A disse a ion submi ed o he
Facul y o Biology, Chemis y and Geosciences
Uni e si y o Bay eu h
o a ain he academic deg ee o
D . e . na .
Si e-speci ic modelling o u bulen luxes on
he Tibe an Pla eau
Wol gang Babel
M.Sc. Global Change Ecology
bo n 3 No embe , 1974
in P on en, Ge many
Bay eu h, Ma ch 2013
supe ised by P o . D . Thomas Foken
Si e-speci ic modelling o u bulen luxes on he Tibe an
Pla eau
supe ised by P o . D . Thomas Foken
i
Die o liegende A bei wu de in de Zei on Ap il 2009 bis M¨a z 2013 in Bay eu h an
de Ab eilung Mik ome eo ologie un e Be euung on He n P o . D . Thomas Foken
ange e ig .
Volls ¨andige Abd uck de on de Fakul ¨a ¨u Biologie, Chemie und Geowissenscha en
de Uni e si ¨a Bay eu h genehmig en Disse a ion zu E langung des akademischen
G ades eines Dok o de Na u wissenscha en (D . e . na .).
Disse a ion einge eich am: 6. M¨a z 2013
Zulassung du ch die P ¨u ungskommission: 13. M¨a z 2013
Wissenscha liches Kolloquium: 15. Mai 2013
Am ie ende Dekanin:
P o . D . Bea e Lohne
P ¨u ungsausschuss:
P o . D . Thomas Foken (E s gu ach e )
D . habil. E a Falge (Zwei gu ach e )
P o . D . And eas Held (Vo si z)
P o . D . Michael Hauhs
P o . D . Be nd Huwe
ii
Con en s
Lis o manusc ip s
Lis o addi ional publica ions i
Acknowledgemen s iii
Summa y ix
Zusammen assung xi
1. In oduc ion 1
1.1. Mo i a ion: Regional es ima es o u bulen luxes on he Tibe an Pla eau 1
1.2. My con ibu ion o he p ojec objec i es . . . . . . . . . . . . . . . . . 2
1.3. Objec i es o he hesis . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2. Backg ound 7
2.1. Obse ed ene gy balance and closu e . . . . . . . . . . . . . . . . . . . 7
2.2. Land su ace modelling on he Tibe an Pla eau . . . . . . . . . . . . . 8
2.3. Lake su ace modelling on he Tibe an Pla eau . . . . . . . . . . . . . . 9
3. Me hods 11
3.1. The Nam Co 2009 expe imen . . . . . . . . . . . . . . . . . . . . . . . 11
3.1.1. Si edesc ip ion........................... 11
3.1.2. Measu emen s............................ 11
3.1.3. Da a pos -p ocessing . . . . . . . . . . . . . . . . . . . . . . . . 12
3.1.4. Ene gy balance closu e and co ec ion . . . . . . . . . . . . . . 14
3.2. Land su ace modelling o Nam Co 2009 . . . . . . . . . . . . . . . . . 16
3.2.1. Model e sions ........................... 16
3.2.2. Model pa ame e s . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3.3. Lake su ace modelling o Nam Co 2009 . . . . . . . . . . . . . . . . . 21
4. Resul s 23
4.1. Da a quali y on he Tibe an Pla eau . . . . . . . . . . . . . . . . . . . 23
4.2. Flux measu emen s a Nam Co . . . . . . . . . . . . . . . . . . . . . . 24
4.3. Land su ace modelling a Nam Co . . . . . . . . . . . . . . . . . . . . 25
iii

4.4. In luence o he ene gy balance co ec ion me hod . . . . . . . . . . . . 27
4.5. Lake su ace modelling a Nam Co . . . . . . . . . . . . . . . . . . . . 30
4.6. Flux he e ogenei y a Nam Co . . . . . . . . . . . . . . . . . . . . . . . 32
5. Conclusions 37
Re e ences 41
A. Indi idual con ibu ions o he join publica ions 54
B. Ge ken e al. (2012) 58
C. Babel e al. (2013) 75
D. Bie mann e al. (2013) 88
E. Cha uchi ipan e al. (2013) 117
F. Li e al. (2013) 146
E kl¨a ung 156
i
Lis o manusc ip s
This disse a ion is p esen ed in a cumula i e o m. I is based on h ee pee - e iewed
publica ions and wo submi ed manusc ip s as lis ed below.
Pee - e iewed publica ions
Bie mann, T., Babel, W., Ma, W., Chen, X., Thiem, E., Ma, Y., and Foken, T.:
Tu bulen lux obse a ions and modelling o e a shallow lake and a we g assland
in he Nam Co basin, Tibe an Pla eau, Theo . Appl. Clima ol., accep ed
Ge ken, T., Babel, W., Ho mann, A., Bie mann, T., He zog, M., F iend, A. D.,
Li, M., Ma, Y., Foken, T., and G a , H.-F.: Tu bulen lux modelling wi h a
simple 2-laye soil model and ex apola ed su ace empe a u e applied a Nam
Co Lake basin on he Tibe an Pla eau, Hyd ol. Ea h Sys . Sci., 16, 1095–1110,
doi:10.5194/hess-16-1095-2012, 2012.
Li, M., Babel, W., Tanaka, K., and Foken, T.: No e on he applica ion o plana - i
o a ion o non-omnidi ec ional sonic anemome e s, A mos. Meas. Tech., 6,
221–229, doi:10.5194/am -6-221-2013, 2013.
Manusc ip s submi ed
Babel, W., Chen, Y., Bie mann, T., Yang, K., Ma, Y., and Foken, T.: Adap a ion
o a land su ace scheme o modeling u bulen luxes on he Tibe an Pla eau
unde di e en soil mois u e condi ions, submi ed o J. Geophys. Res.
Cha uchi ipan, D., Babel, W., Maude , M., Leps, J.-P., and Foken, T.: Ex ension
o he a e aging ime o he eddy-co a iance measu emen and i s e ec on he
ene gy balance closu e, submi ed o Bound.-Lay. Me eo ol.
Lis o addi ional publica ions
The ollowing lis summa ises o he publica ions o mine which a e no included in he
disse a ion. They we e spli up in wo pa s: Publica ions which ha e e e ence o he
hesis and o he publica ions. The i s pa includes wo mas e heses as well, wi h
e e ence o he disse a ion, ini ia ed and supe ised by mysel .
Publica ions wi h e e ence o his hesis
Pee - e iewed publica ions
Zhou, D., Eigenmann, R., Babel, W., Foken, T., and Ma, Y.: The s udy o nea -
g ound ee con ec ion condi ions a Nam Co s a ion on he Tibe an Pla eau,
Theo . Appl. Clima ol., 105, 217–228, doi:10.1007/s00704-010-0393-5, 2011.
Non pee - e iewed publica ions
Babel, W., Eigenmann, R., Ma, Y., and Foken, T.: Analysis o u bulen luxes and
hei ep esen a i eness o he in e ac ion be ween he a mosphe ic bounda y
laye and he unde lying su ace on Tibe an Pla eau, CEOP-AEGIS Deli e able
epo De1.2, Ed. Uni e si y o S asbou g, F ance, ISSN 2118-7843: 35 p.,
2011a.
Babel, W., Li, M., Sun, F., Ma, W., Chen, X., Colin, J., Ma, Y., and Foken, T.:
Ae odynamic and he modynamic a iables o ou s a ions on Tibe an Pla eau
– In oduc ion on he CEOP-AEGIS da abase in Ne CDF, CEOP-AEGIS Deli -
e able epo De1.3, Ed. Uni e si y o S asbou g, F ance, ISSN 2118-7843: 90
p., 2011b.
Babel, W.: An R ou ine o he simpli ied usage o TERRAFEX (Cha ac e isa ion
o a complex measu ing si e o lux measu emen s), Wo k Repo Uni e si y o
Bay eu h, Dep . o Mic ome eo ology, ISSN 1614-8916, in p epa a ion
Bie mann, T., Babel, W., Olesch, J., and Foken, T.: Documen a ion o he Mic ome e-
o ological Expe imen , Nam Tso, Tibe , 25 h o June – 8 h o Augus 2009, Wo k
Repo Uni e si y o Bay eu h, Dep . o Mic ome eo ology, ISSN 1614-8916,
41, 38 pp., URL h p://opus.ub.uni-bay eu h.de/opus4-ubbay eu h/ on doo /
index/index/docId/626, 2009.
Foken, T. and Babel, W.: Au dem “Dach de Wel ” – Die Rolle Tibe s bei de
Wasse e so gung S¨udos asiens, in: Spek um-Magazin de Uni e si ¨a Bay eu h,
Ausgabe 1/2012, 46–48, 2012
i
Ge ken, T., Babel, W., Ho mann, A., Bie mann, T., He zog, M., F iend, A. D.,
Li, M., Ma, Y., Foken, T., and G a , H.-F.: Tu bulen lux modelling wi h a
simple 2-laye soil model and ex apola ed su ace empe a u e applied a Nam
Co Lake basin on he Tibe an Pla eau, Hyd ol. Ea h Sys . Sci. Discuss., 8,
10 275–10 309, doi:10.5194/hessd-8-10275-2011, 2011.
Ge ken, T., Fuchs, K., and Babel, W.: Documen a ion o he A mosphe ic Bounda y
Laye Expe imen , Nam Tso, Tibe , 08 h o July – 08 h o Augus 2012, Wo k
Repo Uni e si y o Bay eu h, Dep . o Mic ome eo ology, ISSN 1614-8916, 53,
48pp., 2013
Li, M., Babel, W., Tanaka, K., and Foken, T.: No e on he applica ion o plana -
i o a ion o non-omnidi ec ional sonic anemome e s, A mos. Meas. Tech.
Discuss., 5, 7323–7340, doi:10.5194/am d-5-7323-2012, 2012.
Mas e heses supe ised by mysel
Baume , M.: Ve gleich zweie Lag ange’sche Modelle zu Bes immung des Foo p in s
¨ube he e ogenem Gel¨ande, Mas e hesis, Uni e si y o Bay eu h, 66pp., 2012.
Thiem, E.: Modelling o he ene gy exchange abo e lake and land su aces, Mas e
hesis, Uni e si y o Bay eu h, 73pp., 2011.
O he publica ions no included in his hesis
I l, S. D. H., S einbaue , M. J., Babel, W., Beie kuhnlein, C., Blume-We y, G.,
Messinge , J., Paloma es Ma ´ınez, ´
A., S ohmeie , S., and Jen sch, A.: An 11- y
exclosu e expe imen in a high-ele a ion island ecosys em: in oduced he bi o e
impac on sh ub species ichness, seedling ec ui men and popula ion dynamics,
J. Veg. Sci., 23, 1114–1125, doi:10.1111/j.1654-1103.2012.01425.x, 2012.
Babel, W., Huneke, S., and Foken, T.: A amewo k o u ilize u bulen lux mea-
su emen s o mesoscale models and emo e sensing applica ions, Hyd ol. Ea h
Sys . Sci. Discuss., 8, 5165–5225, doi:10.5194/hessd-8-5165-2011, 2011.
Babel, W. and Foken, T.: P elimina y oo p in analysis o LAS (La ge ape u e
scin illome e ) measu emen s a Qomolangma s a ion, Tibe an Pla eau, CEOP-
AEGIS echnical epo , 2009, a ailable a he CEOP-AEGIS p ojec o ice, bu
no published ye
ii
1. In oduc ion
Table 1.1. Pe manen measu emen si es on he Tibe an Pla eau as selec ed o he
CEOP-AEGIS p ojec
Si e Coo dina es Al i ude Land co e
Naqu (BJ) 31◦220700N 91◦5305500E 4502 m Alpine s eppe
Nam Co 30◦4602200N 90◦5704700E 4745 m Alpine s eppe
Linzhi 29◦4505600N 94◦4401800E 3327 m Alpine g assland
Qomolangma 28◦2102900N 86◦5604700E 4293 m G a el
o such egional lux maps should be s eng hened , bu an adequa e p ocessing and
quali y con ol, as summa ised by Rebmann e al. (2012), Foken e al. (2012) and
demons a ed on he TP e.g. by Me zge e al. (2006), is a p e equisi e.
1.2. My con ibu ion o he p ojec objec i es
My wo k o he las yea s is mainly ela ed o he p ojec : “Coo dina ed Asia-Eu opean
long- e m Obse ing sys em o Qinghai–Tibe Pla eau hyd ome eo ological p ocesses
and he Asian-monsoon sys Em wi h G ound sa elli e Image da a and nume ical
Simula ions (CEOP-AEGIS)”. I is a collabo a i e p ojec / Small o medium-scale
ocused esea ch p ojec – Speci ic In e na ional Co-ope a ion Ac ion inanced by he
Eu opean Commission unde FP7 opic ENV.2007.4.1.4.2 ”Imp o ing obse ing sys-
ems o wa e esou ce managemen ”, and is coo dina ed by he Uni e si y o S as-
bou g, F ance (www.ceop-aegis.o g). Amongs o he goals i aims a cons uc ing an
obse ing sys em o moni o he pla eau’s wa e yield by a combina ion o g ound
measu emen s and sa elli e based obse a ions. This is o immedia e in e es o wa e
esou ces managemen in Sou h-Eas Asia. The an icipa ed ou come is p o iding in-
as uc u e, i.e. an in e ac i e da a po al will be deli e ed, ea u ing a h ee-yea da a
se wi h obse a ions o he wa e balance e ms. Toge he wi h highe le el p oduc s
and dis ibu ed hyd ological modelling a p o o ype moni o ing sys em is o med in-
cluding ea ly wa ning on loods and d ough s which is in ended o emain in ope a ion
beyond p ojec comple ion.
Wi hin his in as uc u al p ojec I was esponsible o deli e ing quali y checked
g ound based obse a ions o su ace ene gy luxes. The da a se inco po a es luxes
o e h ee yea s a ou s a ions on he Tibe an Pla eau (see Table 1.1), he u -
bulen luxes (sensible hea and la en hea /e apo anspi a ion) we e de i ed by he
eddy-co a iance me hod. Based on p e ious s udies abou quali y assu ance o eddy-
co a iance measu emen s in he communi y and especially in he Depa men o Mi-
2

1.2. My con ibu ion o he p ojec objec i es
c ome eo ology, Uni e si y o Bay eu h, I mysel compiled a wo k low o he p ocessing
o eddy-co a iance da a in h ee le els. These a e:
Le el I Tu bulen luxes, being checked o da a quali y, oo p in , and po en ial ob-
s acles in he icini y o he senso .
Le el II Le el I da a, co ec ed wi h espec o he ene gy balance closu e.
Le el III Gap- illed le el II da a.
In he cou se o his I adap ed espec i e ools o oo p in analysis de eloped by
Ma hias G¨ockede (G¨ockede e al., 2005, 2006, 2008) and me ged hem in o a single,
use -op imised R ou ine (Wo k epo in p ep.). The scheme and i s implemen a ion
up o he c ea ion o Ne CDF da a se s is desc ibed in CEOP-AEGIS echnical epo s
(Babel e al., 2011a,b). As i was no possible o ge access o he Chinese o iginal
da a, I in oduced Chinese colleagues o he me hods and supe ised he p ocessing a
all s ages.
In gene al, he p esen ed wo kload could be s aigh o wa dly handled consul ing he
ele an li e a u e and execu ing he necessa y s eps. Ne e heless, he special condi-
ions on he Tibe an Pla eau men ioned in Sec . 1.1 aise he need o in es iga ions
beyond exis ing s udies. Gap- illing o u bulen luxes o Le el III aims a ep esen -
ing he sys em dynamics adequa ely o his unique en i onmen a he han deli e ing
co ec annual sums o e apo anspi a ion. This canno be achie ed wi h empi ical o
s a is ical app oaches as p esen ed by Falge e al. (2001a,b), bu wi h Soil - Vege a ion
- A mosphe e - T ans e (SVAT) models. The e o e own u bulen lux measu emen s
o e di e en ypes o su aces ha e been conduc ed in o de o de i e he essen ial
inpu pa ame e s as well as u bulen lux obse a ions o alida ion (Bie mann e al.,
2009; Ge ken e al., in p ep.). This was necessa y, because he access o Chinese da a
ga he ed on he Tibe an Pla eau was e y limi ed. The measu emen s, howe e , show a
signi ican non-closu e o he ene gy balance, which is dis inc i e o such he e ogeneous
en i onmen s (Foken, 2008a). Di e en op ions exis o close his gap (e.g. Foken e al.,
2011) and hei applica ion clea ly has an impac on he ou come o model alida ion.
The aim o hese e o s is o p o ide su ace model solu ions o signi ican land
su ace ypes on he Tibe an Pla eau, which se e as a high-s anda d gap- illing ool and
es ima ion o ep esen a i eness o he pe manen lux s a ions wi hin CEOP-AEGIS.
Fu he mo e, inconsis encies in he u bulence measu emen s caused by speci ic senso
ypes had o be in es iga ed in o de o assess he e ec o such p oblems on sensible
and la en hea lux measu emen s. F om hese p oblems I de ined my esea ch asks
u he desc ibed in he nex sec ion.
3
1. In oduc ion
1.3. Objec i es o he hesis
The main ocus o his hesis is he applica ion o p ocess based modelling o es i-
ma e su ace sensible hea luxes and la en hea luxes (e apo anspi a ion) in he
speci ic en i onmen o he Tibe an Pla eau. Al hough se e al su ace model s udies
al eady exis on he Tibe an Pla eau (see Sec . 2.2), a ho ough analysis including
eddy-co a iance measu emen s o u bulen luxes is s ill missing. Mo eo e , he e is a
lack o e apo a ion measu emen s abo e lake su aces. To my bes knowledge, Tobias
Bie mann and me conduc ed he i s eddy-co a iance measu emen s o e a lake su -
ace on he Tibe an Pla eau. This made i possible o alida e a lake su ace model.
Based on hese p econdi ions he ollowing esea ch ques ions ha e been elabo a ed.
•Land su ace modelling unde he speci ic condi ions o he Tibe an Pla eau using
he land su ace scheme SEWAB (Mengelkamp e al., 1999) and alida ion wi h
eddy-co a iance measu emen s.
•Elabo a ion o he necessa y model pa ame e s: equi ed e o s and impac on
model pe o mance.
•In es iga ion o di e en me hods o co ec o he ene gy balance closu e gap
and hei in luence on model pe o mance assessmen .
•Po en ial applica ion o he elabo a ed model e sion.
•Speci ic p oblems o eddy-co a iance da a quali y on he Tibe an Pla eau and
mi iga ion o speci ic measu emen p oblems occu ing wi h some sonic anemome-
e s.
The publica ions and manusc ip s lis ed on page con ibu e o hese esea ch ques-
ions as ollows: Babel e al. (2013, Appendix C) p esen an adap a ion o he land
su ace scheme SEWAB (Mengelkamp e al., 1999) o he Tibe an Pla eau. The adap-
a ion is designed o conside speci ic issues o land su ace modelling as men ioned
in Sec . 2.2. The model pe o mance wi h espec o u bulen luxes is in es iga ed
by using eddy-co a iance measu emen s abo e alpine s eppe om wo nea by si es a
Nam Co lake. Pa ame e se s o igina ing om bo h s anda d alues and labo a o y
and in si u measu emen s a e es ed in his manusc ip . Special emphasis is pu on
he ene gy balance closu e o u bulen lux measu emen s and i s co ec ion. Besides
o using he well known me hod o dis ibu ing he esidual acco ding o he Bowen
a io (Twine e al., 2000), a new co ec ion me hod, sugges ed by Cha uchi ipan e al.
(2013, Appendix E), has been applied: The esidual is dis ibu ed acco ding o he el-
a i e con ibu ion o he u bulen luxes o he buoyancy lux. Cha uchi ipan e al.
(2013, Appendix E) analyse a comp ehensi e da a se om he LITFASS-2003 cam-
paign, Lindenbe g, Ge many. The in luence o he a e aging ime o eddy-co a iance
luxes is in ensely s udied u ilising Ogi e analysis, block ensemble a e aging, wa ele
4
1.3. Objec i es o he hesis
and quad an analysis. The manusc ip sugges s seconda y ci cula ions o be esponsi-
ble o he gap in ene gy balance closu e and a ibu es a dominan ole o he sensible
hea lux. The p oposed new closu e co ec ion me hod is based on hese expe imen-
al indings. As a no el app oach, bo h wo co ec ion me hods ha e been applied o
he da a and he consequences on model pe o mance e alua ion is discussed by Babel
e al. (2013, Appendix C).
The SEWAB model success ully simula es u bulen luxes on he Tibe an Pla eau,
c ea ing se e al bene i s: I can be as a e e ence o a mo e simplis ic pa ame isa ions
as done by Ge ken e al. (2012, Appendix B). The ein he simplis ic land su ace scheme
Hyb id is upda ed wi h a new soil model, aiming o elimina e a delayed su ace esponse
o a mosphe ic o cing in he o iginal e sion. SEWAB is success ully u ilised o
compa isons in example daily cycles. As SEWAB showed no delay in su ace esponse,
i s simula ions ha e been used o e alua e Hyb id’s esponsi eness wi h c oss co ela ion
(Ge ken e al., 2012, Appendix B).
Ano he applica ion is he usage o he simula ed imese ies in o de o assess land-
scape he e ogenei y a Nam Co (Bie mann e al., 2013, Appendix D). The gappy ob-
se a ions o a we alpine s eppe and a shallow lake could be e ec i ely desc ibed by
modelled imese ies o SEWAB and a hyd odynamic mul ilaye model (Foken, 1984)
wi h an ex ension o shallow wa e exchange (Panin and Foken, 2005). Tu bulen
luxes o bo h su ace ypes a e hen compa ed wi h he SEWAB simula ions o he
“s anda d” land su ace a Nam Co, d y alpine s eppe. I could be shown ha he
di e ences among land su ace ypes likely exceed he model unce ain y, so he di e -
ences a e conside able. The e ec o his he e ogenei y is discussed in e ms o using
he eddy-co a iance da a as g ound u h o emo e sensing.
In o de o use eddy-co a iance da a o model e alua ion some issues abou da a
quali y should be cla i ied in ad ance. Al hough no included in he hesis Zhou e al.
(2011)1p o ide a basis o u he usage o eddy-co a iance da a a he Nam Co Mon-
i o ing and Resea ch S a ion o Mul isphe e In e ac ions: In addi ion o s anda d
e alua ion o oo p in and da a quali y he occu ence o nea -g ound ee con ec ion
e en s is in es iga ed. I is shown ha such e en s can be c ea ed on he Tibe an
Pla eau al eady due o changing cloudiness, and hei in luence on da a quali y is as-
sessed. Fu he mo e, a one o he CEOP-AEGIS si es he sonic anemome e DAT
600 TR61A p obe om Kaijo-Denki is in use. The senso is no omnidi ec ional, i.e.
in a ce ain sec o he wind ield is dis u bed by he senso s uc u e and i egula
ic ion eloci ies occu as a consequence. The s udy by Li e al. (2013, Appendix F)
highligh s his p oblem and in es iga es i s in luence on scala luxes and whe he such
p oblems occu also wi h he commonly used CSAT3, Campbell Scien i ic L d. A
sec o -wise plana - i is sugges ed as an app op ia e coo dina e o a ion o mi iga e
1Toge he wi h Ra ael Eigenmann I in oduced Zhou Degang in o he pos -p ocessing o he u bu-
lence da a, including he usage o TK2 and oo p in analysis ools, and in o he in es iga ion and
ele ance o nea g ound ee con ec ion condi ions a Nam Co s a ion. I pe sonally suppo ed
him in he usage and in e p e a ion o he da a quali y ools.
5
1. In oduc ion
such p oblems.
6
2. Backg ound
2.1. Obse ed ene gy balance and closu e
The eddy-co a iance me hod is he only di ec me hod o measu e u bulen exchange
o hea and scala s be ween he a mosphe e and he unde lying su ace and is he e o e
p e e ed in he communi y o quan i y long- e m luxes o wa e apou and ca bon
dioxide (Foken and Wichu a, 1996; Baldocchi e al., 2001). Despi e he gene al us
placed in his me hod, i became appa en ha he ene gy balance canno be closed a
mos expe imen al si es (e.g. Foken, 2008a). The su ace ene gy balance a he su ace
is gi en by
−Rne =QH+QE+QG+ ∆QS(2.1)
wi h he ne adia ion Rne , he sensible hea lux QH, he la en hea lux QE, he
g ound hea lux QG, and he change in ene gy s o age ∆QS. The signs ollow he con-
en ion ha luxes di ec ed owa ds he su ace a e nega i e and ice e sa. Al hough
his heo e ical balance should be ep oduced wi h measu emen s as well, acco ding o
Foken (2008a) mos s udies epo ha he sum o u bulen ene gy (QHand QE) only
yields 70%–100% o he a ailable ene gy (−Rne −QG−∆QS). This esidual is ypically
la ge in complex landscapes (e.g. Aubine e al., 2000; Wilson e al., 2002), while in
homogeneous a eas o dese s he ene gy balance can be closed (e.g. Heusink eld e al.,
2004; Maude e al., 2007).
I is ecognized ha in he measu ed ene gy balance u bulen ene gy is missing
a he han a ailable ene gy being o e es ima ed so long as he ene gy s o age and
he g ound hea lux has been add essed adequa ely (Twine e al., 2000; Foken, 2008a;
Foken e al., 2011; Leuning e al., 2012). Especially when compa ing eddy-co a iance
de i ed u bulen luxes wi h hose land su ace model simula ions, which a e ac ually
cons ained by he ene gy balance equa ion, a sys ema ic misma ch can be expec ed
as poin ed ou by Falge e al. (2005). The e o e a co ec ion o he u bulen luxes
should be conside ed. A widely used co ec ion me hod p oposed by Twine e al.
(2000) dis ibu es he esidual acco ding o he Bowen a io, assuming scala simila i y
be ween la en and sensible hea wi h espec o he missing lux. In con as some
s udies indica e ha he missing ene gy s ems om sensible hea only (Maude and
Foken, 2006; Ingwe sen e al., 2011). The ecen discussion hypo hesizes an in luence
o seconda y ci cula ions in complex landscapes, igge ing nea -g ound ad ec i e and
low- equency lux componen s (S ein eld e al., 2007; Foken e al., 2010, 2011; S oy
e al., 2013; B ¨o z e al., 2013). Such ci cula ion sys ems a e mainly d i en by buoyancy
7

2. Backg ound
sugges ing he buoyancy lux o play a dominan ole o he ene gy balance closu e.
2.2. Land su ace modelling on he Tibe an Pla eau
Due o i s emo eness, egional lux es ima ion on he Tibe an Pla eau does no ha e a
long his o y, see also Babel e al. (2013, Appendix C). Fi s mul i-yea es ima es ac oss
a a ie y o si es on he Tibe an Pla eau ha e been p esen ed by Xu and Haginoya
(2001), calcula ing luxes om s anda d me eo ological measu emen s. Wi hin he
GEWEX Asian Monsoon Expe imen 1998 Takayabu e al. (2001) u ilized ou land
su ace models o a compa ison s udy, epo ed la ge di e ences in u bulen lux
pa i ioning, bu could no alida e he models due o a lack o soil mois u e and lux
measu emen s. Only e y ew s udies exis using eddy-co a iance lux measu emen s
o alida ion (e.g. Yang e al., 2009; Hong and Kim, 2010).
These s udies ind an o e es ima ion o he sensible hea lux as a ypical ea u e
o land su ace modelling on he Tibe an Pla eau. They blame oo high u bulen
di usion coe icien s o his p oblem an d aw a ela ionship o he special condi ions
on he Tibe an Pla eau (see Sec . 1.1), in pa icula he s ong diu nal cycle o su ace
empe a u e o e d y and spa sely ege a ed su aces. Indeed, Yang e al. (2003)
and Ma e al. (2002) obse ed a diu nal a ia ion o he sublaye -S an on numbe
B, desc ibing he loga i hm o he a io be ween ae odynamic and he mal oughness
leng hs κB−1= ln(z0m
z0h ). Fo mula ions o κB−1as a ixed ac ion, o depending on he
ic ion eloci y (Zili inke ich, 1995) canno esol e his diu nal a ia ion. The e o e
(Yang e al., 2008) p opose a new o mula ion wi h an addi ional dependence on he
empe a u e scale T∗and empi ical adap a ion o Tibe an Pla eau obse a ions (see
Sec . 3.2.1). Leading o he de e mina ion o a a iable he mal oughness leng h in
land su ace models, his pa ame isa ion has been success ully a ed in some ecen
s udies in Asian a id egions Yang e al. (2008); Chen e al. (2010, 2011); Liu e al.
(2012); Zhang (2012).
O he de elopmen s include he in luence o soil e ical he e ogenei y which is ound
o be signi ican in case o ema kable s a i ica ion (Yang e al., 2005; an de Velde
e al., 2009). Fo he soil hea lux on he Tibe an Pla eau Yang e al. (2005) adap ed
a pa ame isa ion o he soil he mal conduc i i y, p oposed by Johansen (1975) and
ecommended by Pe e s-Lida d e al. (1998). In his o m i can be easily ans e ed o
any condi ions when d y and sa u a ed he mal conduc i i ies a e known. Fu he mo e,
la en hea luxes can be obse ed on he Tibe an Pla eau e en i soil mois u e d ops
below wil ing poin , a ypical ea u e in dese s o a id landscapes (Agam e al., 2004;
Balsamo e al., 2011; Wallace e al., 1991). Di e en pa ame isa ions o ba e soil
e apo a ion in dependence on soil mois u e a e compa ed by Mihailo i´c e al. (1995),
bu hese a e no es ed on he Tibe an Pla eau ye .
Ano he challenge on he Tibe an Pla eau is a small-scale he e ogenei y o soil mois-
u e (Su e al., 2011). Cold semia id condi ions cha ac e ise he landscape consis ing o
8
2.3. Lake su ace modelling on he Tibe an Pla eau
d y g asslands (Kob esia pas u es, alpine s eppe, pa ly non- ege a ed) and we lands
o g asslands wi h shallow g oundwa e . The ole o he soil mois u e in e ac ing wi h
clima e is highligh ed in a e iew by Sene i a ne e al. (2010). The land–a mosphe e
coupling s eng h (in luence o soil mois u e on p ecipi a ion) on he Tibe an Pla eau,
howe e , is a ed low by Kos e e al. (2004), bu he local pa e ns o soil mois u e and
p ecipi a ion canno be esol ed by such s udies using ensembles o global ci cula ion
models. Land su ace models a e in p inciple able o ake small-scale he e ogenei y o
soil mois u e in o accoun , bu his has o be es ed.
Ob aining lux measu emen s in such a emo e en i onmen as he Tibe an Pla eau
is challenging (Ma e al., 2009b) and land su ace modelling can suppo land su ace
lux es ima ion on local and egional scale. A wide ange o land su ace schemes
being sui able o such a ask ha e been e ol ed in he las decades: S a ing wi h
simple schemes (e.g. Manabe, 1969), second-gene a ion land su ace models e ol ed,
adding e.g. de ailed esis ance schemes o e apo anspi a ion and mo e complexi y
in o he desc ip ion o soil p ocesses (Pi man, 2003). Mo e ecen de elopmen s o
hi d-gene a ion models mainly include a dynamic ege a ion, he “g eening” o land
su ace models (Pi man, 2003).
In he con ex o land su ace modelling his hesis lays he ocus on how he model
desc ip ion o he special Tibe an Pla eau condi ions a ec land su ace luxes (see
Sec . 1.1). The eby a c ucial ea u e is he compa ison wi h si e-speci ic eddy-co a iance
measu emen s. As eedback mechanisms o he land su ace o he a mosphe e a e ou -
o -scope, he mos sui able app oach is an o line o ced land su ace model wi h p e-
sc ibed, si e-speci ic s a e o ege a ion and soil p ope ies. The used model SEWAB
is a ep esen a i e o he second-gene a ion land su ace models and pa icipa ed in
he P ojec o In e compa ison o Land-su ace Pa ame iza ion Schemes (PILPS:
Hende son-Selle s e al., 1996; Chen e al., 1997). F om his expe ience i has been
imp o ed ega ding he desc ip ion o soil and su ace p ocesses (Mengelkamp e al.,
1999, 2001; Wa ach e al., 2001) and is he e o e adequa ely s uc u ed o he in-
ended pu pose. Ini ially de eloped o humid condi ions SEWAB’s pe o mance unde
d y condi ions has o be es ed ye .
2.3. Lake su ace modelling on he Tibe an Pla eau
Lake su aces should be aken in o accoun o egional lux es ima ion on he Tibe an
Pla eau as app oxima ely 45 000 km2is co e ed by lakes (Xu e al., 2009). The impo -
ance o lake su aces o he egional ene gy balance and wa e cycle has been poin ed
ou by Rouse e al. (2005) and No dbo e al. (2011). Some e apo a ion es ima es
al eady exis o lake su aces, modelled wi h simple bulk app oaches based on daily
o mon hly o cing da a om emo e sensing o su ace obse a ions (Haginoya e al.,
2009; Xu e al., 2009; K ause e al., 2010; Yu e al., 2011), in some cases alida ed
wi h pan e apo a ion measu emen s. To my bes knowledge no s udies a e epo ed
9
2. Backg ound
ye , u ilising o conduc ing eddy-co a iance measu emen s abo e lake su aces on he
Tibe an Pla eau.
The e is a huge amoun o lakes on he Tibe an Pla eau (≈1090 lakes la ge han
10 km2, Yu e al., 2011), he e o e a a ie y o lake ex en and dep h can be expec ed o
occu . These ac o s g ea ly in luence su ace empe a u e (and he eby a mosphe ic
s abili y) and su ace oughness (Rouse e al., 2005; Panin e al., 2006a; No dbo e al.,
2011). These a iables in u n ypically exhibi a diu nal a ia ion on he one hand and
hei ela ionship o he su ace luxes is non-linea . The e o e he p ocesses can only
be ep esen ed by esol ing he diu nal cycle, which is no possible wi h he me hods
men ioned abo e.
Fo such a pu pose a hyd odynamic mul ilaye (HM) model (Foken, 1979, 1984) is
a sui able candida e. I is o iginally designed o ene gy exchange abo e he ocean.
Shallow wa e condi ions, howe e , inc ease he wa e heigh , depending on wind e-
loci y, and he e o e enhance he u bulen exchange (Panin and Foken, 2005). The
HM model and he shallow wa e app oach has been alida ed wi h eddy-co a iance
da a om a lake in Ge many and he impac on exchange o e he Caspian Sea has
been discussed (Panin e al., 2006b,a). Ne e heless, i has o be es ed whe he hese
pa ame isa ions wo k unde he condi ions o he Tibe an Pla eau as well.
10
3. Me hods
3.1. The Nam Co 2009 expe imen
As poin ed ou in Sec . 1.2 own expe imen s we e ine i able o de i e necessa y da a
o inpu , pa ame isa ion and alida ion o su ace models. The Nam Co si e (see
desc ip ion in he ollowing) has been chosen, as i is an a ea whe e ypical land su aces
o he Tibe an Pla eau (d y alpine s eppe, mo e we and dense g assland, and lakes)
occu closely oge he . Loca ed a he in e sec ion o he Wes e lies wi h he Asian
Monsoon ci cula ion sys ems he Nam Co basin has been conside ed as a key a ea o
in e es (Haginoya e al., 2009; Keil e al., 2010).
3.1.1. Si e desc ip ion
The Nam Co 2009 expe imen was ca ied ou om 26 June o 8 Augus wi hin he
2009 summe monsoon season. The si e is loca ed 220 km no h o Lhasa in he Nam Co
Basin, Tibe an Pla eau, wi h i s lake su ace a an ele a ion o 4730 km a.s.l. The basin
is domina ed by he lake i sel and he Nyainqen anglha moun ain ange, s e ched
along i s SE side and eaching up o 7270 m a.s.l. wi h an a e age heigh o 5230 m
(Liu e al., 2010). The Ins i u e o Tibe an Pla eau Resea ch (ITP), Chinese Academy
o Sciences is ope a ing he Nam Co Moni o ing and Resea ch S a ion o Mul isphe e
In e ac ions nea a small lake in 1 km dis ance SE o he Nam Co Lake (see Figu e 3.1).
The ege a ion a ound Nam Co e lec s he p e ailing a id, high-al i ude clima e wi h
alpine meadows and s eppe g asses (M¨ugle e al., 2010). Nea he Nam Co S a ion he
ege a ion co e age and composi ion a e highly a iable acco ding o he soil mois u e
condi ions de e mined by opog aphic ea u es: G ass gene a ypical o Alpine s eppe
(S ipa, Ca ex, Helic o ichon, Elymus, Fes uca, Kob esia, Poa, see Bie mann e al.,
2009; Miehe e al., 2011) ha e been obse ed on he hillocks, wi h a o al ege a ion
co e age o 60 % o less (g ass−), while mo e we a eas a e densely co e ed (>90 %)
wi h alpine meadows domina ed by Kob esia species (g ass+, see Figu e 3.1).
3.1.2. Measu emen s
Tu bulen luxes we e ob ained by wo ene gy balance sys ems. One se -up is loca ed
di ec ly a he Nam Co s a ion o e d y alpine s eppe (g ass−), u he called NamITP,
ope a ed by he Ins i u e o Tibe an Pla eau Resea ch. The NamITP complex is se led
11
3. Me hods
•hyd ological modules con aining unable pa ame e s, which canno be de e -
mined, a e disabled (ponding, a iable in il a ion capaci y, ARNO concep o
subsu ace uno and base low, dep h dependency pa ame isa ion o sa u a ed
hyd aulic conduc i i y, see Mengelkamp e al., 1999, 2001)
•o line o cing wi h measu ed p ecipi a ion, ai empe a u e, wind eloci y, ai
p essu e, ela i e humidi y, downwelling sho -wa e and long-wa e adia ion us-
ing he same da a o bo h g ass+and g ass−.
•in e nal model ime s ep o 10 min, in e pola ion o 30-min o cing da a and
agg ega ion o ou pu o 30 min.
•ini ialisa ion o soil mois u e and soil empe a u e p o iles wi h a 3-yea o cing
da a se ex ac ed om he ITPCAS (Ins i u e o Tibe an Pla eau Resea ch,
Chinese Academy o Sciences) g idded o cing da a se (Chen e al., 2011). Tes
simula ions showed easonable simula ions o soil mois u e o he g ass−su ace,
bu could no be used o he g ass+su ace, as he shallow g ound wa e able
a NamUBT could no be ep oduced wi h a single column ealisa ion.
•ini ialisa ion o soil mois u e and soil empe a u e p o iles wi h obse ed p o iles.
This ini ialisa ion showed good ag eemen wi h he 3-yea spin-up a g ass−,
he e o e i has been solely used o all analysis o bo h su ace ypes.
The adap a ion o he Tibe an Pla eau (TP e sion) aims a add essing he issues
men ioned in Sec . 2.2. The changes include:
1. A new calcula ion o he soil he mal conduc i i y λs ollowing Yang e al. (2005)
λs(Θ) = λd y + (λsa −λd y) exp [0.36 ·(1 −Θsa /Θ)] (3.6)
wi h he olume ic soil wa e con en Θ and Θsa as he po osi y. The d y and
sa u a ed he mal conduc i i y limi s we e es ima ed om ield obse a ions as
λd y = 0.15 W m−1K−1and λsa = 0.8 and 1.3 W m−1K−1 o g ass+and g ass−,
espec i ely. This pa ame isa ion eplaced he o iginal o mula ion ea u ing a
weigh ed sum o indi idual he mal conduc i i ies o d y clay/sand, wa e , ice
and ai acco ding o he ac ual s a e.
2. To accoun o diu nal and seasonal a ia ions o he he mal oughness leng h
obse ed on he Tibe an Pla eau (Yang e al., 2003), a o mula ion acco ding o
Yang e al. (2008) has been implemen ed
z0h =70ν
u∗
exp −βu0.5
∗|T∗|0.25(3.7)
wi h he kinema ic iscosi y o ai ν, he ic ion eloci y u∗, he dynamic empe -
a u e scale T∗=−w0T0/u∗and an empi ical cons an β= 7.2 s0.5m−0.5K−0.25. As
18

3.2. Land su ace modelling o Nam Co 2009
T∗depends on z0h, he equa ion has o be sol ed i e a i ely (Yang e al., 2010).
The o iginal o mula ion es ima es z0h as a ixed ac ion o he ae odynamic
oughness leng h z0h = 0.1z0m.
3. Like obse ed in dese landscapes (Agam e al., 2004; Balsamo e al., 2011;
Wallace e al., 1991), la en hea luxes occu on he Tibe an Pla eau e en when
soil mois u e d ops below wil ing poin . The soil ai humidi y con olling ba e
soil e apo a ion is adjus ed in SEWAB wi h a soil mois u e dependen ac o α
(see Table 3.2). To accoun o d y condi ions a o mula ion by Mihailo i´c e al.
(1993) has been implemen ed
α=


1−1−Θ
ΘFC n
,Θ≤ΘFC
1,Θ>ΘFC
(3.8)
wi h he olume ic wa e con en a ield capaci y ΘFC, he ac ual wa e con en
o he opsoil Θ and using n= 2 as exponen . The o iginal pa ame isa ion
α= 0.5h1−cos Θ
ΘFC πi o Θ ≤ΘFC (Noilhan and Plan on, 1989) is e y
p ohibi i e o low Θ as poin ed ou by Mihailo i´c e al. (1995).
3.2.2. Model pa ame e s
In o de o ocus on he impac o he model e sions on he pe o mance, no op imi-
sa ion algo i hms we e applied o cons ain he pa ame e space. Ins ead, wo ways o
de i ing “ easonable” pa ame e se s we e explo ed, de ining a “measu ed” pa ame e
se and a “de aul ” pa ame e se . While he la e can be ob ained wi h a s anda d
knowledge o he su ace and soil ypes in ol ed, he measu ed pa ame e s ep esen
de ailed in si u and labo a o y obse a ions o he ele an si e-speci ic p ope ies. A
summa y o he mos impo an pa ame e s gi es Table 3.3. The lea a ea index,
emissi i y, minimum s oma al esis ance and maximum s oma al esis ance we e no
measu ed and he e o e uni o mly aken o bo h su ace ypes and pa ame e se s (Hu
e al., 2009; Yang e al., 2009; Alapa y e al., 1997).
De aul pa ame e s di e mos conside ably be ween bo h su aces in he desc ip ion
o he soil. The soil ex u e was classi ied as “sand” and “sandy loam”, USDA
ex u al classes, o g ass−and g ass+, espec i ely. The co esponding pa am-
e e s ha e been collec ed o SEWAB by Mengelkamp e al. (1997), o igina ing
om Clapp and Ho nbe ge (1978) in case o he hyd aulic p ope ies. As bo h
land use ypes a e classi ied as sho g assland, he su ace pa ame e s di e solely
in he ac ion o ege a ed a ea and he e o e in he o e -all albedo (albedo o
g assland and d y ba e soil om Foken, 2008b).
Measu ed pa ame e s use me eo ological obse a ions o albedo and oughness leng h
o momen um, ac ion o ege a ed a ea, canopy heigh and oo ing dep h we e
19
3. Me hods
Table 3.3. Mos impo an pa ame e s o he model simula ions: albedo a, emissi i y
ε, ac ion o ege a ed a ea eg, lea a ea index o ege a ed a ea LAI eg, canopy
heigh hc, oo ing dep h z , oughness leng h z0m, minimum s oma al esis ance
Rs,min, maximum s oma al esis ance Rs,max, he mal di usi i y νT, soil hea capac-
i y CG·%G, po osi y Θsa , ma ix po en ial a sa u a ion Ψsa , sa u a ed hyd aulic
conduc i i y Ksa , olume ic wa e con en a ield capaci y ΘFC, olume ic wa-
e con en a wil ing poin ΘWP, and exponen b o ela ionships a e Clapp and
Ho nbe ge (1978).
De aul pa ame e Measu ed pa ame e
Pa ame e Uni NamITP NamUBT NamITP NamUBT
Su ace and ege a ion pa ame e
a- 0.22 0.205 0.196 0.196
ε- 0.97 0.97 0.97 0.97
eg - 0.6 0.9 0.6 0.9
LAI eg - 1.0 1.0 1.0 1.0
hcm 0.15 0.15 0.15 0.07
z m 0.3 0.3 0.3 0.5
z0m m 0.005 0.005 0.005 0.005
Rs,min s m−160.0 60.0 60.0 60.0
Rs,max s m−12500 2500 2500 2500
Soil pa ame e
νTm2s−10.84 ·10−60.84 ·10−61.5 ·10−72.5 ·10−7
CG·%GJ m−3K−12.10 ·1062.10 ·1062.10 ·1062.10 ·106
Θsa m3m−30.395 0.435 0.396 0.63
Ψsa m -0.121 -0.218 -0.51 -0.14
Ksa m s−11.76 ·10−43.47 ·10−52.018 ·10−51.38 ·10−5
ΘFC m3m−30.135 0.150 0.21 0.38
ΘWP m3m−30.068 0.114 0.06 0.19
b- 4.05 4.90 3.61 6.79
20
3.3. Lake su ace modelling o Nam Co 2009
es ima ed in he ield. Soil physical pa ame e s we e deduced om labo a o y
in es iga ion o soil samples aken nea by he measu emen se -up (Chen e al.,
2012) assuming he samples o be ep esen a i e on he scale o he EC oo p in .
Di ec ly measu ed a e soil ex u e, he mal conduc i i y, hyd aulic conduc i i y
a sa u a ion and he soil wa e e en ion cu e, p o iding ma ix po en ial a
sa u a ion and exponen b(Clapp and Ho nbe ge , 1978). Backwa d calcula ion
o he las wo yield he olume ic wa e con en a ield capaci y (pF=2.5 as-
sumed) and a wil ing poin (pF=4.5 assumed). The la e pF alue di e s om
he s anda d 4.2, a easonable assump ion o mesophy ic g ass species (La che ,
2001, p208).
3.3. Lake su ace modelling o Nam Co 2009
Tu bulen luxes o e he shallow lake su ace nea Nam Co we e modelled wi h a hy-
d odynamic mul ilaye model (HM) (Foken, 1979, 1984). As he go e ning p inciple,
su ace – a mosphe e exchange is pa ame ised based on a bulk app oach, bu esol -
ing he molecula bounda y laye , he iscous bu e laye and u bulen laye by an
in eg a ed p o ile coe icien Γ. I accoun s o s a i ica ion by using Monin-Obukho
simila i y heo y.
The model is o ced by measu emen s, using he same da a se as u ilised o SEWAB
(see Sec . 3.2.1). Lake su ace empe a u e is app oxima ed by he measu ed lake
empe a u e, hence he e is no need o adia ion measu emen s and ene gy balance
closu e wi hin he model. The lake su ace empe a u e p obe was shielded agains
di ec adia ion, a adia ion e o due o di use adia ion in he wa e body has been
es ima ed as app oxima ely 0.2 K, see Bie mann e al. (2013, Appendix D). Wendisch
and Foken (1989) in es iga ed which o cing a iables a e mos in luen ial o he model
e o and iden i ied wa e empe a u e (50 %) and wind eloci y, ai empe a u e and
ai humidi y (10 % o 20 % each).
The HM model is designed o u bulen exchange o e he ocean. Shallow wa e ,
howe e , induces la ge wa es leading o highe oughness and an enhanced exchange
depending on wind eloci y and lake dep h H(Panin e al., 2006b). The e o e he
shallow wa e co ec ion p oposed by Panin and Foken (2005) has been implemen ed
in he HM code wi hin a mas e hesis (Thiem, 2011)
QSW
H,E=Qocean
H,E1 + kSW
H,E·h·H−1(3.9)
wi h he coe icien kSW
H,E≈2. As pos ula ed by he heo e ical conside a ion, he
shallow wa e u bulen luxes QSW
H,Ea e always la ge han he co esponding deep
wa e luxes Qocean
H,E. The mean squa e wa e heigh is pa ame ised wi h he empi ical
exp ession h≈0.07u2
10m ·g−1gHu−2
10m0.6, o mula ed by Da idan e al. (1985).
The e o e he in luencial pa ame e s o he shallow wa e ex ension a e wind eloci y
and lake dep h, hei impac on he ela i e inc ease in luxes is displayed in Fig. 3.4.
21
3. Me hods
0 5 10 15
0
2
4
6
8
10
12
u10m in ms−1
lake dep h in m
1.01
1.05
1.1
1.15
1.2
1.3
1.4
1.6
QSW ⋅Qocean
−1
Figu e 3.4. Sensi i i y o he shallow wa e e m on lake dep h Hand wind eloci y in
10 m heigh u10m: Isolines show he ela i e inc ease o deep wa e luxes QSW ·Q−1
ocean
depending on u10m and H
F om he equa ions local sensi i i ies can be de i ed. Using he mean wind eloci y
o 4 m s−1and a wa e dep h o 1.5 m and assuming co esponding ypical e o s o
0.3 m s−1and 1 m would lead o lux unce ain ies o 1 % and 4 %, espec i ely.
22
4. Resul s
4.1. Da a quali y on he Tibe an Pla eau
The quali y o u bulen luxes om eddy-co a iance da a has been analysed on ou s a-
ions on he Tibe an Pla eau (Table 1.1) wi h espec o ul ilmen o eddy-co a iance
equi emen s, ene gy balance closu e, oo p in , as well as obs acles in he icini y o
he senso and po en ially esul ing in e nal bounda y laye s (Babel e al., 2011a,b).
Despi e some si e-speci ic sou ces o dis u bance no discussed he e, wo, mo e gene al
ea u es can be highligh ed.
Fo one hing nea -g ound ee con ec i e condi ions ha e been ound e y equen ly
a Nam Co due o changes in he diu nal land-lake ci cula ion sys em and due o
changing cloud co e inducing sha p con as s in he su ace ene gy budge especially
on he Tibe an Pla eau (Zhou e al., 2011). This comes along wi h a deg ada ion o da a
quali y caused by bo h ins a iona i y and misma ch o heo e ical in eg al u bulence
cha ac e is ics. Zhou e al. (2011) a gue ha da a om hese si ua ions should no be
ou inely ejec ed, as hey desc ibe a ypical day ime phenomenon wi hin a con ec i e
bounda y laye .
Secondly, i egula ic ion eloci ies ha e been equen ly ound in he da a om he
BJ si e (now Naqu s a ion), ela ed o he used sonic anemome e DAT 600 TR61A
p obe om Kaijo-Denki. I egula ic ion means ha momen um lux has been e-
quen ly obse ed wi h he w ong di ec ion, a ing he su ace e oneously as a sou ce
o momen um a he han a sink. The DAT 600 is a non-omnidi ec ional senso wi h
a ela i ely small open sec o o 120◦. I is shown by Li e al. (2013, Appendix F)
ha he p oblem can be educed o da a o he undis u bed (open) sec o by apply-
ing a sec o -wise plana - i . Basically such pa i ions in dis u bed and undis u bed
sec o s a e ele an o all non-omnidi ec ional senso s, he e o e he impac o using a
sec o -wise plana - i is in es iga ed o he CSAT3, Campbell Scien i ic L d. as well.
The ic ion eloci y o he sec o -wise plana - i de ia es up o 10 % (DAT 600) om
he “usual” plana - i applied o he whole sec o o all wind di ec ions. Due o i s
la ge open sec o o 340◦no such di e ences could be ound o CSAT3, bu i egula
ic ion could be sligh ly educed, especially when he on sec o (i.e. he dis u bing
p obe elemen s occu s aigh behind he measu ing pa h) is o a ed sepa a ely. This
disc epancy be ween CSAT3 and DAT 600 is e lec ed by sys ema ic di e ences ound
be ween he ic ion eloci ies de i ed by bo h ins umen s when applying he plana -
i in he usual way. In con as , o sec o -wise plana - i no di e ences be ween bo h
23

4. Resul s
−Res
Rne
QE
QH
QG
g ass−
(a)
−800
−600
−400
−200
0
200
Ene gy luxes in Wm−2
−Res
Rne
QE
QH
QG
g ass+
(b)
Rne
QE
QH
lake(c)
Ts c−land
Tai
0000 0600 1200 1800
5
10
15
20
25
30
Temp in °C
Ts c−land
Tai
0000 0600 1200 1800
Ts c−lake
Tai
0000 0600 1200 1800 2400
Figu e 4.1. Mean diu nal ene gy luxes o he whole measu emen pe iod, sepa a ed
o land (a: g ass−a NamITP, b: g ass+a NamUBT) and lake (c: NamUBT); all
componen s a e measu ed o land luxes (a,b); o lake luxes, he ne adia ion is
calcula ed om measu ed downwelling adia ion and using an albedo o 0.06 and he
lake su ace empe a u e wi h an emissi i y o 0.96; he lowe panel shows diu nal
su ace and ai empe a u e. The ime axis is displayed in Beijing s anda d ime
(CST), mean local sola noon du ing he obse a ion pe iod is a 1400 CST. F om
Bie mann e al. (2013, Appendix D)
ins umen s can be seen. I is also impo an o men ion ha scala luxes we e no
a ec ed by he di e en plana - i o a ions.
4.2. Flux measu emen s a Nam Co
Du ing he monsoon season, he measu ed ene gy luxes a Nam Co exhibi a dis inc
spa ial he e ogenei y co esponding o di e en su ace ypes, see Babel e al. (2013,
Appendix C) and Bie mann e al. (2013, Appendix D). Mean diu nal ene gy luxes o
a d y (g ass−) and a we (g ass+) alpine s eppe and a shallow lake su ace can be seen
in Fig. 4.1. The measu emen s a NamUBT co espond o ei he g ass+o lake su ace,
depending on wind di ec ion (Fig. 3.2). The land su ace luxes (Fig. 4.1a, b) show a
simila diu nal cycle in gene al, wi h la en hea luxes domina ing o e sensible hea
luxes, a ypical ea u e o he monsoon season on he Tibe an Pla eau (e.g. Gu e al.,
2005; Ma and Ma, 2006). Ne e heless, e apo a ion is highe a g ass+on a e age
due o soil mois u e a ailabili y. While g ass+is cons an ly supplied by a shallow
24
4.3. Land su ace modelling a Nam Co
g ound wa e able, wa e a ailabili y is highly a iable a NamITP wi h olume ic
soil mois u es below 5 % mos o he ime, bu wi h sa u a ed soils sho ly a e ain
e en s. In such sho pe iods, la en and sensible hea luxes a e app oxima ely equal,
some example days a e shown by Babel e al. (2013, Appendix C). The NamITP si e
exhibi s he ypical ea u es o d y su aces on he TP wi h a huge diu nal cycle o he
su ace empe a u e (peaks each up o 50 ◦C on d y days), and only mode a e hea
luxes a e no able o edis ibu e his su ace hea con en e ec i ely (Yang e al.,
2009). The p ema u e change o he sign o he g ound hea lux in he ea ly a e noon
indica es a s ongly hea ed shallow soil laye , he mally decoupled om he deepe soil
and supplying ene gy o he su ace well be o e su ace empe a u es d op o he same
magni ude as ai empe a u e.
O e he lake su ace, he u bulen ene gy does no show a diu nal cycle, bu we e
cons an o e he day (Bie mann e al., 2013, Appendix D). The lake body is able o
elease ene gy a any ime, so e apo a ion is mainly limi ed by wind eloci y and apou
p essu e de ici . The shallow lake in pa icula shows compa ably la ge e apo a ion
due o high wind eloci ies o 4 m s−1on a e age and enhanced u bulen exchange
caused by uns able s a i ica ion e en du ing day ime (Fig. 4.1c). In con as , s able
s a i ica ion ypically p e ails o e lakes in day ime (e.g. Bey ich e al., 2006; No dbo
e al., 2011).
4.3. Land su ace modelling a Nam Co
In o de o assess model pe o mance, model uns o g ass−and g ass+we e conduc ed
o measu ed and de aul pa ame e s, using bo h he o iginal e sion and he adap a ion
o he Tibe an Pla eau (TP e sion). The simula ions we e compa ed wi h ene gy
balance co ec ed obse a ions using bo h co ec ion me hods acco ding o he Bowen
a io (EBC-Bo, Twine e al., 2000) and acco ding o he Buoyancy lux (EBC-HB), see
Sec . 3.1.4 and Babel e al. (2013, Appendix C).
In gene al obse ed pa e ns (EBC-Bo co ec ed) we e adequa ely ep oduced show-
ing co ela ion coe icien s o 0.9 o bo h si es, pa ame e se s and model e sions
(Babel e al., 2013, Appendix C). This is a no able ea u e as no op imisa ion algo-
i hm has been applied so a . Di e ences be ween model uns, howe e , can be ound
in model bias B=ξsim −ξobs and he Nash-Su cli e coe icien NS = 1 −PN(ξsim−ξobs)2
PN(ξobs−ξobs)2
(Nash and Su cli e, 1970), which can be in e p e ed simila ly o he common coe i-
cien o de e mina ion R2, bu is sensi i e o bias as well. A la ge and posi i e bias o
u bulen luxes is ound a g ass−, which is no appa en a g ass+(Fig. 4.2a). This is
in pa s connec ed o he es ima ion o he g ound hea lux and a high sensi i i y o
he new he mal conduc i i y o mula ion o changes in soil mois u e in he d y ange
(Babel e al., 2013, Appendix C). Ne e heless, bias in g ound hea lux could be e-
duced a g ass−compa ed o he o iginal e sion. The TP e sion educes he sensible
hea lux ia he new implemen a ion o he mal oughness and he e o e i s bias as
25
4. Resul s
(a) (b)
BQ u b
Bias B in Wm−2
BQH
−40
−20
0
20
40
60
BQE
NamITP NamUBT
od om
Td Tm
od
om
Td
Tm
NSQH
Nash−Su cli e coe icien NS
NSQE
0.0
0.2
0.4
0.6
0.8
1.0
NamITP
NamITP NamUBT
od om Td Tm od
om
Td
Tm
Figu e 4.2. Bias (a) and Nash-Su cli e coe icien NS (b) o u bulen luxes (simu-
la ed s. EBC-Bo co ec ed obse a ions), NamITP co esponds o g ass−, NamUBT
o g ass+; he indi idual blocks show od:o iginal SEWAB e sion, de aul pa ame-
e s; om:o iginal SEWAB e sion, measu ed pa ame e s; Td:TP e sion, de aul
pa ame e s; Tm:TP e sion, measu ed pa ame e s. Modi ied om Babel e al.
(2013, Appendix C)
well, he bias o la en hea is also sligh ly educed. This educ ion in bias o he TP
e sion induces a be e pe o mance wi h he NS coe icien (Fig. 4.2b). Simula ions
wi h de aul pa ame e s yield p edic ions close o he EBC-Bo co ec ed obse a ions.
This can be mainly a ibu ed o la ge ield capaci ies and wil ing poin s in he mea-
su ed pa ame e se (Table 3.3), supp essing e apo anspi a ion on bo h si es. The
new o mula ion o ba e soil e apo a ion pa ly compensa es his e ec a NamITP,
a NamUBT ba e soil e apo a ion akes no e ec as he ac ional a ea o ba e soil is
oo low.
As expec ed, he simula ions pe o m be e o g ass+ han o g ass−in gene al,
as SEWAB o mula ions ha e no been alida ed o such d y condi ions be o e, o
example he s oma al esis ance by Noilhan and Plan on (1989). Fo he g ass−si e
he TP e sion shows be e pe o mance wi hou subs an ially comp omising la en
hea luxes. Simila esul s a e ound o he g ass+side. The e o e his s udy no
only ag ees wi h p e ious wo k o e d y su aces (Yang e al., 2008, 2009; Chen e al.,
2010), bu shows ha he implemen ed scheme o calcula e he mal oughness is no
limi ed o d y su aces. Fu he mo e, he new TP e sion seems o be less sensi i e
o soil pa ame e s as i s pe o mance shows smalle di e ences be ween pa ame e se s
han he o iginal e sion.
The esul s ha e been c oss-checked wi h he g ound hea lux and impo an s a e
a iables as soil mois u e and su ace empe a u e. I could be shown ha he new
TP e sion p edic s he su ace empe a u e mo e accu a ely and is able o educe bias
o he g ound hea lux, e en i he sca e has been inc eased. The soil mois u e is
26
4.4. In luence o he ene gy balance co ec ion me hod
easonably esembled o g ass−wi h bo h pa ame e se . Fo de ails see Babel e al.
(2013, Appendix C).
As SEWAB pe o med well in gene al when being o ced wi h measu ed da a, i has
been deployed as a e e ence ime se ies o e alua e a new soil model inco po a ed in a
simplis ic land su ace model called “Hyb id” (Ge ken e al., 2012, Appendix B). The
new soil model has been in en ed o enhance he esponsi eness o he su ace in hyb id
showing a dis inc ime lag o he obse a ions in i s old e sion. As SEWAB did no
show such a ime lag, i s simula ions we e ideally sui ed o a c oss co ela ion analysis
wi h hyb id, whe e he la ge gaps in he obse ed da a complica e he in e p e a ion
o he esul s, see Ge ken e al. (2012, Appendix B) o de ail.
4.4. In luence o he ene gy balance co ec ion me hod
In he p e ious sec ion model e alua ions we e ca ied ou wi h EBC-Bo co ec ed
obse a ions only. In o de o highligh he ole o he ene gy balance closu e o
he e alua ion o land su ace models he new co ec ion me hod acco ding o he
buoyancy lux (EBC-HB) has been conside ed as well. Table 4.1 summa ises he change
in pe o mance wi h espec o (a) model pa ame e s, (b) model e sion, and (c) me hod
o ene gy balance closu e co ec ion.
Table 4.1a and b con i m he esul s gi en in he p e ious sec ion o EBC-Bo co -
ec ed obse a ions. In con as , EBC-HB co ec ed obse a ions indica e ha mea-
su ed pa ame e s pe o m now subs an ially be e a he g ass+si e. The same hap-
pens he e, o a less ex en , wi h espec o model e sion in case o sensible hea
al hough he posi i e e ec o he TP e sion p e ails in gene al. The eason o his
beha iou is a shi in bias o bo h u bulen luxes, as EBC-HB a ibu es a la ge
ac ion o he esidual o he sensible hea lux. The e o e he choice o he me hod
o close he ene gy balance has a s ong in luence on he decision on he “ igh ” model
pa ame e se o e sion. I should be no ed ha a la ge bias emains o he sum
o u bulen luxes (Fig. 4.2) which is in ac independen o he me hod o ene gy
balance co ec ion and mus be a ibu ed o o he easons.
Swi ching be ween co ec ion me hods (Table 4.1c) yields ambiguous esul s in bias
and NS wi h espec o model e sion and pa ame e s, bu shows ad an age o EBC-
Bo a g ass+and o EBC-HB a g ass−. The pa e n s a isi ics, howe e , o e ano he
pe spec i e: EBC-Bo yields subs an ially highe R2 o he sensible hea lux in any
case and lowe R2 o he la en hea lux a g ass+. The e o e he SEWAB model
is mo e compa ible wi h EBC-Bo, as exempla ily shown in Fig. 4.3. The simula ions
o sensible hea lux show mo e sca e wi h he EBC-HB co ec ed obse a ions han
o EBC-Bo. This migh be caused by an in insic model incompa ibili y o EBC-HB
o by p oblems in es ima ion o he ene gy balance closu e, inco po a ing addi ional
unce ain y in o he u bulen luxes. On he o he hand, unco ec ed obse a ions do
no lead o highe R2 o sensible hea han EBC-Bo co ec ed (no shown). The ed
27
4. Resul s
−−−−−−−−−−−−−−−−−−−−−−−−−
−−−−−−−−−−−−−−−−−−−−−−−−−
QH−EB in Wm−2
g ass−
0
100
200
300
400
−−−−−−−−−−−−−−−−−−−−−−
−−−
−−−−−−−−−−−−−−−−−−−−−−−−−
QE−EB in Wm−2
g ass−
−−−−−−−−−−−−−
−
−−
−−−−−−
−−−−−−−
−−−−
−−−−−−−−−−−
QH−EB in Wm−2
g ass+
0
100
200
300
400
−−−
−−−−
−−−
−
−−
−
−−
−−−−
−
−
−−−−−−−
−−−
−−
−
−−−
−−−
−−
−
QE−EB in Wm−2
g ass+
−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−
−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−
QH in Wm−2
lake
0000 0600 1200 1800
0
100
200
300
400
obs
sim
−−−−−−−
−−−−−
−
−−
−−−−−−−−−−
−−−
−−−−−−−
−
−−−−
−
−
−
−−
−−−−−
−−−−−−−−−−−−−−−−−−−−−−−−−−−−
−
−−−−
−
−−
QE in Wm−2
lake
0000 0600 1200 1800 2400
Figu e 4.6. Diu nal cycle o obse a ions and simula ions a Nam Co 2009. Land
su ace obse a ions ha e been co ec ed acco ding o he Bowen a io (EBC-Bo),
espec i e simula ions ha e been un in he TP e sion using he measu ed pa am-
e e s. Solid lines ep esen mean luxes while ho izon al ba s and g ey shaded a eas
deno e espec i e s anda d de ia ions o obse a ions and simula ions. F om Bie -
mann e al. (2013, Appendix D).
34

4.6. Flux he e ogenei y a Nam Co
g ass−
g ass+
lake
Mean luxes in Wm−2
0
50
100
150
200
250
−Rsw Rlw QGQHQE
Figu e 4.7. Mean luxes om he whole pe iod o he h ee su ace ypes om
obse a ion-based model uns as used in Figs. 4.5 and 4.6. Ne sho wa e adia ion
(Rsw) and ne longwa e adia ion (Rlw) a e pa ame ised as explained in Fig. 4.1.
Fo he land su ace ypes, QG ep esen s he g ound hea lux. In case o lake
su ace, QGdeno es he esidual o he ene gy balance, shown as a ha ched ba ,
including he ene gy luxes no accoun ed o , e.g. s o age change in he wa e body
and lux in o he sedimen . E o ba s indica e 1.96 imes he s anda d e o o he
mean, co esponding o he 95 % con idence in e al. F om Bie mann e al. (2013,
Appendix D).
35
4. Resul s
in compa a i ely high su ace empe a u es. Mean di e ences o sensible and la en
hea lux be ween g ass+ and g ass- a e 24.0 W m−2and −33.5 W m−2, espec i ely,
and be ween g ass+ and lake a e −27.3 W m−2and 22.3 W m−2, espec i ely.
36
5. Conclusions
The o e a ching goal o his hesis is o p og ess he es ima ion o u bulen luxes
o e ypical su ace ypes on he Tibe an Pla eau. Achie emen s ha e been made e-
la ed o eddy-co a iance obse a ions on he Tibe an Pla eau and si e-speci ic land
su ace modelling. An eddy-co a iance da a p ocessing scheme has been de eloped
(Babel e al., 2011a,b), majo ea u es ha e been al eady applied on da a o he Ti-
be an Pla eau (Zhou e al., 2011; Bie mann e al., 2013, Appendix D). Speci ic senso
p oblems wi h he Kaijo-Denki DAT600 TR61A p obe, ins alled a one o he si es
on he Tibe an Pla eau, ha e been encoun e ed wi h a sec o -wise plana - i o a ion
(Li e al., 2013, Appendix F). Fu he mo e, he land su ace scheme SEWAB has been
success ully adap ed o a d y and a we g assland si e o he monsoon season a Nam
Co, Tibe an Pla eau (Babel e al., 2013, Appendix C). Fi s eddy-co a iance measu e-
men s o e a lake su ace ha e been ga he ed and a hyd odynamical mul ilaye model
could esemble hese obse a ions e en esol ing he diu nal cycle (Bie mann e al.,
2013, Appendix D). F om hese achie emen s he ollowing conclusions can be d awn:
•Coo dina e o a ion o non-omnidi ec ional sonic anemome e need a ca e ul
handling, which applies o he Kaijo Denki DAT 600 TR61A p obe in pa ic-
ula and, o a less ex en , o he CSAT3 (Campbell Scien i ic L d.). Bo h in-
s umen show equen occu ences o physical implausible upwa d momen um
luxes, caused by senso dis o ion. I is ecommended o apply he plana - i o-
a ion o only he undis u bed wind sec o , and disca d momen um luxes mea-
su ed in o he wind sec o s. This can mi iga e such p oblems, and he ic ion
eloci y may de ia e up o 10 % om hose de i ed by a con en ional plana - i ,
likely explaining he di e ences be ween DAT 600 and CSAT3 ound in p e ious
s udies as well (Hong e al., 2004). In con as , no in luence o his p oblem on
scala luxes could be shown.
•The adap a ion o SEWAB o he Tibe an Pla eau (TP e sion) shows a sligh ,
bu o e all be e pe o mance han he o iginal e sion. The TP e sion leads
o bias educ ion especially o sensible hea on he d y su ace, which e en holds
when using ene gy balance co ec ed obse a ions acco ding o he buoyancy lux.
I pe o ms easonable also o we su aces al hough he esul s a e ambiguous
wi h espec o he used me hod o co ec he obse a ion o he ene gy balance
closu e gap. Ne e heless his demons a es ha he implemen ed modi ica ions,
including he z0h-scheme by Yang e al. (2008) a e no limi ed o d y o ba e
37
5. Conclusions
soil su aces and he small scale he e ogenei y in soil mois u e, an icipa ed by Su
e al. (2011), can be esol ed wi h his SEWAB e sion.
•When compa ing simula ions wi h eddy-co a iance de i ed u bulen luxes, he
pe o mance is s ongly a ec ed by he measu ed ene gy balance closu e gap.
Thus he selec ion o he su ace model and he choice o he EBC co ec ion
me hod a e in e - ela ed p oblems, which should ne e be o e looked when al-
ida ing a model wi h eddy-co a iance da a. K ache e al. (2009) could show,
ha SEWAB ep oduces he obse ed Bowen a io qui e well. This disse a ion
suppo s hei indings wi h da a om a o al di e en en i onmen . On he
o he hand, he EBC-Bo co ec ion may no be app op ia e due o poo scala
simila i y be ween sensible and la en hea lux o low equency con ibu ions,
as poin ed ou by Ruppe e al. (2006). This is con i med by Cha uchi ipan
e al. (2013, Appendix E), and he new co ec ion acco ding o he buoyancy
lux, EBC-HB, is p oposed in ag eemen wi h indings by e.g. Maude and Foken
(2006); Foken (2008a); Foken e al. (2011); Ingwe sen e al. (2011). The e o e
u u e model de elopmen o u bulen lux pa ame isa ion should ecognize ad-
ec i e luxes supplying seconda y ci cula ions as ecen hypo heses conce ning
he ene gy balance closu e a he han ying o ge he bes i o he unco -
ec ed eddy-co a iance da a. This is a challenging ask, and a meaning ul u u e
s udy o model s uc u e a leas equi es an indi idual pa ame isa ion o all en-
e gy lux componen s and no i em should simply se e as esidual o he ene gy
balance.
•The in eg a ion o he ene gy balance closu e and i s co ec ion demands a high
quali y measu emen s o all ene gy lux componen s. This may no always be
gi en o , in case o he lake su ace, only possible wi h la ge e o s. Especially he
g ound hea lux es ima ion (obse ed and simula ed) is p one o unce ain ies.
In he pa icula case o he g ass−si e a Nam Co p oblems a ise due o a la ge
g a el con en in he soil. Mo e gene al, unde d y condi ions on he Tibe an
Pla eau, a shallow d y uppe soil laye migh he mally decouple om he deepe
soil. Such a ea u e is usually no conside ed in hea lux pa ame isa ions.
•Si e-speci ic land su ace model simula ions based on eddy-co a iance obse a-
ions has been a ely conduc ed on he Tibe an Pla eau. F om his s udy ollows
ha some ield in es iga ions (addi ional o u bulen lux measu emen s and
o cing da a) a e ine i able o de i e a high quali y pa ame e se (measu ed o
de aul pa ame e s), which a e a leas soil mois u e measu emen s, soil empe -
a u e measu emen s and ield based knowledge o he soil ype, de i ed a leas
by a con en ional pedological desc ip ion. The land su ace model SEWAB, con-
s ained by such a da a se , is eliable enough o be used o alida ing mo e
simplis ic models as done by Ge ken e al. (2012, Appendix B).
38
•Tu bulen lux obse a ions and simula ions wi h he HM model pose a unique
da a se on he Tibe an Pla eau, as only s udies conduc ing bulk app oaches on
daily o mon hly basis we e epo ed up o now. Howe e , high quali y es ima es
a e necessa y since he Tibe an Pla eau is co e ed wi h a signi ican lake ac ion
o a ious sizes and he e o e di e en cha ac e is ics. The HM model wi h he
shallow wa e e m p o ed i s sui abili y o es ima e lake e apo a ion on a high
s anda d e en esol ing he diu nal cou se. App op ia e o cing can be achie ed
wi h land de i ed s anda d me eo ological measu emen s, a ep esen a i e su ace
empe a u e is he only measu emen equi ed di ec ly om he lake. Gi en hese
equi emen s, he model can be ans e ed o lakes wi h di e en size and dep h,
o example he la ge Nam Co lake, al hough mo e e o s ha e o be pu in
es ima ing a ep esen a i e su ace empe a u e.
•A he lake sho e si e (NamUBT) he simula ions o land and lake ha e been
in eg a ed by hei ela i e con ibu ion o he measu ed lux acco ding o he
oo p in o each ime s ep. Thus measu emen s and simula ions become compa-
able e en unde condi ions wi h mixed oo p in s. I could be shown, ha he
pe o mance does no de e io a e in such si ua ions and he ile app oach is alid
in his e ain o spa ial in eg a ion. The e o e ep esen a i e lux simula ions
on a g id cell wi h edge leng hs o 1 km o 5 km can be gi en o each ime s ep in
o de o compa e wi h emo e sensing da a. The simula ions o u bulen luxes
o g ass−, g ass+and lake di e in mean alues and empo al cha ac e is ics
beyond model unce ain y, wi h de ia ions occasionally exceeding 200 W m−2on
day ime. In con as , he measu emen s o e d y g assland a he Nam Co Moni-
o ing and Resea ch S a ion o Mul isphe e In e ac ions (g ass−) a e conside ed
o be a e e ence o he land su ace exchange in he Nam Co egion. This s udy
shows, ha he land use dis ibu ion wi hin he espec i e emo e sensing pixel o
g id cell o mesoscale modelling has o be ca e ully de e mined be o e alida ing
wi h he d y g assland s a ion. A po en ial ep esen a ion e o can be educed
by in eg a ing he simula ed luxes o adjacen land use ypes acco ding o hei
con ibu ion o he espec i e g id cell.
The essen ial indings in his hesis conce ning modelling unde he condi ions o he
Tibe an Pla eau, ene gy balance closu e and oo p in applica ions p o ide he basis
o p ocess he eddy-co a iance da a in h ee le els as desc ibed in Sec . 1.2. This is
exempla ily shown wi h he sensible hea lux om Nam Co s a ion in 2009 (Fig. 5.1).
Due o da a quali y il e ing, he le els I and II canno p o ide comple e diu nal cycles.
SEWAB simula ions om g ass− ill mos o he gaps in le el II da a wi h high eliabili y
conce ning magni ude and dynamics (Fig. 5.1c). Whene e o cing da a is missing o
simula ions, some gaps a e s ill le . In o de o ob ain no mo e han seasonal o
annual a e ages, emaining gaps can be elimina ed wi h me hods lis ed by Falge e al.
(2001a,b).
39

5. Conclusions
(a) Le el I: eddy−co a iance obse a ions
Feb Ma Ap May Jun Jul Aug Sep Oc No Dec
0
4
8
12
16
20
24
n.V.
−200
−100
−50
0
50
100
200
300
400
550
(b) Le el II: EBC−Bo co ec ed obse a ions
Feb Ma Ap May Jun Jul Aug Sep Oc No Dec
0
4
8
12
16
20
24
n.V.
−200
−100
−50
0
50
100
200
300
400
550
(c) Le el III: (b) gap illed wi h simula ions
Feb Ma Ap May Jun Jul Aug Sep Oc No Dec
0
4
8
12
16
20
24
n.V.
−200
−100
−50
0
50
100
200
300
400
550
Figu e 5.1. Da a example: Sensible hea lux a Nam Co, 2009, on di e en le els o
p ocessing as de ined o he CEOP-AEGIS p ojec (Sec . 1.2); (a), le el I: quali y
checked eddy-co a iance da a; (b), le el II: ene gy balance closu e co ec ed obse -
a ions acco ding o he Bowen a io, o nigh ime alues, when EBC co ec ion is
no applicable (see Sec . 3.1.4), le el I da a is accep ed unchanged; no e ha day ime
le el II da a canno be p o ided whene e an obse ed ene gy balance componen
is missing (c), le el III: da a om le el II, gap- illed wi h SEWAB simula ions o
g ass−. Remaining gaps indica e missing o cing da a.
40
Re e ences
Agam, N., Be line , P. R., Zang il, A., and Ben-Do , E.: Soil wa e e apo a ion
du ing he d y season in an a id zone, J. Geophys. Res., 109, D16 103–, doi:
10.1029/2004JD004802, 2004.
Alapa y, K., Pleim, J. E., Raman, S., Niyogi, D. S., and Byun, D. W.: Simula ion o
A mosphe ic Bounda y Laye P ocesses Using Local- and Nonlocal-Closu e Schemes,
J. Appl. Me eo ol., 36, 214–233, doi:10.1175/1520-0450(1997)036h0214:SOABLPi2.
0.CO;2, 1997.
Aubine , M., G elle, A., Ib om, A., Rannik, U., Monc ie , J., Foken, T., Kowalski,
A., Ma in, P., Be bigie , P., Be nho e , C., Clemen , R., Elbe s, J., G anie , A.,
G ¨unwald, T., Mo gens e n, K., Pilegaa d, K., Rebmann, C., Snijde s, W., Valen ini,
R., and Vesala, T.: Es ima es o he Annual Ne Ca bon and Wa e Exchange o
Fo es s: The EUROFLUX Me hodology, Ad . Ecol. Res., 30, 113–175, doi:10.1016/
S0065-2504(08)60018-5, 2000.
Babel, W., Eigenmann, R., Ma, Y., and Foken, T.: Analysis o u bulen luxes and
hei ep esen a i eness o he in e ac ion be ween he a mosphe ic bounda y laye
and he unde lying su ace on Tibe an Pla eau, CEOP-AEGIS Deli e able epo
De1.2, Ed. Uni e si y o S asbou g, F ance, ISSN 2118-7843: 35 p., 2011a.
Babel, W., Li, M., Sun, F., Ma, W., Chen, X., Colin, J., Ma, Y., and Foken, T.:
Ae odynamic and he modynamic a iables o ou s a ions on Tibe an Pla eau –
In oduc ion on he CEOP-AEGIS da abase in Ne CDF, CEOP-AEGIS Deli e able
epo De1.3, Ed. Uni e si y o S asbou g, F ance, ISSN 2118-7843: 90 p., 2011b.
Babel, W., Chen, Y., Bie mann, T., Yang, K., Ma, Y., and Foken, T.: Adap a ion o
a land su ace scheme o modeling u bulen luxes on he Tibe an Pla eau unde
di e en soil mois u e condi ions, submi ed o J. Geophys. Res., see Appendix C,
2013.
Baldocchi, D., Falge, E., Gu, L., Olson, R., Hollinge , D., Running, S., An honi, P.,
Be nho e , C., Da is, K., E ans, R., Fuen es, J., Golds ein, A., Ka ul, G., Law, B.,
Lee, X., Malhi, Y., Meye s, T., Munge , W., Oechel, W., Paw, K. T., Pilegaa d,
K., Schmid, H. P., Valen ini, R., Ve ma, S., Vesala, T., Wilson, K., and Wo sy, S.:
41
Re e ences
FLUXNET: A New Tool o S udy he Tempo al and Spa ial Va iabili y o Ecosys em-
Scale Ca bon Dioxide, Wa e Vapo , and Ene gy Flux Densi ies, B. Am. Me eo-
ol. Soc., 82, 2415–2434, doi:10.1175/1520-0477(2001)082h2415:FANTTSi2.3.CO;2,
2001.
Balsamo, G., Bousse a, S., Du a, E., Beljaa s, A., Vi e bo, P., and an den Hu k,
B.: E olu ion o land su ace p ocesses in he IFS, ECMWF Newsle ., 127, 17–22,
2011.
Bey ich, F., Leps, J.-P., Maude , M., Bange, J., Foken, T., Huneke, S., Lohse, H.,
L¨udi, A., Meijninge , W., Mi ono , D., Weisensee, U., and Zi el, P.: A ea-A e aged
Su ace Fluxes O e he Li ass Region Based on Eddy-Co a iance Measu emen s,
Bound.-Lay. Me eo ol., 121, 33–65, doi:10.1007/s10546-006-9052-x, 10.1007/s10546-
006-9052-x, 2006.
Bie mann, T., Babel, W., Olesch, J., and Foken, T.: Documen a ion o he Mic ome-
eo ological Expe imen , Nam Tso, Tibe , 25 h o June – 8 h o Augus 2009, Wo k
Repo Uni e si y o Bay eu h, Dep . o Mic ome eo ology, ISSN 1614-8916, 41,
38 pp., URL h p://opus.ub.uni-bay eu h.de/opus4-ubbay eu h/ on doo /
index/index/docId/626, 2009.
Bie mann, T., Babel, W., Ma, W., Chen, X., Thiem, E., Ma, Y., and Foken, T.:
Tu bulen lux obse a ions and modelling o e a shallow lake and a we g assland in
he Nam Co basin, Tibe an Pla eau, Theo . Appl. Clima ol., accep ed, see Appendix
D, 2013.
B ¨o z, B., Eigenmann, R., D¨o nb ack, A., Foken, T., and Wi h, V.: Ea ly-mo ning
low ansi ion in a alley in low-moun ain e ain, submi ed o Bound.-Lay. Me e-
o ol., 2013.
Cha uchi ipan, D., Babel, W., Maude , M., Leps, J.-P., and Foken, T.: Ex ension o
he a e aging ime o he eddy-co a iance measu emen and i s e ec on he ene gy
balance closu e, submi ed o Bound.-Lay. Me eo ol., see Appendix E, 2013.
Chen, T. H., Hende son-Selle s, A., Milly, P. C. D., Pi man, A. J., Beljaa s, A. C. M.,
Polche , J., Ab amopoulos, F., Boone, A., Chang, S., Chen, F., Dai, Y., Desbo ough,
C. E., Dickinson, R. E., D¨umenil, L., Ek, M., Ga a , J. R., Gedney, N., Guse ,
Y. M., Kim, J., Kos e , R., Kowalczyk, E. A., La al, K., Lean, J., Le enmaie ,
D., Liang, X., Mah ou , J.-F., Mengelkamp, H.-T., Mi chell, K., Nasono a, O. N.,
Noilhan, J., Robock, A., Rosenzweig, C., Schaake, J., Schlosse , C. A., Schulz, J.-
P., Shao, Y., Shmakin, A. B., Ve seghy, D. L., We zel, P., Wood, E. F., Xue, Y.,
Yang, Z.-L., and Zeng, Q.: Cabauw Expe imen al Resul s om he P ojec o In e -
compa ison o Land-Su ace Pa ame e iza ion Schemes, J. Clima e, 10, 1194–1215,
doi:10.1175/1520-0442(1997)010h1194:CERFTPi2.0.CO;2, 1997.
42
Re e ences
Chen, Y., Yang, K., Zhou, D., Qin, J., and Guo, X.: Imp o ing he Noah Land Su -
ace Model in A id Regions wi h an App op ia e Pa ame e iza ion o he The mal
Roughness Leng h, J. Hyd ome eo ol., 11, 995–1006, doi:10.1175/2010JHM1185.1,
2010.
Chen, Y., Yang, K., He, J., Qin, J., Shi, J., Du, J., and He, Q.: Imp o ing land su ace
empe a u e modeling o d y land o China, J. Geophys. Res., 116, D20 104–, doi:
10.1029/2011JD015921, 2011.
Chen, Y., Yang, K., Tang, W., Qin, J., and Zhao, L.: Pa ame e izing soil o ganic ca -
bon’s impac s on soil po osi y and he mal pa ame e s o Eas e n Tibe g asslands,
Sci. China Ea h Sci., 55, 1001–1011, doi:10.1007/s11430-012-4433-0, 2012.
Clapp, R. B. and Ho nbe ge , G. M.: Empi ical Equa ions o some Soil Hyd aulic
P ope ies, Wa e Resou . Res., 14, 601–604, 1978.
Dai, Y., Zeng, X., Dickinson, R. E., Bake , I., Bonan, G. B., Bosilo ich, M. G., Denning,
A. S., Di meye , P. A., House , P. R., Niu, G., Oleson, K. W., Schlosse , C. A., and
Yang, Z.-L.: The Common Land Model, B. Am. Me eo ol. Soc., 84, 1013–1023,
doi:10.1175/BAMS-84-8-1013, 2003.
Da idan, I., Lopa uhin, L., and Rogko , V.: Volny okeane (wa es in he ocean),
Gid ome eoizda , Lening ad, 256pp., 1985.
Duan, A. and Wu, G.: Role o he Tibe an Pla eau he mal o cing in he summe
clima e pa e ns o e sub opical Asia, Clim. Dynam., 24, 793–807, doi:10.1007/
s00382-004-0488-8, 2005.
Falge, E., Baldocchi, D., Olson, R., An honi, P., Aubine , M., Be nho e , C., Bu ba,
G., Ceulemans, R., Clemen , R., Dolman, H., G anie , A., G oss, P., G ¨unwald, T.,
Hollinge , D., Jensen, N.-O., Ka ul, G., Ke onen, P., Kowalski, A., Lai, C. T., Law,
B. E., Meye s, T., Monc ie , J., Moo s, E., Munge , J. W., Pilegaa d, K., Rannik,
U., Rebmann, C., Suyke , A., Tenhunen, J., Tu, K., Ve ma, S., Vesala, T., Wilson,
K., and Wo sy, S.: Gap illing s a egies o de ensible annual sums o ne ecosys em
exchange, Ag . Fo es Me eo ol., 107, 43–69, doi:10.1016/S0168-1923(00)00225-2,
2001a.
Falge, E., Baldocchi, D., Olson, R., An honi, P., Aubine , M., Be nho e , C., Bu ba,
G., Ceulemans, R., Clemen , R., Dolman, H., G anie , A., G oss, P., G ¨unwald, T.,
Hollinge , D., Jensen, N.-O., Ka ul, G., Ke onen, P., Kowalski, A., Lai, C. T., Law,
B. E., Meye s, T., Monc ie , J., Moo s, E., Munge , J. W., Pilegaa d, K., Rannik,
U., Rebmann, C., Suyke , A., Tenhunen, J., Tu, K., Ve ma, S., Vesala, T., Wilson,
K., and Wo sy, S.: Gap illing s a egies o long e m ene gy lux da a se s, Ag .
Fo es Me eo ol., 107, 71–77, doi:10.1016/S0168-1923(00)00235-5, 2001b.
43
Re e ences
Ni, J.: Impac s o clima e change on Chinese ecosys ems: key ulne able e-
gions and po en ial h esholds, Reg. En i on. Change, 11, 49–64, doi:10.1007/
s10113-010-0170-0, 2011.
Noilhan, J. and Plan on, S.: A Simple Pa ame e iza ion o Land Su ace P o-
cesses o Me eo ological Models, Mon. Wea he Re ., 117, 536–549, doi:10.1175/
1520-0493(1989)117h0536:ASPOLSi2.0.CO;2, 1989.
No dbo, A., Launiainen, S., Mamma ella, I., Lepp¨a an a, M., Huo a i, J., Ojala, A.,
and Vesala, T.: Long- e m ene gy lux measu emen s and ene gy balance o e a
small bo eal lake using eddy co a iance echnique, J. Geophys. Res., 116, D02 119,
doi:10.1029/2010JD014542, 2011.
Panin, G., Nasono , A., and Foken, T.: E apo a ion and hea exchange o a body o
wa e wi h he a mosphe e in a shallow zone, Iz . A mos. Ocean. Phys., 42, 337–352,
2006a.
Panin, G. N. and Foken, T.: Ai –sea in e ac ion including a shallow and coas al zone,
J. A mos. Ocean Sci., 10, 289–305, doi:10.1080/17417530600787227, 2005.
Panin, G. N., Nasono , A. E., Foken, T., and Lohse, H.: On he pa ame e isa ion o
e apo a ion and sensible hea exchange o shallow lakes, Theo . Appl. Clima ol.,
85, 123–129, doi:10.1007/s00704-005-0185-5, 2006b.
Pe e s-Lida d, C. D., Blackbu n, E., Liang, X., and Wood, E. F.: The E ec o
Soil The mal Conduc i i y Pa ame e iza ion on Su ace Ene gy Fluxes and Tem-
pe a u es, J. A mos. Sci., 55, 1209–1224, doi:10.1175/1520-0469(1998)055h1209:
TEOSTCi2.0.CO;2, 1998.
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.
Rannik, U., Aubine , M., Ku banmu ado , O., Sabel eld, K. K., Ma kkanen, T., and
Vesala, T.: Foo p in analysis o measu emen s o e a he e ogeneous o es , Bound.-
Lay. Me eo ol., 97, 137–166, doi:10.1023/A:1002702810929, 2000.
Rebmann, C., Kolle, O., Heinesch, B., Queck, R., Ib om, A., and Aubine , M.:
Da a Acquisi ion and Flux Calcula ions, in: Eddy Co a iance: A P ac ical Guide
o Measu emen and Da a Analysis, edi ed by Aubine , M., Vesala, T., and Pa-
pale, D., Sp inge A mosphe ic Sciences, pp. 59–83, Sp inge Ne he lands, doi:
10.1007/978-94-007-2351-1 3, 2012.
Rouse, W. R., Oswald, C. J., Binyamin, J., Spence, C., Sche ze , W. M., Blanken,
P. D., Bussi`e es, N., and Duguay, C. R.: The Role o No he n Lakes in a Regional
Ene gy Balance, J. Hyd ome eo ol., 6, 291–305, doi:doi:10.1175/JHM421.1, 2005.
50

Re e ences
Ruppe , J., Thomas, C., and Foken, T.: Scala Simila i y o Relaxed Eddy Accumu-
la ion Me hods, Bound.-Lay. Me eo ol., 120, 39–63, doi:10.1007/s10546-005-9043-3,
2006.
Scho anus, P., Nieuws ad , F., and De B uin, H.: Tempe a u e measu emen wi h
a sonic anemome e and i s applica ion o hea and mois u e luxes, Bound.-Lay.
Me eo ol., 26, 81–93, doi:10.1007/BF00164332, 1983.
Selle s, P., Randall, D., Colla z, G., Be y, J., Field, C., Dazlich, D., Zhang, C.,
Collelo, G., and Bounoua, L.: A Re ised Land Su ace Pa ame e iza ion (SiB2)
o A mosphe ic GCMs. Pa I: Model Fo mula ion, J. Clima e, 9, 676–705, doi:
10.1175/1520-0442(1996)009h0676:ARLSPFi2.0.CO;2, 1996.
Sene i a ne, S. I., Co i, T., Da in, E. L., Hi schi, M., Jaege , E. B., Lehne , I.,
O lowsky, B., and Teuling, A. J.: In es iga ing soil mois u e–clima e in e ac ions in
a changing clima e: A e iew, Ea h-sci. Re ., 99, 125–161, doi:10.1016/j.ea sci e .
2010.02.004, 2010.
S ein eld, G., Le zel, M., Raasch, S., Kanda, M., and Inagaki, A.: Spa ial ep e-
sen a i eness o single owe measu emen s and he imbalance p oblem wi h eddy-
co a iance luxes: esul s o a la ge-eddy simula ion s udy, Bound.-Lay. Me eo ol.,
123, 77–98, doi:10.1007/s10546-006-9133-x, 2007.
S eppele , J., Doms, G., Scha le , U., Bi ze , H. W., Gassmann, A., Dam a h, U.,
and G ego ic, G.: Meso-gamma scale o ecas s using he nonhyd os a ic model LM,
Me eo ol. A mos. Phys., 82, 75–96, 2003.
S oy, P. C., Maude , M., Foken, T., Ma colla, B., Boegh, E., Ib om, A., A ain, M. A.,
A ne h, A., Au ela, M., Be nho e , C., Cesca i, A., Dellwik, E., Duce, P., Gi-
anelle, D., an Go sel, E., Kiely, G., Knohl, A., Ma golis, H., McCaughey, H.,
Me bold, L., Mon agnani, L., Papale, D., Reichs ein, M., Saunde s, M., Se ano-
O iz, P., So oco nola, M., Spano, D., Vacca i, F., and Va lagin, A.: A da a-
d i en analysis o ene gy balance closu e ac oss FLUXNET esea ch si es: The
ole o landscape scale he e ogenei y, Ag . Fo es Me eo ol., 171–172, 137–152, doi:
10.1016/j.ag o me .2012.11.004, 2013.
Su, Z., Wen, J., Den e, L., an de Velde, R., Wang, L., Ma, Y., Yang, K., and Hu, Z.:
The Tibe an Pla eau obse a o y o pla eau scale soil mois u e and soil empe a u e
(Tibe -Obs) o quan i ying unce ain ies in coa se esolu ion sa elli e and model
p oduc s, Hyd ol. Ea h Sys . Sci., 15, 2303–2316, doi:10.5194/hess-15-2303-2011,
2011.
Takayabu, I., Taka a, K., Yamazaki, T., Ueno, K., Yabuki, H., and Haginoya, S.: Com-
pa ison o he Fou Land Su ace Models D i en by a Common Fo cing Da a P e-
51
Re e ences
pa ed om GAME/Tibe POP’97 P oduc s – Snow Accumula ion and Soil F eezing
P ocesses –, J. Me eo ol. Soc. Jpn., 79, 535–554, doi:10.2151/jmsj.79.535, 2001.
Thiem, E.: Modeling o he ene gy exchange abo e lake and land su aces, Mas e
hesis, Uni e si y o Bay eu h, 73pp., 2011.
Twine, T. E., Kus as, W. P., No man, J. M., Cook, D. R., House , P. R., Meye s,
T. P., P uege , J. H., S a ks, P. J., and Wesely, M. L.: Co ec ing eddy-co a iance
lux unde es ima es o e a g assland, Ag . Fo es Me eo ol., 103, 279–300, doi:10.
1016/S0168-1923(00)00123-4, 2000.
an de Velde, R., Su, Z., Ek, M., Rodell, M., and Ma, Y.: In luence o he modynamic
soil and ege a ion pa ame e iza ions on he simula ion o soil empe a u e s a es
and su ace luxes by he Noah LSM o e a Tibe an Pla eau si e, Hyd ol. Ea h Sys .
Sci., 13, 759–777, doi:10.5194/hess-13-759-2009, 2009.
Wallace, J., W igh , I., S ewa , J., and Holwill, C.: The Sahelian Ene gy Bal-
ance Expe imen (SEBEX): G ound based measu emen s and hei po en ial o
spa ial ex apola ion using sa elli e da a, Ad . Space Res., 11, 131–141, doi:
10.1016/0273-1177(91)90413-E, 1991.
Wa ach, K., Mengelkamp, H.-T., and Raschke, E.: T ea men o ozen soil and snow
co e in he land su ace model SEWAB, Theo . Appl. Clima ol., 69, 23–37, doi:
10.1007/s007040170033, 2001.
Wendisch, M. and Foken, T.: Sensi i i y es o a mul ilaye ene gy exchange model,
Z. Me eo ol., 39, 36–39, 1989.
Wilson, K., Golds ein, A., Falge, E., Aubine , M., Baldocchi, D., Be bigie , P., Be n-
ho e , C., Ceulemans, R., Dolman, H., Field, C., G elle, A., Ib om, A., Law, B. E.,
Kowalski, A., Meye s, T., Monc ie , J., Monson, R., Oechel, W., Tenhunen, J.,
Valen ini, R., and Ve ma, S.: Ene gy balance closu e a FLUXNET si es, Ag . Fo -
es Me eo ol., 113, 223–243, 2002.
Xu, J. and Haginoya, S.: An Es ima ion o Hea and Wa e Balances in he Tibe an
Pla eau, J. Me eo ol. Soc. Jpn., 79, 485–504, doi:10.2151/jmsj.79.485, 2001.
Xu, J., Yu, S., Liu, J., Haginoya, S., Ishigooka, Y., Kuwaga a, T., Ha a, M., and
Yasuna i, T.: The Implica ion o Hea and Wa e Balance Changes in a Lake Basin
on he Tibe an Pla eau, Hyd ol. Res. Le ., 3, 1–5, doi:10.3178/h l.3.1, 2009.
Yanai, M. H., Li, C. F., and Song, Z. S.: Seasonal Hea ing o he Tibe an Pla eau and
i s E ec s on he E olu ion o he Asian Summe Monsoon, J. Me eo ol. Soc. Jpn.,
70, 319–351, 1992.
52
Re e ences
Yang, K. and Wang, J.: A empe a u e p edic ion-co ec ion me hod o es ima ing
su ace soil hea lux om soil empe a u e and mois u e da a, Sci. China Se . D,
51, 721–729, doi:10.1007/s11430-008-0036-1, 2008.
Yang, K., Koike, T., and Yang, D.: Su ace lux pa ame e iza ion in he Tibe an
Pla eau, Bound.-Lay. Me eo ol., 116, 245–262, doi:10.1023/A:1021152407334, 2003.
Yang, K., Koike, T., Ye, B., and Bas idas, L.: In e se analysis o he ole o soil e ical
he e ogenei y in con olling su ace soil s a e and ene gy pa i ion, J. Geophys. Res.,
110, D08101, doi:10.1029/2004JD005500, 2005.
Yang, K., Koike, T., Ishikawa, H., Kim, J., Li, X., Liu, H., Wang, J., Liu, S., and Ma,
Y.: Tu bulen Flux T ans e o e Ba e-Soil Su aces: Cha ac e is ics and Pa am-
e e iza ion, J. Appl. Me eo ol. Clim., 47, 276–290, doi:10.1175/2007JAMC1547.1,
2008.
Yang, K., Chen, Y.-Y., and Qin, J.: Some p ac ical no es on he land su ace
modeling in he Tibe an Pla eau, Hyd ol. Ea h Sys . Sci., 13, 687–701, doi:
10.5194/hess-13-687-2009, 2009.
Yang, K., Guo, X., and Wu, B.: Recen ends in su ace sensible hea lux on
he Tibe an Pla eau, SCIENCE CHINA Ea h Sciences, pp. 1–10, doi:10.1007/
s11430-010-4036-6, 2010.
Ye, D.-Z. and Wu, G.-X.: The ole o he hea sou ce o he Tibe an Pla eau in he
gene al ci cula ion, Me eo ol. A mos. Phys., 67, 181–198, doi:10.1007/BF01277509,
1998.
Yu, S., Liu, J., Xu, J., and Wang, H.: E apo a ion and ene gy balance es ima es
o e a la ge inland lake in he Tibe -Himalaya, En i on. Ea h Sci., 64, 1169–1176,
doi:10.1007/s12665-011-0933-z, 2011.
Zhang, X.: Imp o emen o a soil-a mosphe e- ans e model o he simula ion o ba e
soil su ace ene gy balances in semia id a eas, Asia Pac. J. A mos. Sci., 48, 97–105,
doi:10.1007/s13143-012-0009-3, 2012.
Zhou, D., Eigenmann, R., Babel, W., Foken, T., and Ma, Y.: The s udy o nea -g ound
ee con ec ion condi ions a Nam Co s a ion on he Tibe an Pla eau, Theo . Appl.
Clima ol., 105, 217–228, doi:10.1007/s00704-010-0393-5, 2011.
Zili inke ich, S. S.: Non-local u bulen anspo : Pollu ion dispe sion aspec s o
cohe en s uc u e o con ec i e lows, in: Ai Pollu ion – Volume I. Ai Pollu ion
Theo y and Simula ion, edi ed by Powe , H., Moussiopoulos, N., and B ebbia, C. A.,
pp. 53–60, Compu a ional Mechanics Publica ions, 1995.
53
A. Indi idual con ibu ions o he
join publica ions
This cumula i e hesis consis s o publica ions and manusc ip s lis ed he ea e . O he
au ho s con ibu ed o hese pape s as well. The e o e my own con ibu ion o he
indi idual manusc ip s is speci ied in his sec ion.
Nam Co expe imen 2009 and da a p epa a ion
Some o he ollowing publica ions is based on he Nam Co expe imen in 2009, u -
he desc ibed in sec ion 3.1. Toge he wi h Tobias Bie mann I was esponsible o
he ealisa ion o he expe imen . The se up and design was planned and p epa ed
oge he wi h Tobias Bie mann in equal sha es. We bo h assembled he measu emen
complex a he si e, while Tobias Bie mann alone was esponsible o da a collec ion
and main enance du ing he whole campaign.
The da a om he Nam Co expe imen (NamUBT) as well as da a p o ided om he
pe manen s a ion o he ITP (NamITP) ha e been pos -p ocessed as ollows: Tobias
Bie mann did quali y checks on low equency da a and u bulen lux p ocessing wi h
he TK2/TK3 so wa e package, including sec o -wise plana - i o NamUBT. I mysel
elabo a ed he ac ual land use dis ibu ion and con ibu ed wi h he oo p in analysis
as well as he calcula ion o he g ound hea lux and ene gy balance, including he
EBC co ec ion me hods.
Appendix B
Ge ken, T., Babel, W., Ho mann, A., Bie mann, T., He zog, M., F iend, A. D.,
Li, M., Ma, Y., Foken, T., and G a , H.-F.: Tu bulen lux modelling wi h a
simple 2-laye soil model and ex apola ed su ace empe a u e applied a Nam
Co Lake basin on he Tibe an Pla eau, Hyd ol. Ea h Sys . Sci., 16, 1095–1110,
doi:10.5194/hess-16-1095-2012, 2012.
•Tobias Ge ken de eloped he idea o he manusc ip and coo dina ed indi idual
con ibu ions. He conduc ed he Hyb id modelling and he analysis. He w o e
he whole publica ion and ac ed as co esponding au ho .
54
•I p o ided he NamUBT da a om he Nam Co expe imen oge he wi h Tobias
Bie mann (see p. 54). I mysel conduc ed SEWAB simula ions o ITP and UBT
land su ace and ga he ed espec i e model pa ame e s.
•Alex Ho mann con ibu ed o he Hyb id model de elopmen ele an o his
pape .
•Michael He zog con ibu ed wi h echnical ad ice on he manusc ip .
•And ew F iend is he o iginal au ho o Hyb id. The p o ided assis ance wi h
he handling o Hyb id.
•Li Maoshan and Ma Yaoming p o ided da a om he ITP s a ion and suppo ed
he ield ip.
•Thomas Foken and Hans-F. G a con ibu ed o he manusc ip a a ious s ages
wi h ui ul discussions.
Appendix C
Babel, W., Chen, Y., Bie mann, T., Yang, K., Ma, Y., and Foken, T.: Adap a ion
o a land su ace scheme o modeling u bulen luxes on he Tibe an Pla eau
unde di e en soil mois u e condi ions, submi ed o J. Geophys. Res.
•I mysel de eloped he idea o his manusc ip . I was me who adap ed he
SEWAB code, ga he ed he model pa ame e s and calcula ed all simula ions.
Toge he wi h Tobias Bie mann I p o ided he NamUBT da a om he Nam Co
expe imen (p. 54). I alone conduc ed he whole s a is ical analysis and w o e he
manusc ip . Finally I ac as co esponding au ho o he submi ed manusc ip .
•Chen Yingying p o ided sou ce code o he oughness leng h pa ame isa ion o
he TP e sion and o he g ound hea lux calcula ion a NamITP s a ion. He
conduc ed labo a o y measu emen s o soil samples he ook o e a ious places
on he TP, which I used o de i e he measu ed pa ame e s.
•Yang Kun was he de elope o he new oughness leng h pa ame isa ion and
he new soil he mal conduc i i y calcula ion. Toge he wi h him I had ui ul
discussions abou model simula ion esul s.
•Ma Yaoming p o ided da a om he ITP s a ion and ga e suppo du ing he
ield ip.
•As my supe iso Thomas Foken con ibu ed wi h scien i ic discussions h ough-
ou all s ages o analysis and manusc ip p epa a ion
55

A. Indi idual con ibu ions o he join publica ions
Appendix D
Bie mann, T., Babel, W., Ma, W., Chen, X., Thiem, E., Ma, Y., and Foken, T.:
Tu bulen lux obse a ions and modelling o e a shallow lake and a we g assland
in he Nam Co basin, Tibe an Pla eau, Theo . Appl. Clima ol., accep ed
•The idea o his manusc ip was de eloped in equal pa s by Tobias Bie mann
and me. We bo h ealised he Nam Co expe imen and da a p epa a ion (p. 54)
and w o e he ex in equal sha es. While Tobias Bie mann ocused mo e on
li e a u e and scope o he manusc ip as well as he expe imen al pa , I pu
my emphasis on da a analysis and he modelling pa s. Tobias Bie mann ac s as
co esponding au ho o he submi ed manusc ip .
•Ma Yaoming, Chen Xuelong and Ma Weiqiang suppo ed us du ing he ield ip
in 2009 and ga e access o da a om he ITP s a ion
•Elisabe h Thiem con ibu ed o he manusc ip ’s wo kload wi hin a mas e hesis
unde he supe ision o me and Thomas Foken. The eby she illed gaps in he
model o cing da a, implemen ed he shallow wa e e m in o he lake model HM,
and conduc ed a p elimina y sensi i i y analysis o he lake model.
•Thomas Foken ac ed as supe iso and libe ally sha ed his expe ience in encou -
aging and ui ul discussions.
Appendix E
Cha uchi ipan, D., Babel, W., Maude , M., Leps, J.-P., and Foken, T.: Ex ension
o he a e aging ime o he eddy-co a iance measu emen and i s e ec on he
ene gy balance closu e, submi ed o Bound.-Lay. Me eo ol.
•Doojdao Cha uchi ipan conduc ed he whole da a analysis and w o e he ex o
he manusc ip . She ac s as co esponding au ho o he submi ed manusc ip .
•I de eloped he p oposed new ene gy balance closu e co ec ion algo i hm o-
ge he wi h DC and ealised he espec i e pa in he conclusions.
•Ma hias Maude ini ia ed he use o wa ele s o analyse he long- e m luc ua-
ions.
•Jens-Pe e Leps p o ided da a om se e al LITFASS-2003 s a ions and gene a ed
land-use classi ied da a.
•Thomas Foken encou aged he s uc u e o he manusc ip and con ibu ed wi h
many scien i ic discussions. Fu he mo e, he ini ia ed he p ojec ela ed o he
manusc ip .
56
Appendix F
Li, M., Babel, W., Tanaka, K., and Foken, T.: No e on he applica ion o plana - i
o a ion o non-omnidi ec ional sonic anemome e s, A mos. Meas. Tech., 6,
221–229, doi:10.5194/am -6-221-2013, 2013.
•Li Maoshan was in cha ge o he da a used in his publica ion. He did he da a
p epa a ion and analysis acco ding o my ins uc ions. Fu he mo e he w o e he
desc ip ion o BJ s a ion.
•I mysel coo dina ed he analysis and w o e he main ex passages. Finally I
ac ed as he co esponding au ho .
•Kenji Tanaka conduc ed i s in es iga ions o i egula ic ion occu ing o he
Kaijo Denki DAT 600 a BJ si e. He did his wo k in Bay eu h du ing a sabba -
ical, his esul s a e no published ye .
•Due o his expe ience wi h he Kaijo Denki DAT 600 o e decades, Thomas Foken
ini ia ed he in es iga ion on his opic and con ibu es o i s p og ess wi h se e al
scien i ic discussions.
57
B. Ge ken e al. (2012)
Ge ken, T., Babel, W., Ho mann, A., Bie mann, T., He zog, M., F iend, A. D., Li,
M., Ma, Y., Foken, T., and G a , H.-F.: Tu bulen lux modelling wi h a simple 2-laye
soil model and ex apola ed su ace empe a u e applied a Nam Co Lake basin on he
Tibe an Pla eau, Hyd ol. Ea h Sys . Sci., 16, 1095–1110, doi:10.5194/hess-16-1095-
2012, 2012.
58
Hyd ol. Ea h Sys . Sci., 16, 1095–1110, 2012
www.hyd ol-ea h-sys -sci.ne /16/1095/2012/
doi:10.5194/hess-16-1095-2012
© Au ho (s) 2012. CC A ibu ion 3.0 License.
Hyd ology and
Ea h Sys em
Sciences
Tu bulen lux modelling wi h a simple 2-laye soil model and
ex apola ed su ace empe a u e applied a Nam Co Lake
basin on he Tibe an Pla eau
T. Ge ken1,2, W. Babel2, A. Ho mann1, T. Bie mann2, M. He zog1, A. D. F iend3, M. Li4, Y. Ma5, T. Foken2,6, and
H.-F. G a 1
1Cen e o A mosphe ic Science, Depa men o Geog aphy, Uni e si y o Camb idge, UK
2Depa men o Mic ome eo ology, Uni e si y o Bay eu h, Ge many
3Depa men o Geog aphy, Uni e si y o Camb idge, UK
4Cold and A id Region En i onmen al and Enginee ing Resea ch Ins i u e, Chinese Academy o Sciences, Lanzhou, China
5Ins i u e o Tibe an Pla eau Resea ch, Chinese Academy o Sciences, Beijing, China
6Membe o Bay eu h Cen e o Ecology and En i onmen al Resea ch (BayCEER), Uni e si y o Bay eu h, Ge many
Co espondence o: T. Ge ken ([email p o ec ed])
Recei ed: 19 Oc obe 2011 – Published in Hyd ol. Ea h Sys . Sci. Discuss.: 21 No embe 2011
Re ised: 8 Ma ch 2012 – Accep ed: 27 Ma ch 2012 – Published: 3 Ap il 2012
Abs ac . This pape in oduces a su ace model wi h wo
soil-laye s o use in a high- esolu ion ci cula ion model ha
has been modi ied wi h an ex apola ed su ace empe a u e,
o be used o he calcula ion o u bulen luxes. A quad a ic
empe a u e p o ile based on he laye mean and base em-
pe a u e is assumed in each laye and ex ended o he su -
ace. The model is es ed a wo si es on he Tibe an Pla eau
nea Nam Co Lake du ing ou days du ing he 2009 Mon-
soon season. In compa ison o a wo-laye model wi hou ex-
plici su ace empe a u e es ima e, he e is a g ea ly educed
delay in diu nal lux cycles and he modelled su ace em-
pe a u e is much close o obse a ions. Compa ison wi h
a SVAT model and eddy co a iance measu emen s shows
an o e all easonable model pe o mance based on RMSD
and c oss co ela ion compa isons be ween he modi ied and
o iginal model. A po en ial limi a ion o he model is he
need o ca e ul ini ialisa ion o he ini ial soil empe a u e
p o ile, ha equi es ield measu emen s. We show ha he
modi ied model is capable o ep oducing luxes o simila
magni udes and dynamics when compa ed o mo e complex
me hods chosen as a e e ence.
1 In oduc ion
Tu bulen luxes o momen um, la en hea (QE) and sensi-
ble hea (QH) a e some o he mos impo an in e ac ions
be ween land su ace and a mosphe e. These luxes a e e-
sponsible o he de elopmen o modi ica ion o mesoscale
ci cula ions and he gene a ion o clouds eed back on su -
ace luxes h ough he modi ica ion o sola adia ion. The
e ec s o ege a ion in luencing bounda y laye s uc u e
and mois u e a e widely acknowledged (i.e. F eedman e al.,
2001; an Hee waa den e al., 2009), while he eedback
om sho -li ed clouds is less unde s ood, bu impo an .
Shallow cumulus-su ace in e ac ions we e shown in an LES
(la ge eddy simula ion) s udy o impac su ace empe a-
u e and luxes on e y sho ime scales (Lohou and Pa on,
2011). Fo imp o ed p ocess unde s anding, i is necessa y
o use: (1) a mosphe ic models wi h su icien ly high esolu-
ion (O(100m)) o esol e bounda y laye p ocesses as well
as clouds and (2) su ace models capable o ep oducing he
sys em’s su ace lux dynamics.
Ou esea ch ocuses on su ace-a mosphe e in e ac ions
on he Tibe an Pla eau (TP) in he Nam Co Lake egion.
Wi h mo e han 4700ma.s.l., a semi-a id clima e and wi h a
highly adap ed Kob esia pygmea alpine s eppe (Miehe e al.,
2011), he TP p o es o be a di icul en i onmen o su -
ace models (Yang e al., 2003, 2009). Speci ic p oblems
Published by Cope nicus Publica ions on behal o he Eu opean Geosciences Union.
59
1102 T. Ge ken e al.: A simple wo-laye soil model wi h ex apola ed su ace empe a u e
0 4 8 12 16 20
0
10
20
30
40
50
60
70
80
T2 [°C]
Tbase1
[°C]
a)
Tbase1 = Tbase2 + a2 d22
−30
−20
−10
0
10
20
30
40
50
b) T0 [°C]
Tbase1
[°C]
T1 [°C]
0 4 8 12 16 20
0
4
8
12
16
20
0 4 8 12 16 20
0
4
8
12
16
20
−100
−70
−70
−70
−40 −40 −40
−10 −10 −10
0
00
0
20 20
30 30
40
c) T0 [°C]
T1 [°C]
T2 [°C]
Fig. 5. Dependency o soil empe a u e pa ame e s: (a) ela ion-
ship be ween mean empe a u e o laye 2 ( ¯
T2) and bo om em-
pe a u e o laye 1 [Tbase1, Eq. (7)] – a2is calcula ed acco ding
o Eq. (6); (b) su ace empe a u e (T0) con ou plo as unc ion o
Tbase1and laye 1 mean empe a u e ( ¯
T1) and (c) con ou s o (T0) as
unc ion o ¯
T2and ¯
T1. The black ec angle a he in e sec ion o he
laye empe a u e anges (yellow) indica es he heo e ical pa ame-
e space gi en by he empe a u e alues used in his s udy and he
black c osses ma k he ac ual con igu a ions.
4.2 S a is ical e alua ion measu es
Model quali y was assessed by Roo Mean Squa e De ia ion
(RMSD)
RMSD =
u
u
1
N
N
X
i=1Pp−P 2
i(12)
and C oss Co ela ion acco ding o he coe icien o
de e mina ion (R2):
R2(j) = co Pp(1+j:N),P (1:N−j)
σPp(1+j:N)σP (1:N−j) !2
(13)
wi h R2(j) being he coe icien s o de e mina ion o he
p edic ed (Pp) and e e ence (P ) lux ime se ies shi ed by
jelemen s, he o al numbe o elemen s in each ime se-
ies (N) and σas hei espec i e s anda d de ia ions. Bo h
SEWAB and EC measu emen s p oduce 30-min lux a e -
ages, whe eas Hyb id was se o 10-min a e aged luxes.
The e o e he e e ence luxes we e linea ly in e pola ed o
Hyb id’s ou pu imes be o e s a is ical e alua ion. Pe iods
when no ene gy balance co ec ed EC measu emen s we e
a ailable (see Figs. 6and7 o de ails) we e excluded om
he calcula ion o he s a is ical measu es.
5 Resul s and discussion
The ollowing sec ion p esen s and discusses he imp o e-
men s ha a e achie ed o a simple wo-laye model when a
new algo i hm o he su ace empe a u e was implemen ed.
The o iginal wo-laye model Hyb id ails o ep oduce he
diu nal dynamics obse ed a UBT (Figs. 6and7) due o he
he mal ine ia o he op-laye . The delayed esponse in su -
ace empe a u e leads o a shi in he esul ing u bulen su -
ace luxes. This causes an unde es ima ion o QEand QH
un il ∼18:00BST and la e o an o e es ima ion due o de-
layed su ace cooling. The imp o emen o he modi ied Hy-
b id o e he o iginal o mula ion is discussed in mo e de ail
in Sec s. 5.1 and 5.4.
The la en (Fig. 6 – le column) and sensible hea luxes
( igh column) es ima ed wi h he modi ied Hyb id model a e
gene ally in good ag eemen wi h he e e ence luxes de-
i ed by EC and SEWAB. The diagnos ic su ace empe a-
u e ( igh column) also shows a close ag eemen . In some
ins ances he e emains a small shi in luxes compa ed o
he e e ence alues, bu his has been g ea ly imp o ed com-
pa ed o he o iginal Hyb id. The su ace empe a u es a e
also in good ag eemen a e sun ise, despi e he ac ha
du ing he clea sky days in Augus excessi e nigh - ime su -
ace cooling is simula ed. This is less o an issue du ing he
o e cas nigh s.
Hyd ol. Ea h Sys . Sci., 16, 1095–1110, 2012 www.hyd ol-ea h-sys -sci.ne /16/1095/2012/
B. Ge ken e al. (2012)
66

T. Ge ken e al.: A simple wo-laye soil model wi h ex apola ed su ace empe a u e 1103
0
100
200
300
400
500
600 a) QE UBT
10−Jul−2009
Tu bulen Flux [W m−2]
03:00 06:00 09:00 12:00 15:00 18:00 21:00
−100
0
100
200
300
400
500
600 b) QH UBT
−10
0
10
20
30
40
50
60
T0 [°C]
03:00 06:00 09:00 12:00 15:00 18:00 21:00
0
100
200
300
400
500
600 c) QE UBT
27−Jul−2009
−100
0
100
200
300
400
500
600 d) QH UBT
−10
0
10
20
30
40
50
60
0
100
200
300
400
500
600 e) QE UBT
05−Aug−2009
LHyb,new
LHyb,o g
WCOARE
LEC
LEC,EBC
LSEWAB
−100
0
100
200
300
400
500
600 ) QH UBT
−10
0
10
20
30
40
50
60
LHyb,new
LHyb,o g
WCOARE
LEC
LEC,EBC
LSEWAB
03:00 06:00 09:00 12:00 15:00 18:00 21:00
0
100
200
300
400
500
600 g) QE UBT
06−Aug−2009
Time BST
L+WEC
WEC
WHM
−100
0
100
200
300
400
500
600 h) QH UBT
−10
0
10
20
30
40
50
60
L+WEC
WEC
WHM
T0,Hyb
T0,Re
03:00 06:00 09:00 12:00 15:00 18:00 21:00
Time BST
Fig. 6. Model esul s o he modi ied Hyb id a UBT o 10 July 2009 (a–b), 27 July 2009 (c–d), 5 Augus 2009 (e– ) and 6 Augus 2009
(g–h). Le column: la en hea lux (QE); igh column: sensible hea lux (QH) and su ace empe a u e T0[◦C]. Land W e e o “land”
and “wa e ” as o igin o he luxes. L+W is he comple e a ailable ime se ies. The subsc ip s Hyb,mod and Hyb,o g e e o luxes om he
modi ied and o iginal Hyb id and COARE a e luxes om he lake de i ed by TOGA-COARE whe eas SEWAB is a SVAT model and HM
e e s o a hyd odynamic mul i-laye lake model a e Foken (1984) and Panin e al. (2006). EC and EC,EBC e e o measu emen s by eddy
co a iance me hod whe e in he la e he ene gy balance has been closed by dis ibu ing he esidual acco ding o Bowen- a io ( his equi es
good da a quali y and luxes and can only be done o luxes ha a e a ibu ed o land). The ci cles indica e poo da a quali y o he EC
sys em acco ding o Foken e al. (2004). G ay shading indica es imes whe e he lux oo p in o UBT was o e he lake.
The si ua ion a ITP is qui e simila o UBT. The modi-
ied model ag ees well wi h he EC and SEWAB e e ence
da a. On 5 Augus he u bulen lux dynamics, bu no he
magni ude o he luxes, ma ch he EC measu emen s closely
(Fig. 7), while he o iginal Hyb id showed a s ong delay
in he lux esponse as he soil emained ozen du ing he
mo ning. While he magni ude o he la en hea lux is close
o EC measu emen s, QHp oduced by Hyb id a e o a sim-
ila magni ude as QH om SEWAB. These a e conside ably
la ge han he luxes measu ed by EC and co ec ed o en-
e gy balance closu e. Fo 6 Augus he modelled maximum
o QEis la ge han he maximum QEEC,EBC and much g ea e
h oughou mos o he day compa ed o SEWAB. QHin con-
as shows simila diu nal dynamics as QHEC,EBC, bu wi h i s
magni ude be ween he sensible hea lux de i ed by SEWAB
and QHEC,EBC. A ound 18:00h he QH- luxes om he di e -
en me hods become mo e simila . A la ge nega i e QH- lux
in he mo ning hou s is appa en bu g ea ly imp o ed com-
pa ed o he unmodi ied Hyb id e sion. Figu e 6a and b
also highligh s some limi a ions o ecosys em esea ch as a
la ge po ion o he da a had o be ejec ed due o limi a ions
desc ibed in Sec . 4.1.
Du ing lake b eeze e en s he su ace luxes o e wa e
de i ed om TOGA-COARE a e displayed. Sensible hea
luxes a e in close ag eemen wi h EC da a and luxes de-
i ed by a hyd odynamic mul i-laye lake model (Foken,
1984; Panin e al., 2006). La en hea luxes show a sim-
ila beha iou and a e o simila magni ude on 10 July
www.hyd ol-ea h-sys -sci.ne /16/1095/2012/ Hyd ol. Ea h Sys . Sci., 16, 1095–1110, 2012
67
1104 T. Ge ken e al.: A simple wo-laye soil model wi h ex apola ed su ace empe a u e
0
100
200
300
400
500
600 a) QE ITP
10−Jul−2009
Tu bulen Flux [W m−2]
03:00 06:00 09:00 12:00 15:00 18:00 21:00
−100
0
100
200
300
400
500
600 b) QH ITP
−10
0
10
20
30
40
50
60
T0 [°C]
03:00 06:00 09:00 12:00 15:00 18:00 21:00
0
100
200
300
400
500
600 c) QE ITP
27−Jul−2009
−100
0
100
200
300
400
500
600 d) QH ITP
−10
0
10
20
30
40
50
60
0
100
200
300
400
500
600 e) QE ITP
05−Aug−2009
−100
0
100
200
300
400
500
600 ) QH ITP
−10
0
10
20
30
40
50
60
03:00 06:00 09:00 12:00 15:00 18:00 21:00
0
100
200
300
400
500
600 g) QE ITP
06−Aug−2009
Time BST
LHyb,new
LHyb,o g
LEC
LEC,EBC
LSEWAB
−100
0
100
200
300
400
500
600 h) QH ITP
−10
0
10
20
30
40
50
60
LHyb,new
LHyb,o g
LEC
LEC,EBC
LSEWAB
T0,Hyb
T0,Re
03:00 06:00 09:00 12:00 15:00 18:00 21:00
Time BST
Fig. 7. Same as Fig. 6, bu o ITP. The e a e no con ibu ions om he lake.
and 6 Augus . On 5 Augus he e is a leas a quali a i e
ag eemen be ween COARE and EC measu emen s.
5.1 Discussion o u bulen luxes
The o iginal wo laye model eac s only slowly o he a mo-
sphe ic o cing, delaying he luxes’ esponse. Such a ime
lag leads o a shi in he diu nal cycle and is p oblema ic o
he coupling o a mosphe ic models since su ace luxes a e
one o he main d i e s o egional and local ci cula ion as
well as cloud de elopmen . These will ce ainly be a ec ed
by e oneous su ace lux dynamics. In ou speci ic case, he
dampening o he diu nal empe a u e cycle and he delay
in su ace luxes may educe he in ensi y o he land-lake
b eeze o may delay i s de elopmen h ough a educ ion o
di e en ial hea ing be ween land and lake su ace. Howe e ,
he e is s ill a mino delay isible in he modi ied Hyb id as
he su ace empe a u e is pu ely diagnos ic and dependen
on ¯
T1. This is discussed in mo e de ail in Sec . 5.4.
Table 3 shows he esul s o he RMSD be ween he mod-
elled esul s and he e e ence quan i ies. Wi h he modi ied
Hyb idmodel he e isa 40–60%imp o emen in heRMSDs
compa ed o he o iginal Hyb id, when bo h a e compa ed
agains SEWAB. The only no able excep ion o his is 6 Au-
gus a ITP, whe e a s ong de ia ion o u bulen luxes de-
i ed by SEWAB and measu ed luxes was encoun e ed. This
is due o an unde es ima ion o soil wa e con en by SEWAB
as 6 Augus alls in o a d y in e al be ween ainy pe iods,
whe e SEWAB unde es ima es he soil wa e con en . The
pic u e is mo e di e se o he compa ison be ween he en-
e gy balance co ec ed EC luxes and Hyb id. The e is a e-
duc ion in he e o o all cases, excep QHon 6 Augus a
ITP, bu he educ ions co e a much la ge ange om less
han 1 o 80%. Due o da a quali y conce ns he numbe o
compa able elemen s is much lowe (Ngi en in Table 3) and
p obably oo small o meaning ul s a is ics in case o UBT.
This is especially ue as he day ime lake b eeze in luence
coincides wi h he imes wi h pe iods o usually highe qual-
i y o EC luxes. As lux quali ies a e usually lowe du ing
condi ions wi h limi ed e ical exchange (s able s a i ica-
ion), EC luxes a ITP mainly e lec he day ime model pe -
o mance whe eas he compa ison wi h SEWAB also akes
Hyd ol. Ea h Sys . Sci., 16, 1095–1110, 2012 www.hyd ol-ea h-sys -sci.ne /16/1095/2012/
B. Ge ken e al. (2012)
68
T. Ge ken e al.: A simple wo-laye soil model wi h ex apola ed su ace empe a u e 1105
01234
0
0.25
0.5
0.75
1 a) Hyb−EC Qh(UBT)
R2( )
10−Jul
27−Jul
05−Aug
06−Aug
01234
0
0.25
0.5
0.75
1
c) Hyb−EC QE(UBT) o ig
mod
01234
0
0.25
0.5
0.75
1 e) Hyb−SEWAB QH(UBT)
01234
0
0.25
0.5
0.75
1 g) Hyb−SEWAB QE(UBT)
lag [h]
01234
0
0.25
0.5
0.75
1 b) Hyb−EC QH(ITP)
01234
0
0.25
0.5
0.75
1 d) Hyb−EC QE(ITP)
01234
0
0.25
0.5
0.75
1 ) Hyb−SEWAB QH(ITP)
01234
0
0.25
0.5
0.75
1 h) Hyb−SEWAB QE(ITP)
lag [h]
Fig. 8. C oss co ela ion R2( ) o simula ed luxes agains lux e e ence shi ed by lag as mul iples o 10 minu es o each o he ou days
simula ed wi h he o iginal and mody ied Hyb id. The maximum numbe o elemen s used in he calcula ion o R2 o each cu e can be
aken om Table 3.
in o accoun he nigh - ime, whe e luxes and he e o e ab-
solu e di e ences a e smalle . The small imp o emen o
RMSD o QHand QHEC,EBC a ITP can be explained by he
ac ha he modi ied Hyb id ollows he dynamics o EC, bu
lux es ima es a e la ge and o he same magni ude as luxes
calcula ed by SEWAB. Maude e al. (2006) ha e es ima ed
he e o o EC measu emen s o be 5% o <10Wm2 o
QHand 15% o <30Wm−2 o QE. Addi ional unce ain y
is added o he measu ed luxes by he lack in ene gy balance
closu e. When his is aken in o accoun he e is a signi i-
can di e ence be ween he QHHyb id and QHEC,EBC o ITP on
6 Augus . On 5 Augus (ITP) and 6 Augus (UBT) he de i-
a ion o luxes may s ill be explained by measu emen e o s
and by sho comings in he ene gy balance closu e scheme.
Indeed, he e is no indica ion o assume scala simila i y be-
ween empe a u e and mois u e anspo (Ruppe e al.,
www.hyd ol-ea h-sys -sci.ne /16/1095/2012/ Hyd ol. Ea h Sys . Sci., 16, 1095–1110, 2012
69
1106 T. Ge ken e al.: A simple wo-laye soil model wi h ex apola ed su ace empe a u e
Table 3. Roo mean squa e de ia ion (RMSD) be ween he modelled quan i ies o he o iginal and modi ied Hyb id and e e ence alues.
The e e ence quan i ies used a e ei he measu ed by EC and co ec ed o ene gy balance closu e (EC,EBC) o modelled wi h SEWAB
o luxes o aken om longwa e ou going adia ion o T0. The alues in pa en hesis (N) co espond o he numbe o elemen s used o
calcula ion o RMSD and R2(l =0)in Fig. 8.
RMSD
Si e Da e Run QEQHQEQHT0
EC,EBC [Wm−2] SEWAB [Wm−2] [◦C]
UBT
10 July
o ig
318 117 (8) 94 74 (94) 4.3 (139)
27 July 97 58 (19) 60 59 (139) 4.5 (143)
5 Augus 168 139 (11) 90 64 (110) 4.3 (143)
6 Augus 159 84 (52) 87 71 (128) 3.7 (143)
ITP
10 July
o ig
182 93 (25) 97 69 (143) 3.7 (143)
27 July 43 64 (72) 58 75 (143) 3.8 (143)
5 Augus 224 103 (64) 179 68 (143) 8.3 (143)
6 Augus 118 80 (52) 130 119 (143) 5.1 (143)
UBT
10 July
mod
214 43 (8) 51 36 (94) 2.3 (139)
27 July 79 44 (19) 32 28 (139) 2.9 (143)
5 Augus 93 62 (11) 36 26 (110) 3.4 (143)
6 Augus 78 57 (52) 39 32 (128) 3.2 (143)
ITP
10 July
mod
74 73 (25) 42 32 (143) 1.6 (143)
27 July 42 58 (72) 55 36 (143) 2.6 (143)
5 Augus 44 80 (64) 64 30 (143) 2.6 (143)
6 Augus 68 82 (52) 113 77 (143) 3.5 (143)
UBT all o ig 170 92 (90) 83 67 (471) 4.2 (568)
mod 100 54 (90) 39 31 (471) 3.0 (568)
ITP all o ig 152 84 (213) 125 86 (572) 5.6 (572)
mod 54 73 (213) 74 48 (572) 2.7 (572)
2006; Maude e al., 2007). The e o e, addi ional esea ch,
such as high- esolu ion a mosphe ic modelling s udies, need
o be ca ied ou in o de o de e mine he con ibu ions o
QHand QE o he “missing” ene gy. I should be no ed
ha all modelled luxes and measu emen s ha e e o s, so
ha he e is no absolu e way o knowing which me hod p o-
duces he bes lux es ima es. The inco po a ion o su ace
luxes in o a egional ci cula ion model may gi e some in-
sigh in o whe he modelled su ace a mosphe e in e ac ions
lead o ealis ic a mosphe ic low pa e ns.
The la ge nega i e and po en ially un easonable nigh - ime
QH- luxes ha a e modelled o ITP on 6 Augus a e owed o
a ozen soil and s ong su ace winds ha lead o an o e es-
ima ion o he empe a u e g adien , delayed eac ion o he
su ace model and esul ed in a po en ial unde es ima ion o
modelled su ace empe a u es and hus su ace luxes.
5.2 Discussion o su ace empe a u e
Fo su ace empe a u e he e is a no able dec ease in RMSD
o all cases. Addi ionally, he sou ce o he e o changes.
In he o iginal model he e o in T0was mainly due o he
ime-lag and a gene al unde es ima ion o day ime maximum
su ace empe a u es. In he new model day ime T0ma ches
a lo be e wi h obse a ions excep o ITP 6 Augus , whe e
e apo a i e cooling due o excessi e e apo anspi a ion con-
ibu es o oo small wa ming a es. In e u n, he cooling
du ing he nigh ime is o e es ima ed. This may ei he be due
o e o s in soil mois u e, su ace emissi i y () o due o he
su ace empe a u e ex apola ion unc ion used in his wo k.
5.3 Soil mois u e a ia ion and e apo anspi a ion
A e a 24h un he mois u e con en o he i s model laye
(las wo columns o Table 2) is smalle han measu emen s
sugges . Fo UBT, measu ed soil mois u e con en does
ha dly a y on a day o day scale and is kep well abo e
FC due o g oundwa e in luence. This is no e lec ed by
he model as i lacks he capabili y o include g oundwa e
ables. The ue soil mois u e a ITP has a much la ge a i-
a ion due o i s low FC and compa a i ely la ge po e ol-
ume. Du ing he d y day o 27 July he uppe soil laye loses
1.5mm o wa e whe eas du ing he mois days o Augus
he e is a o al loss om laye one o 6.7 and 5.3mmd−1, e-
spec i ely. Compa ing θ1end o 5 Augus wi h θ1obs o he nex
day shows ha he model would pe o m conside ably wo se
Hyd ol. Ea h Sys . Sci., 16, 1095–1110, 2012 www.hyd ol-ea h-sys -sci.ne /16/1095/2012/
B. Ge ken e al. (2012)
70
T. Ge ken e al.: A simple wo-laye soil model wi h ex apola ed su ace empe a u e 1107
i i we e no es a ed e e y day. This is caused by a e y
limi ed soil hyd ology included in Hyb id. Hu e al. (2008)
ha e es ima ed he summe e apo anspi a ion on a cen al
Tibe an g assland si e o be in he o de o 4–6mmd−1. An
expe imen conduc ed wi hin he amewo k o TiP has es i-
ma ed ba e soil e apo a ion and e apo anspi a ion o a e y
d y soil a Kema in 2010 (∼150km no heas o Nam Co
Lake) a 2mmd−1 ising o a leas 6mmd−1and possibly
mo e o a ege a ed Kob esia pas u e du ing an i iga ion
expe imen (H. Cone s – Uni e si y o G¨
o ingen, pe sonal
communica ion, 27 June 2011). E en hough he soils a e
no di ec ly compa able his sugges s simila dynamics in QE
o he ITP si e. One ac o likely o play a ole in he local
wa e cycle ha is no included is dew all in he ea ly mo n-
ing hou s. Di ec abso p ion o a mosphe ic mois u e on ba e
soil (Agam (Nina i) and Be line , 2004) and dew all a e o -
en conside ed a signi ican mois u e inpu o semi-a id en-
i onmen s (Agam and Be line , 2006). Hea y dew all in
he icini y o Nam Co Lake is equen ly obse ed, bu has,
a leas o ou knowledge, ne e been quan i ied. This addi-
ional sou ce o wa e and he associa ed local ecycling o
wa e may accoun o a signi ican ac ion o he missing
wa e . In addi ion o his, he oo simplis ic ep esen a ion
o soil hyd ology is e y likely esponsible o he emaining
wa e de ici in he uppe laye o he soil model.
5.4 C oss co ela ion o u bulen luxes
A di e en way o looking a he model pe o mance is c oss
co ela ion o he modelled su ace luxes agains EC mea-
su emen s and SEWAB (Fig. 8). These measu es gi e an
insigh in o he easons o he delayed esponse o he su -
ace model and he amoun o lux- a iance explained, bu
does no yield in o ma ion whe he he model and he e -
e ence luxes show a ue one- o-one co ela ion. As wi h
RMSD he quali y o he analysis is limi ed by he numbe
o da a poin s ha can be co ela ed, which is compa a i ely
small o he ene gy balance co ec ed EC measu emen s a
ITP and e en smalle a UBT due o lake b eeze in luences
(Fig. 8a–e). Hence, i is e y di icul o in e p e he c oss
co ela ions o EC. I is p obably ai o say ha he e is a
endency o smalle ime lags du ing he ime se ies wi h
highe numbe o elemen s, no ably UBT 6 Augus and all
days o ITP and ha he o al explained a iances a e a he
same le el o de e mina ion, when compa ing he maximum
R2(j). A no able excep ion is ITP 5 Augus .
Fo he compa ison wi h SEWAB (Fig. 8 –h), i becomes
no able ha o many cases he maximum R2(j) o he modi-
ied Hyb id app oach R2→1 and ha hei maxima a e usu-
ally ound a lags o 10–30min (j=1−3). Sola adia ion
apidly modi ies he skin empe a u e ha is go e ning u -
bulen luxes. As SEWAB has an ins an aneous su ace em-
pe a u e sol e o each model ime s ep, one would expec a
di ec esponse o SEWAB o changes in sola adia ion. This
may e en be as e han in eali y, especially o QE lux ha
is no only dependen on he ac ual skin empe a u e, bu also
on he ege a ion’s esponse. Including nega i e alues o j
in o Fig. 8 would show a g adual dec ease o co ela ions
wi h dec easing j, showing ha he lux dynamics o Hyb id
ne e p ecede EC measu emen s o SEWAB.
5.5 Na u al a iabili y o luxes
A mosphe ic quan i ies and u bulen su ace luxes ha e a
la ge na u al a iabili y ha is di icul o measu e o o
model. The EC app oach is dependen on a e aging p oce-
du es and mos s anda d measu emen s will yield mean al-
ues. In o de o use high- equency measu emen s o lux
es ima ion, less common echniques such as condi ional sam-
pling o wa ele -spec a ha e o be used. E en i models
a e capable o ep oducing a iabili y on ealis ic scales i is
di icul o supply o cing da a wi h simila esolu ion. The
o cing da a used in his s udy, sampled and a e aged 10 o
30min means, a e used o SEWAB. Running Hyb id a ime
s eps compa able o a high- esolu ion mesoscale model e-
qui es in e pola ion o he o cing da a and he e o e po en-
ially causes a smoo hing o he model’s esponse compa ed
o he ac ual wea he o cing as i would be p o ided by a
coupled model. As su ace models sha e a simila app oach
o he pa ame e isa ion o su ace luxes and close he su ace
ene gy balance locally, SEWAB and Hyb id luxes a e mo e
simila o each o he han hey a e o ield measu emen s.
6 Conclusions
The accu a e gene a ion o su ace luxes is a necessa y p e-
equisi e o s udies o su ace-a mosphe e in e ac ions and
local o mesoscale ci cula ions. In o de o gain a be e p o-
cess unde s anding o he in e ac ion be ween a mosphe ic
ci cula ion, clouds, adia ion and su ace luxes, he gene -
a ed diu nal lux cycles ha e o be o ealis ic magni ude and
wi hou empo al shi . The o iginal wo-laye su ace model
wi hou a speci ic o mula ion o T0p oduced bo h a consid-
e able ime lag and ailed o cap u e he ull diu nal dynamics
due o i s un esponsi eness.
We ha e demons a ed ha he in oduc ion o an ex apo-
la ed su ace empe a u e enables e en a qui e simplis ic soil
model o ealis ically simula e skin empe a u es and hus o
gene a e mo e ealis ic su ace luxes. The delay o luxes
du ing he daily cycle was g ea ly educed, making he model
usable o diu nal p ocess s udies. The o al magni ude o
luxes is also much imp o ed, when ew and compu a ionally
cheap addi ional physically based p ocesses a e in oduced.
Compa ing SEWAB wi h Hyb id, he RMSD o bo h luxes
and su ace empe a u e is dec eased by gene ally 40–60%.
The imp o emen in quali y was somewha mo e a ied in
compa ison o EC measu emen s, as compa ison o models
and measu emen s is no s aigh o wa d. The imp o ed
R2(j) o smalle alues o jshows ha empo al shi s o
www.hyd ol-ea h-sys -sci.ne /16/1095/2012/ Hyd ol. Ea h Sys . Sci., 16, 1095–1110, 2012
71

1108 T. Ge ken e al.: A simple wo-laye soil model wi h ex apola ed su ace empe a u e
he lux ime se ies ha e been g ea ly educed and he o e -
all co ela ions a e high. As wi h any na u al sys em i is
impossible o ob ain comple e da a se s ha cap u e he ull
amoun o na u al a iabili y. Howe e , he modi ied model
has been es ed o a la ge spec um o en i onmen al condi-
ions on he TP and p oduced easonable esul s o bo h d y
and mois condi ions.
We ha e shown ha a a he simple soil su ace model can
e icien ly calcula e u bulen luxes a a high empo al eso-
lu ion when d i en by ealis ic a mosphe ic condi ions. Ne -
e heless, i is qui e clea ha such an app oach wi h ex apo-
la ed su ace empe a u e needs ca e ul model ini ialisa ion.
The ini ial soil hea con en s and he e o e knowledge o soil
empe a u e p o iles is necessa y. Due o he ac ha he
su ace empe a u e in his s udy is a pu ely diagnos ic quan-
i y, he e may s ill be some limi a ions such as a delayed
o smoo hed esponse o a mosphe ic o cing on e y sho
imescales, such as he eedback be ween passing bounda y-
laye clouds and he su ace luxes. The in luence o su -
ace luxes and hei dynamics o egional ci cula ion will be
in es iga ed in a u u e s udy.
Acknowledgemen s. This esea ch was unded by he Ge man
Resea ch Founda ion (DFG) P io i y P og amme 1372 “Tibe an
Pla eau: Fo ma ion, Clima e, Ecosys ems” as pa o he A mo-
sphe e – Ecology – Glaciology Clus e (TiP-AEG). ITP da a was
p o ided h ough CEOP-AEGIS, which is a EU-FP7 Collabo a i e
P ojec / Small o medium-scale ocused esea ch p ojec – Speci ic
In e na ional Co-ope a ion Ac ion coo dina ed by he Uni e si y o
S asbou g, F ance unde he call ENV.2007.4.1.4.2: “Imp o ing
obse ing sys ems o wa e esou ce managemen .” The au ho s
would like o hank e e yone who has con ibu ed o he collec ion
o ield da a in one o he wo ld’s mos emo e egions. ADF
acknowledges suppo om he Eu opean Communi y’s Se en h
F amewo k P og amme (FP7/2007-2013) unde g an ag eemen
no. 238366. AH acknowledges unding by he Fonds Na ional de
la Reche che (FNR-Luxembou g), unde he g an BFR07-089 and
suppo by he Camb idge Eu opean T us (CET-UK).
Edi ed by: C. de Michele
Re e ences
Agam (Nina i), N. and Be line , P. R.: Diu nal Wa e
Con en Changes in he Ba e soil o a coas al dese ,
J. Hyd ome eo ol., 5, 922–933, doi:10.1175/1525-
7541(2004)005<0922:DWCCIT>2.0.CO;2, 2004.
Agam, N. and Be line , P. R.: Dew o ma ion and wa e apo ad-
so p ion in semi-a id en i onmen s – A e iew, J. A id En i on.,
65, 572–590, doi:10.1016/j.ja iden .2005.09.004, 2006.
Aubine , M., G elle, A., Ib om, A., Rannik, U., Monc ie , J., Fo-
ken, T., Kowalski, A., Ma in, P., Be bigie , P., Be nho e , C.,
Clemen , R., Elbe s, J., G anie , A., G ¨
unwald, T., Mo gen-
s e n, K., Pilegaa d, K., Rebmann, C., Snijde s, W., Valen ini,
R., Vesala, T., Fi e , A., and Ra aelli, D.: Es ima es o he an-
nual ne ca bon and wa e exchange o o es s: The EUROFLUX
Me hodology, Ad . Ecol.Res., 30, 113–175, doi:10.1016/S0065-
2504(08)60018-5, 1999.
Bie mann, T., Babel, W., Olesch, J., and Foken, T.: Mesoscale ci -
cula ions and ene gy and gas exchange o e he Tibe an Pla eau
– Documen a ion o he mic ome eo ological expe imen , Nam
Tso, Tibe – 25 June–8 Augus 2009, A bei se gebnisse 41, Uni-
e si y o Bay eu h, ISSN 1614-8616, Bay eu h, 2009.
Blackada , A.: High esolu ion models o he plane a y bounda y
laye , in: Ad ances in En i onmen al Science and Enginee ing,
edi ed by: P a lin, J. and Ziegle , E., Vol. 1, 50–85, Go don and
B each, New Yo k, 1979.
Camillo, P. J. and Gu ney, R. J.: A esis ance pa ame e o ba e-soil
e apo a ion models, Soil Sci., 141, 95–105, 1986.
Cong, Z., Kang, S., Smi no , A., and Holben, B.: Ae osol op ical
p ope ies a Nam Co, a emo e si e in cen al Tibe an Pla eau,
A mos. Res., 92, 42–48, doi:10.1016/j.a mos es.2008.08.005,
2009.
Cui, X., Langmann, B., and G a , H.: Summe monsoonal ain all
simula ion on he Tibe an Pla eau wi h a egional clima e model
using a one-way double-nes ing sys em, SOLA, 3, 49–52, 2007.
Dea do , J. W.: Dependence o ai -sea ans e coe icien s on bulk
s abili y, J. Geoph. Res., 73, 2549–2557, 1968.
Fai all, C. W., B adley, E. F., God ey, J. S., Wick, G. A., Edson,
J. B., and Young, G. S.: Cool-skin and wa m-laye e ec s on sea
su ace empe a u e, J. Geophys. Res., 101, 1295–1308, 1996a.
Fai all, C. W., B adley, E. F., Roge s, D. P., Edson, J. B., and
Young, G. S.: Bulk pa ame e iza ion o ai -sea luxes o T opi-
cal Ocean-Global A mosphe e Coupled-Ocean A mosphe e Re-
sponse Expe imen , J. Geophy. Res., 101, 3747–3764, 1996b.
Foken, T.: The pa ame e isa ion o he ene gy exchange ac oss
he ai -sea in e ace, Dynam. A mos. Ocean, 8, 297–305,
doi:10.1016/0377-0265(84)90014-9, 1984.
Foken, T.: The ene gy balance closu e p oblem: An o e iew, Ecol.
Appl., 18, 1351–1367, 2008.
Foken, T., G¨
ockede, M., Maude , M., Mah , L., Ami o, B., and
Munge , J.: Pos ield da a quali y con ol, in: Handbook o
Mic ome eo ology: A guide o su ace lux measu emen s and
analysis, edi ed by: Lee, X., Massman, W., and Law, B., 181–
208, Kluwe , Do d ech , 2004.
Foken, T., Aubine , M., Finnigan, J. J., Lecle c, M. Y., Maude , M.,
and Paw U, K. T.: Resul s O A Panel Discussion Abou The En-
e gy Balance Closu e Co ec ion Fo T ace Gases, B. Am. Me-
eo ol. Soc., 92, ES13–ES18, doi:10.1175/2011BAMS3130.1,
2011.
F eedman, J. M., Fi zja ald, D. R., Moo e, K. E., and Sakai,
R. K.: Bounda y laye clouds and ege a ion-a mosphe e
eedbacks, J. Clima e, 14, 180–197, doi:10.1175/1520-
0442(2001)013<0180:BLCAVA>2.0.CO;2, 2001.
F iend, A. D.: Te es ial plan p oduc ion and clima e change, J.
Exp. Bo ., 61, 1293–1309, doi:10.1093/jxb/e q019, 2010.
F iend, A. D. and Kiang, N. Y.: Land su ace model de elopmen
o he GISS GCM: e ec s o imp o ed canopy physiology on
simula ed clima e, J. Clima e, 18, 2883–2902, 2005.
F iend, A. D., S e ens, A. K., Knox, R. G., and Cannell, M. G. R.:
A p ocess-based, e es ial biosphe e model o ecosys em dy-
namics (Hyb id 3.0), Ecol. Model., 95, 249–287, 1997.
Hansen, J., Russell, G., Rind, D., S one, P., Lacis, A., Lebede ,
S., Ruedy, R., and T a is, L.: E icien h ee-dimensional global
models o clima e s udies: Models I and II, Mon. Wea he Re .,
Hyd ol. Ea h Sys . Sci., 16, 1095–1110, 2012 www.hyd ol-ea h-sys -sci.ne /16/1095/2012/
B. Ge ken e al. (2012)
72
T. Ge ken e al.: A simple wo-laye soil model wi h ex apola ed su ace empe a u e 1109
111, 609–662, 1983.
He zog, M., G a , H., Tex o , C., and Obe hube , J. M.: The e -
ec o phase changes o wa e on he de elopmen o olcanic
plumes, J. Volcanol. Geo h. Res., 87, 55–74, doi:10.1016/S0377-
0273(98)00100-0, 1998.
Hu, Z., Yu, G., Fu, Y., Sun, X., Li, Y., Shi, P., Wang, Y., and
Zheng, Z.: E ec s o ege a ion con ol on ecosys em wa e
use e iciency wi hin and among ou g assland ecosys ems in
China, Glob. Change Biol., 14, 1609–1619, doi:10.1111/j.1365-
2486.2008.01582.x, 2008.
Hu, Z., Yu, G., Zhou, Y., Sun, X., Li, Y., Shi, P., Wang, Y., Song, X.,
Zheng, Z., Zhang, L., and Li, S.: Pa i ioning o e apo anspi a-
ion and i s con ols in ou g assland ecosys ems: Applica ion
o a wo-sou ce model, Ag ic. Fo . Me eo ol., 149, 1410–1420,
doi:10.1016/j.ag o me .2009.03.014, 2009.
Kanda, M., Inagaki, A., Le zel, M. O., Raasch, S., and Wa an-
abe, T.: LES S udy o he ene gy imbalance p oblem wi h
eddy co a iance luxes, Bound.-Lay. Me eo ol., 110, 381–404,
doi:10.1023/B:BOUN.0000007225.45548.7a, 2004.
Keil, A., Be king, J., M¨
ugle , I., Sch¨
u , B., Schwalb, A., and S eeb,
P.: Hyd ological and geomo phological basin and ca chmen
cha ac e is ics o Lake Nam Co, Sou h-Cen al Tibe , Qua . In .,
218, 118–130, 2010.
K ache , D., Mengelkamp, H., and Foken, T.: The esidual
o he ene gy balance closu e and i s in luence on he e-
sul s o h ee SVAT models, Me eo ologische Z., 18, 647–661,
doi:10.1127/0941-2948/2009/0412, 2009.
Kuwaga a, T., Kondo, J., and Sumioka, M.: The mal e ec o
he sea b eeze on he s uc u e o he bounda y laye and he
hea budge o e land, Bound.-Lay. Me eo ol., 67, 119–144,
doi:10.1007/BF00705510, 1994.
Li, M., Ma, Y., Hu, Z., Ishikawa, H., and Oku, Y.: Snow dis i-
bu ion o e he Namco lake a ea o he Tibe an Pla eau, Hy-
d ol. Ea hSys .Sci., 13, 2023–2030, doi:10.5194/hess-13-2023-
2009, 2009.
Lohou, F. and Pa on, E. G.: Land-su ace esponse o shal-
low cumulus, EGU Gene al Assembly, Vienna, Aus ia, 3–
8 Ap il 2011, EGU2011-10280-1, 2011.
Ma, Y., Wang, Y., Wu, R., Hu, Z., Yang, K., Li, M., Ma, W., Zhong,
L., Sun, F., Chen, X., Zhu, Z., Wang, S., and Ishikawa, H.: Re-
cen ad ances on he s udy o a mosphe e-land in e ac ion ob-
se a ions on he Tibe an Pla eau, Hyd ol. Ea h Sys . Sci., 13,
1103–1111, doi:10.5194/hess-13-1103-2009, 2009.
Maude , M. and Foken, T.: Documen a ion and ins uc ion man-
ual o he eddy co a iance so wa e package TK3, A bei se geb-
nisse 46, Uni e si y o Bay eu h, ISSN 1614-8916, Bay eu h,
2011.
Maude , M., Oncley, S. P., Vog , R., Weidinge , T., Ribei o,
L., Be nho e , C., Foken, T., Kohsiek, W., B uin, H. A. R.,
and Liu, H.: The ene gy balance expe imen EBEX-2000. Pa
II: In e compa ison o eddy-co a iance senso s and pos - ield
da a p ocessing me hods, Bound.-Lay. Me eo ol., 123, 29–54,
doi:10.1007/s10546-006-9139-4, 2006.
Maude , M., Desja dins, R. L., and MacPhe son, I.: Scale anal-
ysis o ai bo ne lux measu emen s o e he e ogeneous e -
ain in a bo eal ecosys em, J. Geophys. Res., 112, D13112,
doi:10.1029/2006JD008133, 2007.
Maude , M., Foken, T., Clemen , R., Elbe s, J. A., Eugs e ,
W., G ¨
unwald, T., Heusink eld, B., and Kolle, O.: Quali y
con ol o Ca boEu ope lux da a – Pa 2: In e -compa ison
o eddy-co a iance so wa e, Biogeosciences, 5, 451–462,
doi:10.5194/bg-5-451-2008, 2008.
Mengelkamp, H., Wa ach, K., and Raschke, E.: SEWAB – a pa-
ame e iza ion o he su ace ene gy and wa e balance o a mo-
sphe ic and hyd ologic models, Ad . Wa . Resou ., 23, 165–175,
doi:10.1016/S0309-1708(99)00020-2, 1999.
Me zge , S., Ma, Y., Ma kkanen, T., G¨
ockede, M., Li, M., and Fo-
ken, T.: Quali y assessmen o Tibe an Pla eau Eddy co a iance
measu emen s u ilizing oo p in modeling, Ad . Ea h Sci., 21,
1260–1267, 2006.
Miehe, G., Miehe, S., Bach, K., N¨
olling, J., Hanspach, J., Reuden-
bach, C., Kaise , K., Wesche, K., Mosb ugge , V., Yang, Y., and
Ma, Y.: Plan communi ies o cen al Tibe an pas u es in he
Alpine S eppe/Kob esia pygmaea eco one, J. A id En i on., 75,
711–723, doi:10.1016/j.ja iden .2011.03.001, 2011.
Obe hube , J. M., He zog, M., G a , H., and Schwanke, K.: Vol-
canic plume simula ion on la ge scales, J. Volcanol. Geo h. Res.,
87, 29–53, doi:10.1016/S0377-0273(98)00099-7, 1998.
Panin, G. N., Te zla , G., and Raabe, A.: Inhomogenei y o he
land su ace and p oblems in he pa ame e iza ion o su ace
luxes in na u al condi ions, Theo . Appl. Clima ol., 60, 163–178,
doi:10.1007/s007040050041, 1998.
Panin, G. N., Nasono , A. E., and Foken, T.: E apo a ion and
hea exchange o a body o wa e wi h he a mosphe e in a
shallow zone, Iz es iya, A mos. Ocean. Phys., 42, 337–352,
doi:10.1134/S0001433806030078, 2006.
Ruppe , J., Thomas, C., and Foken, T.: Scala simila i y o elaxed
eddy accumula ion me hods, Bound.-Lay. Me eo ol., 120, 39–
63, doi:10.1007/s10546-005-9043-3, 2006.
Selle s, P. J., Min z, Y., Sud, Y. C., and Dalche , A.: A Simple Bio-
sphe e Model (SIB) o use wi hin gene al ci cula ion models, J.
A mos. Sci., 43, 505–531, 1986.
Su, Z., Wen, J., Den e, L., an de Velde, R., Wang, L., Ma, Y.,
Yang, K., and Hu, Z.: The Tibe an Pla eau obse a o y o pla eau
scale soil mois u e and soil empe a u e (Tibe -Obs) o quan i-
ying unce ain ies in coa se esolu ion sa elli e and model p od-
uc s, Hyd ol. Ea h Sys . Sci., 15, 2303–2316, doi:10.5194/hess-
15-2303-2011, 2011.
Twine, T. E., Kus as, W. P., No man, J. M., Cook, D. R.,
House , P. R., Meye s, T. P., P uege , J. H., S a ks, P. J.,
and Wesely, M. L.: Co ec ing eddy-co a iance lux unde es-
ima es o e a g assland, Ag ic. Fo . Me eo ol., 103, 279–300,
doi:10.1016/S0168-1923(00)00123-4, 2000.
an Hee waa den, C. C., de A ellano, J. V., Moene, A. F., and Hol -
slag, A. A. M.: In e ac ions be ween d y-ai en ainmen , su ace
e apo a ion and con ec i e bounda y-laye de elopmen , Q. J.
Roy. Me eo ol. Soc., 135, 1277–1291, doi:10.1002/qj.431, 2009.
Xue, Y., Zeng, F. J., and Schlosse , C. A.: SSiB and i s sensi-
i i y o soil p ope ies – a case s udy using HAPEX-Mobilhy
da a, Glob. Plane . Change, 13, 183–194, doi:10.1016/0921-
8181(95)00045-3, 1996.
Yang, K., Koike, T., and Yang, D.: Su ace lux pa ame e iza ion
in he Tibe an Pla eau, Bound.-Lay. Me eo ol., 106, 245–262,
doi:10.1023/A:1021152407334, 2003.
Yang, K., Chen, Y.-Y., and Qin, J.: Some p ac ical no es on he land
su ace modeling in he Tibe an Pla eau, Hyd ol. Ea h Sys . Sci.,
13, 687–701, doi:10.5194/hess-13-687-2009, 2009.
www.hyd ol-ea h-sys -sci.ne /16/1095/2012/ Hyd ol. Ea h Sys . Sci., 16, 1095–1110, 2012
73
1110 T. Ge ken e al.: A simple wo-laye soil model wi h ex apola ed su ace empe a u e
Yee, S. Y. K.: The o ce- es o e me hod e isi ed, Bound.-Lay. Me-
eo ol., 43, 85–90, doi:10.1007/BF00153970, 1988.
You, Q. L., Kang, S. C., Li, C. L., Li, M. S., and Liu, J. S.: Fea-
u es o me eo ological pa ame e s a Nam Co s a ion, Tibe an
Pla eau, in: Annual Repo o Nam Co Moni o ing and Resea ch
S a ion o Mul isphe e In e ac ions, edi ed by: Nam Co Mon-
i o ing and Resea ch S a ion o Mul isphe e In e ac ion, 1–8,
Chinese Academy o Sciences, Beijing, 2006 (in Chinese wi h
English abs ac ).
Zhou, D., Eigenmann, R., Babel, W., Foken, T., and Ma, Y.: The
s udy o nea -g ound ee con ec ion condi ions a Nam Co s a-
ion on he Tibe an Pla eau, Theo . Appl. Clima ol., 105, 217–
228, doi:10.1007/s00704-010-0393-5, 2011.
Hyd ol. Ea h Sys . Sci., 16, 1095–1110, 2012 www.hyd ol-ea h-sys -sci.ne /16/1095/2012/
B. Ge ken e al. (2012)
74
C. Babel e al. (2013)
Babel, W., Chen, Y., Bie mann, T., Yang, K., Ma, Y., and Foken, T.: Adap a ion
o a land su ace scheme o modeling u bulen luxes on he Tibe an Pla eau unde
di e en soil mois u e condi ions, submi ed o J. Geophys. Res.
75
BABEL ET AL.: ADAPTATION OF A LAND SURFACE SCHEME X - 7
NamITP NamUBT
(a) (b)
TP e sion
0 10 20 30 40 50
0
10
20
30
40
50
x
n = 1872
Ts c in °C, NamITP
modeled
obse ed
−200 0 100 200 300 400
−200
−100
0
100
200
300
400
x
n = 1872
QG in Wm−2, NamITP
obse ed
0 10 20 30 40 50
0
10
20
30
40
50
x
n = 1585
Ts c in °C, NamUBT
modeled
obse ed
−200 0 100 200 300 400
−200
−100
0
100
200
300
400
x
n = 1528
QG in Wm−2, NamUBT
obse ed
(c) (d)
o iginal e sion
0 10 20 30 40 50
0
10
20
30
40
50
x
n = 1872
Ts c in °C, NamITP
modeled
obse ed
−200 0 100 200 300 400
−200
−100
0
100
200
300
400
x
n = 1872
QG in Wm−2, NamITP
obse ed
0 10 20 30 40 50
0
10
20
30
40
50
x
n = 1585
Ts c in °C, NamUBT
modeled
obse ed
−200 0 100 200 300 400
−200
−100
0
100
200
300
400
x
n = 1528
QG in Wm−2, NamUBT
obse ed
Figu e 6. Obse a ions o su ace empe a u e and g ound hea lux e sus model simula ions, mea-
su ed pa ame e s a NamITP (a,c) and NamUBT (b,d); simula ions a e displayed o TP e sion (a,b)
and o iginal e sion (c,d).
1
obse ed simula ed
−10cm
−20cm
−40cm
−7cm
−20cm
−40cm
NamITP
1
obse ed simula ed
−10cm
−30cm
−50cm
−7cm
−20cm
−40cm
NamUBT
0
5
10
15
20
da es
Θ in %
(a) measu ed pa ame e
0
10
20
30
40
50
da es
Θ in %
(b) measu ed pa ame e
Jul 05 Jul 15 Jul 25 Aug 04
0
5
10
15
20
Θ in %
(c) de aul pa ame e
Jul 05 Jul 15 Jul 25 Aug 04
0
10
20
30
40
50
Θ in %
(d) de aul pa ame e
Figu e 7. Timese ies o obse ed and simula ed olume ic soil mois u e con en a a ious dep hs, TP
e sion, a NamITP (a,c) and NamUBT (b,d); simula ions a e displayed o measu ed pa ame e s (a,b)
and de aul pa ame e s (c,d). No e, ha simula ed soil dep hs do no ma ch he obse ed soil dep hs in
case o NamUBT.
The Taylo diag am handles measu es o pe o mance
which la gely igno e model bias. The e o e an o e iew o
he model bias B=ξsim −ξobs is gi en in Figu e 4, dis-
playing bias o sensible hea , la en hea and he sum o
u bulen luxes, o he same combina ions as in Figu e 3.
I should be no ed ha he bias is calcula ed wi h EBC-
Bo co ec ed obse a ions, he e o e “obse ed” u bulen
luxes also equal he obse ed a ailable ene gy.
The u bulen lux bias is in gene al la ge and posi i e a
NamITP, while i is nea ly anishing a NamUBT. This can
be a ibu ed o he g ound hea lux, which is discussed in
he nex sec ion. In gene al he TP e sion educes he sen-
sible hea lux and he e o e i s bias as well as, o a smalle
ex en , he bias o la en hea . Simula ions wi h de aul
pa ame e s show less bias han hose wi h measu ed pa am-
e e s, which is mo e p onounced a NamUBT han NamITP.
This is caused by pa ame e di e ences o ield capaci y and
wil ing poin : The high measu ed wil ing poin supp esses
e apo anspi a ion compa ed o he lowe de aul alues a
NamUBT. A NamITP, his e ec o di e en ield capac-
i y and wil ing poin can be pa ly compensa ed by ba e
soil e apo a ion, which is e en highe due o he changed
pa ame e iza ion in he TP e sion.
In Figu e 5 he model pe o mance is summed up wi h
heNash-Su cli e coe icien [Nash and Su cli e, 1970], u -
C. Babel e al. (2013)
82

X - 8 BABEL ET AL.: ADAPTATION OF A LAND SURFACE SCHEME
he abb e ia ed as NS. The NS coe icien can be simila ly
in e p e ed as he common coe icien o de e mina ion R2,
bu in con as , he NS is sensi i e o bo h co ela ion and
bias. Again, obse a ions a e EBC-Bo co ec ed. In gen-
e al, model simula ions pe o m be e a NamUBT han a
NamITP, as a esul o smalle bias (Figu e 4). This can be
a ibu ed o SEWAB o mula ions which ha e no been al-
ida ed o such d y condi ions be o e, such as he scheme o
calcula e s oma al esis ance by Noilhan and Plan on [1989].
I is u he shown o NamITP ha he TP e sion pe -
o ms conside ably be e wi h he sensible hea lux, and
nega i e e ec s on la en hea a e e y small and can be
neglec ed. This esul ag ees wi h p e ious s udies o e d y
su aces [Yang e al., 2008, 2009; Chen e al., 2010]. Fu he -
mo e, u bulen luxes a NamUBT a e no comp omised
by he TP e sion. P edic ions o sensible hea lux e en
imp o e, pa ly a he expense o la en hea lux pe o -
mance. This sugges s ha he implemen ed z0h-scheme is
no limi ed o d y su aces. The simula ions wi h he de-
aul pa ame e s pe o m signi ican ly be e han he ones
wi h he measu ed pa ame e s, bu his di e ence is la ge
o he o iginal e sion han o he TP e sion. Thus he
TP e sion seems o be less sensi i e o soil pa ame e s.
3.2.2. Soil Mois u e and G ound Hea Flux
In o de o judge i “meaning ul” pa ame e s ha e been
se , i is impo an o also look in o accompanying a iables.
Mos p ominen a e g ound hea lux and su ace empe -
a u e, ende ing he budge o u bulen luxes, and soil
mois u e, which de e mines he wa e a ailabili y o e ap-
o anspi a ion.
Figu e 6 compa es simula ions o soil hea lux and su ace
empe a u e (o iginal and TP e sion, measu ed pa ame-
e s) wi h obse a ions. As expec ed, he d ie NamITP
si e exhibi s highe su ace empe a u es han NamUBT
wi h maximum empe a u es a ound 50 ◦C and 35 ◦C, e-
spec i ely. In gene al, simula ions e lec hese obse a ions
easonably well, excep ha he maximum su ace empe -
a u e a NamITP in he TP e sion (Figu e 6a) pe o ms
sligh ly be e han he o iginal e sion (Figu e 6c). How-
e e , he soil hea lux is mode a ely o e en poo ly ep-
esen ed. I clea ly explains he obse ed bias o u bu-
len luxes o measu ed pa ame e s in Figu e 4: Unde es-
ima ion o su ace empe a u e (and he e o e long-wa e
upwelling adia ion) as well as unde es ima ion o he soil
hea lux lead o an o e es ima ion o a ailable ene gy a
NamITP, while he o e es ima ion o soil hea lux a Na-
mUBT leads o an unde es ima ion o u bulen luxes. A
NamITP, he new pa ame e iza ion o he mal conduc i i y
(TP e sion) shows less bias, bu mo e sca e (Figu e 6a).
This can be a ibu ed o he di e en model o mula ions
o he he mal conduc i i y (Figu e 8). Wi hin he ob-
se ed soil mois u e angea NamITP he TP e sion shows
0.0 0.2 0.4 0.6
0.2
0.4
0.6
0.8
1.0
1.2
Θ in m3m−3
1
NamITP
o ig
TP
λs in Wm−1K−1
0.0 0.2 0.4 0.6
Θ in m3m−3
1
NamUBT
Figu e 8. Simula ed he mal conduc i i y λsin depen-
dence on soil mois u e Θ; g ey ec angles indica e he
obse ed ange o uppe soil mois u e o he espec i e
s a ion.
a much la ge sensi i i y han he o iginal e sion, he e o e
e o s in soil mois u e p oduce a la ge sca e . Ano he
handicap o g ound hea lux simula ion a NamITP migh
be he he mal decoupling o he uppe soil laye om deepe
laye s unde e y d y condi ions. A NamUBT, he TP e -
sion p edic s he su ace empe a u e accu a ely, bu o e es-
ima es he soil hea lux, as a consequence o highe he mal
conduc i i ies wi hin he ele an soil mois u e ange (Fig-
u e 8). Howe e , he measu ed he mal conduc i i y i s
well o calcula ed alues in he TP e sion, he e o e his
o e es ima ion is somewha su p ising.
Looking a soil mois u e, bo h measu ed and de aul pa-
ame e s sligh ly o e es ima e he obse ed soil mois u e a
NamITP (Figu e 7). This illus a es he la ge ange o e-
alis ic es ima es o soil p ope ies: When using a ield ca-
paci y o 0.05 and wil ing poin o 0.02, aken om a gen-
e al soil wa e e en ion cu e o sandy soil [Blume e al.,
2010, p. 228], he model esembles he measu ed soil mois-
u e much be e (no shown). Howe e we s ick he e o
he pa ame e s om he labo a o y measu emen s o con-
sis ency. He e soil mois u e and empe a u e was ini ialized
by he measu ed p o ile. Addi ional simula ions, using a 3-
yea spin-up o cing da a se , yield simila esul s o soil
mois u e a NamITP (no shown). A NamUBT, measu ed
pa ame e s oughly app oxima e he special mois u e p o ile
nea he lake. De aul pa ame e s, howe e , show a di e en
p o ile, despi e he be e pe o mance o u bulen luxes.
This cha ac e is ic p o ile could no be ealized wi h longe
spin-up pe iods because o he nea , bu unknown, posi ion
o he g oundwa e able and i s seasonal a ia ion. This
p oblem has o be sol ed when using he model o a longe
pe iod a such a loca ion. Fu he mo e, modeling wi h de-
ailed pa ame e s o di e en soil laye s may imp o e he
mois u e p o ile, as measu ed soil pa ame e s di e subs an-
ially wi hin he i s 50 cm dep h.
3.3. In luence o he Me hod o Ene gy Balance
Closu e Co ec ion
Model simula ions we e compa ed o obse a ions co -
ec ed o ene gy balance closu e acco ding o he Bowen
a io (EBC-Bo). This co ec ion assumes scala simila i y o
sensible and la en hea luxes wi h espec o he obse ed
esidual. Some s udies sugges , howe e , ha a la ge pa
o he esidual should be a ibu ed o he sensible hea lux
[Maude and Foken, 2006; Ingwe sen e al., 2011]. Recen
discussion hypo hesizes an in luence o seconda y ci cula-
ions on he ene gy balance closu e [Foken e al., 2011, 2010;
Foken, 2008a].
I seconda y ci cula ions in he longwa e pa o he u -
bulence spec a – no measu ed wi h he eddy-co a iance
NSQH
Nash−Su cli e coe icien NS
NSQE
0.0
0.2
0.4
0.6
0.8
1.0
NamITP
NamITP NamUBT
od om Td Tm od
om
Td
Tm
Figu e 9. Simila o Figu e 5, bu o EBC-HB co -
ec ed obse a ions.
83
BABEL ET AL.: ADAPTATION OF A LAND SURFACE SCHEME X - 9
QH
σsim σobs
0.0 0.5 1.0 1.5
0.0 0.5 1.0 1.5
0.2
0.4
0.6
0.8
1
1.2
1.4
●
0.1 0.2 0.3 0.4 0.5
0.6
0.7
0.8
0.9
0.95
0.99
Co ela ion
●
●
●
●
QE
σsim σobs
S anda d de ia ion
0.0 0.5 1.0 1.5
0.0 0.5 1.0 1.5
0.2
0.4
0.6
0.8
1
1.2
1.4
●
0.1 0.2 0.3 0.4 0.5
0.6
0.7
0.8
0.9
0.95
0.99
Co ela ion
●
●
●
●
S a ion
NamITP
NamUBT
●●
Ve sion, pa
od
om Td
Tm
Figu e 10. No malized Taylo diag am simila o Figu e 3, bu o EBC-HB co ec ed obse a ions.
The esul s wi h EBC-Bo co ec ed obse a ions a e displayed in ligh colo s o compa ison.
NamITP NamUBT
(a) (b)
EBC-Bo co ec ion
0 100 200 300 400 500 600
0
100
200
300
400
500
600
Qh_EB[index.Qh]
sim_Qh[index.Qh]
●
●
●
●●
●
●
●●
●
●●
●●
●●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●●
●
●
●●
●
●
●
●
●
●
●
●●
●
●
●
●●
●
●●
●
●●
●●
●
●●
●
NS coe = 0.269
n = 623
QH in Wm−2, NamITP
modelled
EB−co ec ed obs
0 100 200 300 400 500 600
Qe_EB[index.Qe]
sim_Qe[index.Qe]
●
●●
●
●
●
●
●
●●
●●
●
●●
●●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●●
●
●
●
●
●●
●
●
●
●●
●●
●
●
●
●
●
●
●
●●
●
●
●
●●
●
NS coe = 0.703
n = 623
QE in Wm−2, NamITP
EB−co ec ed obs
0 100 200 300 400 500 600
0
100
200
300
400
500
600
Qh_EB[index.Qh]
sim_Qh[index.Qh]
●●
●
●
●
●
●●
●●●
●
●
●●●
●●
NS coe = 0.605
n = 81
QH in Wm−2, NamUBT
modelled
EB−co ec ed obs
0 100 200 300 400 500 600
Qe_EB[index.Qe]
sim_Qe[index.Qe]
●
●●
●
●●●
●
●
●●●●
●
●
●
●●
NS coe = 0.598
n = 81
QE in Wm−2, NamUBT
EB−co ec ed obs
(c) (d)
EBC-HB co ec ion
0 100 200 300 400 500 600
0
100
200
300
400
500
600
Qh_EB[index.Qh]
sim_Qh[index.Qh]
●
●●
●●
●
●
●
●●
●
●
●●
●●
●
●●
●
●
●●●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●●
●
●
●
●
●●
●
●●
●
●
●
●
●
●
●
●●
●●
●●
●●
●
●●●
●●
●●
●●
●
●●
●
NS coe = 0.359
n = 572
QH in Wm−2, NamITP
modelled
EB−co ec ed obs
0 100 200 300 400 500 600
Qe_EB[index.Qe]
sim_Qe[index.Qe]
●
●●
●
●
●
●●
●●●
●●
●●
●
●●
●
●●
●
●
●●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●●●
●
●●
●
●
●
●
●●
●
●
●
●●
●
●●
●
●
●
●
●
●●
●
●●
●●
●
NS coe = 0.677
n = 572
QE in Wm−2, NamITP
EB−co ec ed obs
0 100 200 300 400 500 600
0
100
200
300
400
500
600
Qh_EB[index.Qh]
sim_Qh[index.Qh]
●
●●
●
●●
●●
●●●
●●
●●●
●●
NS coe = 0.376
n = 71
QH in Wm−2, NamUBT
modelled
EB−co ec ed obs
0 100 200 300 400 500 600
Qe_EB[index.Qe]
sim_Qe[index.Qe]
●●●
●
●●●
●
●
●
●●●
●
●
●
●
●
NS coe = 0.817
n = 71
QE in Wm−2, NamUBT
EB−co ec ed obs
Figu e 11. Tu bulen lux obse a ions e sus model simula ions, measu ed pa ame e s, TP e sion,
a NamITP (a,c) and NamUBT (b,d); Obse a ions a e displayed using he EBC-Bo co ec ion (a,b) and
he EBC-HB co ec ion (c,d). The ed poin s indica e da a wi h −Res > 150 W m−2(no included in
he analysis).
me hod – due o con ec ion a e he eason o he unclosed
ene gy balance, he esidual should no simply be added o
he sensible hea lux. Because densi y is also a ec ed by
mois u e Cha uchi ipan e al. [2013] p opose a co ec ion
wi h he buoyancy lux QHB, u he named EBC-HB
QEBC−HB
H=QH+ HB ·Res (5)
QEBC−HB
E=QE+ (1 − HB)·Res, wi h (6)
HB =QH
QHB
=1 + 0.61Tcp
λ·Bo−1(7)
whe e cpis he ai hea capaci y and λis he hea o e ap-
o a ion. The e is a weak dependency o EBC-HB o ai
empe a u e T. Fo Bo = 1, mo e han 90 % o he esidual
is added o he sensible hea lux, while o Bo = 0.1 i is
app oxima ely 60 %.
A gene al e ec on model pe o mance in e ms o pa -
e n s a is ics is summa ized in a Taylo diag am in Fig-
C. Babel e al. (2013)
84
X - 10 BABEL ET AL.: ADAPTATION OF A LAND SURFACE SCHEME
u e 10. The EBC-HB co ec ion co ela es less wi h he
simula ions han EBC-Bo in he case o sensible hea , bu
he co ela ion is simila o la en hea luxes. The inc ease
o obse ed sensible hea luxes due o he EBC-HB co -
ec ion leads o la ge “obse ed” s anda d de ia ions and
he e o e smalle no malized s anda d de ia ions (and, ice
e sa, he dec ease o la en hea luxes leads o la ge no -
malized s anda d de ia ions).
The NS coe icien s s ill a ibu e be e pe o mance o
he TP e sion wi h espec o he o iginal e sion (Fig-
u e 9). Bu in con as , hemeasu ed pa ame e s now show
sligh ly highe NS alues o NamITP han he de aul pa-
ame e s and signi ican ly be e pe o mance o NamUBT,
which is he opposi e esul han is ob ained om Figu e 5.
The EBC-HB co ec ion in gene al shi s sensible hea lux
bias om EBC-Bo co ec ed obse a ions (Figu e 4) in he
nega i e di ec ion and la en hea lux bias in he posi i e
di ec ion. This leads o he pa e n change o he NS.
Conside ing an example in de ail, Figu e 11 shows ob-
se ed s. simula ed luxes (TP e sion, measu ed pa-
ame e s) a bo h si es and o bo h co ec ion me hods.
Looking a he black do s, he simula ions o sensible hea
lux show mo e sca e wi h he EBC-HB co ec ed obse -
a ions han o EBC-Bo. This p oblem is no so p o-
nounced o he la en hea lux; a NamUBT simula ions i
e en be e . The ed poin s indica e alues wi h esiduals
−Res > 150 W m−2, which we e no used in he analysis.
Ob iously, hese poin s we e mos s ongly a ec ed by he
choice o he co ec ion me hod, pa icula ly because hese
la ge esiduals occu nea ly exclusi ely o low Bowen a-
ios (no shown). This is no su p ise o NamUBT, whe e
only we condi ions occu , bu somehow unexpec ed o Na-
mITP. The eason migh be unce ain y in he g ound hea
lux calcula ion a ec ing he s o age e m [Leuning e al.,
2012], as al eady men ioned in Sec . 2.4.
Also K ache e al. [2009] show wi h ano he da a se ,
ha he ene gy balance implemen a ion in SEWAB oughly
p ese es he Bowen a io measu ed by eddy-co a iance.
F om hei calcula ions i ollows ha models which de e -
mine he luxes independen ly, such as TERRA [pa o he
“Lokalmodell” LM, S eppele e al., 2003] and REMO [Jacob
and Podzun, 1997], would ollow he p oposed algo i hm be -
e , bu a he expense o he g ound hea lux, which is no
de e mined by hese models and in ac equals he esidual
o hei ene gy balance. O he amilia land su ace models
such as, o example, he common land model CLM [Dai
e al., 2003], o he simple biosphe e model SiB2 [Selle s
e al., 1996] sol e his s ep in a way simila o SEWAB.
4. Discussion and Conclusions
Two da a se s o ene gy luxes o e a we and a d y g ass-
land si e on he Tibe an Pla eau we e de i ed o a six-week
pe iod in summe 2009. The ca e ully quali y-checked eddy-
co a iance measu emen s ha e been co ec ed o ene gy bal-
ance closu e. Fu he mo e, he u bulen luxes ha e been
modeled wi h he land su ace model SEWAB in i s o iginal
e sion and a e sion adap ed o Tibe an Pla eau condi ions
(TP e sion). Bo h e sions ha e been un wi h wo di e -
en pa ame e se s, a de aul se and one wi h soil physical
pa ame e s om labo a o y in es iga ions o ield samples.
Bo h pa ame e se s – de aul and measu ed –, pe o m
easonably well, aking in o accoun ha no calib a ion al-
go i hm has been applied. Howe e , measu ed pa ame e s
do no p o e o gi e be e esul s; a he he opposi e is
ue. This migh be a ibu ed o he high spa ial a iabil-
i y o soil p ope ies in he ield, which a e no adequa ely
e lec ed by he aken sample. Ano he aspec is ha a
gi en model s uc u e migh equi e e ec i e pa ame e al-
ues o yield op imal pe o mance, which de ia e om mea-
su ed ones. Ne e heless, some ield in es iga ions will be
ine i able o de i e a high quali y pa ame e se (measu ed
o de aul pa ame e s), namely soil mois u e measu emen s
and ield based knowledge o he soil ype, de i ed a leas
by a con en ional pedological desc ip ion.
SEWAB has been adap ed o Tibe an Pla eau condi ions,
modi ying o mula ions o he soil he mal conduc i i y,
ba e soil e apo a ion and z0h. In ag eemen wi h p e ious
s udies, his e sion imp o es p edic ions o sensible hea
lux on he d y su ace, al hough he o iginal model al eady
pe o ms qui e well. In e es ingly, p edic ions o luxes o e
he we su ace a e no (o no subs an ially) comp omised,
hus he obse ed spa ial di e ences in su ace luxes due
o small scale he e ogenei y in soil mois u e could be ep o-
duced wi h a consis en model e sion. The e o e his s udy
sugges s ha he z0h-scheme by Yang e al. [2008] is no
limi ed o d y o ba e soil su aces. Mo eo e , he sensi i -
i y due o soil pa ame e s seems o dec ease wi h he TP
e sion, implying mo e obus esul s.
In es iga ions conce ning he ene gy balance closu e co -
ec ion clea ly show ha SEWAB is mo e compa ible wi h
he obse a ions co ec ed by p ese ing he Bowen a io.
Howe e , he e is no p oo ha his closu e me hod e lec s
eali y. In con as , he ecen discussion o he closu e p ob-
lem ends o a ibu e he esidual mo e – o in o al – o
he sensible hea lux as he d i e o ad ec i e luxes no
esol ed by EC sys ems [Maude and Foken, 2006; Foken,
2008a; Foken e al., 2011; Ingwe sen e al., 2011; Cha uchi -
ipan e al., 2013]. While he TP e sion ou pe o ms he
o iginal e sion o bo h co ec ion me hods, he anking
be ween he pa ame e se s change. In summa y, his case
s udy yields huge di e ences in model pe o mance depen-
den on he closu e me hod. Thus i ollows ha he closu e
issue should no be neglec ed when models we e alida ed
wi h EC measu emen s.
Typically o case s udies, his wo k is limi ed by he
amoun o da a se s and si es enclosed, and in he i s in-
s ance he esul s a e ep esen a i e o he Nam Co egion.
They could be s eng hened by ex ending he in es iga ions
o e a la ge a ie y o measu emen pe iods and si es o
es he obus ness o he model pe o mance. Fu he mo e,
he e o s o obse a ions a e no explici ly aken in o ac-
coun . Beside he usual e o s o be expec ed, he e a e
some di icul ies a he NamITP si e due o a la ge g a el
con en in he soil and he su ace as well as missing op soil
empe a u e measu emen s. Fi s ly, his comp omises qual-
i y and ep esen a i eness o he measu ed a ailable ene gy;
secondly, labo a o y in es iga ions o soil physical p ope ies
become e y unce ain and di icul o ca y ou . E en “ e-
alis ic” pa ame e s ha e ob iously o be conside ed wi hin
a ange due o model and measu emen unce ain y as well
as soil he e ogenei y. Especially o he we si e, soil p op-
e ies change a lo wi hin he e ical p o ile, which should
be aken in o accoun by land su ace models.
This s udy shows ha a success ul modeling o u bu-
len luxes equi es a ca e ul model selec ion and pa ame e
in es iga ion. Howe e , when compa ing simula ions wi h
eddy-co a iance de i ed u bulen luxes, he pe o mance
is s ongly a ec ed by he measu ed ene gy balance closu e
gap. This is especially impo an in a sensi i e egion like
he Tibe an Pla eau, in which Kob esia pas u es and alpine
ege a ion a e supposed o be highly suscep ible o clima e
change and li es ock g azing [Miehe e al., 2008, 2011]. The
selec ed da a se s also imply ha su ace lux he e ogenei y,
gene a ed by landscape he e ogenei y, could be ema kable
in a sho dis ance. This should be aken in o accoun when
ega ding lux measu emen s as ep esen a i e o a la ge
a ea. This s udy p epa es he SEWAB model o be usable
o such issues.
Acknowledgmen s. The au ho s acknowledge he s a
membe s o he ITP Nam Co Moni o ing and Resea ch S a ion o
85
BABEL ET AL.: ADAPTATION OF A LAND SURFACE SCHEME X - 11
hei wo k on he Tibe an Pla eau, and Heinz-Theo Mengelkamp
o making SEWAB a ailable. The wo k desc ibed in his pub-
lica ion has been suppo ed by he Eu opean Commission (Call
FP7-ENV-2007-1 G an n . 212921) as pa o he CEOP-AEGIS
p ojec (h p://www.ceop-aegis.o g/) coo dina ed by he Uni e -
si y o S asbou g. Fu he mo e, his wo k has been unded by
he DFG P io i y P og amme 1372 “Tibe an Pla eau: Fo ma ion,
Clima e, Ecosys ems”. D . Yingying Chen g a e ully acknowl-
edges he suppo o he Na ional Na u al Science Founda ion o
China (G an No. 41105003).
Re e ences
Agam, N., P. R. Be line , A. Zang il, and E. Ben-Do (2004),
Soil wa e e apo a ion du ing he d y season in an a id
zone, J. Geophys. Res.,109(D16), D16,103–, doi:10.1029/
2004JD004802.
Alapa y, K., J. E. Pleim, S. Raman, D. S. Niyogi, and D. W.
Byun (1997), Simula ion o a mosphe ic bounda y laye p o-
cesses using local- and nonlocal-closu e schemes, J. Appl. Me-
eo ol.,36(3), 214–233, doi:10.1175/1520-0450(1997)036h0214:
SOABLPi2.0.CO;2.
Balsamo, G., S. Bousse a, E. Du a, A. Beljaa s, P. Vi e bo, and
B. an den Hu k (2011), E olu ion o land su ace p ocesses
in he IFS, ECMWF Newsle .,127, 17–22.
Blume, H.-P., G. W. B ¨umme , R. Ho n, E. Kandele , I. K¨ogel-
Knabne , R. K e zschma , K. S ah , and B.-M. Wilke (2010),
Sche e /Schach schabel Leh buch de Bodenkunde, 16 ed., 569
pp., Spek um Akademische Ve lag Heidelbe g.
B illi, F., L. H¨o nagl, A. Hamme le, A. Haslwan e , A. Hansel,
F. Lo e o, and G. Wohl ah (2011), Lea and ecosys em
esponse o soil wa e a ailabili y in moun ain g asslands,
Ag . Fo es Me eo ol.,151(12), 1731–1740, doi:10.1016/j.
ag o me .2011.07.007.
Cha uchi ipan, D., W. Babel, M. Maude , J.-P. Leps, and T. Fo-
ken (2013), Ex ension o he a e aging ime o he eddy-
co a iance measu emen and i s e ec on he ene gy balance
closu e, submi ed o Bound.-Lay. Me eo ol.
Chen, Y., K. Yang, D. Zhou, J. Qin, and X. Guo (2010), Imp o -
ing he Noah land su ace model in a id egions wi h an ap-
p op ia e pa ame e iza ion o he he mal oughness leng h, J.
Hyd ome eo ol.,11(4), 995–1006, doi:10.1175/2010JHM1185.
1.
Chen, Y., K. Yang, J. He, J. Qin, J. Shi, J. Du, and Q. He
(2011), Imp o ing land su ace empe a u e modeling o d y
land o china, J. Geophys. Res.,116(D20), D20,104–, doi:
10.1029/2011JD015921.
Chen, Y., K. Yang, W. Tang, J. Qin, and L. Zhao (2012), Pa am-
e e izing soil o ganic ca bon’s impac s on soil po osi y and
he mal pa ame e s o eas e n ibe g asslands, Sci. China
Ea h Sci.,55, 1001–1011, doi:10.1007/s11430-012-4433-0.
Clapp, R. B., and G. M. Ho nbe ge (1978), Empi ical equa ions
o some soil hyd aulic p ope ies, Wa e Resou . Res.,14 (4),
601–604.
Dai, Y., e al. (2003), The common land model, B. Am. Me eo ol.
Soc.,84(8), 1013–1023, doi:10.1175/BAMS-84-8-1013.
Duan, A., and G. Wu (2005), Role o he ibe an pla eau he mal
o cing in he summe clima e pa e ns o e sub opical asia,
Clim. Dynam.,24, 793–807, doi:10.1007/s00382-004-0488-8.
Foken, T. (2008a), The ene gy balance closu e p oblem: An
o e iew, Ecol. Appl.,18(6), 1351–1367.
Foken, T. (2008b), Mic ome eo ology, 308 pp., Sp inge , Be lin,
Heidelbe g.
Foken, T., e al. (2010), Ene gy balance closu e o he LITFASS-
2003 expe imen , Theo . Appl. Clima ol.,101, 149–160, doi:
10.1007/s00704-009-0216-8.
Foken, T., M. Aubine , J. J. Finnigan, M. Y. Lecle c, M. Maude ,
and K. T. Paw U (2011), Resul s o a panel discussion
abou he ene gy balance closu e co ec ion o ace gases,
B. Am. Me eo ol. Soc.,92(4), ES13–ES18, doi:10.1175/
2011BAMS3130.1.
Foken, T., R. Leuning, S. R. Oncley, M. Maude , and M. Aubine
(2012), Co ec ions and da a quali y con ol, in Eddy Co a i-
ance: A P ac ical Guide o Measu emen and Da a Analy-
sis, edi ed by M. Aubine , T. Vesala, and D. Papale, Sp inge
A mosphe ic Sciences, pp. 85–131, Sp inge Ne he lands, doi:
10.1007/978-94-007-2351-1 4.
Gao, Y. X. (1985), Tibe an soil, Science P ess, Beijing, pp 2–15.
Ge ken, T., e al. (2012), Tu bulen lux modelling wi h a
simple 2-laye soil model and ex apola ed su ace empe -
a u e applied a nam co lake basin on he ibe an pla eau,
Hyd ol. Ea h Sys . Sci.,16(4), 1095–1110, doi:10.5194/
hess-16-1095-2012.
G¨ockede, M., C. Rebmann, and T. Foken (2004), A combina-
ion o quali y assessmen ools o eddy co a iance measu e-
men s wi h oo p in modelling o he cha ac e isa ion o
complex si es, Ag . Fo es Me eo ol.,127(3-4), 175–188, doi:
10.1016/j.ag o me .2004.07.012.
G¨ockede, M., e al. (2008), Quali y con ol o ca boeu ope lux
da a - pa 1: Coupling oo p in analyses wi h lux da a qual-
i y assessmen o e alua e si es in o es ecosys ems, Biogeo-
sciences,5(2), 433–450, doi:10.5194/bg-5-433-2008.
Hong, J., and J. Kim (2010), Nume ical s udy o su ace en-
e gy pa i ioning on he Tibe an Pla eau: compa a i e anal-
ysis o wo biosphe e models, Biogeosciences,7(2), 557–568,
doi:10.5194/bg-7-557-2010.
Hsu, H.-H., and X. Liu (2003), Rela ionship be ween he ibe an
pla eau hea ing and eas asian summe monsoon ain all, Geo-
phys. Res. Le .,30(20), 2066, doi:10.1029/2003GL017909.
Hu, Z., e al. (2009), Pa i ioning o e apo anspi a ion and i s
con ols in ou g assland ecosys ems: Applica ion o a wo-
sou ce model, Ag . Fo es Me eo ol.,149(9), 1410–1420, doi:
10.1016/j.ag o me .2009.03.014.
Ingwe sen, J., e al. (2011), Compa ison o noah simula ions
wi h eddy co a iance and soil wa e measu emen s a a win-
e whea s and, Ag . Fo es Me eo ol.,151 (3), 345–355, doi:
10.1016/j.ag o me .2010.11.010.
Jacob, D., and R. Podzun (1997), Sensi i i y s udies wi h he
egional clima e model REMO, Me eo ol. A mos. Phys.,63,
119–129.
Kang, S., Y. Xu, Q. You, W.-A. Fl¨ugel, N. Pepin, and T. Yao
(2010), Re iew o clima e and c yosphe ic change in he i-
be an pla eau, En i on. Res. Le .,5(1), 015,101, doi:10.1088/
1748-9326/5/1/015101.
K ache , D., H. T. Mengelkamp, and T. Foken (2009), The esid-
ual o he ene gy balance closu e and i s in luence on he
esul s o h ee s a models, Me eo ol. Z.,18 (6), 1–15, doi:
10.1127/0941-2948/2009/0412.
La che , W. (2001), ¨
Okophysiologie de P lanzen: Leben, Leis-
ung und S eßbew¨al igung de P lanzen in ih e Umwel , 6 h
ed., 408 pp., Ve lag Eugen Ulme S u ga .
Leuning, R., E. an Go sel, W. J. Massman, and P. R. Isaac
(2012), Re lec ions on he su ace ene gy imbalance p oblem,
Ag . Fo es Me eo ol.,156(0), 65–74, doi:10.1016/j.ag o me .
2011.12.002.
Liebe hal, C., B. Huwe, and T. Foken (2005), Sensi i i y analysis
o wo g ound hea lux calcula ion app oaches, Ag . Fo es
Me eo ol.,132, 253–262, doi:10.1016/j.ag o me .2005.08.001.
Liu, J., S. Kang, T. Gong, and A. Lu (2010), G ow h o a high-
ele a ion la ge inland lake, associa ed wi h clima e change
and pe ma os deg ada ion in ibe , Hyd ol. Ea h Sys . Sci.,
14(3), 481–489, doi:10.5194/hess-14-481-2010.
Liu, Y., Q. He, H. Zhang, and A. Mam imin (2012), Imp o ing
he colm in aklimakan dese hin e land wi h accu a e key
pa ame e s and an app op ia e pa ame e iza ion scheme, Ad .
A mos. Sci.,29 (2), 381–390, doi:10.1007/s00376-011-1068-6.
Louis, J. (1979), A pa ame ic model o e ical eddy luxes
in he a mosphe e, Bound.-Lay. Me eo ol.,17, 187–202, doi:
10.1007/BF00117978.
Ma, Y., O. Tsukamo o, J. Wang, H. Ishikawa, and I. Tamagawa
(2002), Analysis o ae odynamic and he modynamic pa ame-
e s on he g assy ma shland su ace o ibe an pla eau, P og.
Na . Sci.,12, 36–40.
Maude , M., and T. Foken (2004), Documen a ion and ins uc-
ion manual o he eddy co a iance so wa e package TK2,
Wo k Repo Uni e si y o Bay eu h, Dep . o Mic ome eo-
ology, ISSN 1614-8916, 26, 42 pp.
Maude , M., and T. Foken (2006), Impac o pos - ield da a
p ocessing on eddy co a iance lux es ima es and ene gy bal-
ance closu e, Me eo ol. Z.,15 (13), 597–609, doi:10.1127/
0941-2948/2006/0167.
Maude , M., and T. Foken (2011), Documen a ion and ins uc-
ion manual o he eddy-co a iance so wa e package TK3,
Wo k Repo Uni e si y o Bay eu h, Dep . o Mic ome eo-
ology, ISSN 1614-8916, 46, 58 pp.
C. Babel e al. (2013)
86
X - 12 BABEL ET AL.: ADAPTATION OF A LAND SURFACE SCHEME
Mengelkamp, H. T., K. Wa ach, and E. Raschke (1997), A
land su ace scheme o a mosphe ic and hyd ologic models:
SEWAB (Su ace Ene gy and Wa e Balance), GKSS Repo
97/E/69, 40pp., GKSS Gees hach .
Mengelkamp, H. T., K. Wa ach, and E. Raschke (1999), SEWAB
- a pa ame e iza ion o he su ace ene gy and wa e balance
o a mosphe ic and hyd ologic models, Ad . Wa e Resou .,
23(2), 165–175, doi:10.1016/S0309-1708(99)00020-2.
Mengelkamp, H. T., G. Kiely, and K. Wa ach (2001), E alua ion
o he hyd ological componen s added o an a mosphe ic land-
su ace scheme, Theo . Appl. Clima ol.,69(3-4), 199–212, doi:
10.1007/s007040170025.
Miehe, G., S. Miehe, K. Kaise , L. Jianquan, and X. Zhao
(2008), S a us and dynamics o he kob esia pygmaea ecosys-
em on he ibe an pla eau, Ambio,37(4), 272–279, doi:
10.1579/0044-7447(2008)37[272:SADOTK]2.0.CO;2.
Miehe, G., K. Bach, S. Miehe, J. Kluge, Y. Yongping, L. Duo,
S. Co, and K. Wesche (2011), Alpine s eppe plan commu-
ni ies o he ibe an highlands, Applied Vege a ion Science,
14(4), 547–560, doi:10.1111/j.1654-109X.2011.01147.x.
Mihailo i´c, D. T., R. A. Pielke, B. Rajko i´c, T. J. Lee, and
M. Je i´c (1993), A esis ance ep esen a ion o schemes o
e apo a ion om ba e and pa ly plan -co e ed su aces o use
in a mosphe ic models, J. Appl. Me eo ol.,32(6), 1038–1054,
doi:10.1175/1520-0450(1993)032h1038:ARROSFi2.0.CO;2.
Mihailo i´c, D. T., B. Rajko i´c, B. Lali´c, and L. Deki´c (1995),
Schemes o pa ame e izing e apo a ion om a non-plan -
co e ed su ace and hei impac on pa i ioning he su -
ace ene gy in land-ai exchange pa ame e iza ion, J. Appl.
Me eo ol.,34(11), 2462–2475, doi:10.1175/1520-0450(1995)
034h2462:SFPEFAi2.0.CO;2.
M¨ugle , I., e al. (2010), A mul i-p oxy app oach o econs uc
hyd ological changes and Holocene clima e de elopmen o
Nam Co, Cen al Tibe , J. Paleolimnol.,43, 625–648, doi:
10.1007/s10933-009-9357-0.
Nash, J. E., and J. V. Su cli e (1970), Ri e low o ecas ing
h ough concep ual models pa I – A discussion o p inci-
ples, J. Hyd ol.,10(3), 282–290, doi:10.1016/0022-1694(70)
90255-6.
Noilhan, J., and S. Plan on (1989), A simple pa ame e iza ion
o land su ace p ocesses o me eo ological models, Mon.
Wea he Re .,117(3), 536–549, doi:10.1175/1520-0493(1989)
117h0536:ASPOLSi2.0.CO;2.
Rannik, U., M. Aubine , O. Ku banmu ado , K. K. Sabel eld,
T. Ma kkanen, and T. Vesala (2000), Foo p in analysis o
measu emen s o e a he e ogeneous o es , Bound.-Lay. Me e-
o ol.,97(1), 137–166, doi:10.1023/A:1002702810929.
Selle s, P., D. Randall, G. Colla z, J. Be y, C. Field, D. Da-
zlich, C. Zhang, G. Collelo, and L. Bounoua (1996), A e ised
land su ace pa ame e iza ion (SiB2) o a mosphe ic GCMs.
pa I: Model o mula ion, J. Clima e,9(4), 676–705, doi:
10.1175/1520-0442(1996)009h0676:ARLSPFi2.0.CO;2.
S eppele , J., G. Doms, U. Scha le , H. W. Bi ze , A. Gassmann,
U. Dam a h, and G. G ego ic (2003), Meso-gamma scale o e-
cas s using he nonhyd os a ic model LM, Me eo ol. A mos.
Phys.,82, 75–96.
Su, Z., J. Wen, L. Den e, R. an de Velde, L. Wang, Y. Ma,
K. Yang, and Z. Hu (2011), The Tibe an Pla eau obse a o y
o pla eau scale soil mois u e and soil empe a u e ( ibe -obs)
o quan i ying unce ain ies in coa se esolu ion sa elli e and
model p oduc s, Hyd ol. Ea h Sys . Sci.,15(7), 2303–2316,
doi:10.5194/hess-15-2303-2011.
Takayabu, I., K. Taka a, T. Yamazaki, K. Ueno, H. Yabuki,
and S. Haginoya (2001), Compa ison o he ou land su -
ace models d i en by a common o cing da a p epa ed om
GAME/Tibe POP’97 p oduc s – snow accumula ion and soil
eezing p ocesses –, J. Me eo ol. Soc. Jpn.,79 (1B), 535–554,
doi:10.2151/jmsj.79.535.
Taylo , K. E. (2001), Summa izing mul iple aspec s o model
pe o mance in a single diag am, J. Geophys. Res.,106(D7),
7183–7192, doi:10.1029/2000JD900719.
Twine, T. E., W. P. Kus as, J. M. No man, D. R. Cook, P. R.
House , T. P. Meye s, J. H. P uege , P. J. S a ks, and M. L.
Wesely (2000), Co ec ing eddy-co a iance lux unde es ima es
o e a g assland, Ag . Fo es Me eo ol.,103(3), 279–300, doi:
10.1016/S0168-1923(00)00123-4.
an de Velde, R., Z. Su, M. Ek, M. Rodell, and Y. Ma
(2009), In luence o he modynamic soil and ege a ion pa-
ame e iza ions on he simula ion o soil empe a u e s a es
and su ace luxes by he Noah LSM o e a Tibe an Pla eau
si e, Hyd ol. Ea h Sys . Sci.,13(6), 759–777, doi:10.5194/
hess-13-759-2009.
Wallace, J., I. W igh , J. S ewa , and C. Holwill (1991), The
sahelian ene gy balance expe imen (SEBEX): G ound based
measu emen s and hei po en ial o spa ial ex apola ion
using sa elli e da a, Ad . Space Res.,11(3), 131–141, doi:
10.1016/0273-1177(91)90413-E.
Wilczak, J., S. Oncley, and S. S age (2001), Sonic anemome e
il co ec ion algo i hms, Bound.-Lay. Me eo ol.,99, 127–150,
doi:10.1023/A:1018966204465.
Xu, J., and S. Haginoya (2001), An es ima ion o hea and wa-
e balances in he Tibe an Pla eau, J. Me eo ol. Soc. Jpn.,
79(1B), 485–504, doi:10.2151/jmsj.79.485.
Yanai, M. H., C. F. Li, and Z. S. Song (1992), Seasonal hea -
ing o he Tibe an Pla eau and i s e ec s on he e olu ion o
he Asian Summe Monsoon, J. Me eo ol. Soc. Jpn.,70(1B),
319–351.
Yang, K., and J. Wang (2008), A empe a u e p edic ion-
co ec ion me hod o es ima ing su ace soil hea lux om
soil empe a u e and mois u e da a, Sci. China Se . D,51,
721–729, doi:10.1007/s11430-008-0036-1.
Yang, K., T. Koike, and D. Yang (2003), Su ace lux pa ame e -
iza ion in he Tibe an Pla eau, Bound.-Lay. Me eo ol.,116,
245–262, doi:10.1023/A:1021152407334.
Yang, K., T. Koike, B. Ye, and L. Bas idas (2005), In e se anal-
ysis o he ole o soil e ical he e ogenei y in con olling su -
ace soil s a e and ene gy pa i ion, J. Geophys. Res.,110 (D8),
D08101, doi:10.1029/2004JD005500.
Yang, K., T. Koike, H. Ishikawa, J. Kim, X. Li, H. Liu, J. Wang,
S. Liu, and Y. Ma (2008), Tu bulen lux ans e o e ba e-
soil su aces: Cha ac e is ics and pa ame e iza ion, J. Appl.
Me eo ol. Clim.,47, 276–290, doi:10.1175/2007JAMC1547.1.
Yang, K., Y.-Y. Chen, and J. Qin (2009), Some p ac i-
cal no es on he land su ace modeling in he Tibe an
Pla eau, Hyd ol. Ea h Sys . Sci.,13(5), 687–701, doi:10.
5194/hess-13-687-2009.
Yang, K., X. Guo, and B. Wu (2010), Recen ends in su ace
sensible hea lux on he ibe an pla eau, Sci. China Ea h
Sci., pp. 1–10, doi:10.1007/s11430-010-4036-6.
Ye, D.-Z., and G.-X. Wu (1998), The ole o he hea sou ce o he
Tibe an Pla eau in he gene al ci cula ion, Me eo ol. A mos.
Phys.,67(1), 181–198, doi:10.1007/BF01277509.
Zhang, X. (2012), Imp o emen o a soil-a mosphe e- ans e
model o he simula ion o ba e soil su ace ene gy balances
in semia id a eas, Asia Pac. J. A mos. Sci.,48(1), 97–105,
doi:10.1007/s13143-012-0009-3.
Zhou, D., R. Eigenmann, W. Babel, T. Foken, and Y. Ma (2011),
The s udy o nea -g ound ee con ec ion condi ions a nam
co s a ion on he ibe an pla eau, Theo . Appl. Clima ol.,105,
217–228, doi:10.1007/s00704-010-0393-5.
W. Babel, Ab . Mik ome eo ologie, Uni e si ¨a Bay eu h,
95440 Bay eu h, Ge many. (wol gang.bab[email p o ec ed])
T. Bie mann, Ab . Mik ome eo ologie, Uni e si ¨a Bay eu h,
95440 Bay eu h, Ge many.
Y. Chen, Ins i u e o Tibe an Pla eau Resea ch, Chinese
Academy o Sciences, 4A Da un Road, Chaoyang Dis ic , Beijing
100101, China.
T. Foken, Ab . Mik ome eo ologie, Uni e si ¨a Bay eu h,
95440 Bay eu h, Ge many.
Y. Ma, Ins i u e o Tibe an Pla eau Resea ch, Chinese
Academy o Sciences, 4A Da un Road, Chaoyang Dis ic , Beijing
100101, China.
K. Yang, Ins i u e o Tibe an Pla eau Resea ch, Chinese
Academy o Sciences, 4A Da un Road, Chaoyang Dis ic , Beijing
100101, China.
87

D. Bie mann e al. (2013)
Bie mann, T., Babel, W., Ma, W., Chen, X., Thiem, E., Ma, Y., and Foken, T.:
Tu bulen lux obse a ions and modelling o e a shallow lake and a we g assland in
he Nam Co basin, Tibe an Pla eau, Theo . Appl. Clima ol., accep ed
88
1
Tu bulen lux obse a ions and modelling
o e a shallow lake and a we g assland in he
Nam Co basin, Tibe an Pla eau
Tobias Bie mann
1
, Wol gang Babel
1
, Weiqiang Ma
2
, Xuelong Chen
3
, Elisabe h
Thiem
4
, Yaoming Ma
5
, Thomas Foken
1,6
5
1
Depa men o Mic ome eo ology, Uni e si y o Bay eu h, Bay eu h, Ge many
2
Cold and A id Regions En i onmen and Enginee ing Resea ch Ins i u e,
Chinese Academy o Sciences, Lanzhou, China
3
Facul y o Geo-In o ma ion Science and Ea h Obse a ion, Uni e si y o
Twen e, Enschede, The Ne he lands 10
4
Depa men o Geog aphy, Ludwig-Maximilian Uni e si y, Munich, Ge many
5
Labo a o y o Tibe an En i onmen Changes and Land Su ace P ocesses,
Ins i u e o Tibe an Pla eau Resea ch, Chinese Academy o Sciences, Beijing,
China
6
Membe o Bay eu h Cen e o Ecology and Ecosys em Resea ch (BayCEER) 15
Co esponding au ho add ess:
Tobias Bie mann, Ab . Mik ome eo ologie, Uni e si ä Bay eu h, Uni e si ä ss .
30 , 95440 Bay eu h, Ge many
Phone: +49 921 552180
20
Fax: +49 921 552366
E-mail: obias.bie mann@uni-bay eu h.de
89
2
Abs ac
The Tibe an Pla eau plays an impo an ole in he global wa e cycle and is s ongly in luenced by
clima e change. While ene gy and ma e luxes ha e been mo e in ensely s udied o e land
25
su aces, a la ge p opo ion o lakes ha e been ei he neglec ed o pa ame e ised wi h simple bulk
app oaches. The e o e u bulen luxes we e measu ed o e we g assland and a shallow lake wi h
a single eddy-co a iance complex a he sho eline in he Nam Co basin in summe 2009. Foo p in
analysis was used o spli obse a ions acco ding o he unde lying su ace, and wo sophis ica ed
su ace models we e u ilised o de i e gap- ee ime se ies. Resul s we e hen compa ed wi h
30
obse a ions and simula ions om a nea by eddy-co a iance s a ion o e d y g assland, yielding
p onounced di e ences. Obse a ions and oo p in in eg a ed simula ions compa ed well, e en
o si ua ions wi h lux con ibu ions including g assland and lake. I is shown ha he accessibili y
p oblem o EC measu emen s on lakes can be o e come by combining s anda d me eo ological
measu emen s a he sho eline wi h model simula ions, only equi ing ep esen a i e es ima es o
35
lake su ace empe a u e.
Keywo ds: Eddy-co a iance, Lake, Nam Co, Su ace modelling, Tibe an Pla eau,
Tu bulen luxes
D. Bie mann e al. (2013)
90
3
1. In oduc ion
40
The ole o he Tibe an Pla eau in he global wa e cycle and i s eac ion o clima e
change has become a opic o s ong scien i ic in e es (e.g. Imme zeel e al. 2011,
Ni 2011). Rep esen ing a unique geological o ma ion, he Tibe an Pla eau is
conside ed he la ges and highes pla eau on ea h, wi h an a e age ele a ion
g ea e han 4000 m a.s.l.. Fu he mo e, he Tibe an Pla eau is he sou ce o a 45
la ge numbe o majo i e s in Asia. I s ole in he modula ion o he Asian
Monsoon and he clima e o la ge pa s o Asia, due o i s hea budge caused by
i s ele a ion in conjunc ion wi h he bo de ing Himalayan moun ain ange, has
been o majo esea ch in e es (Molna e al. 2010, Boos and Kuang 2010).
To unde s and he ole o he Tibe an Pla eau o he global hea and wa e budge , 50
much e o has been pu in o he es ima ion o ene gy balance and u bulen lux
measu emen s wi hin in e na ional campaigns like GAME/Tibe (GEWEX-
Global Ene gy and Wa e cycle Expe imen Asian Monsoon Expe imen ) and
(CAMP -Coo dina ed Enhanced Obse ing Pe iod Asia-Aus alia Monsoon
P ojec ) (Xu and Haginoya 2001; Ma e al. 2003; Ma e al. 2005) and in he 55
amewo k o he Tibe an Obse a ion and Resea ch Pla o m, TORP (Ma e al.
2009).
Despi e hese e o s, obse a ions on he Tibe an Pla eau a e spa se due o i s
emo e loca ion (F auen eld e al. 2005, Kang e al. 2010, Maussion e al. 2011).
The impo ance o e apo a ion o he hyd ological cycle unde he in luence o 60
clima e change has been highligh ed by Yang e al. (2011).
Mos long- e m obse a ion s a ions ocus on he majo land co e ypes such as
alpine s eppe, Kob esia pas u es and we lands (Zhao e al. 2010), howe e
app oxima ely 45.000 km² o he pla eau a e co e ed by lakes (Xu e al. 2009).
This lake a ea has been subjec o changes in he las decades, he easons a e no 65
well unde s ood due o lack o obse a ional da a (Xu e al. 2009). Al hough
Huang e al. (2008) epo a gene al dec ease o lake olume in Qinghai-Tibe
Pla eau, he Nam Co lake a ea has been inc easing (Liu e al. 2010, Wu and Zhu
2008, Zhu e al. 2010). They a ibu e his change o inc easing p ecipi a ion as
well as hawing pe ma os and glacial mel due o ising mean annual 70
empe a u es , ne e heless he ela i e con ibu ion o he balance componen s,
especially he ole o e apo a ion, is discussed con o e sially. Consequen ly
luxes o e lake su aces on he Tibe an Pla eau should no be neglec ed, since
a ious s udies ha e shown he con ibu ion o lakes o he egional ene gy
balance and wa e cycle in di e en ca chmen s a ound he wo ld (Rouse e al. 75
2005, No dbo e al. 2011). Un il now, es ima ions o e apo a ion o e lake
su aces on he Tibe an Pla eau ha e been modelled using emo e sensing o land
su ace obse a ions as o cing (Xu e al. 2009, Haginoya e al. 2009), whe eas no
di ec measu emen s o u bulen luxes o e a lake su ace ha e been conduc ed
so a . The ins alla ion o a lux s a ion in a lake on he Tibe an Pla eau is nea ly 80
impossible, due o p oblems o accessibili y, s ong winds and wa es du ing he
summe , as well as ice co e du ing win e .
Ne e heless, i is known om model es ima ions ha e apo a ion o e lake
su aces di e s om e apo anspi a ion o e land h oughou he yea due o he
hea s o age capaci y o he lakes, and has a s ong e ec on con ec ion and hus 85
on local clima es (Haginoya e al. 2009). The landscape on he Tibe an Pla eau is
ai ly he e ogeneous, including alpine s eppe, Kob esia pygmea ma s, we lands
and open wa e su aces in a ious sizes. The e o e high quali y e apo a ion
measu emen s o e wa e su aces on he Tibe an Pla eau need o be conside ed
91
10
Table 2. Go e ning equa ions o he hyd odynamic mul ilaye (HM) model (Foken 1979, 1984)
wi h shallow wa e ex ension (Panin and Foken 2005)
Va iable/componen Equa ion
Sensible hea lux




Γ






wi h
Γ

!
"
#
$
"
%
&
!
"
P


)
*
"
+
,
-
.
5
.
ln
1
$
"

23
4
5


La en hea lux Analogue o



, assuming
+
,
-
6
+
7
-
,
Δ

-
6
Δ
9
-
, and
eplacing P wi h Sc
S abili y dependence Monin-Obukho Simila i y Theo y, uni e sal unc ion
a e Foken and Skeib (1983)
Shallow wa e e m


,
;
<=



,
;


"

1
.
>

,
;
<=
"
?
@


wi h mean squa e
wa e heigh
?
6
0
.
07
#

B

C@
"
D
#

B
E
3
.
)
"
C

(Da idan
e al., 1985) and
>

,
;
<=
6
2
(Panin e al. 2006b)
Symbols
C
g a i y accele a ion [ms
-
2
]
@
lake dep h [m]
? mean squa e wa e heigh [m]
>
,;
<=
empi ical co ec ion ac o [-]
FG P and l numbe [-]

,;

sensible (H) and la en (L) hea lux wi hou shallow wa e
co ec ion [Wm
-2
]

,;
<=
sensible (H) and la en (L) hea lux wi h shallow wa e
co ec ion [Wm
-2
]
9 speci ic humidi y [-]
HI Schmid numbe [-]
 empe a u e [K]

-
dimensionless empe a u e [-]
#
$
ic ion eloci y [ms
-1
]
#

wind eloci y in heigh z [ms
-1
]
 measu emen heigh [m]
Γ p o ile coe icien [ms
-1
]
+
,
-
dimensionless hickness o he molecula empe a u e bounda y
laye [-]
! on Ká mán cons an [-]
J
kinema ic iscosi y [m
2
s
-1
]
2.3.1. Desc ip ion o he models used
Fo he lake su ace a hyd odynamic mul ilaye model (HM) by Foken (1979, 300
1984) was u ilised. In o de o accoun o mul iple laye s wi hin he su ace laye ,
u bulen luxes a e pa ame e ised in HM using an in eg a ed p o ile coe icien .
As opposed o a single bulk coe icien he in eg a ed p o ile coe icien esol es
he molecula bounda y laye , he iscous bu e laye , and he u bulen laye .
The e o e nea -su ace exchange condi ions a e e lec ed acco ding o 305
hyd odynamic heo y. O iginally designed o exchange o e he ocean, a
co ec ion e m o shallow wa e (Panin and Foken 2005) was added, esul ing in
inc eased u bulen luxes due o an enhanced mixing by highe wa es in shallow
wa e . The model has been success ully applied o simula e luxes abo e ocean
su aces and lakes wi h a la ge e ch as well as o e a c ic snow ields (Panin e al. 310
2006b, Foken 1986, Lüe s and Ba eiss 2010). De ails o he go e ning equa ions
can be ound in Table 2.
D. Bie mann e al. (2013)
98

11
Tu bulen luxes o e he land su ace we e simula ed wi h he one-dimensional
Su ace Ene gy and WA e Balance scheme (SEWAB, Mengelkamp e al. 1999, 315
2001), a soil- ege a ion-a mosphe e- ans e model. All ene gy balance
componen s a e gi en sepa a ely. Tu bulen luxes a e o mula ed wi h bulk
app oaches, a mosphe ic s abili y is conside ed. The main ea u es a e
summa ized in Table 3. The ene gy balance is hen closed by i e a ion o he
su ace empe a u e. E apo anspi a ion om ege a ion is calcula ed wi h a 320
single lea concep in a Ja is- ype scheme a e Noilhan and Plan on (1989).
Emphasis is placed on he desc ip ion o soil p ocesses. Soil empe a u e
dis ibu ion and e ical soil wa e mo emen a e desc ibed by he di usion
equa ion and he Richa ds equa ion, espec i ely. Soil mois u e cha ac e is ics a e
in e - ela ed ollowing Clapp and Ho nbe ge (1978). 325
Bo h models we e o ced wi h s anda d me eo ological in-si u measu emen s. In
o de o p o ide gap- ee inpu da a, he small gaps wi hin he o cing da a om
NamUBT we e illed by linea in e pola ion while la ge gaps we e illed by
linea eg ession using he da a om NAMORS. 330
2.3.2. Applica ion o he HM model o a shallow lake
The o cing da a se o he HM model includes he s anda d me eo ological
pa ame e s wind eloci y, ai empe a u e, humidi y and ai p essu e. Radia ion
measu emen s a e no equi ed o he HM model, ins ead wa e su ace
empe a u e has o be supplied ins ead. In his s udy we used he measu ed wa e 335
empe a u e (Table 1) as an es ima e o he wa e su ace empe a u e.
Wendisch and Foken (1989) in es iga ed he ela i e e o con ibu ion o model
pa ame e , amongs o he s wa e empe a u e, ai empe a u e, ai humidi y and
wind eloci y, o he model ou pu wi h a sensi i i y analysis (Fou ie Ampli ude 340
Sensi i i y Tes by Cukie e al., 1978). Assuming ypical measu emen e o s o
he ini ial pa ame e dis ibu ion hey es ima ed ha wa e empe a u e con ibu ed
up o 50% o he o e all e o while he in luence o wind eloci y, ai
empe a u e and humidi y a e compa a i ely small, each con ibu ing 10-20% o
he e o . Since he empe a u e p obe was only shielded agains di ec 345
(downwa d) adia ion, he e ec o di use adia ion on he accu acy o he wa e
empe a u e measu emen s was e alua ed. The adia ion e o has been es ima ed
wi h a g aphical analysis o sho e m empe a u e pe u ba ions as ela ed o
apid changes in downwa d sho wa e adia ion. Caused by he small ac ion o
di use adia ion in he low ai densi y o he Tibe an Pla eau and sudden cloud 350
co e changes, he sho wa e adia ion obse a ions occasionally d op om
1000 Wm
-2
o 150 Wm
-2
(o inc ease in e e se) wi hin a ew minu es. The
co esponding shi s in wa e empe a u e sugges a possible adia ion e o o
app oxima ely 0.2 K. The e o e wa e empe a u e measu emen s ha e been
accep ed o model o cing wi hou co ec ion. 355
The shallow wa e pa ame e isa ion included in he cu en e sion o he HM
model accoun s o an enhanced u bulen exchange due o inc eased wa e heigh s
in shallow wa e . Consequen ly, he u bulen luxes inc ease wi h he mean
squa e wa e heigh (Table 2). Toge he wi h he wa e heigh pa ame e isa ion 360
a e Da idan e al. (1985), addi ional pa ame e s in luence he model esul s.
These a e he wind eloci y, lake dep h and an empi ical coe icien , which was
99
12
se o 2 in his s udy ollowing Panin e al. (2006b). In his s udy he wa e dep h
has been es ima ed as 1.5 m wi hin he a e age oo p in a ea o he measu emen
pe iod. 365
Table 3. Go e ning equa ions o SEWAB (Mengelkamp e al. 1999, 2001) and adap a ions o he
Tibe an Pla eau as used in Babel e al. (2013).
Va iable/componen Equa ion
Ne adia ion
K
L


K
MN
D
1

O
E

K
PMN
.
QR


S
K
MN
and
K
PMN
in o cing da a se
G ound hea lux

T

U






<

"
Δ

<



Sensible hea lux



V

W
I
X
#
D

E





D

E
La en hea lux Composed o ba e soil
Y

, we oliage
Y

and plan
anspi a ion
Y
LZ
a e Noilhan and Plan on (1989)
Y

V
;
W#DE[9



9DE
Y


V
;
W#
D

E

9





9
D

E
Y
LZ

D
K


K

E


W

9





9
D

E
S abili y dependence
V

a e Louis (1979),
V
;

V

Adap a ions o TP:
I) Soil he mal
conduc i i y
U

D
E

U
NZ]
.

U
L

U
NZ]
exp
a
>
,
D
1

Θ
L
/
Θ
E
d
,
>
,

0
.
36
(Yang e al. 2005)
II) he mal
oughness leng h

3

70
J
"
#
$


"
exp
D

g
#
$
3
.
h
|

$
|
3
.
Bh
E
g

7
.
2
s
0.5
m
-0.5
K
-0.25
(Yang e al. 2008)
III) ba e soil
e apo a ion
[

j
1

&
1

k
k
lm
*
B
,
Θ

Θ
no
1
,
Θ

Θ
no
p
(Mihailo ić e al. 1993)
Symbols
O
albedo [-]
V

S an on numbe [-]
V
;
Dal on numbe [-]
I
X
ai hea capaci y [J kg
-1
K
-1
]
9 speci ic humidi y [-]
9

sa u a ion speci ic humidi y [-]
K

u bulen a mosphe ic esis ance [s m
-1
]
K

s oma a esis ance [s m
-1
]
K
PMN
long wa e downwa d adia ion [Wm
-2
]
K
MN
sho wa e downwa d adia ion [Wm
-2
]
 empe a u e [K]

$
dynamic empe a u e scale [K]


su ace empe a u e [K]

<
empe a u e in i s soil laye [K]
#
$
ic ion eloci y [ms
-1
]
 measu emen heigh [m]
Δ
<
hickness o i s soil laye [m]
[ dependence ac o o soil ai humidi y o soil wa e con en [-]
Q emissi i y [-]
Θ olume ic soil wa e con en [-]
Θ
L
olume ic soil wa e con en a sa u a ion, po osi y [-]
Θ
no
olume ic soil wa e con en a ield capaci y [-]
U

soil he mal conduc i i y [Wm
-1
K
-1
]
U
NZ]
soil he mal conduc i i y o d y soil [Wm
-1
K
-1
]
U
L
soil he mal conduc i i y, soil mois u e a sa u a ion [Wm
-1
K
-1
]
W ai densi y [kg m
-3
]
R
S e an Bol zmann cons an [Wm
-2
K
-4
]
J
kinema ic iscosi y [m
2
s
-1
]
D. Bie mann e al. (2013)
100
13
The in luence o he shallow wa e e m becomes dominan wi h inc easing wind
eloci y and dec easing wa e dep h. Calcula ion o he shallow wa e equa ions
desc ibed in Table 2 wi h he es ima ed wa e dep h o 1.5 m and he a e age wind 370
eloci y o 4 ms
-1
yields an inc ease in u bulen luxes o 14.5 % compa ed o
deep wa e condi ions. Conside a ion o small changes in wind eloci y and wa e
dep h yields local sensi i i ies o oughly 2.9% o he deep wa e luxes pe ms
-1
and -3.9% pe m wa e dep h, espec i ely. Fo high wind eloci ies (10 ms
-1
) he
shallow wa e ex ension causes an inc ease o 30.1% wi h sensi i i ies o 2.4% 375
pe ms
-1
and -8.0% pe m wa e dep h. Assuming a ypical e o o 0.3 ms
-1
o
wind eloci y and a iabili y o he lake dep h up o 1 m wi hin he oo p in leads
o lux unce ain ies o 1% and 4%, espec i ely. These e o s, al hough no
negligible, a e wi hin he unce ain y ange o he EC lux measu emen s.
2.3.3. Adap a ion o SEWAB 380
On he Tibe an Pla eau a s ong diu nal cycle o he su ace empe a u e du ing
d y pe iods o e ba e soil and sho g assland ha e been obse ed, which ypically
leads o an o e es ima ion o su ace sensible hea lux (Yang e al. 2009, Hong
and Kim 2010). To accoun o hese condi ions, SEWAB has been adap ed o
he Tibe an Pla eau by (i) a e ised calcula ion o he soil he mal conduc i i y as 385
used in Yang e al. (2005), (ii) a di e en o mula ion o he he mal oughness
leng h a e Yang e al. (2008) and (iii), by changing he pa ame e isa ion o ba e
soil e apo a ion acco ding o Mihailo ić e al. (1993). The o mula ions can be
seen in Table 3.
These changes ha e been implemen ed using lux da a om NamITP s a ion and 390
lux da a om NamUBT co esponding o land su ace (Babel e al. 2013), who
e alua ed his adap a ion as an imp o emen compa ed o he o iginal e sion.
SEWAB has been un o line, o ced by measu emen s o p ecipi a ion, ai
empe a u e, wind eloci y, ai p essu e, ela i e humidi y and downwelling 395
sho wa e and longwa e adia ion. The espec i e pa ame e s o bo h land
su ace ypes we e es ima ed by a combina ion om he in-si u measu emen s and
labo a o y in es iga ion o soil cha ac e is ics (Chen e al., 2012). Su ace
emissi i y, lea a ea index and minimum s oma al esis ence ha e been de i ed
om a ious sou ces (Yang e al. 2009, Hu e al. 2009, Alapa y e al. 1997) 400
2.4. S a is ics
Fo e alua ion o model pe o mance, simple compa isons we e ca ied ou using
he bias q

∑DF
s

s
E
u
w
and he mean absolu e e o xyY
q

∑|F
s

s
|
u
w
, wi h O as he obse a ions and P he model p edic ions.
In equi alence o he MAE he di e ences be ween wo ime se ies o p edic ions 405
can be quan i ied and we de ine he desi ed measu e as
+
sz
{
q


|
|
F

,
s

F
B
,
s
|
u
w

(1)
wi h P
1
and P
2
as p edic ions om he espec i e land use ypes 1 and 2. The
Nash-Su cli e coe icien NS se es as a goodness o i measu e
qH

1

∑
D
F
s

s
E
B
}
w

∑
D
s

~
E
B
}
w

(2)
wi h
~ as he mean o he obse a ions.
101
14
410
Fig. 5 Mean diu nal ene gy luxes o he whole measu emen pe iod, sepa a ed o (a) g ass
-
, (b)
g ass
+
and (c) lake; o land su aces, all componen s a e measu ed; lake luxes: he ne adia ion is
calcula ed om measu ed downwelling adia ion and using an albedo o 0.06 and he lake su ace
empe a u e wi h an emissi i y o 0.96; he lowe panel shows diu nal su ace and ai empe a u e.
415
The ime axis is displayed in Beijing s anda d ime (CST), mean local sola noon du ing he
obse a ion pe iod is a 1400 CST
3. Resul s
3.1. Flux measu emen s 420
The measu ed ene gy luxes o e he lake su ace and land (g ass
+
) show
p onounced di e ences in hei magni ude and dynamics. The day ime ne
adia ion is subs an ially highe o e he lake su ace, caused by a lowe albedo
and dec eased upwelling longwa e adia ion due o damped su ace empe a u es
o e he lake (Fig. 5c). Howe e , upwa d adia ion componen s we e only 425
measu ed o e he land su ace; o he lake su ace hey we e pa ame e ised
using an albedo o 0.06 and he lake su ace empe a u e wi h an emissi i y o
0.96.
As expec ed o he monsoon season on he Tibe an Pla eau, he la en hea lux 430
o e he land su ace was la ge han he sensible hea lux (Fig. 5a,b). This
obse a ion is in ag eemen wi h e.g. Gu e al. (2005) and Ma and Ma (2006). The
mean diu nal cycles o su ace and ai empe a u e also show he ypical dynamics
abo e land su ace, wi h uns able s a i ica ion du ing day ime bu highe su ace
empe a u es a e obse ed o e g ass
-
. G ound hea lux and sensible hea lux a e 435
in he same o de o magni ude o each land su ace again wi h highe alues o
g ass
-
. In consequence he la en hea lux is lowe o e his land su ace.
The u bulen luxes o e he lake, howe e , do no show a diu nal cycle, bu
emain cons an o e he day. The ene gy inpu om adia ion is s o ed in he lake 440
body and is a ailable a any ime as indica ed by he lake su ace empe a u e in
D. Bie mann e al. (2013)
102
15
Table 4: Model pe o mance o u bulen luxes o he h ee land use ypes and wo ene gy
balance co ec ion me hods o he land obse a ions: n (numbe o obse a ions), bias, MAE
(mean absolu e e o ), o se and slope om linea eg ession (mean geome ic eg ession) as well
as NS (Nash-Su cli e coe icien ) and R
2
.
445
Flux
Land use EBC n Bias MAE O se Slope NS R²
[Wm
-
2
] [Wm
-
2
] [Wm
-
2
] [-] [-] [-]
Q
H
g ass
-
Bo 627
52.2 55.5 36.8 1.13 0.26
0.80
HB 572
38.3 55.5 36.9 1.01 0.36
0.62
g ass
+
Bo 81 18.5 23.8 17.2 1.02 0.61
0.78
HB 71 -24.4 40.5 12.9 0.7 0.38
0.52
lake - 327
-2.7 7.6 5.3 0.72 0.75
0.79
Q
E
g ass
-
Bo 627
-10.8 45.3 -8.6 0.99 0.70
0.73
HB 572
-0.4 42.0 -14.3 1.1 0.68
0.74
g ass
+
Bo 81 -28.6 50.8 13.5 0.77 0.60
0.69
HB 71 1.7 23.4 5.4 0.98 0.82
0.82
lake - 392
-23.3 30.3 -8.5 0.9 0.50
0.64
Fig. 5c. No comple e ene gy balance could be es ima ed o e he lake su ace, as
no measu emen s exis o he hea s o age in he wa e body and hea luxes in o
he sedimen .
450
E apo a ion is compa ably high o lake su aces due o high wind eloci ies o 4
ms
-1
on a e age. In addi ion, high lake su ace empe a u es, caused by he
shallow wa e able and he small ex en o his lake, lead o uns able s a i ica ion
e en du ing day ime (Fig. 5c). The e o e u bulen exchange is enhanced
compa ed o s able s a i ica ion ypically ound o e lake su aces du ing day ime 455
(e.g. Bey ich e al. 2006, No dbo e al. 2011).
3.2. Model pe o mance
Di e en measu es o model pe o mance a e summa ised in Table 4. The esul s
om g ass
+
simula ions show easonable pe o mance, al hough he e a e only
ew obse a ions le a e il e ing, sepa a ion and ene gy balance closu e 460
co ec ion (Fig. 6). In case o EBC-Bo co ec ed obse a ions, he la en hea lux
is sligh ly unde es ima ed while he simula ion o he sensible hea lux esembles
he measu emen s qui e well. The opposi e is ue when using he buoyancy lux
me hod o ene gy balance closu e co ec ion (EBC-HB). The co ela ion is
a ec ed in a simila way. Good R
2
alues a e achie ed o he sensible hea lux 465
co ec ed using EBC-Bo and la en hea lux co ec ed using EBC-HB, and hey
dec ease o he o he wo cases. Lake su ace modelling yields easonable
cohe ence o he EC obse a ions wi hin he oo p in o he measu emen s, wi h a
bias o -23.3 W m
-2
and -2.7 W m
-2
o he la en hea lux Q
E
and he sensible
hea lux Q
H
, espec i ely. Fo g ass
-
a NamITP, mean absolu e e o s o 470
sensible hea lux a e la ge , mainly caused by he bias, al hough a good
co ela ion is ob ained in he case o EBC-Bo co ec ed obse a ions. Aside om
model de iciencies, he eason o he emaining bias can be a ibu ed o
unce ain ies in es ima ion o he obse ed g ound hea lux due o high g a el
con en in he soil and a lack o empe a u e measu emen s in he opmos soil 475
laye .
103

16
Fig. 6 Sca e plo s o modelled and obse ed u bulen luxes. Fo land su ace lux obse a ions
(g ass
+
) s. SEWAB model simula ions, obse a ions a e ene gy balance co ec ed wi h he
480
Bowen a io me hod (a,b) and wi h he buoyancy me hod (c,d). Tu bulen luxes wi hou EBC
co ec ion o e lake s. HM model uns (e, ). Model pe o mance is indica ed wi h he Nash-
Su cli e coe icien (NS coe ), bias and he squa ed Pea son co ela ion coe icien . Red c osses
indica ed da a excluded due o esiduals K•€150 •‚
B
485
D. Bie mann e al. (2013)
104
17
3.3. Foo p in and spa ial in eg a ion
In he p e ious sec ion we ha e shown ha eddy-co a iance measu emen s,
selec ed acco ding o hei oo p in as pu e luxes om each su ace ype, can be
ep esen ed by SEWAB in case o g assland and by he HM model in case o he
lake su ace. Howe e , a pa o he measu emen s show con ibu ions om mo e 490
han one land use ype as well. The oo p in concep enables us o link he
simula ions e en wi h such obse a ions. Fo each ime s ep, he oo p in
app oach p o ides he ela i e con ibu ion o all in ol ed su aces o he
measu ed luxes. The simula ions a e hen ela ed o he obse a ions by
calcula ing a weigh ed mean om he ou pu o bo h models acco ding o he 495
ac ual land use con ibu ion. This is shown wi h he oo p in in eg a ed
simula ions o lake and g ass
+
oge he wi h he EC obse a ions a NamUBT in
Fig. 7 o h ee di e en si ua ions: 17 July – changing condi ions unde mode a e
wind eloci ies, 5 Augus – ypical day wi h land - lake ci cula ion and mode a e
winds o abou 2-6 ms
-1
, and 6 Augus – si ua ion wi h la ge han a e age wind 500
speeds o abou 6 ms
-1
. In all selec ed si ua ions he eddy-co a iance
measu emen s can be closely modelled by he oo p in in eg a ed simula ion.
This also holds o measu emen s wi h con ibu ions om bo h su aces, seen in
some e en s on 17 July and 5 Augus . Due o he di e en exchange o each
su ace wi h i s a mosphe e o e he cou se o he day, ins an aneous u bulen 505
luxes can show di e ences o up o 200 Wm
-2
. The pe o mance o he oo p in
in eg a ed simula ion is displayed in Fig. 8 o he whole pe iod. Si ua ions wi h
con ibu ions om bo h su ace ypes la ge han 20% (misc) a e highligh ed. The
simula ions o such si ua ions ollow he same pa e n as simula ions o he pu e
su ace ypes (con ibu ion o a single su ace g ea e han 80%). Miscellaneous 510
oo p in s, howe e , did no occu o si ua ions wi h e y high luxes.
3.4. Flux he e ogenei y a Nam Co
I is well known ha he e ogeneous su aces a ec he landscape scale luxes. The
p esen ed measu emen s ha e shown ha he luxes o e land and lake su aces
beha e di e en ly. To conside he mos abundan su aces nea Nam Co s a ion, 515
g ass
-
a NamITP has also been included in addi ion o he obse a ions and
simula ions o g ass
+
and lake a NamUBT. Fig. 9 shows he mean diu nal cycles
o measu ed luxes, co ec ed wi h EBC-Bo o land su aces, and modelled
luxes. The model simula ions esemble he obse ed cha ac e is ics o he
di e en su aces in a easonable sense. Since simula ions o e es ima ed he 520
sensible hea lux o bo h land su aces, he di e ences be ween land use ypes
we e main ained.
The ob ious di e ences in cha ac e is ics o he in es iga ed su aces, especially
be ween land and lake, a e also e lec ed in mean luxes o he whole pe iod (Fig. 525
10). The wo land su ace ypes al eady di e in he longwa e adia ion balance.
As expec ed, he mean la en hea lux became mo e dominan wi h inc easing soil
mois u e o he land su aces. The e apo a ion o e he small lake is e en highe ,
due o i s shallow wa e able esul ing in compa a i ely high su ace
empe a u es. Mean di e ences o sensible and la en hea lux be ween g ass
+
530
and g ass
-
a e 24.0 Wm
-2
and -33.5 Wm
-2
, espec i ely, and be ween g ass
+
and
lake a e -27.3 Wm
-2
and 22.3 Wm
-2
, espec i ely.
105
18
Fig. 7 Sou ce weigh in eg a ed modelled luxes a NamUBT, 17 July (a,b), 5 Augus (c,d), 6
Augus (e, ). Displayed a e simula ed luxes wi h SEWAB (dashed line) and HM (solid g ey line)
535
and in eg a ed simula ions (solid black line) acco ding o con ibu ions o lake o land wi hin he
oo p in . Obse a ions (no ene gy balance co ec ed) a e shown as black ci cles. The land use
con ibu ion in % is indica ed as ba plo , wi h upwind si ua ions om he land in g een and
upwind si ua ions om lake in blue. The ime axis is displayed in Beijing s anda d ime (CST),
mean local sola noon du ing he obse a ion pe iod is a 1400 CST
540
D. Bie mann e al. (2013)
106
19
Fig. 8 Obse a ions s. oo p in in eg a ed model simula ions a NamUBT. Th ee classes o da a
545
a e p esen ed, cases wi h a g ea e con ibu ion han 80% o one land use ype a e conside ed as
ep esen a i e. The misc cases con ain all luxes wi h a con ibu ion less han 80% o bo h land
use ypes. Since no EBC co ec ion could be pe o med o he lake da a, all da a is shown wi hou
co ec ion o be e in e compa ison
550
Fig. 9 Mean diu nal cycles o he whole measu emen pe iod. Obse ed luxes (co ec ed wi h
EBC-Bo) a e deno ed by black solid lines, he ho izon al ba s indica e he espec i e s anda d
de ia ion; g ey lines show he modelled luxes wi h s anda d de ia ions gi en by he g ey shaded
a ea. The ime axis is displayed in Beijing s anda d ime (CST), mean local sola noon du ing he
555
obse a ion pe iod is a 1400 CST
107
26
Maude M, Foken T (2006) Impac o pos - ield da a p ocessing on eddy co a iance lux es ima es
and ene gy balance closu e. Me eo ol Z 15(13):597–609, DOI 10.1127/0941-2948/2006/0167
Maude M, Foken T (2011) Documen a ion and ins uc ion manual o he eddy-co a iance
so wa e package TK3. Wo k Repo Uni e si y o Bay eu h, Dep . o Mic ome eo ology, ISSN
1614-8916, 46, 58 pp., URL h p://opus.ub.uni-bay eu h.de/opus4-
800
ubbay eu h/ on doo /index/index/docId/681
Maude M, Foken T, Clemen R, Elbe s JA, Eugs e W, G ünwald T, Heusink eld B, Kolle O
(2008) Quali y con ol o Ca boEu ope lux da a – Pa 2: In e -compa ison o eddy-co a iance
so wa e. Biogeosciences 5(2):451–462. doi:10.5194/bg-5-451-2008
Maussion F, Sche e D, Finkelnbu g R, Rich e s J, Yang W, Yao T (2011) WRF simula ion o a
805
p ecipi a ion e en o e he Tibe an Pla eau, China – an assessmen using emo e sensing and
g ound obse a ions. Hyd ol Ea h Sys Sci 15(6):1795–1817, DOI 10.5194/hess-15-1795-2011
Mengelkamp HT, Wa ach K, Raschke E (1999) SEWAB - a pa ame e iza ion o he su ace
ene gy and wa e balance o a mosphe ic and hyd ologic models. Ad Wa e Resou 23(2):165–
175, DOI 10.1016/S0309-1708(99)00020-2
810
Mengelkamp HT, Kiely G, Wa ach K (2001) E alua ion o he hyd ological componen s added o
an a mosphe ic land-su ace scheme. Theo Appl Clima ol 69(3-4):199–212, DOI
10.1007/s007040170025
Me zge S, Ma Y, Ma kkanen T, Göckede M, Li M, Foken T (2006) Quali y Assessmen o
Tibe an Pla eau Eddy Co a iance Measu emen s U ilizing Foo p in Modeling. Ad ances in Ea h
815
Sciences 21(12):1260–1267
Mihailo ić DT, Pielke RA, Rajko ić B, Lee TJ, Je ić M (1993) A esis ance ep esen a ion o
schemes o e apo a ion om ba e and pa ly plan -co e ed su aces o use in a mosphe ic
models. J Appl Me eo ol 32(6):1038–1054, DOI 10.1175/1520-
0450(1993)032<1038:ARROSF>2.0.CO;2
820
Molna P, Boos WR, Ba is i DS (2010) O og aphic Con ols on Clima e and Paleoclima e o
Asia: The mal and Mechanical Roles o he Tibe an Pla eau. Annu. Re . Ea h Plane . Sci.
38(1):77–102, DOI 10.1146/annu e -ea h-040809-152456
Mügle I, Gleixne G, Gün he F, Mäusbache R, Dau G, Schü B, Be king J, Schwalb A,
Schwa k L, Xu B, Yao T, Zhu L, Yi C (2010) A mul i-p oxy app oach o econs uc hyd ological
825
changes and Holocene clima e de elopmen o Nam Co, Cen al Tibe . J Paleolimnol 43(4):625–
648, DOI 10.1007/s10933-009-9357-0
Ni J (2011) Impac s o clima e change on Chinese ecosys ems: key ulne able egions and
po en ial h esholds. Regional En i onmen al Change 11(1):49-64, DOI 10.1007/s10113-010-
0170-0
830
Noilhan J, Plan on S (1989) A simple pa ame e iza ion o land su ace p ocesses o
me eo ological models. Mon Wea he Re 117(3):536–549, DOI 10.1175/1520-
0493(1989)117<0536:ASPOLS>2.0.CO;2
No dbo A, Launiainen S, Mamma ella I, Leppa an a M, Huo a i J, Ojala A, Vesala T (2011) Long-
e m ene gy lux measu emen s and ene gy balance o e a small bo eal lake using eddy co a iance
835
echnique. J. Geophys. Res 116(D2):D02119
D. Bie mann e al. (2013)
114

27
Panin GN, Foken T (2005) Ai –sea in e ac ion including a shallow and coas al zone. Jou nal o
A mosphe ic & Ocean Science 10(3):289–305, DOI 10.1080/17417530600787227
Panin GN, Nasono AE, Foken T (2006a) E apo a ion and hea exchange o a body o wa e wi h
he a mosphe e in a shallow zone. Iz . A mos. Ocean. Phys. 42(3):337–352.
840
doi:10.1134/S0001433806030078
Panin GN, Nasono AE, Foken T, Lohse H (2006b) On he pa ame e iza ion o e apo a ion and
sensible hea exchange o shallow lakes. Theo Appl Clima ol 85:123–129, DOI 10.1007/s00704-
005-0185-5
Paw U K, Baldocchi D, Meye s T, Wilson K (2000) Co ec ion o eddy-co a iance measu emen s
845
inco po a ing bo h ad ec i e effec s and densi y luxes. Bound-Lay Me eo ol 97:487–511, DOI
10.1023/A:1002786702909
Rannik U, Aubine M, Ku banmu ado O, Sabel eld KK, Ma kkanen T, Vesala T (2000) Foo p in
analysis o measu emen s o e a he e ogeneous o es . Bound-Lay Me eo ol 97(1):137–166, DOI
10.1023/A:1002702810929
850
Rebmann C, Kolle O, Heinesch B, Queck R, Ib om A, Aubine M (2012) Da a Acquisi ion and
Flux Calcula ions. In: Aubine M, Vesala T, Papale D (eds) Eddy Co a iance. Sp inge
Ne he lands, pp 59-83
Rouse WR, Oswald CJ, Binyamin J, Spence C, Sche ze WM, Blanken PD, Bussiè es N, Duguay
CR (2005) The Role o No he n Lakes in a Regional Ene gy Balance. Jou nal o
855
Hyd ome eo ology. J. Hyd ome eo 6(3):291–305, DOI 10.1175/JHM421.1
Twine TE, Kus as WP, No man JM, Cook DR, House PR, Meye s TP, P uege JH, S a ks PJ,
Wesely ML (2000) Co ec ing eddy-co a iance lux unde es ima es o e a g assland. Ag Fo es
Me eo ol 103(3):279–300. doi:10.1016/S0168-1923(00)00123-4
Wendisch M, Foken T (1989) Sensi i i y es o a mul ilaye ene gy exchange model. Z Me eo ol
860
39:36–39
Wilczak J, Oncley S, S age S (2001) Sonic anemome e il co ec ion algo i hms. Bound-Lay
Me eo ol 99:127–150, DOI 10.1023/A:1018966204465
Wu Y, Zhu L (2008) The esponse o lake-glacie a ia ions o clima e change in Nam Co
Ca chmen , cen al Tibe an Pla eau, du ing 1970–2000. J. Geog . Sci. 18(2):177–189, DOI
865
10.1007/s11442-008-0177-3
Xu J, Haginoya S (2001) An es ima ion o hea and wa e balances in he ibe an pla eau. J
Me eo ol Soc Jpn 79(1B):485–504, DOI 10.2151/jmsj.79.485
Xu J, Yu S, Liu J, Haginoya S, Ishigooka Y, Kuwaga a T, Ha a M, Yasuna i T (2009) The
Implica ion o Hea and Wa e Balance Changes in a Lake Basin on he Tibe an Pla eau.
870
Hyd ological Resea ch Le e s 3:1–5, DOI 10.3178/h l.3.1
Yang K, Koike T, Ye B, Bas idas L (2005) In e se analysis o he ole o soil e ical
he e ogenei y in con olling su ace soil s a e and ene gy pa i ion. J Geophys Res
110(D8):D08101, DOI 10.1029/2004JD005500
Yang K, Chen YY, Qin J (2009) Some p ac ical no es on he land su ace modeling in he Tibe an
875
Pla eau. Hyd ol Ea h Sys Sci 13(5):687–701, DOI 10.5194/hess-13-687-2009
115
28
Yang K, Koike T, Ishikawa H, Kim J, Li X, Liu H, Wang J, Liu S, Ma Y (2008) Tu bulen lux
ans e o e ba e-soil su aces: Cha ac e is ics and pa ame e iza ion. J Appl Me eo ol Clim
47:276–290, DOI 10.1175/2007JAMC1547.1
Yang K, Ye B, Zhou D, Wu B, Foken T, Qin J, Zhou Z (2011) Response o hyd ological cycle o
880
ecen clima e changes in he Tibe an Pla eau. Clima ic Change 109(3-4):517–534, DOI
10.1007/s10584-011-0099-4
Yu S, Liu J, Xu J, Wang H (2011) E apo a ion and ene gy balance es ima es o e a la ge inland
lake in he Tibe -Himalaya. En i on Ea h Sci 64(4):1169–1176. doi:10.1007/s12665-011-0933-z
Zhao L, Li J, Xu S, Zhou H, Li Y, Gu S, Zhao X (2010) Seasonal a ia ions in ca bon dioxide
885
exchange in an alpine we land meadow on he Qinghai-Tibe an Pla eau. Biogeosciences
7(4):1207–1221, DOI 10.5194/bg-7-1207-2010
Zhou D, Eigenmann R, Babel W, Foken T, Ma Y (2011) The s udy o nea -g ound ee con ec ion
condi ions a Nam Co s a ion on he Tibe an Pla eau. Theo Appl Clima ol 105(1-2):217-228, DOI
10.1007/s00704-010-0393-5
890
Zhu L, Xie M, Wu Y (2010) Quan i a i e analysis o lake a ea a ia ions and he in luence ac o s
om 1971 o 2004 in he Nam Co basin o he Tibe an Pla eau. Chin. Sci. Bull. 55(13):1294–1303.
DOI 10.1007/s11434-010-0015-8
D. Bie mann e al. (2013)
116
E. Cha uchi ipan e al. (2013)
Cha uchi ipan, D., Babel, W., Maude , M., Leps, J.-P., and Foken, T.: Ex ension o
he a e aging ime o he eddy-co a iance measu emen and i s e ec on he ene gy
balance closu e, submi ed o Bound.-Lay. Me eo ol.
117
Noname manusc ip No.
(will be inse ed by he edi o )
Ex ension o he a e aging ime o he eddy-co a iance1
measu emen and i s e ec on he ene gy balance closu e2
Doojdao Cha uchi ipan ·Wol gang Babel ·3
Ma hias Maude ·Jens-Pe e Leps ·4
Thomas Foken5
6
Recei ed: da e / Accep ed: da e7
Abs ac In his s udy, he modi ied ogi e analysis and block ensemble a e age we e8
employed o in es iga e he impac o he a e aging ime ex ension on he ene gy bal-9
ance closu e o e six di e en land use ypes. The modi ied ogi e analysis sugges s10
ha he s anda d a e aging ime o 30 minu es is s ill gene ally enough o he eddy-11
co a iance measu emen . The block ensemble a e age e eals ha he a e aging ime12
ex ension o e se e al days can imp o e ene gy balance closu e o some si es and13
o e some speci ic ime, when seconda y ci cula ions exis in he icini y o he sen-14
so . These nea -su ace seconda y ci cula ions mainly anspo sensible hea , and15
when nea -g ound wa m ai is anspo ed upwa d, he sensible hea lux obse ed by16
he block ensemble a e age will inc ease a longe a e aging imes. The close ela-17
ionship be ween nea -su ace seconda y ci cula ions and sensible hea lux sugges s18
an al e na i e ene gy balance co ec ion o a nea -su ace eddy-co a iance measu e-19
men by using he buoyancy lux a io, which is he la ge ac ion o he esidual20
a ibu e o he sensible hea lux.21
Keywo ds Ene gy balance closu e ·Ensemble a e age ·LITFASS ·Ogi e analysis22
D. Cha uchi ipan (B)·W. Babel ·T. Foken
Depa men o Mic ome eo ology, Uni e si y o Bay eu h, D-95440 Bay eu h, Ge many
Tel.: +49-921-552176
Fax: +49-921-552366
E-mail: [email p o ec ed]
T. Foken
Membe o Bay eu h Cen e o Ecology and En i onmen al Resea ch (BayCEER), Uni e si y o Bay eu h,
Ge many
M. Maude
Ins i u e o Me eo ology and Clima e Resea ch, Ka ls uhe Ins i u e o Technology, D-82467 Ga misch-
Pa enki chen, Ge many
J.-P. Leps
Ge man Me eo ological Se ice, Richa d-Aßmann-Obse a o y, D-15848 Lindenbe g, Ge many
E. Cha uchi ipan e al. (2013)
118
2 D. Cha uchi ipan e al.
1 In oduc ion23
The imbalance o he measu ed luxes a he ea h’s su ace is known as he ene gy24
balance closu e p oblem and mic ome eo ologis s ha e been awa e o i since he la e25
1980s (Foken, 2008a; Leuning e al., 2012). Many mic ome eo ological expe imen s26
o e low ege a ion e eal ha he a ailable ene gy, which is he sum o he ne 27
adia ion and he g ound hea lux, is la ge han he sum o he sensible and la en 28
hea luxes. These expe imen s include he EBEX expe imen (Oncley e al., 2007),29
which was especially designed o s udy he ene gy balance a he ea h’s su ace, and30
he LITFASS-2003 expe imen (Bey ich and Mengelkamp, 2006), which aimed o31
s udy he e ec o su ace he e ogenei y. To conse e ene gy, he esidual was added32
o he ene gy budge equa ion o e low ege a ion a he ea h’s su ace,33
Res =−Q∗−(QG+QH+QE),(1)
whe e Res is he esidual o missing ene gy, Q∗is he ne adia ion, QGis he g ound34
hea lux, QHis he sensible hea lux, and QEis he la en hea lux. Each e m in Eq.35
1 is posi i e, as he ene gy is anspo ed away om he g ound.36
In he pas ew yea s, despi e imp o emen s in measu ing and da a p ocessing37
echniques, his closu e p oblem s ill emains. I is belie ed ha he esidual is caused38
by la ge scale eddies o seconda y ci cula ions. These seconda y ci cula ions a e gen-39
e a ed by su ace he e ogenei y and no mally mo e away om he g ound (Kanda40
e al., 2004; Inagaki e al., 2006). Due o hei la ge size and slow mo ion, hei con-41
ibu ions o he low equency pa o he u bulen spec um canno be de ec ed by42
he eddy-co a iance (EC) measu emen , which is ypically a e aged o e a pe iod o 43
30 minu es. This esul s in he unde es ima ion o QHand QE, which a e no mally44
measu ed by he EC echnique.45
An ex ension o he a e aging ime was sugges ed and expec ed o esul in a46
g ea e con ibu ion om he low equency pa s. Two di e en app oaches we e47
in oduced o his ask: he ogi e analysis (Desja dins e al., 1989; Oncley e al.,48
1990) and he block ensemble a e age (Finnigan e al., 2003). The ogi e analysis uses49
he u bulen spec um o es ima e he u bulen luxes a di e en equency anges,50
allowing assessmen o he con ibu ion o he low equency pa s o he u bulen 51
luxes measu ed by he EC me hod. In Foken e al. (2006), he ogi e analysis was52
applied o he da a measu ed o e a maize ield o he LITFASS-2003 expe imen . I 53
was ocused mainly on da a om h ee selec ed days, whe e he a e aging ime was54
ex ended up o 4 hou s. I was ound ha he ime ex ension would no signi ican ly55
inc ease he u bulen luxes o e all.56
Fo he block ensemble a e age, low equency con ibu ions we e added o he57
u bulen luxes om long e m luc ua ions o e se e al hou s o days. In Maude 58
and Foken (2006), his ea men was also applied o he da ase om he same maize59
ield o he LITFASS-2003 expe imen . The leng h o he selec ed da ase was 1560
days, while he block ensemble a e aging pe iod a ied om 5 minu es o 5 days.61
This s udy showed ha he block ensemble a e age can close he ene gy balance62
a longe a e aging ime. Ex ensi e discussion o he ene gy balance closu e o he63
LITFASS-2003 expe imen can be ound in Foken e al. (2010).64
119

Ex ension o he a e aging ime o eddy-co a iance measu emen s 3
In ou s udy, o in es iga e whe he he a e aging ime ex ension would ha e he65
same impac o e di e en ypes o su ace, we ex ended bo h ogi e analysis and66
block ensemble a e age o co e mo e land use ypes o he LITFASS-2003 expe i-67
men . Since he LITFASS a ea is composed o many dis inc ag icul u al ields, his68
su ace he e ogenei y could induce seconda y ci cula ions, some o which may s ill69
exis in he icini y o he measu ing s a ions. Howe e , since he da a ob ained om70
di e en measu ing s a ions do no always ha e he same sampling a e, some mino 71
modi ica ions in bo h ogi e analysis and block ensemble a e age we e made. These72
modi ica ions we e alida ed by epea ing he ogi e analysis in Foken e al. (2006)73
and he block ensemble a e age in Maude and Foken (2006). Then we con inued74
ou in es iga ions, which was mo e o ien ed owa d he low equency con ibu ions.75
Finally, we came up wi h he app op ia e me hod o co ec ing he ene gy balance o 76
a nea -su ace eddy-co a iance measu emen .77
2 Ma e ial and Me hod78
2.1 LITFASS-2003 expe imen and da a p ocessing79
The LITFASS-2003 expe imen was pe o med be ween 19 May 2003 and 18 June80
2003 nea he me eo ological obse a o y o he Ge man Me eo ological Se ice in81
Lindenbe g, Ge many. The local ime zone in his a ea is UTC+1. Du ing he ex-82
pe imen , he e we e 14 g ound-based mic ome eo ological measu ing s a ions o e 83
13 si es, and 2 ele a ed measu ing s a ions on he owe a 50 and 90 me e heigh s.84
This expe imen co e ed an a ea o 20x20 km2and made up o 5 majo land use85
ypes: g ass, maize, ape, ce eals (include ye, ba ley and i icale), lake and o es .86
Mo e in o ma ion abou he LITFASS-2003 expe imen can be ound in Bey ich and87
Mengelkamp (2006).88
To co e he mos impo an land use ypes o he LITFASS-2003 expe imen ,89
we selec ed he ollowing measu ing s a ions o ou s udy: g ass (NV2 and NV4),90
maize (A6), ape (A7), ye (A5), lake (FS) and o es (HV). No e ha NV2 and NV491
we e ac ually ins alled on he same ield. They we e o ien ed o di e en wind sec o s92
o moni o u bulence a his ield om all wind di ec ions. We combined hese wo93
s a ions acco ding o he wind di ec ion and hey we e epo ed as he single s a ion94
NV. In o ma ion o hese selec ed s a ions can be ound in Table 1.95
All selec ed s a ions we e equipped wi h EC sys ems as lis ed in Table 1. Fou -96
componen ne adiome e s and soil hea lux pla es we e also ins alled in each s a ion.97
Each measu ing s a ion can measu e all he ene gy balance componen s in Eq. 1. De-98
ails o hese measu emen s we e well desc ibed in Maude e al. (2006) and Liebe hal99
e al. (2005). These measu emen s allow es ima ion o he esidual, which—on a e -100
age— eached i s maximum du ing 1000 -1200 UTC. Fo low ege a ion, i s a e age101
alue du ing his ime anged om 75 o 145 W m−2(o 20% - 30% o he a ailable102
ene gy as shown in Table 1).103
O e he o es , an addi ional e m, o canopy hea s o age, needs o be added o104
Eq. 1. Fo all plan s, he canopy hea s o age has wo main con ibu ions, he plan 105
ma e ial (o biomass) and he ai be ween plan s. O e low ege a ion, e.g. co on106
E. Cha uchi ipan e al. (2013)
120
4 D. Cha uchi ipan e al.
Table 1 Summa y o selec ed measu ing s a ions om he LITFASS-2003 expe imen du ing 20 May
2003, 1200 UTC - 18 June 2003, 0000 UTC. No a ions: hc=canopy heigh ; zm=measu emen heigh ;
θ
=accep ed wind di ec ion;
∆
= imes ep o sho e m ime se ies; Res = Mean esidual be ween
1000-1200 UTC, his is when he esidual is no mally each i s maximum; %Res = 100·Res/(−Q∗−QG)
be ween 1000-1200 UTC. Full de ails can be ound in Bey ich and Mengelkamp (2006) and Maude e al.
(2006)
S a ion Canopy hczm
θ
Tu bulence
∆
Res %Res
(m) (m) (deg ee) Senso s (minu es) (Wm−2) (%)
HV Pine o es 14 30.5 30 - 330 USA-1/LI-7500 10 133 24%
A5 Rye 0.75-1.50 2.80 60 - 30 USA-1/KH20 5 144 30%
A6 Maize 0.05-0.70 2.70 90 - 270 CSAT3/LI-7500 5 122 31%
A7 Rape 0.70-0.90 3.40 30 - 240 CSAT3/KH20 5 87 22%
NV2 G ass 0.05-0.25 2.40 60 - 180 USA-1/LI-7500 5 75 22%
NV4 G ass 0.05-0.25 2.40 150 - 330 USA-1/LI-7500 5 82 24%
FS Lake 0 3.85 180 - 30 USA-1/LI-7500 10 247 63%
M50 G ass 0.05-0.25 50.7 90 -300 USA-1/LI-7500 5 - -
M90 G ass 0.05-0.25 90.7 90 - 300 USA-1/LI-7500 5 - -
ields, bo h con ibu ions o canopy hea s o age a e ela i ely small and negligible107
(Oncley e al., 2007). The canopy hea s o age becomes signi ican o e o es and108
mus be included in he ene gy budge equa ion (Lind o h e al., 2010). Un o una ely,109
we did no collec all equi ed biomass p ope ies o he o es du ing he LITFASS-110
2003 expe imen , so he o es ’s canopy hea s o age could no be p ecisely es ima ed.111
Hence, all analyses o his si e we e done wi hou a canopy hea s o age e m. O e 112
he lake, due o i s la ge hea capaci y, QGwas calcula ed acco ding o No dbo e al.113
(2011).114
Du ing he LITFASS-2003 campaign, he aw da a we e p ocessed and a e aged115
o e 30 minu es. Fo his ask, all he pa icipa ing g oups ag eed o use he so wa e116
package TK2 (Maude and Foken, 2004), which has been es ed and compa ed in e -117
na ionally (Maude e al., 2008). Du ing lux calcula ion p ocesses, se e al lux co -118
ec ions we e applied. C oss-co ela ion analysis was used o ixing he ime delay119
be ween he sonic anemome e and hyg ome e . Moo e co ec ion was used o co ec 120
he spec al loss in he high equency ange (Moo e, 1986). The plana - i o a ion121
was used o align he sonic anemome e wi h a long e m mean s eamline (Wilczak122
e al., 2001). SND co ec ion was used o con e he sonic empe a u e, as eco ded123
by he sonic anemome e , o he ac ual empe a u e (Scho anus e al., 1983). WPL124
co ec ion was used o co ec he densi y luc ua ion (Webb e al., 1980). C osswind125
co ec ion was used o accoun o a di e en ype o sonic anemome e (Liu e al.,126
2001). Tanne co ec ion was used o co ec he c oss sensi i i y be ween H2O and127
O2molecules (Tanne e al., 1993), which was only applied o he K yp on Hyg om-128
e e KH20 (deployed in A5 and A7). Mo e de ails o hese co ec ions can be ound129
in Maude e al. (2006) and Foken e al. (2012).130
A e all hese lux co ec ions, quali y lags we e assigned o each 30 minu e131
pe iod. These quali y lags a e he s eady s a e lag, he in eg al u bulence cha ac-132
e is ic (ITC) lag (Foken and Wichu a, 1996) and combined lag. The s eady s a e133
121
Ex ension o he a e aging ime o eddy-co a iance measu emen s 5
lag is a esul o he s eady s a e es and ep esen s he s a iona i y o he da a. The134
ITC lag ep esen s he de elopmen o u bulen condi ions, which is he esul o he135
lux a iance simila i y es . The combined lag is he combina ion o he s eady s a e136
and ITC lags. All hese lags ange om 1-9 ( om bes o wo s ). High quali y da a,137
conside ed sui able o undamen al scien i ic esea ch, ha e lag alues o 1-3. Mo e138
de ails o he da a quali y analysis can be ound in Foken e al. (2012, 2004).139
Besides lux calcula ions, lux co ec ions and assignmen o da a quali y lags,140
TK2 can also gene a e sho e m a e ages and co a iances a 5 o 10 minu e in e -141
als. Due o he limi ed s o age capaci y, hese sho e m a e age da a poin s we e142
s o ed ins ead o he aw da a in some measu ing s a ions. Howe e , he s a is ics143
o longe pe iods can be econs uc ed om hese sho e m in o ma ion wi h he144
ollowing ela ions (Foken, 2008b),145
a′b′=1
M−1"(U−1)
N
∑
j=1a′b′j+UN
∑
j=1
ajbj−U2
M−1
N
∑
j=1
aj
N
∑
j=1
bj#,(2)
whe e a′b′is he long e m co a iance and Mis he numbe o measu emen poin s146
o he long e m ime se ies. This long e m ime se ies consis s o Nsho e m147
ime se ies, whose numbe o measu emen poin s is U.a′b′jis he sho e m148
co a iance, and ajand bja e he sho e m a e ages. These sho e m a e ages a e149
de i ed om aw da a, o which no lux co ec ions ha e been applied. The e o e, any150
necessa y lux co ec ions mus be included when using hese sho e m a e ages o 151
lux calcula ions. These sho e m a e age da a poin s om selec ed s a ions we e152
used o bo h ogi e analysis and block ensemble a e age calcula ions. Sho e m153
a e aging in e als o selec ed s a ions a e shown in Table 1.154
2.2 Da a selec ion155
In mos selec ed measu ing s a ions, g ound hea lux and adia ion da a a e only156
a ailable since 20 May 2003, 1200 UTC, so he pe iod du ing 20 May 2003, 1200157
UTC - 18 June 2003, 0000 UTC was used in his s udy. To ensu e high da a quali y158
as well as o minimize he i ele an ac o s which migh in luence u bulen luxes,159
we imposed se s o da a selec ion c i e ia o he ogi e analysis and block ensemble160
a e age sepa a ely. Fo he ogi e analysis, we inc eased he a e aging ime o up o161
4 hou s. This 4 hou pe iod consis s o 8 consecu i e subpe iods (o blocks) o 30162
minu es. We pe o med he ogi e analysis o e any 4 hou pe iod only i all blocks163
sa is ied he selec ion c i e ia.164
The i s selec ion c i e ion is iden ical o Maude e al. (2006), which is ha he165
sonic anemome e s mus no be dis u bed by ei he he in e nal bounda y laye esul -166
ing om he he e ogenei y o he su ace, o he low dis o ion caused by obs acles.167
The in e nal bounda y laye heigh was es ima ed om168
zm≤
δ
=0.3√x,(3)
(Raabe, 1983) whe e zmis he measu emen heigh ,
δ
is he in e nal bounda y laye 169
heigh and xis he dis ance om he senso o bounda y o he nex land use class.170
E. Cha uchi ipan e al. (2013)
122
6 D. Cha uchi ipan e al.
To keep he measu emen undis u bed, zmmus no exceed
δ
. Hence, we ejec ed any171
wind di ec ion whose co esponding xdid no sa is y Eq. 3. The undis u bed wind172
sec o s (
θ
) o each measu ing s a ion, om bo h in e nal bounda y laye and low173
dis o ion, a e lis ed in Table 1. Addi ionally, oo p in analysis was used o con i m174
ha he a ge land use ype has a signi ican con ibu ion o ou measu emen . This175
con ibu ion a ied o e he s abili y ange. We u he ejec ed any wind sec o s176
whose con ibu ion om a ge land use ype is less han 80%.177
The nex da a selec ion c i e ion is a s eady s a e condi ion o he ime se ies,178
which is indica ed by he s eady s a e lag (sec ion 2.1). We only accep ed da a wi h179
high quali y lags ( lag 1-3). In his s udy, we did he ogi e analysis o he ene gy180
balance componen s (QHand QE) and CO2 lux (Fc=w′c′CO2) sepa a ely. Fo he181
ene gy balance componen s, we only conside ed he s eady s a e lags o ic ion e-182
loci y (u∗), QHand QE. We only pe o med he ogi e analysis on any pe iods du ing183
which hese h ee s eady s a e lags quali ied simul aneously. Fo Fc, we conside ed184
only s eady s a e lags o u∗and CO2 lux, and pe o med he ogi e analysis on any185
pe iods whe e hese wo s eady s a e lags we e accep ed simul aneously.186
We a oided he ansi ion pe iod by excluding om ou analysis he ime pe iod187
co e ing one hou be o e o one hou a e bo h sun ise and sunse . We also speci ied188
he h eshold alue o each u bulen lux as a minimum equi emen in ou analy-189
sis. Fo u∗, which indica es he le el o u bulence (Massman and Lee, 2002), he190
h eshold alue is 0.1 m s−1. This was se o ule ou e y small u bulen luxes,191
which can esul om ins umen a ion noise. This limi no mally excludes pe iods192
wi h e y weak wind as well. Fo QH,QEand Fc, h eshold alues we e o mula ed193
o a oid complica ion wi h hei measu emen e o s. Acco ding o Maude e al.194
(2006), based on 30 minu es a e aging ime, he measu emen e o s o QHand QE
195
a e 10% - 20% o he u bulen lux a 30 minu es, o 10 - 20 W m−2, whiche e is196
la ge . Fo u∗and Fc, he measu emen e o s a e 0.02 - 0.04 m s−1and 0.5 - 1
µ
mol197
m−2s−1, espec i ely (Meek e al., 2005). The e o e, we se he h eshold alues o 198
QHand QE o be 20 W m−2, and he h eshold alue o Fc o be 1
µ
mol m−2s−1.199
Unusually la ge unce ain y o Fcdu ing he nigh ime was aken in o accoun by200
using only da a pe iods wi h u∗g ea e han 0.25 m s−1(Hollinge and Richa dson,201
2005).202
Simila selec ion c i e ia canno apply o he block ensemble a e age, as i in-203
ol es a e aging imes o se e al hou s o days. The e o e, he quali y con ol o his204
pa was done by disca ding any pe iods wi h mo e han 10% o missing aw da a.205
This missing da a could esul om a ious ac o s, e.g. elec ical black ou .206
2.3 Modi ied ogi e analysis207
The ogi e analysis was in oduced by Desja dins e al. (1989) and Oncley e al. (1990)208
o in es iga e he lux con ibu ion om each equency ange as well as o de e mine209
sui able a e aging pe iods o cap u e mos o he u bulen luxes. The ogi e unc ion210
o he u bulen lux (ogw,c) is de ined as he cumula i e in eg al o he cospec um o 211
123
Ex ension o he a e aging ime o eddy-co a iance measu emen s 13
Table 3 Resul s om he modi ied ogi e analysis o he ene gy balance componen s (QHand QE) om
selec ed s a ions o he LITFASS-2003 expe imen be ween 20 May 2003, 1200 UTC - 18 June 2003, 0000
UTC. No a ions:
η
is he wid h o e o band, which is se o be 10% and 20% o he u bulen lux a 30
minu e pe iod and has a minimum alue equals o he measu emen e o o each u bulen lux; F30 is
he a e age size o u bulen lux a 30 minu e pe iod;
∆
max is he a e age o maximum lux di e ence;
Runs(%) is numbe o uns in each ogi e case, which is also a ailable in pe cen age in he pa en hesis.
S a ion Flux
η
Case 1 Case 2 Case 3
F30 Runs(%) F30
∆
max Runs(%) F30
∆
max Runs(%)
Fo es QH10% 261 92(74.8%) 205 -33 4(3.3%) 224 33 27(22.0%)
(Wm−2) 20% 252 119(96.7%) 237 -56 1(0.8%) 217 70 3(2.4%)
(HV) QE10% 107 53(43.1%) 128 -33 12(9.8%) 119 27 58(47.2%)
(Wm−2) 20% 112 93(75.6%) 126 -45 6(4.9%) 125 40 24(19.5%)
Rye QH10% 148 192(88.1%) 99 -15 6(2.8%) 85 19 20(9.2%)
(Wm−2) 20% 143 213(97.7%) - - 0(0.0%) 61 36 5(2.3%)
(A5) QE10% 145 196(89.9%) 118 -20 10(4.6%) 131 23 12(5.5%)
(Wm−2) 20% 143 212(97.2%) 116 -26 2(0.9%) 132 30 4(1.8%)
Maize QH10% 106 99(84.6%) 98 -12 3(2.6%) 116 28 15(12.8%)
(Wm−2) 20% 108 111(94.9%) - - 0(0.0%) 92 39 6(5.1%)
(A6) QE10% 134 97(82.9%) 77 -20 14(12.0%) 80 18 6(5.1%)
(Wm−2) 20% 127 112(95.7%) 91 -37 3(2.6%) 57 22 2(1.7%)
Rape QH10% 127 85(90.4%) 83 -13 8(8.5%) 94 12 1(1.1%)
(Wm−2) 20% 123 94(100.0%) - - 0(0.0%) - - 0(0.0%)
(A7) QE10% 181 93(98.9%) - - 0(0.0%) 141 16 1(1.1%)
(Wm−2) 20% 181 94(100.0%) - - 0(0.0%) - - 0(0.0%)
G ass QH10% 117 187(93.0%) 101 -15 12(6.0%) 132 23 2(1.0%)
(Wm−2) 20% 116 200(99.5%) 99 -27 1(0.5%) - - 0(0.0%)
(NV) QE10% 131 173(86.1%) 95 -19 4(2.0%) 118 19 24(11.9%)
(Wm−2) 20% 140 196(97.5%) 94 -31 1(0.5%) 114 27 4(2.0%)
Lake QH10% 40 69(95.8%) - - 0(0.0%) 31 14 3(4.2%)
(Wm−2) 20% 40 72(100.0%) - - 0(0.0%) - - 0(0.0%)
(FS) QE10% 197 69(95.8%) 93 -15 1(1.4%) 121 14 2(2.8%)
(Wm−2) 20% 193 72(100.0%) - - 0(0.0%) - - 0(0.0%)
and 3 o bo h QHand QE om ye, g ass, maize and— ema kably— o es s a ions.413
These pe iods o Case 2 and 3 o ye, g ass and maize si es we e closely ela ed o he414
s a iona i y o QHand QEo e a 4 hou pe iod. Fo hese h ee si es, pe iods o Case415
1 usually had a ou hou s eady s a e lag o 1 o QHand QE, while Case 2 and 3416
usually had s eady s a e lags o 2 o mo e. This ela ion was no eadily appa en in417
he o es si e, implying ha he a e aging ime ex ension has imposed uns eadiness418
on he u bulence o e low ege a ion. I we es ic ou conside a ion o ye, g ass,419
maize and o es si es, we ound ha he numbe o Case 3s was no mally g ea e han420
he numbe o Case 2s in bo h QHand QE. This would ell us ha he a e aging ime421
ex ension mos likely inc eases QHand QE. The a e age maximum lux di e ence422
(
∆
max) o QHwas mos ly highe han o QE.
∆
max is inc eased wi h la ge size o 423
E. Cha uchi ipan e al. (2013)
130

14 D. Cha uchi ipan e al.
Table 4 Resul s om he modi ied ogi e analysis o ic ion eloci y (u∗) and CO2 lux (Fc). The desc ip-
ion is simila o Table 3
S a ion Flux
η
Case 1 Case 2 Case 3
F30 Runs(%) F30
∆
max Runs(%) F30
∆
max Runs(%)
Fo es u∗10% 0.64 191(99.5%) - - 0(0.0%) 0.38 0.06 1(0.5%)
(ms−1) 20% 0.64 192(100.0%) - -2.48 0(0.0%) - - 0(0.0%)
(HV) Fc10% 8.68 112(58.3%) 8.25 -1.57 24(12.5%) 7.43 1.54 56(29.2%)
µ
mol m−2s−120% 8.29 171(89.1%) 7.73 -2.48 8(4.2%) 8.21 3.23 13(6.8%)
Maize u∗10% 0.31 111(97.4%) 0.26 -0.03 1(0.9%) 0.15 0.03 2(1.8%)
(ms−1) 20% 0.31 114(100.0%) - -1.70 0(0.0%) - - 0(0.0%)
(A6) Fc10% 9.09 71(62.3%) 7.13 -1.56 16(14.0%) 7.34 2.40 27(23.7%)
µ
mol m−2s−120% 8.69 90(78.9%) 7.10 -1.70 12(10.5%) 7.52 4.09 12(10.5%)
G ass u∗10% 0.33 183(88.8%) - - 0(0.0%) 0.27 0.04 23(11.2%)
(ms−1) 20% 0.33 199(96.6%) - - 0(0.0%) 0.22 0.05 7(3.4%)
(NV) Fc10% 9.95 153(74.3%) 8.80 -1.67 29(14.1%) 7.65 1.27 24(11.7%)
µ
mol m−2s−120% 9.57 195(94.7%) 8.47 -2.74 8(3.9%) 9.41 2.82 3(1.5%)
an e o band (
η
), while lowe numbe s o Case 2 and 3 we e obse ed. This would424
indica e ha he ewe pe iods le had la ge
∆
max. Howe e , e en wi h he g ea es 425
∆
max added on op o lux co ec ions, he ene gy inc ease is s ill no enough o close426
he ene gy balance. Fu he mo e, om scala simila i y o QHand QE, we expec ed427
hese luxes o inc ease o dec ease oge he . This means we should see Case 2 o 428
Case 3 in bo h QHand QEsimul aneously, which was a ely obse ed.429
F30 o u∗had he highes alue o e he o es and he smalles alue o e he lake,430
and hey we e closely g ouped oge he o e low ege a ion. Ou MOG classi ied431
mos pe iods om all si es as Case 1. This sugges s ha he ime ex ension has almos 432
no impac on u∗ ega dless o canopy ypes.433
Fo Fc, all si es ga e compa ible alues o F30. Case 1 was s ill in he majo i y,434
wi h a la ge ac ion o Case 2 and 3 han he ene gy balance componen s. Fo es 435
also had la ge ac ion o Case 2 and 3 han did low ege a ion. O e all, he numbe 436
o Case 3s was g ea e han numbe o Case 2s, and
∆
max was also inc eased wi h
η
.437
The ou hou s eady s a e lags we e no mally 1 o Case 1 and highe o Case 2 and438
Case 3. Howe e , Case 2s gene ally had highe s eady s a e lags han Case 3. This439
sugges s ha i he a e aging ime ex ension does no impose much uns eadiness, i 440
ends o inc ease Fc, and dec ease i when mo e uns eadiness has been imposed.441
3.2 Block ensemble a e age442
The block ensemble a e age (Eq. 17) o all selec ed si es du ing 2 June 2003, 1800443
UTC - 18 June 2003, 0000 UTC, a e shown in Fig. 2. We chose his pe iod as ou 444
obse a ion pe iod NP o epea Maude and Foken (2006) wi h some mino mod-445
i ica ions (sec ion 2.4). We ound ha ou esul om he maize s a ion (Fig. 2 d)446
di e ed om he o iginal by less han he measu emen e o s o QHand QE. The e-447
131
Ex ension o he a e aging ime o eddy-co a iance measu emen s 15
o e, hese modi ica ions s ill gi e he same esul s and we can con iden ly apply hem448
wi h o he selec ed measu ing s a ions.449
log(a e aging ime in minu e)
Ene gy lux densi y (W m−2)
101102103104
−100
−50
0
50
100
150
200
(a)Lake
101102103104
0
20
40
60
80
100
(b)Fo es
101102103104
−20
0
20
40
60
80
100
(c)Rye
101102103104
0
20
40
60
80
(d)Maize
101102103104
0
50
100
(e)G ass
Qh Qe Res
101102103104
−20
0
20
40
60
80
100
120
( )Rape
Fig. 2 Block ensemble a e ages o sensible hea lux and la en hea lux (Eq. 17 wi h cis empe a u e and
absolu e humidi y), and hei co esponding esiduals, du ing 2 June 2003, 1800 UTC - 18 June 2003, 0000
UTC o selec ed si es in he LITFASS-2003 expe imen : (a) lake, (b) o es , (c) ye, (d) maize ( ep oduc ion
o Maude and Foken (2006)), (e) g ass and ( ) ba ley.
The ou come o he block ensemble a e age was qui e unexpec ed o us. I could450
close he ene gy balance only o e maize, ye and ape si es. Fo maize and ye si es,451
he closu es we e a a ound 15 - 30 hou s, which is close o he esul s ob ained in452
Maude and Foken (2006). These closu es we e mainly caused by he inc easing o 453
E. Cha uchi ipan e al. (2013)
132
16 D. Cha uchi ipan e al.
hQHiwi h longe block ensemble a e aging pe iod P. Fo he ape si e, bo h hQHi
454
and hQEiwe e app oxima ely cons an a all P. Du ing he obse a ion pe iod, his455
si e was also in luenced by ain e en s in he sou he n pa o he LITFASS a ea.456
The e o e, i s closu e a e y long Pwas no enhanced by he block ensemble a e age.457
Fo g assland and lake, hQHiwas dec eased wi h longe P, which was compensa ed458
by he inc ease in hQEi, and caused he esidual o be app oxima ely cons an a all459
P. Fo lake and o es si es, we mus in e p e he esul s ca e ully, because he lake460
has di e en cha ac e is ics om o he e ain si es and we canno p ecisely es ima e461
he canopy hea s o age (sec ion 2.1) o he o es om ou da a.462
A all si es, bo h hQHiand hQEiwe e app oxima ely cons an wi hin he i s 463
ew hou s. O e longe P,hQEiwas mo e s eady han hQHi. The in lec ion a he464
diu nal scale was ound a all si es o bo h hQHiand hQEi. As all hese selec ed si es465
a e p ac ically in he same 20x20 km2a ea, he diu nal e ec s should no be much466
di e en and he deg ee o in lec ion should be compa ible. The e o e, he s onge 467
in lec ion o e some si es and luxes may no be en i ely caused by he diu nal e ec s.468
As he block ensemble a e age could no close he ene gy balance o all selec ed469
si es om 2 June 2003, 1800 UTC o 18 June 2003, 0000 UTC, we need o de e mine470
he eason behind his and whe he i would be he same in a di e en obse a ion471
pe iod NP. We know ha he ˜w˜c e m o he block ensemble a e age is ela ed o472
he low equency lux con ibu ion. In p inciple, ˜w˜c ep esen s he lux con ibu ions473
beyond he a e aging pe iod P. I we se P o be 30 minu es, ˜w˜cwould ep esen 474
addi ional lux a e he 30 minu e a e aging ime. Hence, long e m obse a ion o ˜w˜c475
would show a ia ion o addi ional luxes om low equency con ibu ions, which476
may be ela ed o obse ed block ensemble a e age luxes. These a ia ions can be477
obse ed mo e clea ly when he obse a ion pe iod NP is long enough o supp ess478
any ansien e ec s in he block ensemble a e age luxes.479
Ou obse a ion pe iod NP, which co e ed an en i e pe iod o he LITFASS-2003480
expe imen , was 20 May 2003, 1200 UTC - 18 June 2003, 0000 UTC. We used ˜w˜c481
om all 30 minu e non-o e lapping blocks (P=30 minu es) wi hin his pe iod NP482
o cons uc he Ho mølle diag ams o ˜
QH( ˜w˜
Tin ene ge ic uni s, Tis empe a u e)483
and ˜
QE( ˜w˜ain ene ge ic uni s, ais absolu e humidi y ). These diag ams would show484
he a ia ion o addi ional luxes beyond 30 minu e a e aging ime. Acco ding o485
sec ion 2.4, ˜w˜ccan be e y la ge in any a bi a y blocks. The e o e, we expec ed o486
obse e some andom la ge ˜
QHand ˜
QEin hese diag ams.487
TheHo mølle diag ams o ˜
QH o ye and g asslanda e shown in Fig.3. Th ough-488
ou he en i e expe imen , we ound la ge ˜
QHmo e o en han la ge ˜
QE. We i s ly489
s a ed wi h he pe iod du ing 2 June 2003, 1800 UTC - 18 June 2003, 0000 UTC.490
Wi hin his pe iod, la ge ˜
QHwe e mainly posi i e o he ye (Fig. 3 a) and maize (no 491
shown) si es, and mainly nega i e o e he g assland (Fig. 3 b). These obse a ions492
a e consis en wi h he obse ed block ensemble a e age luxes, in which hQHiwas493
inc easing a longe P o he ye and maize si es, and ice e sa o he g assland494
(Fig. 2). The lake si e is mo e domina ed by la ge nega i e ˜
QH(no shown), which495
is consis en wi h he dec easing o hQHia longe P. The e we e only ew la ge496 ˜
QEa all si es, which a e consis en wi h app oxima ely cons an hQEia all Po e 497
ye, maize, ape and o es si es. Howe e , o e lake and g assland, hese ew la ge498
133
Ex ension o he a e aging ime o eddy-co a iance measu emen s 17
˜
QEwe e ex emely la ge when compa ed o hei own block ensemble a e ages, and499
caused hei hQEi o inc ease a longe P.500
Da e
Time o day
24/05 29/05 03/06 08/06 13/06
0
3
6
9
12
15
18
21
0
−800
−600
−400
−200
0
200
400
600
800
(a) Rye
Da e
Time o day
24/05 29/05 03/06 08/06 13/06
0
3
6
9
12
15
18
21
0
−800
−600
−400
−200
0
200
400
600
800
(b) G ass
Da e
Time o day
24/05 29/05 03/06 08/06 13/06
0
3
6
9
12
15
18
21
0
−800
−600
−400
−200
0
200
400
600
800
(c) 90 m owe
Fig. 3 The Ho mølle diag ams o ˜
QH om (a) Rye, (b) G ass and (c) 90 m owe , hey ep esen ˜w˜
To
each 30 minu e block in he ene ge ic uni s. Se ies o la ge ˜
QHwe e obse ed in ye (posi i e) and g ass
(mainly nega i e) du ing 1 June 2003 - 5 June 2003. These pa e ns a e ela ed o seconda y ci cula ions,
which is consis en wi h equen obse a ions o la ge ˜
QHa 90 m heigh . Each colou depic s he lux in
W m−2.
Mo e in e es ingly, la ge ˜
QHwe e obse ed consecu i ely o a ew days du ing501
1 June 2003 - 5 June 2003, o e ye, maize, g ass and lake. This pe iod was he d y502
pe iod be ween he ain e en s and was no in luenced by any signi ican synop ic503
E. Cha uchi ipan e al. (2013)
134
18 D. Cha uchi ipan e al.
e en s. These la ge ˜
QHwe e posi i e o ye and maize, and mainly nega i e o 504
g ass and lake. La ge ˜
QEwas no ound in his same pe iod. As dec ibed in sec ion505
2.4, la ge ˜
QH(o la ge ˜w˜
T) could compensa e a s ong ho izon al di e gence in an506
indi idual block. Howe e , consecu i e occu ences indica e ha hey we e ce ainly507
no ansien e ec s. A s ong ho izon al di e gence would imply o a s ong ho i-508
zon al ad ec ion, which is ela ed o seconda y ci cula ions. Hence, we belie e ha 509
hese pa e ns o la ge ˜
QHwe e caused by nea -su ace seconda y ci cula ions. To510
suppo his s a emen , we inspec ed he Ho mølle diag am o ˜
QHob ained om he511
measu emen a 90 me e heigh (M90). A his heigh , he e always exis seconda y512
ci cula ions, which means we should obse e se ies o la ge ˜
QHmo e o en han in513
g ound measu emen s. We did ac ually obse e se ies o la ge posi i e and nega i e514 ˜
QH h oughou he en i e pe iod o he LITFASS-2003 expe imen (Fig. 3 c).515
To obse e he e ec o nea -su ace seconda y ci cula ions mo e clea ly, we516
chose 1 June 2003, 1500 UTC - 5 June 2003, 1500 UTC as he new obse a ion517
pe iod NP. We used he beginning and ending ime o 1500 UTC o a oid gaps in he518
da a om he maize ield on he mo ning o 1 June 2003 and he ain e en on he519
e ening o 5 June 2003. Addi ionally, we wan ed o comple e a daily cycle as well.520
Since his long pe iod only las ed o 4 days, he block ensemble a e aging pe iod521
Pwas a ied om 10 minu es o 3 days. The block ensemble a e ages o his new522
obse a ion pe iod a e shown in Fig. 4. As expec ed, we ound ha du ing his pe iod,523
he ene gy balance is closed using a e aging imes o hal a day o ye and maize524
si es. O e g assland and lake, hQHiwas dec eased a longe P, which co esponds o525
he la ge nega i e ˜
QH.hQEiwe e app oxima ely cons an a all Pa all si es, which526
we e consis en wi h he absence o la ge ˜
QE.527
Accep ed models s a e ha seconda y ci cula ions can only each down o le els528
nea he ea h’s su ace unde he ee con ec ion condi ion, which occu s when he529
buoyancy e m domina es he shea p oduc ion e m as z/L≤ −1. This si ua ion is530
also accompanied by low ic ion eloci y (Eigenmann e al., 2009). As we did no 531
obse e any ee con ec ion du ing 1 June 2003, 1500 UTC - 5 June 2003, 1500 UTC,532
we belie e ha hese nea -su ace seconda y ci cula ions we e caused by he he mal533
he e ogenei y be ween di e en land use ypes (S oy e al., 2013).534
3.3 Scale analysis535
We used he wa ele analysis o esol e he scales o mo ion du ing 1 June 2003, 1500536
UTC - 5 June, 2003 1500 UTC wi h da a om ye, maize and g assland s a ions.537
The wa ele analysis o ye and g assland s a ions a e shown in Fig. 5 and Fig. 6538
espec i ely. Resul s o he maize ield and he ye ield a e e y simila . F om hese539
wa ele c oss-scalog ams, we ound small and la ge scales o mo ion. The size o he540
small one is a ound a ew minu es, which should be cap u ed by he eddy-co a iance541
measu emen o e 30 minu e a e aging ime. I appea s du ing he day ime a all542
si es and anspo s bo h QHand QE. The size o he la ge scale is app oxima ely543
a day, and mainly anspo s QH. I ends o inc ease QHin he maize and ye si es,544
while dec easing QHo e g ass. This con o ms o he pa e ns o ˜
QHand he block545
135

Ex ension o he a e aging ime o eddy-co a iance measu emen s 19
log(a e aging ime in minu e)
Ene gy lux densi y (W m−2)
101102103104
−200
−100
0
100
200
(a)Lake
101102103104
−20
0
20
40
60
80
100
120
140
(b)Fo es
101102103104
−50
0
50
100
(c)Rye
101102103104
−50
0
50
100
150
(d)Maize
101102103104
−20
0
20
40
60
80
(e)G ass
Qh Qe Res
101102103104
0
20
40
60
80
( )Rape
Fig. 4 Block ensemble a e ages o sensible hea lux and la en hea lux (Eq. 17 wi h cis empe a u e and
absolu e humidi y), and hei co esponding esiduals, du ing 1 June 2003, 1500 UTC - 5 June 2003, 1500
UTC o selec ed si es in he LITFASS-2003 expe imen : (a) lake, (b) o es , (c) ye, (d) maize, (e) g ass
and ( ) ba ley.
ensemble a e age luxes. This scale o mo ion would no be cap u ed by he eddy-546
co a iance measu emen a e aging o e he 30 minu e pe iod.547
E. Cha uchi ipan e al. (2013)
136
20 D. Cha uchi ipan e al.
Fig. 5 Wa ele c oss-scalog ams o e he ye ield du ing 1 June 2003, 1500 UTC - 5 June 2003, 1500
UTC o (a) sensible hea lux and (b) la en hea lux. The colou ep esen s he alue in W m−2. The black
solid line ep esen s he cone o in luence.
137
Ex ension o he a e aging ime o eddy-co a iance measu emen s 21
Fig. 6 Wa ele c oss-scalog ams o e he g assland du ing 1 June 2003, 1500 UTC - 5 June 2003, 1500
UTC o (a) sensible hea lux and (b) la en hea lux. The colou ep esen s he alue in W m−2. The black
solid line ep esen s he cone o in luence.
E. Cha uchi ipan e al. (2013)
138
22 D. Cha uchi ipan e al.
Bo h pa e ns om he Ho mølle diag am and wa ele analysis show he in-548
c ease o dec ease o ˜
QH. Howe e , hey do no ac ually show wha con ibu es o549
hese changes. Fo he u bulen luxes (w′c′), which a e caused by ins an aneous550
luc ua ions, we can ca y ou a quad an analysis by di iding ins an aneous con i-551
bu ions o w′c′in o ou quad an s o w′and c′(Shaw, 1985). Ou indings sugges 552
ha he main con ibu ion o closing he ene gy balance is ˜
QH, which is caused by553
block o block luc ua ions ( ˜wand ˜c). We he e o e di ided block o block con ibu-554
ions o ˜w˜cin o ou quad an s o ˜wand ˜c. We used ˜
T( empe a u e) and ˜a(absolu e555
humidi y) as ho izon al axes and ˜w( e ical eloci y) as a e ical axis, which ga e556
ou ou quad an s (Qi,i=1,...,4) as557
Q1: ˜w>0 and ˜
T>0 o ˜a>0 wa m ai ising o mois ai ising,558
Q2: ˜w>0 and ˜
T<0 o ˜a<0 cold ai ising o d y ai ising,559
Q3: ˜w<0 and ˜
T<0 o ˜a<0 cold ai sinking o d y ai sinking,560
Q4: ˜w<0 and ˜
T>0 o ˜a>0 wa m ai sinking o mois ai ising.561
Q1and Q3con ibu e o he posi i e lux, while Q2and Q4con ibu e o he nega i e562
lux. We hen no malized each axis by i s s anda d de ia ion and se he hype bolic563
hole size o 0.5 (H=0.5). We can neglec he weak con ibu ion by excluding he564
con ibu ion inside he hole and only conside ing any poin s which sa is y565

˜w˜
T
σ
˜w
σ
˜
T
o 
˜w˜a
σ
˜w
σ
˜a
>H.(18)
Wi h he quad an analysis, we expec o see which ypes o u bulence ac ually566
con ibu e o he inc easing o dec easing o ˜w˜c. To make i consis en wi h ou 567
Ho mølle diag ams, we used he same obse a ion pe iod NP, which is 20 May568
2003,1200 UTC -18 June 2003,0000 UTC, and se P o 30 minu es (non-o e lapped).569
The e o e, any poin s on he quad an analysis diag am ep esen he no malized ˜w˜c570
om each non-o e lapped 30 minu e pe iod.571
The esul s o he quad an analysis o ye and g assland s a ions a e shown in572
Fig. 7. In his igu e, we dis inguished all poin s du ing 1 June 2003, 1500 UTC -573
5 June 2003, 1500 UTC om he es by he use o ed colou do s. By conside ing574
only s ong con ibu ion ou side a hype bolic hole (blue line), we ound ha du ing575
his pe iod, ˜
QH( ia ˜w˜
T) has mo e con ibu ion om Q1(wa m ai ising) o he ye576
si es (Fig. 7 a), while he e a e mo e con ibu ions om Q4(wa m ai sinking) o 577
he g assland (Fig. 7 b). The e was no signi ican con ibu ion ou side he hype bolic578
hole o ˜
QE( ia ˜w˜a) in bo h ye and g ass s a ions. O e he maize ield, he quad an 579
analysis is simila o ha o he ye ield, while he lake is simila o he g assland.580
Fo bo h ape and o es (no shown), Q1and Q4equally con ibu ed o ˜
QH, wi h no581
signi ican con ibu ion ou side he hole o ˜
QE. These esul s ell us ha he inc ease582
o hQHia longe Po ye and maize ields we e caused by wa m ai nea he su ace583
ising, while he dec easing o hQHio g assland and lake we e caused by wa m584
ai alo sinking. Fo o es and ape s a ions, bo h con ibu ions om Q1and Q4
585
canceled each o he and keep hQHiapp oxima ely cons an a all P. The absence586
o signi ican con ibu ions ou side he hype bolic hole keeps hQEiapp oxima ely587
cons an a all si es.588
139