Geosci. Model De ., 9, 2951–2972, 2016
www.geosci-model-de .ne /9/2951/2016/
doi:10.5194/gmd-9-2951-2016
© Au ho (s) 2016. CC A ibu ion 3.0 License.
E alua ing he pe o mance o land su ace model
ORCHIDEE-CAN 1.0 on wa e and ene gy lux es ima ion wi h a
single- and mul i-laye ene gy budge scheme
Yiying Chen1,a, James Ryde 1, Vladisla Bas iko 1, Ma hew J. McG a h1, Kim Naud s1,b, Juliane O o1,c,
Ca he ine O lé1, Philippe Peylin1, Jan Polche 2, Aude Valade3, And ew Black4, Jan A. Elbe s5, Eddy Moo s5,
Thomas Foken6, E a an Go sel7, Vanessa Ha e d7, Be na d Heinesch8, F ank Tiedemann9, Alexande Knohl9,
Samuli Launiainen10, Denis Lous au11, Jé ôme Ogée11, Timo Vessala12,13, and Sebas iaan Luyssae 1,d
1Labo a oi e des Sciences du Clima e de l’En i onnemen , LSCE/IPSL, CEA-CNRS-UVSQ, Uni e si é Pa is-Saclay,
91191 Gi -su -Y e e, F ance
2Labo a oi e de Mé éo ologie Dynamique (LMD, CNRS), Ecole Poly echnique, Palaiseau, F ance
3Ins i u Pie e Simon Laplace, Place Jussieu 4, 75010 Pa is, F ance
4Land and Food Sys ems, Uni e si y o B i ish Columbia, Vancou e , BC, Canada
5Al e a, Wageningen UR, Wageningen, he Ne he lands
6Depa men o Mic ome eo ology Uni e si y o Bay eu h, Bay eu h Cen e o Ecology and En i onmen al Resea ch,
Bay eu h, Ge many
7CSIRO, Ma ine and A mosphe ic Resea ch, Canbe a, Aus alia
8Dep . Biosys em Enginee ing (BIOSE), Uni e si y o Liege, Gembloux, Belgium
9Dep . Bioclima ology, Geo g-Augus Uni e si y o Gö ingen, Büsgenweg, Gö ingen, Ge many
10Na u al Resou ces Ins i u e Finland, Van aa, Finland
11INRA UMR 1391 ISPA Cen e de Bo deaux Aqui aine, Bo deaux, F ance
12Depa men o Physics, Uni e si y o Helsinki, Helsinki, Finland
13Depa men o Fo es Sciences, Uni e si y o Helsinki, Helsinki, Finland
anow a : Resea ch Cen e o En i onmen al Changes (RCEC), Academia Sinica, Taipei, Taiwan
bnow a : Depa men o Land in he Ea h Sys em, Max Planck Ins i u e o Me eo ology, Hambu g, Ge many
cnow a : Clima e Se ice Cen e Ge many (GERICS), Helmhol z-Zen um Gees hach , Hambu g, Ge many
dnow a : Depa men o Ecological Sciences, VU Uni e si y, Ams e dam, he Ne he lands
Co espondence o: Yiying Chen ([email p o ec ed])
Recei ed: 2 Feb ua y 2016 – Published in Geosci. Model De . Discuss.: 25 Feb ua y 2016
Re ised: 5 Augus 2016 – Accep ed: 8 Augus 2016 – Published: 2 Sep embe 2016
Abs ac . Canopy s uc u e is one o he mos impo an
ege a ion cha ac e is ics o land–a mosphe e in e ac ions,
as i de e mines he ene gy and scala exchanges be ween
he land su ace and he o e lying ai mass. In his s udy
we e alua ed he pe o mance o a newly de eloped mul i-
laye ene gy budge in he ORCHIDEE-CAN 1.0 land su -
ace model (O ganising Ca bon and Hyd ology In Dynamic
Ecosys ems – CANopy), which simula es canopy s uc u e
and can be coupled o an a mosphe ic model using an im-
plici coupling p ocedu e. We aim o p o ide a se o accep -
able pa ame e alues o a ange o o es ypes. Top-canopy
and sub-canopy lux obse a ions om eigh si es we e col-
lec ed in o de o conduc his e alua ion. The si es c ossed
clima e zones om empe a e o bo eal and he ege a ion
ypes included deciduous, e e g een b oad-lea ed and e e -
g een needle-lea ed o es wi h a maximum lea a ea index
(LAI; all-sided) anging om 3.5 o 7.0. The pa ame iza ion
app oach p oposed in his s udy was based on h ee selec ed
physical p ocesses – namely he di usion, ad ec ion, and
u bulen mixing wi hin he canopy. Sho - e m sub-canopy
Published by Cope nicus Publica ions on behal o he Eu opean Geosciences Union.
2952 Y. Chen e al.: E alua ing he pe o mance o ORCHIDEE-CAN 1.0
obse a ions and long- e m su ace luxes we e used o cali-
b a e he pa ame e s in he sub-canopy adia ion, u bulence,
and esis ance modules wi h an au oma ic uning p ocess.
The mul i-laye model was ound o cap u e he dynamics
o sub-canopy u bulence, empe a u e, and ene gy luxes.
The pe o mance o he new mul i-laye model was u he
compa ed agains he exis ing single-laye model. Al hough
he mul i-laye model simula ion esul s showed ew o no
imp o emen s o bo h he nigh ime ene gy balance and en-
e gy pa i ioning du ing win e compa ed wi h a single-laye
model simula ion, he inc eased model complexi y does p o-
ide a mo e de ailed desc ip ion o he canopy mic ome eo-
ology o a ious o es ypes. The mul i-laye model links o
po en ial u u e en i onmen al and ecological s udies such as
he assessmen o in-canopy species ulne abili y o clima e
change, he clima e e ec s o dis u bance in ensi ies and e-
quencies, and he consequences o biogenic ola ile o ganic
compound (BVOC) emissions om he e es ial ecosys em.
1 In oduc ion
Today’s Ea h sys em models (ESMs) in eg a e ocean, ice
shee , a mosphe e, and land su ace in o de o p o ide a
powe ul ool o simula e he Ea h’s pas , p esen , and u-
u e clima es (D obinski e al., 2012). In such a model, he
land su ace sub-model p o ides he su ace luxes o he a -
mosphe ic sub-model, a ec s he dynamics o he plane a y
bounda y laye , and exe s a s ong in luence on he clima e.
The dynamics o he simula ed su ace luxes ely on he land
su ace sub-model ha , o e he pas 40 yea s, has e ol ed
om a simple bucke model app oach owa ds sophis ica ed
soil– ege a ion–a mosphe e ans e (SVAT) schemes (Pi -
man, 2003; S öckli and Vidale, 2005).
Al hough p esen -day land su ace models di e om
each o he in hei o mula ion and de ails, hei pe o -
mances show simila de iciencies. Fo example, imposing he
same land co e changes o se en land su ace models e-
sul ed in di e ging clima e e ec s. Amongs o he ac o s,
his di e gence was due o he pa ame iza ion o albedo and
he ep esen a ion o e apo anspi a ion o di e en land
co e ypes (Pi man e al., 2009). Di icul ies in ep oduc-
ing luxes o sensible and la en hea o a wide ange o
ege a ion ypes ha e been asc ibed o he so-called “big-
lea ” app oach (Bonan, 1996; Selle s e al., 1996; Dickinson
e al., 1998; Jiménez e al., 2011) which ea s he su ace as
a iso he mal la ge lea . Po en ially, ep esen ing he e ical
canopy s uc u e in de ail and simula ing adia ion pa i ion-
ing and u bulen anspo wi hin he ege a ion will esul
in an imp o ed de e mina ion o sensible and la en hea lux
es ima es (Baldocchi and Wilson, 2001; Ogée e al., 2003;
Bonan e al., 2014). Fo example, se e al mul i-laye SVAT
schemes ha e been p oposed and alida ed wi h si e-le el
obse a ions (Ogée e al., 2003; S aud e al., 2011; Ha e d
e al., 2012; Launiainen e al., 2015). These s udies demon-
s a ed ha bo h op-canopy and wi hin-canopy luxes and
mic ome eo ological p o iles could be cap u ed by means o
a sophis ica ed pa ame iza ion scheme o desc ibe he eg-
e a ion dynamics and he coupling be ween he a mosphe e
and he canopy.
Because he s anda d e sion o ORCHIDEE (O ganis-
ing Ca bon and Hyd ology In Dynamic Ecosys ems) makes
use o a big-lea app oach (Ducoud é e al., 1993; K in-
ne e al., 2005), imp o ed model capaci y and pe o mance
we e aimed o by implemen a ion o a mul i-laye ene gy
budge scheme (Ryde e al., 2016) ha was in eg a ed wi h
e ically disc e e e lec i i y, pho osyn hesis, s oma al esis-
ance, and ca bon alloca ion schemes. This new design e-
sul ed in a new e sion o ORCHIDEE named ORCHIDEE-
CAN (ORCHIDEE-CANopy, e ision 2290) (Naud s e al.,
2015). Despi e i s code including a mul i-laye ene gy bud-
ge scheme (Ryde e al., 2016), ORCHIDEE-CAN is cu -
en ly applied using a single-laye ene gy budge , due o a
lack o alida ed pa ame e s o he mul i-laye ene gy bud-
ge scheme.
In Ryde e al. (2016), he model was de eloped and
es ed o a single si e. In his s udy, we compiled a se o
wi hin-canopy and abo e-canopy measu emen s o ene gy,
wa e , and CO2 luxes and used hese da a o pa ame ize
and alida e he new mul i-laye ene gy budge scheme in
he ORCHIDEE-CAN 1.0 ( e ision 2754) global-scale land
su ace model. The da a se allowed us o es he model
unde di e se en i onmen al condi ions in o de o demon-
s a e ha he nume ics can deal wi h he a ia ion ha can
be ound in global ecosys ems. Fo his we g an ed ou sel es
he eedom o de i e a sepa a e pa ame e se o each si e.
Model pe o mance o he new mul i-laye pa ame iza ion
was compa ed agains he exis ing single-laye model. By do-
ing so we lea ned abou he s eng hs and weaknesses o he
model and i s pa ame e s. In subsequen s udies, we will ha e
o de i e a single pa ame e se o each plan unc ional ype
(PFT) and es how well he model ep oduces global pa e ns
in, o example, e apo anspi a ion.
2 Me hodology
2.1 Mul i-laye ene gy budge scheme
The mul i-laye ene gy budge scheme used in his s udy
was de eloped o global land su ace models (Ryde e al.,
2016) and he calcula ions di e om he mo e common
big-lea ene gy budge scheme in h ee aspec s. The new
scheme calcula es he ollowing: (a) wi hin-canopy longwa e
and sho wa e adia ion based on a e ical lea a ea in-
dex (LAI; m2m−2) p o ile, (b) a wi hin-canopy and below-
canopy wind p o ile based on he e ical LAI p o ile, and
(c) he dependency o s oma al esis ance and ae odynamic
esis ance based on he mic oclima ological condi ions along
Geosci. Model De ., 9, 2951–2972, 2016 www.geosci-model-de .ne /9/2951/2016/
Y. Chen e al.: E alua ing he pe o mance o ORCHIDEE-CAN 1.0 2953
Table 1. Symbolic no a ion used h oughou he pape .
Symbol Desc ip ion Uni
a1,a2,a3,a4,a5 uning coe icien s o CDe uni less
a6 ac o ceiling o he slope uni less
a7c i ical ic ion eloci y in he middle poin o he S-shape unc ion uni less
a8 ac o o cons ain he S-shape unc ion uni less
a9 h eshold o ege a ion co e uni less
a10 linea weigh ing ac o uni less
Aassimila ion a e µmolm−2s−1
CDe e ec i e d ag coe icien uni less
CSconcen a ion o CO2a lea su ace ppm
CD,i e ically disc e ized es ima e o canopy d ag coe icien uni less
Dh,ai hea di usi i y o ai cm2s−1
Dh,H2Ohea di usi i y o wa e apou cm2s−1
dlcha ac e is ic lea leng h m
Pgap o e -s o y gap p obabili y om Pgap ac ion m2m−2
G eg logic a iable o indica e he g ow h s a us o he ege a ion uni less
g0 esidual s oma al conduc ance i he i adiance app oaches ze o ms−1
hs ela i e humidi y a lea su ace %
hccanopy heigh m
kidi usi i y o le el im2s−1
k∗
imodi ied di usi i y o le el im2s−1
ksu conduc ance o he su ace–a mosphe e in e ace ms−1
LAIilea a ea index a le el im2m−2
Nu Nussel numbe uni less
Pm,i momen um shielding ac o uni less
PAI plan a ea index m2m−2
Rco ela ion coe icien be ween he simula ion and he obse a ion uni less
R0maximum co ela ion coe icien uni less
Rb,i bounda y laye esis ance a le el i o hea sm−1
R0
b,i bounda y laye esis ance a le el i o wa e apou sm−1
Rs,i s oma al esis ance a le el ism−1
Re Reynold’s numbe uni less
SLA speci ic lea a ea m2g−1
STTaylo skill sco e uni less
Tweek weekly mean ai empe a u e K
Tg empe a u e h eshold o unde -s o y phenology K
TLLag angian imescale s
u∗ ic ion eloci y ms−1
ui eloci y a le el ims−1
Vcmax ca boxyla ion capaci y µmolm−2s−1
Wb weigh ing pa ame e o bounda y laye esis ance uni less
Wn nea - ield weigh ing ac o uni less
Ws weigh ing pa ame e o soil–a mosphe e conduc ance uni less
Ws linea educ ion pa ame e o s oma al esis ance uni less
β3 ac ion o po en ial plan anspi a ion ealized uni less
β4 ac ion o soil e apo a ion ealized uni less
µkinema ic iscosi y o ai cm2s−1
ˆσ a io o he a iance o he simula ions o e he a iances o obse a ions uni less
σws anda d de ia ion in e ical eloci y ms−1
he LAI p o ile. All symbols a e explained in Table 1. In
he ollowing pa ag aphs hese calcula ions a e u he de-
sc ibed.
a. The mul i-laye ene gy budge scheme makes use o he
longwa e adia ion ans e scheme p oposed by Gao
e al. (1989) and Gu e al. (1999). The scheme simula es
longwa e adia ion anspo , as well as sca e ing and
abso p ion, along a e ically laye ed lea a ea dis ibu-
ion. The simula ed longwa e adia ion wi hin a laye
depends on he emi ed longwa e adia ion by all o i s
www.geosci-model-de .ne /9/2951/2016/ Geosci. Model De ., 9, 2951–2972, 2016
2954 Y. Chen e al.: E alua ing he pe o mance o ORCHIDEE-CAN 1.0
neighbou ing laye s. The sho wa e adia ion ans e
scheme, de eloped by Pin y e al. (2006), was applied
o he albedo calcula ion. The scheme compu es he ab-
so p ion, ansmission, and e lec ion o incoming adia-
ion by ege a ion canopies, which depends on he sola
zeni h angle, he ype o illumina ion (di ec o di use),
he ege a ion ype, and he ege a ion s uc u e. This
scheme conside s sho wa e adia ion bo h om isible
and nea -in a ed bands and was o iginally de eloped
o single-laye canopies, bu has since been ex ended
o use wi h laye ed canopies (McG a h e al., 2016).
b. The wind p o ile and he e ical eddy di usi i y
(k; m2s−1) a e calcula ed using he one-dimensional
second-o de closu e model o Massman and Weil
(1999), which makes use o he LAI p o ile o he s and.
I calcula es wind p o ile and e ical eddy di usi i y
based on Lag angian heo y.
c. The ae odynamic esis ance (Rb; sm−1) is calcu-
la ed based upon he lea bounda y-laye esis ance,
which is es ima ed acco ding o Baldocchi (1988).
The s oma al esis ance (Rs; sm−1) is calcula ed using
a Fa quha – on Caemme e –Be y- ype C3 (Fa quha
e al., 1980) and Colla z- ype C4 pho osyn hesis model
(Colla z e al., 1992) which simul aneously sol es ca -
bon assimila ion and s oma al conduc ance a he lea
le el bu excludes mesophyll conduc ance calcula ion.
ORCHIDEE-CAN 1.0 uses an analy ical app oach as
desc ibed by Yin and S uik (2009) o calcula e laye ed
s oma al esis ances which depend on he ambien ai
empe a u e, humidi y, wi hin-canopy CO2concen a-
ion, ege a ion-speci ic maximum ca boxyla ion a e,
and wa e supply om he oo s o he s oma a.
Reade s a e e e ed o Ryde e al. (2016) o a com-
p ehensi e desc ip ion o he mul i-laye ene gy budge , i s
assump ions, ma hema ical de ails, and a p oo o concep .
No e ha in ORCHIDEE-CAN 1.0 LAI is calcula ed om
a p ognos ic lea mass by making use o a ege a ion-speci ic
speci ic lea a ea (SLA; m2g−1). The calcula ion o he e -
ical and ho izon al dis ibu ions o he lea mass, and hus
he ege a ion canopy, depends on plan phenology, in a-
s and compe i ion, o es managemen , and allome ic ela-
ionships, and is de ailed in Naud s e al. (2015).
2.2 Obse a ional da a
Fo his s udy o es si es we e e ained i he ollowing
da a we e a ailable: (a) sho bu in ensi e campaigns ak-
ing lux and p o ile measu emen s wi hin and/o below he
ee canopy, and (b) mul i-yea moni o ing o op-canopy
luxes. Th ough nume ous egional p ojec s such as CAR-
BOEUROPE, AMERIFLUX, Fluxne Canada, OZFLUX,
ICOS, and NEON, and e o s such as FLUXNET (Baldoc-
chi and Wilson, 2001), mul iple yea -long ime se ies a e
now commonly a ailable, especially o he empe a e and
bo eal zones in Eu ope, Japan, Aus alia, and No h Ame -
ica. Si e selec ion was hus mos ly limi ed by he a ailabili y
o wi hin-canopy and below-canopy measu emen s.
Eigh lux obse a ion si es (Table 2) me he a o emen-
ioned c i e ia, and ep esen ed a ious clima es om he
Medi e anean o he bo eal zone and di e en ege a ion
ypes including b oad-lea ed summe g een, b oad-lea ed
e e g een and needle-lea ed e e g een. Da a we e hus miss-
ing om needle-lea ed summe g een ege a ion such as
la ch (La ix sp.) and opical ege a ion, so i was no pos-
sible o co e all o he o es ypes ha a e conside ed in
ORCHIDEE-CAN.
The sho in ensi e campaigns, aking measu emen s
wi hin-canopy and below-canopy, usually ex ended o pe-
iods anging om se e al days o a ew weeks (Pe iod I;
Table 3). Du ing in ensi e campaigns, e ical p o ile mea-
su emen s o wind speed, empe a u e, and a mosphe ic hu-
midi y we e ypically conduc ed. Such measu emen s we e
some imes complemen ed wi h p o ile measu emen s o sen-
sible and la en hea luxes, as well as sub-canopy adia-
ion measu emen s (Pe iod II and III; Table 3). Fu he mo e,
ou pa ame iza ion and alida ion se -up equi ed ha op-
canopy obse a ions had o be a ailable o pe iods exceed-
ing 1 yea (Pe iod IV; Table 3). A ypical long- e m se -up
measu ed sensible and la en hea luxes, longwa e and sho -
wa e incoming adia ion, wind speed, a mosphe ic empe a-
u e, and humidi y.
Pa ame iza ion and alida ion u ilize he ORCHIDEE-
CAN 1.0 model simula ions, and so clima e o cing da a
we e equi ed o d i e he simula ions. Si e-le el wea he ob-
se a ion, i.e. sho wa e incoming adia ion, longwa e in-
coming adia ion, wo-dimensional wind speed, p ecipi a-
ion, snow, nea -su ace ai p essu e, and speci ic humidi y
we e e o ma ed and gap- illed using he me hod p oposed
by Vuicha d and Papale (2015). Wea he obse a ions a e an
in eg al pa o bo h in ensi e campaigns and mul i-yea op-
canopy lux moni o ing. Hence, wi hin a measu emen si e,
lux, p o ile, and wea he da a we e usually a ailable a he
same empo al esolu ion and o e he same ime pe iods.
Finally, he o cing iles we e comple ed wi h he obse ed
e ical LAI p o iles. Howe e , he empo al esolu ion o
LAI was much lowe han he esolu ion o he me eo o-
logical a iables. When he o al LAI was measu ed a a
highe ime esolu ion han i s e ical p o ile, he obse ed
o al LAI was e ically dis ibu ed acco ding o he ob-
se ed ela i e e ical LAI dis ibu ion. Model pa ame iza-
ion (Sec . 2.3) and model expe imen s ha aimed a es -
ing he pe o mance o only he mul i-laye ene gy budge
(see EXP1 and EXP3 in Sec . 2.5) made use o he obse ed
LAI p o iles. Fo he emaining wo model expe imen s (see
EXP2 and EXP4 in Sec . 2.5), ORCHIDEE-CAN 1.0 cal-
cula ed he e ical LAI p o iles ollowing he ca bon allo-
ca ion and ca bon u no e schemes, as desc ibed in Naud s
e al. (2015).
Geosci. Model De ., 9, 2951–2972, 2016 www.geosci-model-de .ne /9/2951/2016/
Y. Chen e al.: E alua ing he pe o mance o ORCHIDEE-CAN 1.0 2955
Table 2. S and s uc u e and da a a ailabili y o he expe imen al si es. The maximum obse ed lea a ea (LAI; m2m−2) o he o e -s o y and unde -s o y LAI (all-sided) a e epo ed
sepa a ely. The heigh o he o e -s o y is exp essed in me es. Udeno es wind speed, Tadeno es a mosphe ic empe a u e, and qadeno es a mosphe ic humidi y. LE, H, and Rndeno e
he la en hea lux, he sensible hea lux, and he ne adia ion, espec i ely. +indica es ha p o ile measu emen s we e a ailable. −indica es ha no p o ile measu emen s we e
a ailable.
Si e Code FI-Hyy FR-LB NL-Loo DE-Bay CA-Oas AU-Tum DE-Hai BE-Vie
Species Pinus syl es is Pinus pinas e Pinus syl es is Picea abies Populus sp. Eucalyp us sp. Fagus syl a ica Fagus syl a icaa
Lea ype Needlelea ed Needlelea ed Needlelea ed Needlelea ed B oadlea ed B oadlea ed B oadlea ed B oadlea ed
G ow h o m E e g een E e g een E e g een E e g een Deciduous E e g een Deciduous Mixed
ORCHIDEE PFT 18 5 6 7 20 15 13 13
O e -s o y LAI 6.5 2.0 1.9 4.8 2.9 2.5 5.8 5.1
Unde -s o y LAI 0.5 1.5 1.5 0.5 2.8 1.0 0.1 0.1
Heigh 17.0 23.0 15.0 15.0 22.0 50.0 30.0 25.0
Up o ile + − + + + + + +
Tap o ile + + + + + + + +
qap o ile + + + + + + − +
LE p o ile + + + +b+ − − −
Hp o ile + + + + + + + −
Rnp o ile − + + +b− − − −
Re e ence Launiainen e al.
(2007)
Ogée e al. (2003),
Po e e al. (2000)
Dolman e al. (2002),
Moo s (2012)
Foken e al. (2012),
S aud e al. (2011)
Ba e al. (2004) Ha e d e al.
(2012),
Lo ell e al. (2012)
Knohl e al. (2003) Aubine e al. (2001),
Lai a e al. (1998)
aThis si e is pa ially mixed wi h Pseudo suga menziesii.bThe LE p o ile was a ailable o he 2007 and 2008 pe iod, bu no 2011, and he Rnp o ile was pa ly a ailable in 2007.
2.3 Model pa ame iza ion
A he s a o his s udy he mul i-laye ene gy budge did
no ye ha e a wo king se o pa ame e s o ORCHIDEE-
CAN 1.0. The e o e, we e ained om pe o ming a sen-
si i i y analysis p io o op imizing he model pa ame e s
(Kuppel e al., 2014; MacBean e al., 2015), bu ins ead se-
lec ed h ee p ocesses desc ibed by a o al o 10 pa ame e s
o op imiza ion. The selec ed p ocesses we e ela ed o he
physical p ocesses wi hin he canopy – ha is o say, di u-
sion, ad ec ion, and u bulen mixing.
2.3.1 E ec i e d ag coe icien CDe (uni less)
The canopy s uc u e is a e y impo an cha ac e is ic o
he land–a mosphe e in e ac ion, which can now be simu-
la ed by he ORCHIDEE-CAN 1.0 land su ace model. We
assumed ha he d ag coe icien is scala independen and
can be pa ame ized by he canopy s uc u e. The e ec i e
d ag coe icien is used in he one-dimensional second-o de
closu e wind p o ile model (Massman and Weil, 1999) ha
was used o es ima e he e ical wi hin-canopy wind p o ile.
In his wind p o ile model (Massman and Weil, 1999), he
d ag coe icien is assumed o be a cons an h oughou he
canopy laye , bu i can also be ea ed as a unc ion o he
e ical canopy s uc u e.
In his s udy, we made use o a p o o ype pa ame iza-
ion app oach p oposed by Wohl ah and Ce nusca (2002).
Wohl ah and Ce nusca (2002) p o ided he basic idea o
conside ing he e ec i e d ag coe icien in g asslands ha
can be a ied due o changes in canopy s uc u e, such as
bending e ec s. Thus, we adap ed his pa ame iza ion o
ou model; howe e , we le he i s wo uning coe icien s
(a1and a2) as cons an s. This modi ica ion allows he e ec-
i e d ag o dec ease om a la ge alue o a cons an while
mo ing om he op o he canopy o he soil su ace laye .
Hence, we applied he ideas de i ed in g assland esea ch o
a o es canopy. This app oach equi es an e ec i e d ag co-
e icien , which ela es o he e ically disc e ized es ima e
o he canopy d ag coe icien (CD,i ; uni less) and he mo-
men um shielding ac o (Pm,i; uni less) as ollows:
CDe ,i =CD,i/Pm,i.(1)
Bo h he wi hin-canopy d ag and he momen um shielding
we e pa ame ized using a unc ion o he cumula i e lea
a ea index (LAIcum; m2m−2) om he op canopy laye o
he bo om laye , which was modi ied om he o iginal unc-
ion (Wohl ah and Ce nusca, 2002) as below:
CDe ,i =a−LAIcum,i /a2
1+a−LAIcum,i /a4
3+a5,(2)
whe e he subsc ip ideno es he index o laye ing om he
bo om laye (i=1) o he op-canopy laye (i=n). a1 o
a5a e uning coe icien s (uni less). The de aul pa ame e
alues o a1 o a5a e p esen ed in Table 4.
www.geosci-model-de .ne /9/2951/2016/ Geosci. Model De ., 9, 2951–2972, 2016

2956 Y. Chen e al.: E alua ing he pe o mance o ORCHIDEE-CAN 1.0
Table 3. Obse a ion pe iods o he di e en da a uses in his s udy. Da e o ma : dd/mm/yy. The in o ma ion o he ene gy closu e gap o
each si e o e di e en selec ed pe iods was also calcula ed based on Chen and Li (2012). EXP1: single-laye scheme wi h a p esc ibed LAI
p o ile; EXP2: single-laye scheme wi h a long- e m simula ed LAI p o ile; EXP3: mul i-laye scheme wi h a p esc ibed LAI p o ile; EXP4:
mul i-laye scheme wi h a simula ed LAI p o ile.
Si e code FI-Hyy FR-LB NL-Loo DE-Bay CA-Oas AU-Tum DE-Hai BE-Vie
Pe iod o sho - e m
pa ame e op imiza-
ion (Pe iod I)
01/08/06
14/08/06
31/07/06
05/08/06
08/07/97
12/07/97
04/07/11
17/07/11
16/08/94
22/08/94
08/11/06
11/11/06
10/05/01
19/05/01
01/08/02
07/08/02
Closu e gap (W m−2) 43.34 41.56 10.48 18.97 19.82 18.40 29.89 28.19
Pe iod o long- e m
pa ame e op imiza-
ion (Pe iod II)
01/01/02
31/12/02
01/01/03
31/12/03
01/01/02
31/12/02
01/01/97
31/12/97
01/01/05
31/12/05
01/06/01
31/06/02
01/01/05
31/12/05
01/01/97
31/12/97
Closu e gap (W m−2) 11.47 21.59 15.38 42.47 2.89 7.12 27.83 42.43
Pe iod o single-yea
EXP1 and EXP3
alida ion (Pe iod III)
01/01/05
31/12/05
01/01/06
31/12/06
01/01/97
31/12/97
01/01/99
31/12/99
01/01/04
31/12/04
01/06/04
31/06/05
01/01/01
31/12/01
01/01/02
31/12/02
Closu e gap (W m−2) 10.99 13.20 16.61 50.24 4.13 7.73 23.49 42.43
Pe iod o mul i-yea
EXP2 and EXP4
alida ion (Pe iod IV)
01/01/02
31/12/06
01/01/03
31/12/06
01/01/02
31/12/06
01/01/97
31/12/99
01/01/04
31/12/05
01/06/01
31/06/05
01/01/00
31/12/06
01/01/97
31/12/06
Closu e gap (W m−2) 10.68 17.03 22.65 48.14∗3.51 9.40 23.69 33.77
∗The o es in 1997–1999 was s ongly a ec ed by o es decline; in 2011 he o es was again in a good s a e.
Table 4. Desc ip ion o pa ame e s, code e e ence, ini ial alues, and uning anges used in he mul i-laye ene gy budge model in his
wo k.
Pa ame e Physical pa ame e Empi ical ep esen a ion o ORCHIDAS De aul Tuning ange
name name alue
a1E ec i e su ace d ag Bending o ee b anches a_1 6.410 Use de aul
a2E ec i e su ace d ag Bending o ee b anches a_2 0.001 Use de aul
a3E ec i e su ace d ag Bending o ee b anches a_3 0.434 0.1 o 0.8
a4E ec i e su ace d ag Bending o ee b anches a_4 −0.751 −0.9 o −0.1
a5E ec i e su ace d ag Bending o ee b anches a_5 0.071 0.05 o 0.1
a6Eddy di usi i y Inne -canopy u bulen mixing k_eddy_slope 5.0 1.0 o 20.0
a7Eddy di usi i y Inne -canopy u bulen mixing k_eddy_us a 0.3 0.0 o 0.6
a8Su ace–a mosphe e conduc ance Inne -canopy u bulen mixing ks_slope 5.0 1.0 o 20.0
a9Su ace–a mosphe e conduc ance Unde -s o y phenology ks_ ege 0.5 0.0 o 1.0
a10 Su ace–a mosphe e conduc ance Unde -s o y phenology ks_ une 1.0 0.5 o 1.5
Wb Laye bounda y esis ance Upscaling he lea coupling b _ ac 1.0 0.1 o 10.0
Ws Laye s oma al esis ance Upscaling he lea coupling s _ ac 1.0 0.1 o 10.0
2.3.2 Eddy di usi i y o e ical ene gy and wa e
anspo k(m2s−1)
A e he e ical wind p o ile was de i ed om he one-
dimensional second-o de closu e wind p o ile model, he
ic ion eloci y (u∗, ms−1), he e ical wind eloci y a i-
ance (σw; ms−1), and he Lag angian imescale (TL; s) we e
calcula ed ollowing he app oach by Raupach (1989). In his
app oach he e ical eddy di usi i y is a unc ion o σwand
TL. Subsequen ly, he e ical eddy di usi i y down he ai
column o he o es loo was calcula ed as ollows:
ki=σ2
w,i TL,i.(3)
He e we ollowed he app oach p oposed by Ha e d e al.
(2009) o he Lag angian imescale calcula ion. The La-
g angian imescale is hus calcula ed as
TL,i =0.661−e−4.86(z/hc)
1−e−4.86
hc
u∗
.(4)
A p e ious e o o alida e his model agains in si u
obse a ions esul ed in a bias o he ai empe a u e p o-
Geosci. Model De ., 9, 2951–2972, 2016 www.geosci-model-de .ne /9/2951/2016/
Y. Chen e al.: E alua ing he pe o mance o ORCHIDEE-CAN 1.0 2957
0.0 0.5 1.0 1.5
0.0 0.2 0.4 0.6 0.8 1.0
0
Weigh ing ac o , Wn (uni less)
F ic ion eloci y (m s−1 )
a6=5 , a7=0.5
a6=10 , a7=0.5
a6=25 , a7=0.5
(a)
0.0 0.2 0.4 0.6 0.8 1.0
0.0 0.2 0.4 0.6 0.8 1.0
0
Weigh ing ac o , Ws (uni less)
Vege a ion co e , 1− Pgap (m2 m−2 )
a8=10 , a9=0.5 , a10 =1 , Tweek ≥Tg
a8=10 , a9=0.5 , a10 =1 , Tweek =278.15
(b)
Figu e 1. Weigh ing unc ions o eddy di usi i y and su ace conduc ance. (a) Weigh ing unc ion o he eddy di usi i y (k) wi hin he
ai column (Eq. 3). The weigh ing is a unc ion o he ic ion eloci y (u∗) and was op imized by uning he pa ame e s a6and a7. Th ee
di e en pa ame e se s show he esponse o he weigh ing unc ion o di e en pa ame e alues. (b) The weigh ing unc ion o he su ace
conduc ance is a unc ion o he ege a ion co e and ai empe a u e (Eq. 7). This weigh ing unc ion was op imized by uning he pa ame e s
a8 o a10. Two examples ha e he ollowing pa ame e alues: a8=10.0, a9=0.5, a10 =1.0, Tweek ≥Tg, and Tweek =278.15. Bo h o he
wo cases demons a e he seasonal cycle o he weigh ing which will be used o scale he alue o ksu . Values o he le o he de lec ion
poin show he e ec o an inc easing/dec easing o e -s o y co e wi h an inc easing/dec easing empe a u e in sp ing/au umn. In sp ing
and au umn, unde -s o y g ow h, and hus i s con ibu ion o e apo anspi a ion, was assumed o be empe a u e limi ed. Values igh o he
de lec ion poin (a9=0.5) show he dependency o he e apo anspi a ion on he soil su ace laye on he o e -s o y canopy co e when ai
empe a u e is no longe limi ing unde -s o y g ow h.
ile wi hin he canopy laye du ing nigh ime (Ryde e al.,
2016). These issues ha e been well documen ed in he scien-
i ic li e a u e (Gao e al., 1989; Dolman and Wallace, 1991;
Maka e al., 1999; Wol e e al., 2011). One possible, al-
hough empi ical, solu ion is o apply a di e en scaling o
ki, acco ding o he ime o he day. He e we build on a sim-
ila app oach bu , a he han using he ime o he day, we
used he calcula ed ic ion eloci y (u∗=u(hc)×(0.32 −
0.264e−15.1ζ(hc)), whe e ζis he cumula i e unc ion o CDe
and hcis he canopy heigh ) o accoun o he obse ed
di e ences in e ical anspo wi hin he canopy be ween
day ime and nigh ime by applying a weigh ing ac o (Wn ;
uni less). The e o e he modi ied di usi i y o le el i(k∗
i;
m2s−1) was de ined as
k∗
i=Wn σ2
w,iTL,i,(5)
whe e Wn was calcula ed as
Wn =1
1+e(−a6(u∗−a7)) .(6)
This unc ion has a sigmoidal shape, whe e a6is he ceiling
ac o o he slope and a7is he c i ical ic ion eloci y a
he in lec ion poin o he sigmoid unc ion (Fig. 1a). Con-
sequen ly, a mosphe ic di usi i y is educed i u∗is low,
which ep esen s s able a mosphe ic condi ions. Unde u -
bulen a mosphe ic condi ions, which a e ep esen ed by a
high u∗,Wn is close o 1 and he simula ed di usi i y will
closely ollow he ela ionship p oposed by Raupach (1991).
The de aul pa ame e alues o a6and a7a e p esen ed in
Table 4. As an al e na i e o using u∗, i has been p oposed o
use a mixing leng h scale o classi y low egimes in o de o
gi e a be e desc ip ion o he coupling p ocess below and
abo e he o es canopy (Thomas and Foken, 2007; S aud
e al., 2011; Foken e al., 2012). The nume ical scheme o his
app oach elies on i e a ions. Since ORCHIDEE-CAN 1.0
is designed o be coupled o egional o global a mosphe ic
models, i s nume ics has been designed o a oid i e a ions in
o de o un e icien ly.
Fu u e s udies may ocus on eplacing his empi ical solu-
ion by a mo e mechanis ic solu ion. In he con ex o OR-
CHIDEE and i s coupling o he a mosphe ic model, his im-
plies ha we will ha e o sea ch o an implici solu ion o
he nea - ield a - ield heo y by Raupach (1989).
2.3.3 Conduc ance o he soil–a mosphe e in e ace
ksu (ms−1)
Equa ion (7) desc ibes he seasonali y o he soil–a mosphe e
in e ace, which we belie e is d i en by he unde -s o y and
i s phenology (Launiainen e al., 2015). Cu en ly, he model
does no simula e he p oduc ion o he phenology o he
unde -s o y. As a subs i u e o his a he complex p ocess,
we made use o a weigh ing coe icien o he conduc ance
o he soil–a mosphe e in e ace (ksu ) o , in o he wo ds, he
calcula ion o he wa e apou exchange be ween he soil
www.geosci-model-de .ne /9/2951/2016/ Geosci. Model De ., 9, 2951–2972, 2016
2958 Y. Chen e al.: E alua ing he pe o mance o ORCHIDEE-CAN 1.0
laye and he i s ai column (see he φλE and Ksu in Fig. 1
o Ryde e al., 2016, and he o mal desc ip ion o using
Ksu , which is gi en in he Supplemen o Ryde e al., 2016,
in Eqs. S4.30 and S4.31).
A ela ionship be ween unde -s o y phenology and he
conduc ance o he soil–a mosphe e in e ace has been ob-
se ed in bo eal o es (Launiainen e al., 2015). In win-
e , when he unde -s o y is senescen , he cha ac e is ics in
e ms o he e apo anspi a ion a he in e ace will closely
esemble he e apo anspi a ion o a ba e soil. In summe ,
howe e , an unde -s o y will be p esen and i s densi y e-
la es o he gap ac ion o he o e -s o y canopy. Hence, he
summe ime e apo anspi a ion o he in e ace will be close
o he e apo anspi a ion o a ege a ion canopy. The e o e,
we in oduced β0(uni less) as a weigh ing unc ion anging
om ze o o uni y, in o de o scale he su ace conduc i -
i y as a unc ion o o e -s o y phenology. Unde -s o y phe-
nology was desc ibed as a unc ion o he o e -s o y canopy
co e age (1 − Pgap), he mean ai empe a u e du ing he
p e ious week (Tweek), and a h eshold empe a u e (Tg):
β0=













a10
1+e(−a8((1− Pgap)−a9)) ,
when G eg = ue,
a10
1+e(−a8((1− Pgap)−a9))
Tg−Tweek
Tg−273.15 ,
when G eg = alse,
(7)
whe e a8is a ac o ha cons ains he slope o he unc ion
and a9is a ege a ion co e h eshold. a10 is a linea weigh -
ing ac o . Tgis a empe a u e h eshold se o 283.15 K. G eg
is a logic a iable o indica e he g ow h s a us o he ege a-
ion. G eg is an exis ing a iable in ORCHIDEE-CAN 1.0
and depends on a h eshold o soil wa e con en and em-
pe a u e Tg. G ow h can be expec ed and he e o e G eg
is se o ue when he weekly a e aged soil wa e con en
and empe a u e exceed he h esholds. Pgap is calcula ed
in ORCHIDEE-CAN 1.0 and desc ibes he o e -s o y gap
p obabili y, which is a unc ion o he canopy s uc u e o
he ege a ion and he sola zeni h angle and is calcula ed in
ORCHIDEE-CAN 1.0.
Fo he lowes laye in he ai column, i.e. he laye adja-
cen o he su ace, he su ace conduc ance is hen calcula ed
as
ksu =(Ws β3+(1−Ws )β4)u1CDe ,1,(8)
whe e β3and β4a e coe icien s desc ibing espec i ely he
ac ions o he po en ial plan anspi a ion and soil e apo a-
ion ha a e ealized. The de ini ion o hese coe icien s and
he nume ical app oaches a e p esen ed in Ryde e al. (2016)
and Du esne and Gha as (2009). u1is he wind speed a he
lowes canopy laye hus close o he o es loo and is de-
i ed om he one-dimensional second-o de closu e model.
CDe is he e ec i e d ag coe icien calcula ed acco ding o
Eq. (2). Ws is he weigh ing ac o o he soil–a mosphe e
in e ace, which is desc ibed as he condi ional unc ion o
o e -s o y canopy co e ac ion (1− Pgap). Ws =β0when
(1− Pgap)>a9, and Ws =1−β0when (1− Pgap ≤a9(see
Fig. 1b). The de aul pa ame e alues o a8,a9,a10, and Ws
a e p esen ed in Table 4.
2.3.4 Bounda y-laye esis ance o he lea su ace Rb
(sm−1)
The bounda y-laye esis ance o he lea su ace Rb,i is de-
sc ibed acco ding o he exp ession om Baldocchi (1988):
Rb=


Wb dl
Dh,ai Nu , o sensible hea ,
Wb dl
Dh,H20Sh , o la en hea ,(9)
whe e Wb accoun s o he ac ha he lea leng h o he
species unde s udy di e s om he cha ac e is ic lea leng h
(uni less), dlis he cha ac e is ic lea leng h (0.001 m was
used as he de aul alue), Dh,ai is he hea di usi i y o s ill
ai (m2s−1), Dh,H2Ois he hea di usi i y o wa e apou
(m2s−1), Sh is he She wood numbe (uni less), and Nu is he
Nussel numbe (uni less). The She wood numbe was cal-
cula ed as Sh =0.66Re0.5Sc0.33 o lamina low and Sh =
0.03Re0.8Sc0.33 o u bulen low, whe e Sc is he Schmid
numbe (0.63 o wa e apou ; uni less). The ansi ion om
lamina o u bulen low akes place in he model when he
Reynolds numbe exceeds a alue o 8000. The Nussel num-
be was calcula ed as Nu =0.66ReP 0.33, whe e P is he
P and l numbe (0.7 o ai ; uni less) (G ace, 1978) and Re is
he Reynolds numbe (uni less) which was calcula ed as
Re =dlui
µ,(10)
whe e uiis he ho izon al eloci y a le el i(ms−1) and µis
he kinema ic iscosi y o ai and was se o 0.0015 (m2s−1)
(Ga a , 1992). The de aul pa ame e alue o Wb is p o-
ided in Table 4.
2.3.5 S oma al esis ance Rs(sm−1)
The s oma al esis ance o he lea es was calcula ed o each
canopy laye based on he pa ame e s wi hin he laye unde
conside a ion. Two s oma al esis ances we e calcula ed wi h
he concu en assimila ion a e: (a) he s oma al esis ance
assuming unlimi ed soil wa e a ailabili y ( he a mosphe ic
demand) and (b) he s oma al esis ance ha exac ly sa is ies
he amoun o wa e he plan can anspo om i s oo s o
i s s oma a ( he plan supply). ORCHIDEE-CAN 1.0 calcu-
la es he plan supply o he wa e a ailable o anspi a ion
as he p essu e di e ence be ween he soil and he lea es di-
ided by he sum o hyd aulic esis ances o ine oo s, sap-
wood, and lea es (see Eq. 20 in Naud s e al., 2015). The
a mosphe ic demand o wa e o anspi a ion is calcula ed
as he apou p essu e di e ence be ween he lea es and a -
mosphe e di ided by he sum o bounda y-laye esis ance
Geosci. Model De ., 9, 2951–2972, 2016 www.geosci-model-de .ne /9/2951/2016/
Y. Chen e al.: E alua ing he pe o mance o ORCHIDEE-CAN 1.0 2959
(Rb) and s oma al esis ance (Rs) (see Eqs. 9 and 13 in Ryde
e al., 2016). When he supply can sa is y he demand, he e
is no wa e s ess and pho osyn hesis (A) is calcula ed. When
he demand is limi ed by he supply e m, Aand Rsa e ecal-
cula ed such ha hey sa is y he supply. Wa e s ess hus en-
e s Eq. (11) in he alue o A. ORCHIDEE-CAN 1.0 scales
s oma al esis ance o accoun o he pa o he canopy ha
is coupled o he a mosphe e and hus con ibu es o he la-
en hea lux. In his s udy, his weigh ing was o malized
h ough a linea pa ame e Ws :
Rs,i =Ws 

1
g0+Aihs
CsLAIi

,(11)
whe e g0is he esidual s oma al conduc ance i he sola i -
adiance app oaches ze o, Csis he concen a ion o CO2a
he lea su ace, and hsis he ela i e humidi y a he lea su -
ace. Ais he CO2assimila ion a e which is sol ed analy i-
cally ollowing Yin and S uik (2009). In Eq. (11) he ela i e
humidi y used is he op-canopy o cing ins ead o a laye ed
ela i e humidi y in o de o a oid an i e a i e p ocess. The
de aul pa ame e alue o Ws is p esen ed in Table 4.
2.4 Model op imiza ion
Op imiza ion p ocedu e
Pa ame izing he scaling coe icien s and weigh ing ac o s
enabled us o simul aneously imp o e he ma ch be ween he
simula ed and obse ed sub-canopy mic ome eo ology, in-
cluding empe a u e and speci ic humidi y when a ailable,
and be ween he simula ed and obse ed op-canopy hea
luxes (LE and H). Wi hin-canopy luxes we e also sim-
ula ed bu a e no usually measu ed. The pa ame iza ion
made use o an in-house op imiza ion package called OR-
CHIDAS (ORCHIDEE Da a Assimila ion Sys ems; h p://
o chidas.lsce.ipsl. /). ORCHIDAS p o ides a ange o nu-
me ical app oaches o assimila ing mul iple da a s eams in
ORCHIDEE.
We used he maximum g adien app oach o une he pa-
ame e s a3 o a10,Wb , and Ws o each s udy si e indepen-
den ly. O e he cou se o se e al i e a ions, he op imiza ion
app oach minimized he misma ch be ween he model ou pu
and he obse a ions, using a g adien -based algo i hm called
L-BFGS-B (Limi ed-memo y B oyden–Fle che –Gold a b–
Shanno algo i hm wi h Bound cons ain s), which p o ides
he possibili y o p esc ibe bounda ies o each pa ame e
(By d e al., 1995). The ange assigned o each pa ame e is
epo ed in Table 4. Fu he mo e, his app oach allowed o
measu emen unce ain ies in he eddy co a iance LE mea-
su emen by educing i s weigh in he cos unc ion om 1.0
o 0.66. This alue o 0.66 was se based on he ou come o
a pai ed owe expe imen o es ima e he andom e o s o
he eddy co a iance measu emen s (Richa dson e al., 2006).
Fo he op imiza ion, he LAI in ORCHIDEE-CAN 1.0 was
se o ma ch he obse ed e ical LAI p o ile.
A h ee-s ep op imiza ion p ocedu e was ca ied ou in his
s udy. Fi s ly, he wi hin-canopy and below-canopy obse a-
ions om he sho - e m in ensi e measu emen campaigns
(Pe iod I in Table 3) we e used o op imize a3 o a7,Wb , and
Ws . Du ing his s ep, he pa ame e s o he soil–a mosphe e
in e ace (ksu , i.e. a8 o a10 and Ws ) we e se o hei de aul
alues. Since hese campaigns ook place du ing summe , pa-
ame e s ela ed o he wi hin-canopy e ec i e d ag p o iles,
eddy di usi i y, bounda y laye esis ance, and s oma al e-
sis ance (CDe ;k;Rb;Rs) we e biased owa ds he summe .
Secondly, he seasonal dynamics o ksu was pa ame ized
by ying o imp o e he co espondence be ween he simu-
la ed and obse ed op-canopy luxes o e 1 yea (Pe iod II
in Table 4). In his s ep, a3 o a7,Wb , and Ws we e se o he
alues ob ained om he i s s ep o he op imiza ion and
a8 o a10 and Ws we e uned. Finally, he pe o mance o
he calib a ed model was e alua ed based on a second single
yea o op-canopy obse a ions (Pe iod III in Table 3).
Al hough he spin-up was s opped on 30 June (Table S1 in
he Supplemen ) and all simula ions hus used he 30 June
soil wa e con en as hei ini ial condi ion, his app oach
does no gua an ee ha his ypical summe soil wa e con-
en ma ches he soil wa e con en in he yea o he in ensi e
measu emen campaign. The e ec o his possible misma ch
was quan i ied by unning a sensi i i y analysis in which
he whole pa ame iza ion app oach, which was epea ed o
se en di e en ini ial soil wa e con en s, a ied om −30 o
30 % in inc emen s o 10 % o he 30 June alue.
2.5 A ibu ion o changes in model pe o mance
The mul i-laye ene gy budge scheme (Ryde e al., 2016)
ha was pa ame ized and es ed in his s udy equi ed ealis-
ic spa ial and empo al soil wa e con en and a alue o he
g ound hea lux om su ace le el as ini ial condi ions. This
need was sa is ied by implemen ing his scheme wi hin he
newly enhanced ORCHIDEE-CAN 1.0 land su ace model
(Naud s e al., 2015). In eg a ion o he mul i-laye ene gy
budge in o ORCHIDEE-CAN 1.0, howe e , complica ed
he design o he alida ion s udy, as i was now necessa y o
sepa a e, as much as possible, he pe o mance o he mul i-
laye ene gy budge scheme om he pe o mance o he es
o he model. To his aim, ou expe imen s we e designed
in o de o be e unde s and he pe o mance o he new
scheme (Table S1 in he Supplemen ).
– Expe imen 1 (EXP1): single-laye scheme wi h a
p esc ibed canopy
The i s expe imen was un a he si e le el and made
use o he de aul single-laye ene gy budge scheme.
The ene gy budge scheme was d i en by he obse ed
clima e o cing and he obse ed o al LAI (Table 2).
In his expe imen , he e ical LAI p o ile was only
www.geosci-model-de .ne /9/2951/2016/ Geosci. Model De ., 9, 2951–2972, 2016
2966 Y. Chen e al.: E alua ing he pe o mance o ORCHIDEE-CAN 1.0
Rn H LE G
One−laye LAI p esc ibed (EXP1)
One−laye LAI simula ed (EXP2) Mul i−laye LAI p esc ibed (EXP3)
Mul i−laye LAI simula ed (EXP4)
S , Taylo sco e (0 o 1)
T
0.0 0.2 0.4 0.6 0.8 1.0
Figu e 7. Change in model pe o mance, exp essed as Taylo skill
sco e, wi h inc easing expe imen al complexi y o bo h he single-
laye and mul i-laye ene gy budge schemes o all eigh s udy
si es. EXP1: single-laye scheme wi h a p esc ibed LAI p o ile;
EXP2: single-laye scheme wi h a simula ed LAI p o ile; EXP3:
mul i-laye scheme wi h a p esc ibed LAI p o ile; EXP4: mul i-
laye scheme wi h a simula ed LAI p o ile.
ula e land su ace luxes dynamically and demons a ed ha
his app oach has di icul ies in he ep oduc ion o su ace
ene gy luxes.
In his s udy, we ied o o e come hese di icul ies by im-
plemen ing a mul i-laye ene gy budge scheme. The mul i-
laye ene gy and wa e calcula ions make use o a e i-
cally esol ed adia ion ans e scheme o sho wa e and
longwa e adia ion ( eplacing p esc ibed sho wa e e lec-
ion alues), a wi hin-canopy wind eloci y p o ile ( eplac-
ing empi ical o mula ions o oughness leng h), a e ical
p ognos ic LAI p o ile ( eplacing a p esc ibed LAI alue),
wi hin-canopy lea bounda y-laye esis ance p o iles o en-
e gy and wa e anspo , a wi hin-canopy s oma al esis ance
p o ile, a e ical disc e e eddy di usi i y p o ile, and a soil–
a mosphe e laye conduc i i y.
This app oach esul ed in small imp o emen s in simula -
ing ene gy pa i ioning du ing nigh ime o dense canopies,
small losses in model pe o mance in e ms o ene gy pa i-
ioning o spa se canopies, and yea - ound gains in model
pe o mance o simula ion o he g ound hea lux. As such,
he mul i-laye ene gy and wa e apou lux scheme did no
sol e he long-s anding issues ela ed o simula ing nigh ime
ene gy pa i ioning (Jo dan and Smi h, 1994; P ihodko e al.,
2008; Wild, 2009; He e al., 2011), bu i succeeded in ob ain-
ing a simila model pe o mance, while much o he empi i-
cism o he big-lea app oach was eplaced by a mo e eal-
is ic p ocess desc ip ion. A mo e ealis ic model desc ip ion
opens new a enues o esea ch (see Sec . 4.3).
4.2 Pa ame iza ion app oach
Despi e he di ec ion o he land su ace model communi y
owa ds he de elopmen o mo e mechanis ic models, all
la ge-scale land su ace models con ain an impo an le el o
empi icism. When he model is ca e ully de eloped and al-
ida ed, he empi ical pa ame e s mimic an o e ly complex
( o he pu pose o he model) o incomple ely unde s ood
p ocess. As we ied o ollow his philosophy, we belie e
ha ou pa ame e s ha e a plausible na u al backg ound (Ta-
ble 4), bu his does no o e come he issue o equi inali y
o he model. Ideally, u u e de elopmen s should aim a e-
placing such pa ame e s by a mo e mechanis ic app oach i
he empi ical module ep esen s a p ocess ha is a he co e
o he objec i es o he model. In his s udy, he pa ame iza-
ion o he new scheme and i s unde lying p ocesses e ealed
s eng hs and weaknesses o he model as well as a enues o
u u e expe imen al wo k.
1. Wi hin-canopy d ag
Fo he inne -canopy d ag pa ame iza ion, we modi-
ied an app oach (Eq. 2) ha has p e iously only been
es ed and alida ed a g assland si es (Wohl ah and
Ce nusca, 2002). In ha s udy, LAI was ea ed as equal
o he plan a ea index (PAI), which is a sepa a e mea-
su e ha accoun s no only o lea es, bu also o o he
ege a ion ma e ial such as s ems and seed heads. In
o es s, howe e , he di e ence be ween LAI and PAI is
made up o he b anches and unks and becomes espe-
cially impo an in win e in deciduous s ands as canopy
d ag s ill exis s. As a i s pa ame iza ion his simpli-
ica ion allowed a be e compa ison wi h he obse a-
ions and wi h he single-laye model. We applied a o -
mula ion ha makes use o LAI and, by doing so, some
model e o s migh ha e been in oduced, especially o
he deciduous o es si es. ORCHIDEE-CAN 1.0 now
simula es bo h LAI and PAI, and so his enhanced ap-
p oach could be adop ed. Resul s con i med ha sub-
s i u ing PAI by LAI is accep able du ing he lea -on
seasons.
Al e na i e app oaches ha e been p oposed by Cesca i
and Ma colla (2004). Fo example, he inne -canopy
d ag could also be modelled as he unc ion o he
pe cen age o ho izon al gaps in he o es canopy –
a canopy cha ac e is ic ha is p esen ly simula ed in
ORCHIDEE-CAN 1.0. Measu emen si es such as
DE-Bay o AU-Tum ha e de ailed wind and e ical
LAI p o ile obse a ions and could hus be used in a pi-
lo s udy o de eloping a sui able pa ame iza ion ap-
p oach linking inne -canopy d ag and shielding o he
canopy gaps. Such a de elopmen would also mee he
equi emen s o calcula ing d ag and shielding ollow-
ing small-scale mo ali y om o es managemen , i es,
wind damages, and pes s.
Geosci. Model De ., 9, 2951–2972, 2016 www.geosci-model-de .ne /9/2951/2016/

Y. Chen e al.: E alua ing he pe o mance o ORCHIDEE-CAN 1.0 2967
●
●
0 60 120 180 240 300
0 0.2 0.4 0.6 0.8 1
La en hea lux (W m−2 )
No mailzed heigh (z hc
−1 )
●Obs. (13:00)
Sim. β3
Sim. β4
Sim. combine
(a)
●
●
0 60 120 180 240 300
0 0.2 0.4 0.6 0.8 1
Sensible hea lux (W m−2 )
No mailzed heigh (z hc
−1 )
●Obs. (13:00)
Sim. β3
Sim. β4
Sim. combine
(b)
Figu e 8. E ec o unde -s o y phenology on he e ical p o ile o he la en and sensible hea luxes a he FR-LB si e. (a) Simula ed
la en hea lux assuming ha he in e ace be ween he soil and he lowes a mosphe ic laye beha es as a ba e soil (do ed line), a ully
ege a ed su ace (dashed line), o a pa ly ege a ed, pa ly ba e su ace whe e he a io be ween ba e soil and ege a ed soil depends on
he unde -s o y phenology ( ull line). The obse ed p o ile is shown as black do s whe e he e o ba s deno e he 5-day empo al a iance.
(b) Simula ed sensible hea lux assuming ha he in e ace be ween he soil and he lowes a mosphe ic laye beha es as a ba e soil (do ed
line) o a ully ege a ed su ace (dashed line), o depends on he unde -s o y phenology ( ull line). The obse ed p o ile is shown as black
do s whe e he e o ba s deno e he 5-day empo al a iance.
2. Wi hin-canopy anspo
In his s udy, wi hin-canopy anspo was pa ame ized
by K- heo y. A one-dimensional second-o de closu e
model was applied o de i e he wi hin-canopy u -
bulence s a is ics, based bo h on he LAI p o ile and
he canopy heigh . This app oach has been epo ed o
p oduce a easonable app oxima ion o abo e-canopy
luxes es ima ion, e en i he wi hin-canopy empe a u e
and humidi y g adien s a e no always well cap u ed
(Raupach, 1989). As p e ious s udies ha e demon-
s a ed, inco ec es ima ion on g adien s may be ac-
commoda ed o some ex en by in oducing a scaling
ac o (Eq. 6) o cons ain he wi hin-canopy anspo
(Maka e al., 1999; Wol e e al., 2011; Ryde e al.,
2016). Al e na i ely, such a scaling ac o migh a y in
e ms o he o m o he canopy s uc u e o openness,
hough he de e mina ion o he ac o has ye o be ad-
equa ely desc ibed due o a es ic ed ange o measu e-
men s (McNaugh on and Van Den Hu k, 1995; S oud
e al., 2005).
A spa se o es si es, he empe a u e measu emen s
showed a gene al posi i e g adien du ing he day ime
(Fig. S5 in he Supplemen ) and a nega i e g adien du -
ing he nigh ime (no shown). Fo he spa se o es s,
he empe a u e g adien is e en mo e complex, ha ing
a nega i e o e e sed g adien h oughou he e ical
p o iles. By using he cu en pa ame iza ion app oach,
mos o he spa se o es si es equi ed a highe shea
s ess (a s onge h eshold ic ion eloci y a7) o he
wi hin-canopy mixing, compa ed o dense o es si es
(Table S2 in he Supplemen ), in o de o eplica e he
measu emen esul s. This obse a ion ela es o a gen-
e al di icul y in being able o simula e canopy anspo
based on limi ed gene al measu emen s (S oud e al.,
2005).
3. Sub-canopy and su ace–a mosphe e condi ions
In his s udy, we ea ed he unde -s o y and o e -s o y
as he same species o cons uc he e ical LAI p o-
ile based on he obse ed LAI p o ile. This ea men
only allowed he unde -s o y g ow h o ollow o e -
s o y canopy phenology. In ac , he o es loo is o en
occupied by plan s wi h e y di e en ai s, o which
one o he mos ob ious is he di e ence in lea onse
and/o lea all (Ba e al., 2004). Gi en he a o emen-
ioned model o mula ion, simula ion o he unde -s o y
phenology and ai s could be u he imp o ed in he
u u e. Fo example, o e -s o y and unde -s o y ege-
a ion could be simula ed as di e en plan unc ional
ypes o plan species wi hin he same ene gy budge
www.geosci-model-de .ne /9/2951/2016/ Geosci. Model De ., 9, 2951–2972, 2016
2968 Y. Chen e al.: E alua ing he pe o mance o ORCHIDEE-CAN 1.0
column. Also, he mic oclima e c ea ed by he o e -
s o y could be used as an inpu o simula e he en i-
onmen al condi ions in he unde -s o y.
S a ing om he poin o iew o he in e ac ion be-
ween ecosys ems and he clima e, we in oduced a
weigh ing ac o (Ws ) as a unc ion o a long- e m a -
e age empe a u e and ligh condi ions (gap ac ion), a
anspi a ion ac ions desc ibed as β3in he model code
and a soil e apo a ion ac ion (β4) as en i onmen al
ac o s o pa ame ize su ace conduc ance (Fig. 8 and
consequen ly con ol he su ace la en hea lux. This
app oach demons a ed he model’s capabili y o simu-
la e he lux p o ile in ag eemen wi h obse a ions. I
may, howe e , no be alid o he sa anna ecosys em
because he unde -s o y phenology o his ecosys em
elies on wa e a ailabili y in he op soil laye (Bal-
docchi and Wilson, 2001; Hu ley e al., 2000), which is
an en i onmen al condi ion no accoun ed o in ou ap-
p oach. Fu he mo e, accoun ing o ecosys em-speci ic
di e ences in oo densi y p o iles and ae ial co e o
he unde -s o y migh also help in he simula ion o wa-
e and ene gy luxes (El Mas i e al., 2015; Launiainen
e al., 2015). F om his pe spec i e, de ailed soil mois-
u e p o ile obse a ions would be e y use ul in de-
eloping a mo e ad anced su ace–a mosphe e in e ace
pa ame iza ion.
4. Misma ch be ween low- esolu ion d i e da a and e i-
cally esol ed ege a ion laye s
In his s udy an appa en misma ch was p esen be ween
he low esolu ion o he d i e da a ha con ain in o -
ma ion de i ed om se e al di e en land co e ypes
and he highly esol ed e ical laye ing o he canopy.
When low- esolu ion d i e da a a e used, he bene-
i om eplacing he big-lea app oach in a ou o a
mul i-laye app oach becomes ques ionable.
In his s udy he spin-up o he soil wa e con en made
use o low- esolu ion d i e da a, bu he simula ions
hemsel es we e d i en by spa ially and empo ally
high- esolu ion si e obse a ions. Ne e heless, he ap-
pa en misma ch ouches upon an in e es ing issue: how
does one accoun o he a e age su ace luxes om
he con ibu ion o di e en subg id-scale land co e
ypes? The p esen ORCHIDEE single-laye model cal-
cula es a weigh ed a e age o di e en PFTs ac oss a
g id squa e o calcula e a o al ep esen a i e lux. An
al e na i e app oach, and one ha we a e in es iga ing
using his mul i-laye model in ORCHIDEE-CAN 1.0,
is o calcula e he hea luxes o each ege a ion ype
sepa a ely (sub-g id-scale modelling) so ha he mixing
occu s abo e he canopy.
5. The p oposed pa ame iza ion app oach and u u e wo k
In gene al, we p o ide a simple bu use ul pa ame iza-
ion app oach o he mul i-laye ene gy budge scheme
in he ORCHIDEE-CAN 1.0 global land su ace
model. Compa ing wi h o he s udies (Ogée e al., 2003;
S aud e al., 2011; Launiainen e al., 2015), ou ap-
p oach di ec ly de e mines he ene gy and wa e luxes
and success ully a oids he i e a i e p ocesses o mee
he nume ical equi emen . In o al, a se o 12 pa am-
e e s need o be p esc ibed and calib a ed ega ding he
empi ical ep esen a ion o su ace d ag, u bulen mix-
ing, sub-canopy phenology, and lea –a mosphe e cou-
pling p ocesses. Ou app oach p esen s a good pe o -
mance a all s udy si es, hough we may ha e some
de ici s in wind speed es ima ion.
In his s udy he model had been es ed o se e al en-
i onmen al condi ions and demons a ed ha he nu-
me ics can deal wi h he a ia ion ha can be ound in
global ecosys ems. A sepa a e pa ame e se o each
si e has been p o ided. Nex , we will ha e o de i e a
single pa ame e se o each PFT and es how well he
model ep oduces global pa e ns in, o example, e ap-
o anspi a ion. Only hen will we be able o lea n abou
he ans e abili y o he pa ame e s om he si e le el
o he PFT le el.
4.3 Inc eased model capaci y
The inno a ion o he mul i-laye ene gy and wa e scheme
is he capaci y o simula e he beha iou o luxes wi hin
he canopy and he sepa a ion o he soil-le el empe a u e
om he empe a u e o he ege a ion le els. The mul i-
laye scheme helps o add ess how o es managemen such
as hinning o shel e wood cu ing may al e he o es –
a mosphe e coupling and esul ing luxes. I also pa es he
way o he conside a ion o mixed o es s whe e di e en
plan species o unc ional ypes can be in a di e en mi-
c oclima ic en i onmen o ha o he high canopy. This ca-
paci y is essen ial o he ollowing ypes o u u e po en ial
applica ions.
1. The simula ion o emission o biogenic ola ile o -
ganic compounds (BVOCs) om plan s, linking clima e
change, a mosphe ic chemis y, and he e es ial bio-
sphe e. The implemen ed mul i-laye ene gy and wa-
e budge calcula es he lea empe a u e and wi hin-
canopy adia ion, and he e o e would po en ially al-
low us o simula e he emission o BVOCs, such as
isop ene o mono e pene om plan s (Guen he e al.,
1995, 2006).
2. Na u al dis u bances, such as i es, pes s, and wind all,
can esul in inc eases in lea all, indi idual ee mo -
ali y, o comple e s and des uc ion (Lugo, 2008; Seidl
e al., 2011; Yue e al., 2014), which in u n de e mine
he e ical LAI p o ile. The implemen ed mul i-laye
ene gy and wa e budge scheme calcula es he e ical
eddy di usi i y and e ec i e d ag coe icien as a unc-
ion o he e ical LAI p o ile; hence, he new scheme
Geosci. Model De ., 9, 2951–2972, 2016 www.geosci-model-de .ne /9/2951/2016/
Y. Chen e al.: E alua ing he pe o mance o ORCHIDEE-CAN 1.0 2969
allows he s udy o e ec s o changes in dis u bance in-
ensi y on he ene gy budge and hus he clima e sys-
em.
3. Fo es canopy s uc u e plays an impo an ole in eg-
ula ing he p o ision o o es ecosys em se ices such
as main aining biodi e si y (Sche e s e al., 2013; De-
aeye e al., 2014) o egula ing s eam low (Jack-
son, 2005). The e o e, s uc u al changes o he o es
canopy, h ough, o example, o es hinning o species
changes, will educe he bu e ing e ec o he canopy.
I is only wi h models including a mul i-laye ene gy
budge ha an in o med p edic ion o he long- e m con-
sequences o land-managemen policies can be made.
4. This wo k akes he i s s ep in explo ing he use o e -
ical canopy p o iles in coupled ege a ion/a mosphe ic
models, pa icula ly in ela ion o he calcula ion o
GPP, which is sensi i e o he e ical p o iles o ligh ,
wa e and ni ogen (Bonan e al., 2012, 2014). To un
a a egional o global scale, i is essen ial o i s
pa ame ize he model a he si e le el.
5 Conclusion
Al hough he i s pa ame iza ion o a mul i-laye ene gy
and wa e budge scheme did no g ea ly imp o e he model
pe o mance o e he use o he so-called big-lea app oach
o ene gy and wa e calcula ions, i p o ides a mo e de-
ailed desc ip ion o he wi hin-canopy mic ome eo ology o
a ious o es ypes. A mo e de ailed p ocess desc ip ion is
essen ial when linking clima e change o s udies add ess-
ing, o example, species ulne abili y o clima e change,
he clima e eedbacks om di e en dis u bance in ensi-
ies, changes in unde -s o y habi a ollowing managemen
changes, and BVOCs as a esul o clima e change.
In his s udy, mul iple-si e calib a ion and op imiza ion
we e pe o med in o de o be e unde s and he unc ion-
ali y o he newly implemen ed mul i-laye ene gy budge
in ORCHIDEE-CAN 1.0. De eloping he mul i-laye en-
e gy budge equi es accu a e ield measu emen s o model
calib a ion and alida ion. He e we we e able o collec and
make use o many o he ew da a se s ha exis o in en-
si e in-canopy p o ile ime se ies measu emen s. We sug-
ges ha mo e in ensi e ield campaigns, wi h soil wa e con-
en obse a ions, especially du ing he win e season, would
help in he de elopmen o a mo e eliable pa ame iza-
ion scheme o he wi hin-canopy eddy di usi i y and soil–
a mosphe e in e ace conduc ance. Fo u u e model de el-
opmen s, adding an ex a soil–a mosphe e in e ace ep e-
sen a ion such as moss o he bs on he o es loo would
be bene icial o a mo e comple e mul i-laye ene gy budge
wi h he objec i e o desc ibing he su ace–a mosphe e in-
e ace gas and wa e apou exchanges.
6 Code and da a a ailabili y
The code and he un en i onmen a e open sou ce. Ne e -
heless, eade s in e es ed in unning ORCHIDEE-CAN 1.0
( e ision 2754) a e encou aged o con ac he co espond-
ing au ho o ull de ails and he la es bug ixes. The
ORCHIDEE-CAN b anch is a ailable ia he ollow web
link: h ps:// o ge.ipsl.jussieu. /o chidee/b owse /b anches/
ORCHIDEE-DOFOCO/ORCHIDEE.
The Supplemen ela ed o his a icle is a ailable online
a doi:10.5194/gmd-9-2951-2016-supplemen .
Au ho con ibu ions. Yiying Chen, James Ryde , and Sebas i-
aan Luyssae de eloped he pa ame iza ion scheme. Yiying Chen,
Sebas iaan Luyssae , and Philippe Peylin designed he s udy
and Yiying Chen w o e he manusc ip wi h con ibu ions om
all co-au ho s. James Ryde , Ma hew J. McG a h, Juliane O o,
Kim Naud s, Sebas iaan Luyssae , and Aude Valade helped Yiy-
ing Chen wi h in eg a ing he pa ame iza ion scheme o he mul i-
laye ene gy budge in ORCHIDEE-CAN 1.0. Vladisla Bas-
iko and Philippe Peylin p o ided he op imiza ion ools and
helped wi h he con igu a ion o hese ools. E a an Go sel,
Vanessa Ha e d, Be na d Heinesch, Alexande Knohl, Samuli Lau-
niainen, Denis Lous au, Eddy Moo s, Jé ôme Ogée, Thomas Foken,
and Timo Vessala p o ided ield obse a ions o all s udy si es.
Acknowledgemen s. Yiying Chen, James Ryde , Ma hew J. Mc-
G a h, Juliane O o, Kim Naud s and Sebas iaan Luyssae we e
unded h ough ERC s a ing g an 242564 (DOFOCO), and
Aude Valade was unded h ough ADEME (BiCaFF). We hank he
wo anonymous e iewe s o hei e y help ul obse a ions and
sugges ions.
Edi ed by: J. Kala
Re iewed by: wo anonymous e e ees
Re e ences
Aubine , M., Che manne, B., Vandenhau e, M., Longdoz, B., Ye -
naux, M., and Lai a , E.: Long e m ca bon dioxide exchange
abo e a mixed o es in he Belgian A dennes, Ag . Fo es
Me eo ol., 108, 293–315, doi:10.1016/S0168-1923(01)00244-1,
2001.
Baldocchi, D.: A Mul i-laye model o es ima ing sul u dioxide
deposi ion o a deciduous oak o es canopy, A mos. En i on.,
22, 869–884, doi:10.1016/0004-6981(88)90264-8, 1988.
Baldocchi, D. D. and Wilson, K. B.: Modeling CO2and wa-
e apo exchange o a empe a e b oadlea ed o es ac oss
hou ly o decadal ime scales, Ecol. Model., 142, 155–184,
doi:10.1016/S0304-3800(01)00287-3, 2001.
Ba , A. G., Black, T. A., Hogg, E. H., Kljun, N., Mo gen-
s e n, K., and Nesic, Z.: In e -annual a iabili y in he lea
www.geosci-model-de .ne /9/2951/2016/ Geosci. Model De ., 9, 2951–2972, 2016
2970 Y. Chen e al.: E alua ing he pe o mance o ORCHIDEE-CAN 1.0
a ea index o a bo eal aspen-hazelnu o es in ela ion o ne
ecosys em p oduc ion, Ag . Fo es Me eo ol., 126, 237–255,
doi:10.1029/2002JD003011, 2004.
Bonan, G. B.: A land su ace model (LSM e sion 1.0) o ecologi-
cal, hyd ological, and a mosphe ic s udies, Technical desc ip ion
and use ’s guide, NCAR Tech. No e NCAR/TN-417+STR, Tech.
ep., Boulde , Colo ado, USA, 1996.
Bonan, G. B., Oleson, K. W., Fishe , R. A., Lasslop, G., and Reich-
s ein, M.: Reconciling lea physiological ai s and canopy lux
da a: Use o he TRY and FLUXNET da abases in he Com-
muni y Land Model e sion 4, J. Geophys. Res.-Biogeo., 117,
G02026, doi:10.1029/2011JG001913, 2012.
Bonan, G. B., Williams, M., Fishe , R. A., and Oleson, K. W.:
Modeling s oma al conduc ance in he ea h sys em: linking
lea wa e -use e iciency and wa e anspo along he soil–
plan –a mosphe e con inuum, Geosci. Model De ., 7, 2193–
2222, doi:10.5194/gmd-7-2193-2014, 2014.
By d, R. H., Lu, P., Nocedal, J., and Zhu, C.: A Limi ed Memo y
Algo i hm o Bound Cons ained Op imiza ion, SIAM J. Sci.
Compu ., 16, 1190–1208, doi:10.1137/0916069, 1995.
Cesca i, A. and Ma colla, B.: D ag coe icien and u bulence in-
ensi y in coni e canopies, Ag . Fo es Me eo ol., 121, 197–206,
doi:10.1016/j.ag o me .2003.08.028, 2004.
Chen, Y.-Y. and Li, M.-H.: De e mining Adequa e A e aging Pe-
iods and Re e ence Coo dina es o Eddy Co a iance Mea-
su emen s o Su ace Hea and Wa e Vapo Fluxes o e
Moun ainous Te ain, Te . A mos. Ocean. Sci., 23, 685–701,
doi:10.3319/TAO.2012.05.02.01(Hy), 2012.
Colla z, G., Ribas-Ca bo, M., and Be y, J.: Coupled
Pho osyn hesis-S oma al Conduc ance Model o Lea es
o C4Plan s, Aus . J. Plan Physiol., 19, 519–538,
doi:10.1071/PP9920519, 1992.
De aeye, T., De ome, D., Ve bo en, P., Ca melie , J., and Nico-
lai, B.: C oss-scale modelling o anspi a ion om s om-
a a ia he lea bounda y laye , Ann. Bo ., 114, 711–723,
doi:10.1093/aob/mc 313, 2014.
de Noble -Ducoud é, N., Boisie , J.-P. P., Pi man, A., Bonan, G. B.,
B o kin, V., C uz, F., Deli e, C., Gayle , V., an den Hu k, B. J.
J. M., Law ence, P. J., an de Molen, M. K., Mülle , C., Reick,
C. H., S enge s, B. J., and Voldoi e, A.: De e mining obus im-
pac s o land-use-induced land co e changes on su ace clima e
o e No h Ame ica and Eu asia: Resul s om he i s se o LU-
CID expe imen s, J. Clima e, 25, 3261–3281, doi:10.1175/JCLI-
D-11-00338.1, 2012.
Dickinson, R. E., Shaikh, M., B yan , R., and G aumlich, L.: In e -
ac i e Canopies o a Clima e Model, J. Clima e, 11, 2823–2836,
doi:10.1175/1520-0442(1998)011<2823:ICFACM>2.0.CO;2,
1998.
Dolman, A. J. and Wallace, J. S.: Lag angian andK-
heo y app oaches in modelling e apo a ion om spa se
canopies, Q. J. Roy. Me eo ol. Soc., 117, 1325–1340,
doi:10.1002/qj.49711750210, 1991.
Dolman, A. J., Moo s, E. J., and Elbe s, J. A.: The ca bon up ake
o a mid la i ude pine o es g owing on sandy soil, Ag . Fo es
Me eo ol., 111, 157–170, doi:10.1016/S0168-1923(02)00024-2,
2002.
D obinski, P., Ana , A., Lebeaupin B ossie , C., Samson, G., S é-
anon, M., Bas in, S., Baklou i, M., Bé ange , K., Beu ie ,
J., Bou dallé-Badie, R., Coqua , L., D’And ea, F., de Noble -
Ducoud é, N., Diaz, F., Du ay, J. C., E he, C., Foujols,
M. A., Kh o os yano , D., Madec, G., Mancip, M., Mas-
son, S., Menu , L., Palmie i, J., Polche , J., Tu que y, S., Val-
cke, S., and Vio y, N.: Model o he Regional Coupled Ea h
sys em (MORCE): Applica ion o p ocess and clima e s ud-
ies in ulne able egions, En i on. Modell. So w., 35, 1–18,
doi:10.1016/j.en so .2012.01.017, 2012.
Ducoud é, N. I., La al, K., and Pe ie , A.: SECHIBA, a New Se
o Pa ame e iza ions o he Hyd ologic Exchanges a he Land-
A mosphe e In e ace wi hin he LMD A mosphe ic Gene al
Ci cula ion Model, J. Clima e, 6, 248–273, doi:10.1175/1520-
0442(1993)006<0248:SANSOP>2.0.CO;2, 1993.
Du esne, J.-L. and Gha as, J.: Desc ip ion du schema de la couche
limi e u bulen e e l’in e ace a ec la su ace plane ai e dans
LMDZ, Tech. ep., LMDZ, Pa is, F ance, 2009.
El Mas i, B., Shu, S., and Jain, A. K.: Implemen a ion o a dy-
namic oo ing dep h and phenology in o a land su ace model:
E alua ion o ca bon, wa e , and ene gy luxes in he high
la i ude ecosys ems, Ag . Fo es Me eo ol., 211–212, 85–99,
doi:10.1016/j.ag o me .2015.06.002, 2015.
Fa quha , G. D., on Caemme e , S., and Be y, J. A.: A biochem-
ical model o pho osyn he ic CO2assimila ion in lea es o C3
species, Plan a, 90, 78–90, doi:10.1007/BF00386231, 1980.
Foken, T., Meixne , F. X., Falge, E., Ze zsch, C., Se a imo ich, A.,
Ba gs en, A., Beh end , T., Bie mann, T., B euninge , C., Dix,
S., Ge ken, T., Hunne , M., Lehmann-Pape, L., Hens, K., Joche ,
G., Kesselmeie , J., Lüe s, J., Maye , J.-C., Mo a ek, A., Plake,
D., Riede e , M., Rü z, F., Scheibe, M., Siebicke, L., Sö gel, M.,
S aud , K., T ebs, I., Tsokankunku, A., Welling, M., Wol , V.,
and Zhu, Z.: Coupling p ocesses and exchange o ene gy and
eac i e and non- eac i e ace gases a a o es si e – esul s
o he EGER expe imen , A mos. Chem. Phys., 12, 1923–1950,
doi:10.5194/acp-12-1923-2012, 2012.
Gao, W., Shaw, R. H., and Paw U, K. T.: Obse a ion o o ganized
s uc u e in u bulen low wi hin and abo e a o es canopy,
Bound.-Lay. Me eo ol., 47, 349–377, doi:10.1007/BF00122339,
1989.
Ga a , J. R.: The A mosphe ic Bounda y Laye , Camb idge Uni-
e si y P ess, New Yo k, USA, 1992.
G ace, J.: The u bulen bounda y laye o e a lapping Pop-
ulus lea , Plan Cell En i on., 1, 35–38, doi:10.1111/j.1365-
3040.1978. b00743.x, 1978.
Gu, L., Shuga , H. H., Fuen es, J. D., Black, T., and Shewchuk,
S. R.: Mic ome eo ology, biophysical exchanges and NEE de-
composi ion in a wo-s o y bo eal o es – de elopmen and es
o an in eg a ed model, Ag . Fo es Me eo ol., 94, 123–148,
doi:10.1016/S0168-1923(99)00006-4, 1999.
Guen he , A., Hewi , C. N., E ickson, D., Fall, R., Ge on, C.,
G aedel, T., Ha ley, P., Klinge , L., Le dau, M., Mckay, W. A.,
Pie ce, T., Scholes, B., S einb eche , R., Tallam aju, R., Tay-
lo , J., and Zimme man, P.: A global model o na u al ola ile
o ganic compound emissions, J. Geophys. Res., 100, 8873,
doi:10.1029/94JD02950, 1995.
Guen he , A., Ka l, T., Ha ley, P., Wiedinmye , C., Palme , P. I.,
and Ge on, C.: Es ima es o global e es ial isop ene emissions
using MEGAN (Model o Emissions o Gases and Ae osols om
Na u e), A mos. Chem. Phys., 6, 3181–3210, doi:10.5194/acp-6-
3181-2006, 2006.
Geosci. Model De ., 9, 2951–2972, 2016 www.geosci-model-de .ne /9/2951/2016/
Y. Chen e al.: E alua ing he pe o mance o ORCHIDEE-CAN 1.0 2971
Ha e d, V., Leuning, R., G i i h, D., Go sel, E. V., and Cun z,
M.: The Tu bulen Lag angian Time Scale in Fo es Canopies
Cons ained by Fluxes, Concen a ions and Sou ce Dis ibu ions,
Bound.-Lay. Me eo ol., 130, 209–228, doi:10.1007/s10546-008-
9344-4, 2009.
Ha e d, V., Lo ell, J. L., Cun z, M., Jupp, D. L. B., Newn-
ham, G. J., and Sea, W.: The Canopy Semi-analy ic P
gap And Radia i e T ans e (CanSPART) model: Fo mula-
ion and applica ion, Ag . Fo es Me eo ol., 160, 14–35,
doi:10.1016/j.ag o me .2012.01.018, 2012.
He, Y., De Wekke , S. F., Fuen es, J. D., and D’Odo ico, P.: Coupled
land-a mosphe e modeling o he e ec s o sh ub enc oachmen
on nigh ime empe a u es, Ag . Fo es Me eo ol., 151, 1690–
1697, doi:10.1016/j.ag o me .2011.07.005, 2011.
Hu ley, L. B., O’G ady, A. P., and Eamus, D.: E apo anspi a ion
om eucalyp open- o es sa anna o no he n aus alia, Func .
Ecol., 14, 183–194, doi:10.1046/j.1365-2435.2000.00416.x,
2000.
Jackson, R. B.: T ading Wa e o Ca bon wi h Biolog-
ical Ca bon Seques a ion, Science, 310, 1944–1947,
doi:10.1126/science.1119282, 2005.
Jiménez, C., P igen , C., Muelle , B., Sene i a ne, S. I., Mc-
Cabe, M. F., Wood, E. F., Rossow, W. B., Balsamo, G., Be s,
A. K., Di meye , P. A., Fishe , J. B., Jung, M., Kanami su,
M., Reichle, R. H., Reichs ein, M., Rodell, M., She ield, J.,
Tu, K., and Wang, K.: Global in e compa ison o 12 land su -
ace hea lux es ima es, J. Geophys. Res.-A mos., 116, D02102,
doi:10.1029/2010JD014545, 2011.
Jo dan, D. and Smi h, W.: Ene gy balance analysis o nigh ime
lea empe a u es and os o ma ion in a subalpine en i on-
men , Ag . Fo es Me eo ol., 71, 359–372, doi:10.1016/0168-
1923(94)90020-5, 1994.
Knohl, A., Schulze, E. D., Kolle, O., and Buchmann, N.: La ge
ca bon up ake by an unmanaged 250-yea -old deciduous o -
es in Cen al Ge many, Ag . Fo es Me eo ol., 118, 151–167,
doi:10.1016/S0168-1923(03)00115-1, 2003.
K inne , G., Vio y, N., de Noble -Ducoud é, N., Ogée, J., Polche ,
J., F iedlings ein, P., Ciais, P., Si ch, S., and P en ice, I. C.:
A dynamic global ege a ion model o s udies o he cou-
pled a mosphe e-biosphe e sys em, Global Biogeochem. Cy., 19,
GB1015, doi:10.1029/2003GB002199, 2005.
Kuppel, S., Peylin, P., Maignan, F., Che allie , F., Kiely, G.,
Mon agnani, L., and Cesca i, A.: Model–da a usion ac oss
ecosys ems: om mul isi e op imiza ions o global simula ions,
Geosci. Model De ., 7, 2581–2597, doi:10.5194/gmd-7-2581-
2014, 2014.
Kus as, W. P., P uege , J. H., Ha ield, J. L., Ramalingam, K., and
Hipps, L. E.: Va iabili y in soil hea lux om a mesqui e dune
si e, Ag . Fo es Me eo ol., 103, 249–264, doi:10.1016/S0168-
1923(00)00131-3, 2000.
Lai a , E., Che manne, B., and Po ie , B.: Biomass, ca bon and ni-
ogen alloca ion in open op chambe s unde ambien and ele-
a ed CO2and in a mixed o es s and A en a i e app oach o
scaling up om he expe imen s o Vielsalm, in: Fo es Ecosys-
em Modelling, Upscaling and Remo e Sensing, 33–59, Aca-
demic Publishing, The Hague, he Ne he lands, 1998.
Launiainen, S., Vesala, T., Mölde , M., Mamma ella, I., Smolande ,
S., Rannik, Ü., Kola i, P., Ha i, P., Lind o h, A., and Ka ul, G. G.:
Ve ical a iabili y and e ec o s abili y on u bulence cha ac e -
is ics down o he loo o a pine o es , Tellus B, 59, 919–936,
doi:10.1111/j.1600-0889.2007.00313.x, 2007.
Launiainen, S., Ka ul, G. G., Lau en, A., and Kola i, P.:
Coupling bo eal o es CO2, H2O and ene gy lows by
a e ically s uc u ed o es canopy – Soil model wi h
sepa a e b yophy e laye , Ecol. Model., 312, 385–405,
doi:10.1016/j.ecolmodel.2015.06.007, 2015.
Lo ell, J., Ha e d, V., Jupp, D., and Newnham, G.: The Canopy
Semi-analy ic Pgap And Radia i e T ans e (CanSPART) model:
Valida ion using g ound based lida , Ag . Fo es Me eo ol., 158–
159, 1–12, doi:10.1016/j.ag o me .2012.01.020, 2012.
Lugo, A. E.: Visible and in isible e ec s o hu icanes on o es
ecosys ems: an in e na ional e iew, Aus al Ecol., 33, 368–398,
doi:10.1111/j.1442-9993.2008.01894.x, 2008.
MacBean, N., Maignan, F., Peylin, P., Bacou , C., B éon, F.-M., and
Ciais, P.: Using sa elli e da a o imp o e he lea phenology o
a global e es ial biosphe e model, Biogeosciences, 12, 7185–
7208, doi:10.5194/bg-12-7185-2015, 2015.
Maignan, F., B éon, F.-M., Che allie , F., Vio y, N., Ciais, P., Ga -
ec, C., T ules, J., and Mancip, M.: E alua ion o a Global Veg-
e a ion Model using ime se ies o sa elli e ege a ion indices,
Geosci. Model De ., 4, 1103–1114, doi:10.5194/gmd-4-1103-
2011, 2011.
Maka , P. A., Fuen es, J. D., Wang, D., S aeble , R. M., and Wiebe,
H. A.: Chemical p ocessing o biogenic hyd oca bons wi hin and
abo e a empe a e deciduous o es , J. Geophys. Res., 104, 3581,
doi:10.1029/1998JD100065, 1999.
Massman, W. J. and Weil, J. C.: An analy ical one-dimensional
second-o de closu e model o u bulence s a is ics and he
Lag angian ime scale wi hin and abo e plan canopies
o a bi a y s uc u e, Bound.-Lay. Me eo ol., 91, 81–107,
doi:10.1023/A:1001810204560, 1999.
Mayocchi, C. and B is ow, K.: Soil su ace hea lux: some gene al
ques ions and commen s on measu emen s, Ag . Fo es Me eo-
ol., 75, 43–50, doi:10.1016/0168-1923(94)02198-S, 1995.
McG a h, M. J., Pin y, B., Ryde , J., O o, J., and Luyssae , S.:
A mul ile el canopy adia i e ans e scheme based on a do-
maina e aged s uc u e ac o , in p epa a ion, 2016.
McNaugh on, K. G. and Van Den Hu k, B. J. J. M.: A “Lag angian”
e ision o he esis o s in he wo-laye model o calcula ing
he ene gy budge o a plan canopy, Bound.-Lay. Me eo ol., 74,
261–288, doi:10.1007/BF00712121, 1995.
Moo s, E. J.: Wa e Use o Fo es s in he Ne he lands, PhD hesis,
Wageningen, he Ne he lands, 2012.
Naud s, K., Ryde , J., McG a h, M. J., O o, J., Chen, Y., Valade,
A., Bellasen, V., Be honga ay, G., Bönisch, G., Campioli, M.,
Gha as, J., De G oo e, T., Ha e d, V., Ka ge, J., MacBean,
N., Maignan, F., Me ilä, P., Penuelas, J., Peylin, P., Pin y, B.,
P e zsch, H., Schulze, E. D., Solyga, D., Vuicha d, N., Yan, Y.,
and Luyssae , S.: A e ically disc e ised canopy desc ip ion o
ORCHIDEE (SVN 2290) and he modi ica ions o he ene gy,
wa e and ca bon luxes, Geosci. Model De ., 8, 2035–2065,
doi:10.5194/gmd-8-2035-2015, 2015.
Ogée, J., B une , Y., Lous au, D., Be bigie , P., and Delzon,
S.: MuSICA, a CO2, wa e and ene gy mul ilaye , mul i-
lea pine o es model: E alua ion om hou ly o yea ly ime
scales and sensi i i y analysis, Glob. Change Biol., 9, 697–717,
doi:10.1046/j.1365-2486.2003.00628.x, 2003.
www.geosci-model-de .ne /9/2951/2016/ Geosci. Model De ., 9, 2951–2972, 2016

2972 Y. Chen e al.: E alua ing he pe o mance o ORCHIDEE-CAN 1.0
Pin y, B., La e gne, T., Dickinson, R. E., Widlowski, J. L., Gob-
on, N., and Ve s ae e, M. M.: Simpli ying he in e ac ion o
land su aces wi h adia ion o ela ing emo e sensing p od-
uc s o clima e models, J. Geophys. Res.-A mos., 111, D02116,
doi:10.1029/2005JD005952, 2006.
Pi man, A. J.: The e olu ion o , and e olu ion in, land su ace
schemes designed o clima e models, In . J. Clima ol., 23, 479–
510, doi:10.1002/joc.893, 2003.
Pi man, A. J., De Noble -Ducoud é, N., C uz, F. T., Da in, E. L.,
Bonan, G. B., B o kin, V., Claussen, M., Deli e, C., Ganze eld,
L., Gayle , V., Van Den Hu k, B. J. J. M., Law ence, P. J., Van
De Molen, M. K., Mülle , C., Reick, C. H., Sene i a ne, S. I.,
S engen, B. J., and Voldoi e, A.: Unce ain ies in clima e e-
sponses o pas land co e change: Fi s esul s om he LU-
CID in e compa ison s udy, Geophys. Res. Le ., 36, L14814,
doi:10.1029/2009GL039076, 2009.
Po e, A., Bosc, A., Champion, I., and Lous au, D.: Es ima -
ing he oliage a ea o Ma i ime pine (Pinus pinas e Ai .)
b anches and c owns wi h applica ion o modelling he o-
liage a ea dis ibu ion in he c own, Ann. Fo . Sci., 57, 73–86,
doi:10.1051/ o es :2000110, 2000.
P ihodko, L., Denning, A., Hanan, N., Bake , I., and Da is, K.: Sen-
si i i y, unce ain y and ime dependence o pa ame e s in a com-
plex land su ace model, Ag . Fo es Me eo ol., 148, 268–287,
doi:10.1016/j.ag o me .2007.08.006, 2008.
Raupach, M. R.: Applying Lag angian Fluid-Mechanics To In-
e Scala Sou ce Dis ibu ion Concen a ion P o iles in Plan
Canopies, Ag . Fo es Me eo ol., 47, 85–108, 1989.
Raupach, M. R.: Vege a ion-a mosphe e in e ac ion in ho-
mogeneous and he e ogeneous e ain: some implica-
ions o mixed-laye dynamics, Vege a io, 91, 105–120,
doi:10.1007/BF00036051, 1991.
Richa dson, A. D., Hollinge , D. Y., Bu ba, G. G., Da is, K. J.,
Flanagan, L. B., Ka ul, G. G., Munge , J. W., Ricciu o, D. M.,
S oy, P. C., Suyke , A. E., Ve ma, S. B., and Wo sy, S. C.: A
mul i-si e analysis o andom e o in owe -based measu emen s
o ca bon and ene gy luxes, Ag . Fo es Me eo ol., 136, 1–18,
doi:10.1016/j.ag o me .2006.01.007, 2006.
Ryde , J., Polche , J., Peylin, P., O lé, C., Chen, Y., an Go sel,
E., Ha e d, V., McG a h, M. J., Naud s, K., O o, J., Valade,
A., and Luyssae , S.: A mul i-laye land su ace ene gy budge
model o implici coupling wi h global a mosphe ic simula ions,
Geosci. Model De ., 9, 223–245, doi:10.5194/gmd-9-223-2016,
2016.
Sche e s, B. R., Phillips, B. L., Lau ance, W. F., Sodhi, N. S., Dies-
mos, A., and Williams, S. E.: Inc easing a bo eali y wi h al i ude:
a no el biogeog aphic dimension, P. Roy. Soc. B Biol. Sci., 280,
1–9, doi:10.1098/ spb.2013.1581, 2013.
Seidl, R., Fe nandes, P. M., Fonseca, T. F., Gille , F., Jönsson, A. M.,
Me ganiˇ
co á, K., Ne he e , S., A paci, A., Bon emps, J.-D., Bug-
mann, H., González-Olaba ia, J. R., Lasch, P., Me edieu, C.,
Mo ei a, F., Schelhaas, M.-J., and Moh en, F.: Modelling na u al
dis u bances in o es ecosys ems: a e iew, Ecol. Model., 222,
903–924, doi:10.1016/j.ecolmodel.2010.09.040, 2011.
Selle s, P. J., Los, S. O., Tucke , C. J., Jus ice, C. O., Dazlich,
D. A., Colla z, G. J., and Randall, D. A.: A e ised land su -
ace pa ame e iza ion (SiB2) o a mosphe ic GCMs. Pa II: The
gene a ion o global ields o e es ial biophysical pa ame e s
om sa elli e da a, J. Clima e, 9, 706–737, doi:10.1175/1520-
0442(1996)009<0706:ARLSPF>2.0.CO;2, 1996.
S aud , K., Se a imo ich, A., Siebicke, L., Pyles, R. D., and
Falge, E.: Ve ical s uc u e o e apo anspi a ion a a o -
es si e (a case s udy), Ag . Fo es Me eo ol., 151, 709–729,
doi:10.1016/j.ag o me .2010.10.009, 2011.
S öckli, R. and Vidale, P. L.: Modeling diu nal o seasonal wa e and
hea exchanges a Eu opean Fluxne si es, Theo . Appl. Clima o.,
80, 229–243, doi:10.1007/s00704-004-0102-3, 2005.
S oud, C., Maka , P., Ka l, T., Guen he , A., Ge on, C., Tu nipseed,
A., Nemi z, E., Bake , B., Po osnak, M., and Fuen es, J. D.:
Role o canopy-scale pho ochemis y in modi ying biogenic-
a mosphe e exchange o eac i e e pene species: Resul s om
he CELTIC ield s udy, J. Geophys. Res.-A mos., 110, 149–162,
doi:10.1029/2005JD005775, 2005.
Taylo , K. E.: Summa izing mul iple aspec s o model pe o -
mance in a single diag am, J. Geophys. Res., 106, 7183,
doi:10.1029/2000JD900719, 2001.
Thomas, C. and Foken, T.: Flux con ibu ion o cohe en s uc-
u es and i s implica ions o he exchange o ene gy and ma e
in a all sp uce canopy, Bound.-Lay. Me eo ol., 123, 317–337,
doi:10.1007/s10546-006-9144-7, 2007.
Vuicha d, N. and Papale, D.: Filling he gaps in me eo ological con-
inuous da a measu ed a FLUXNET si es wi h ERA-In e im e-
analysis, Ea h Sys . Sci. Da a, 7, 157–171, doi:10.5194/essd-7-
157-2015, 2015.
Wang, T., Peng, S., K inne , G., Ryde , J., Li, Y., Dan ec-
Nédélec, S., and O lé, C.: Impac s o Sa elli e-Based
Snow Albedo Assimila ion on O line and Coupled Land
Su ace Model Simula ions, PLOS ONE, 10, e0137275,
doi:10.1371/jou nal.pone.0137275, 2015.
Wild, M.: How weil do IPCC-AR4/CMIP3 clima e models simu-
la e global dimming/b igh ening and wen ie h-cen u y day ime
and nigh ime wa ming?, J. Geophys. Res.-A mos., 114, 1–10,
doi:10.1029/2008JD011372, 2009.
Wohl ah , G. and Ce nusca, A.: Momen um ans e by a moun-
ain meadow canopy: A simula ion analysis based on Mass-
man’s (1997) model, Bound.-Lay. Me eo ol., 103, 391–407,
doi:10.1023/A:1014960912763, 2002.
Wol e, G. M., Tho n on, J. A., Bou ie -B own, N. C., Golds ein, A.
H., Pa k, J.-H., McKay, M., Ma oss, D. M., Mao, J., B une, W.
H., LaF anchi, B. W., B owne, E. C., Min, K.-E., Woold idge,
P. J., Cohen, R. C., C ounse, J. D., Faloona, I. C., Gilman, J.
B., Kus e , W. C., de Gouw, J. A., Huisman, A., and Keu sch,
F. N.: The Chemis y o A mosphe e-Fo es Exchange (CAFE)
Model – Pa 2: Applica ion o BEARPEX-2007 obse a ions,
A mos. Chem. Phys., 11, 1269–1294, doi:10.5194/acp-11-1269-
2011, 2011.
Yin, X. and S uik, P. C.: C3and C4pho osyn hesis models:
An o e iew om he pe spec i e o c op modelling, NJAS –
Wagen. J. Li e Sc., 57, 27–38, doi:10.1016/j.njas.2009.07.001,
2009.
Yue, C., Ciais, P., Cadule, P., Thonicke, K., A chibald, S., Poul-
e , B., Hao, W. M., Han son, S., Mouillo , F., F iedlings ein, P.,
Maignan, F., and Vio y, N.: Modelling he ole o i es in he
e es ial ca bon balance by inco po a ing SPITFIRE in o he
global ege a ion model ORCHIDEE – Pa 1: simula ing his o -
ical global bu ned a ea and i e egimes, Geosci. Model De ., 7,
2747–2767, doi:10.5194/gmd-7-2747-2014, 2014.
Geosci. Model De ., 9, 2951–2972, 2016 www.geosci-model-de .ne /9/2951/2016/