Discussion



Competing interests

Acknowledgements

References

References

Abstract

Introduction

Model description



Albedo

Soil energy balance

Plant physiology

Photosynthesis

Phenology

Productivity


Allocation

Vegetation dynamics



Decomposition

Water balance

Soil water balance

Evapotranspiration

River routing

Irrigation and dams


Land use

Sowing dates


Managed grassland

Biomass plantations

Modelling protocol



Land use input

Standard outputs

Discussion



Competing interests

Acknowledgements

References

Abstract

Introduction

Model description



Albedo

Soil energy balance

Plant physiology

Photosynthesis

Phenology

Productivity


Allocation

Vegetation dynamics



Decomposition

Water balance

Soil water balance

Evapotranspiration

River routing

Irrigation and dams


Land use

Sowing dates


Managed grassland

Biomass plantations

Modelling protocol



Land use input

Standard outputs

Discussion



Competing interests

Acknowledgements

References

Abstract

Introduction

Model description



Albedo

Soil energy balance

Plant physiology

Photosynthesis

Phenology

Productivity


Allocation

Vegetation dynamics



Decomposition

Water balance

Soil water balance

Evapotranspiration

River routing

Irrigation and dams


Land use

Sowing dates


Managed grassland

Biomass plantations

Modelling protocol



Land use input

Standard outputs

Discussion



Competing interests

Acknowledgements

References

Abstract

Introduction

Model description



Albedo

Soil energy balance

Plant physiology

Photosynthesis

Phenology

Productivity


Allocation

Vegetation dynamics



Decomposition

Water balance

Soil water balance

Evapotranspiration

River routing

Irrigation and dams


Land use

Sowing dates


Managed grassland

Biomass plantations

Modelling protocol



Land use input

Standard outputs

Discussion



Competing interests

Acknowledgements

References

Abstract

Introduction

Model description



Albedo

Soil energy balance

Plant physiology

Photosynthesis

Phenology

Productivity


Allocation

Vegetation dynamics



Decomposition

Water balance

Soil water balance

Evapotranspiration

River routing

Irrigation and dams


Land use

Sowing dates


Managed grassland

Biomass plantations

Modelling protocol



Land use input

Standard outputs

Discussion



Competing interests

Acknowledgements

References

Abstract

Introduction

Model description



Albedo

Soil energy balance

Plant physiology

Photosynthesis

Phenology

Productivity


Allocation

Vegetation dynamics



Decomposition

Water balance

Soil water balance

Evapotranspiration

River routing

Irrigation and dams


Land use

Sowing dates


Managed grassland

Biomass plantations

Modelling protocol



Land use input

Standard outputs

Discussion



Competing interests

Acknowledgements

References

Abstract

Introduction

Model description



Albedo

Soil energy balance

Plant physiology

Photosynthesis

Phenology

Productivity


Allocation

Vegetation dynamics



Decomposition

Water balance

Soil water balance

Evapotranspiration

River routing

Irrigation and dams


Land use

Sowing dates


Managed grassland

Biomass plantations

Modelling protocol



Land use input

Standard outputs

Discussion



Competing interests

Acknowledgements

References

Abstract

Introduction

Model description



Albedo

Soil energy balance

Plant physiology

Photosynthesis

Phenology

Productivity


Allocation

Vegetation dynamics



Decomposition

Water balance

Soil water balance

Evapotranspiration

River routing

Irrigation and dams


Land use

Sowing dates


Managed grassland

Biomass plantations

Modelling protocol



Land use input

Standard outputs

Discussion



Competing interests

Acknowledgements

References

Documents

New LPJmL4 – a dynamic global vegetation model with managed … · 2020. 6. 23. · LPJmL4 – a dynamic global vegetation model with managed land ... DGVMs were further developed