Geosci. Ins um. Me hod. Da a Sys ., 5, 417–426, 2016
www.geosci-ins um-me hod-da a-sys .ne /5/417/2016/
doi:10.5194/gi-5-417-2016
© Au ho (s) 2016. CC A ibu ion 3.0 License.
Digi al pho og aphy o assessing he link be ween ege a ion
phenology and CO2exchange in wo con as ing no he n
ecosys ems
Maiju Linkosalmi1, Mika Au ela1, Juha-Pekka Tuo inen1, Mikko Pel oniemi2, Cemal M. Tanis1, Ali N. A slan1,
Pasi Kola i3, K is in Bö che 4, Tuula Aal o1, Juuso Rainne1, Juha Ha akka1, and Tuomas Lau ila1
1Finnish Me eo ological Ins i u e, Helsinki, Finland
2Na u al Resou ces Ins i u e Finland (LUKE), Van aa, Finland
3Facul y o Biosciences, Uni e si y o Helsinki, Helsinki, Finland
4Finnish En i onmen Ins i u e (SYKE), Helsinki, Finland
Co espondence o: Maiju Linkosalmi (maiju.link[email p o ec ed])
Recei ed: 7 Decembe 2015 – Published in Geosci. Ins um. Me hod. Da a Sys . Discuss.: 11 Ma ch 2016
Re ised: 25 July 2016 – Accep ed: 3 Augus 2016 – Published: 12 Sep embe 2016
Abs ac . Digi al epea pho og aphy has become a widely
used ool o assessing he annual cou se o ege a ion phe-
nology o di e en ecosys ems. By using he g een ch o-
ma ic coo dina e (GCC) as a g eenness measu e, we exam-
ined he easibili y o digi al epea pho og aphy o assess-
ing he ege a ion phenology in wo con as ing high-la i ude
ecosys ems. Ecosys em–a mosphe e CO2 luxes and a ious
me eo ological a iables we e con inuously measu ed a bo h
si es. While he seasonal changes in GCC we e mo e ob ious
o he ecosys em ha is domina ed by annual plan s (open
we land), clea seasonal pa e ns we e also obse ed o he
e e g een ecosys em (coni e ous o es ). Daily and seasonal
ime pe iods wi h su icien sola adia ion we e de e mined
based on images o a g ey e e ence pla e. The a iabili y in
cloudiness had only a mino e ec on GCC, and GCC did
no depend on he sun angle and di ec ion ei he . The daily
GCC o we land co ela ed well wi h he daily pho osyn he ic
capaci y es ima ed om he CO2 lux measu emen s. A he
o es si e, he co ela ion was high in 2015 bu he e we e
disce nible de ia ions du ing he cou se o he summe o
2014. The yea - o-yea di e ences we e mos likely gene -
a ed by me eo ological condi ions, wi h highe empe a u es
coinciding wi h highe GCCs. In addi ion o depic ing he
seasonal cou se o ecosys em unc ioning, GCC was shown
o espond o en i onmen al changes on a imescale o days.
O e all, moni o ing o phenological a ia ions wi h digi al
images p o ides a powe ul ool o linking g oss p ima y
p oduc ion and phenology.
1 In oduc ion
Phenology is an impo an ac o in he ecology o ecosys-
ems. The mos dis inc i e phenomena comp ising ege a-
ion phenology a e he changes in plan physiology, biomass
and lea a ea (Miglia acca e al., 2011; Sonnen ag e al.,
2011, 2012; Baue le e al., 2012). In pa , hese changes d i e
he ca bon cycle o ecosys ems, and hey ha e a ious eed-
backs o he clima e sys em h ough e ec s on su ace albedo
and ae odynamic oughness, and ecosys em–a mosphe e ex-
changes o a ious gases (e.g. H2O, CO2and ola ile o ganic
compounds) (A ne h e al., 2010). Besides lea a ea, gas ex-
change is modula ed by seasonal a ia ions in pho osyn he-
sis and espi a ion (Richa dson e al., 2013). Globally, hese
a ia ions con ibu e o he luc ua ions in he a mosphe ic
CO2concen a ion (Keeling e al., 1996). In he long e m,
possible ends in ege a ion phenology can ha e a sys em-
a ic e ec on he mean CO2le el. Phenology u he plays
a ole in he compe i i e in e ac ions, ophic dynamics, e-
p oduc i e biology, p ima y p oduc ion and nu ien cycling
(Mo ise e e al., 2009). Phenological phenomena a e la gely
con olled by abio ic ac o s such as empe a u e, wa e a ail-
abili y and day leng h (B yan and Bai d, 2003; Kö ne and
Published by Cope nicus Publica ions on behal o he Eu opean Geosciences Union.
418 M. Linkosalmi e al.: Digi al pho og aphy o linking phenology and CO2exchange
Basle , 2010), and hus hey a e sensi i e o clima e change
(Richa dson e al., 2013; Rosenzweig e al., 2007; Migli-
a acca e al., 2012).
Se e al s udies ha e epo ed an ad anced onse o he
g owing season du ing ecen decades (Linkosalo e al.,
2009; Delba e al., 2008; No dli e al., 2008; Pudas e al.,
2008). An ea lie onse o g ow h has been obse ed o play a
signi ican ole in he annual ca bon budge o empe a e and
bo eal o es s, while leng hening au umns ha e a less clea
e ec (Goulden e al., 1996; Be ninge , 1997; Black e al.,
2000; Ba e al., 2007; Richa dson e al., 2009). This can
be explained by he apid C accumula ion ha s a s as soon
as condi ions u n a ou able o pho osyn hesis and g ow h
in sp ing, while he opposing e ec , i.e. ecosys em espi a-
ion, becomes inc easingly impo an in summe and au umn
(Whi e and Nemani, 2003; Dunn e al., 2007).
In gene al, moni o ing o ege a ion changes by digi al
came as has become easible wi h he de elopmen o ad-
anced bu inexpensi e came as ha p oduce au oma ed and
con inuous eal- ime da a. I has been shown ha simple
ime-lapse pho og aphy can acili a e de ec ion o ege a-
ion phenophases and e en he ela ed a ia ions in CO2ex-
change (Winga e e al., 2015; Richa dson e al., 2007, 2009).
This p o ides new possibili ies o moni o ing and modelling
o ecosys em unc ioning, o e i ica ion o emo e sensing
p oduc s, and o analysis o ecosys em CO2exchange luxes
and ela ed balances. Especially dynamic ege a ion models
and simula ions o C cycle could be imp o ed by mo e ac-
cu a e in o ma ion on he iming o budbu s and lea senes-
cence, as simple empi ical pa ame e iza ions, ypically based
on deg ee days o he onse and o se da es o C up ake, a e
p esen ly used as indica o s o he g owing season s a and
end (Baldocchi e al., 2005; Delpie e e al., 2009; Richa d-
son e al., 2013).
Digi al came as p oduce ed-g een-blue (RGB) colou
channel in o ma ion, om which di e en g eenness indices
can be calcula ed. Fo example, canopy g eenness has been
exp essed in e ms o he so-called g een ch oma ic coo di-
na e (GCC), which has been ela ed o ege a ion ac i i y
and u he o ca bon up ake o o es s (Richa dson e al.,
2007, 2009; Ah ends e al., 2009; Ide e al., 2011) and pea -
lands (Sonnen ag e al., 2011; Peichl e al., 2015). In decidu-
ous o es s, he main d i e o gas exchange is lea a ea ha
changes apidly in sp ing and au umn, which is easy o de-
ec . In e e g een coni e o es s he lea a ea changes a e
much smalle , so i is no ob ious whe he a simila ela ion-
ship can be es ablished o hem. In a pea land en i onmen ,
epea images ha e been used o map he mean g eenness o
mi e ege a ion o e a wide a ea (Peichl e al., 2015). Fo
pea land ecosys ems wi h a he e ogeneous ege a ion co e ,
i may be possible o simul aneously de ec seasonali y e -
ec s o di e en ege a ion ypes. Thus digi al epea im-
ages o di e en ially de eloping ege a ion ypes could po-
en ially help decompose an in eg a ed CO2 lux obse a ion
in o componen s alloca ed o hese ege a ion ypes.
Compa isons o phenological obse a ions made in con-
as ing ecosys ems a e needed o highligh ing he pheno-
logical ea u es ha can be ex ac ed om came a moni o ing
a di e en si es (Winga e e al., 2015; Keenan e al., 2014;
Toomey e al., 2015; Sonnen ag e al., 2012). Di e ences in
he ecosys em cha ac e is ics may also a ec he ideal se -up
o came as and he in e p e a ion o images, o example in
conjunc ion wi h su ace lux da a.
The objec i es o his s udy we e o (1) e alua e he digi-
al epea pho og aphy as a me hod o moni o ing he phe-
nology o bo eal ege a ion a high la i udes, (2) in es iga e
he di e ences in he phenology be ween wo adjacen bu
con as ing ecosys ems (pine o es and we land) loca ed in
no he n Finland, and (3) assess whe he he da a ob ained
om such came as can suppo he in e p e a ion o he mi-
c ome eo ological measu emen s o CO2 luxes conduc ed a
he si es.
This pape is s uc u ed as ollows: Sec . 2 in oduces he
measu emen si es, came a se -up, image analysis, and he
CO2 lux and me eo ological da a employed; Sec . 3 p o ides
he esul s and ela ed discussion, including es s o he mon-
i o ing sys em and an analysis o he obse ed phenologi-
cal de elopmen in ela ion o CO2exchange; inally, Sec . 4
p esen s he conclusions eme ging om his s udy.
2 Ma e ials and me hods
2.1 Measu emen si es
The s udy si es we e loca ed a Sodankylä in no he n Fin-
land, 100 km no h o he A c ic Ci cle. They ep esen wo
con as ing ecosys ems, a Sco s pine (Pinus syl es is) o es
(67◦21.7080N, 26◦38.2900E; 179 m a.s.l.) and an open p is-
ine we land (67◦22.1170N, 26◦39.2440E; 180 m a.s.l.). The
long- e m (1981–2010) mean empe a u e and p ecipi a ion
wi hin he a ea a e −0.4 ◦C and 527 mm, espec i ely (Pi i-
nen e al., 2012).
The Sco s pine s and is loca ed on lu ial sandy podzol
and has a dominan ee heigh o 13 m and a ee densi y
o 2100 ha−1. The age o he ees wi hin he came a scope
is abou 50 yea s. A single-sided lea a ea index (LAI) o
1.2 m2m−2has been es ima ed o he s and based on a o es
in en o y in 2000. The spa se g ound ege a ion consis s o
lichens (73 %), mosses (12 %) and e icaceous sh ubs (15 %).
The we land si e is loca ed on a meso ophic en ha ep-
esen s ypical no he n aapa mi e. The ege a ion a his si e
mainly consis s o low species (Ca ex spp., Menyan hes i-
olia a, And omeda poli olia, Be ula nana, Vaccinium oxy-
coccos, Sphagnum spp.). The e a e no all ees, only some
B. pubescens and a ew isola ed Sco s pines. Di e en ypes
o ege a ion a e loca ed on d ie (s ings) and we e ( la ks)
pa s o he we land.
The physical su ace s uc u e (ae odynamic oughness
leng h) di e s be ween he pine o es and we land si es.
Geosci. Ins um. Me hod. Da a Sys ., 5, 417–426, 2016 www.geosci-ins um-me hod-da a-sys .ne /5/417/2016/
M. Linkosalmi e al.: Digi al pho og aphy o linking phenology and CO2exchange 419
Also, he mic oclima e and su ace exchange o CO2and
sensible and la en hea di e due o di e en ege a ion and
soil cha ac e is ics.
2.2 Came a se -up
The images analysed in his s udy we e aken au oma i-
cally wi h S a Do Ne cam SC 5 digi al came as. The se -
up included a wea he -p oo housing and connec ions o line
cu en and a web se e . The pic u es we e s o ed in he
8 bi JPEG o ma e e y 30 min wi h 2592 ×1944 esolu ion
and ans e ed au oma ically o a emo e se e . The daily
collec ing pe iod a ied acco ding o he ime o he yea
oughly co e ing he dayligh hou s.
A he o es si e, he came as we e moun ed o a owe
a wo di e en heigh s: 29 m (“canopy came a”) and 13 m
(“c own came a”). The iewing angle o he canopy came a
was 45◦ om he ho izon al plane, while he c own came a
was posi ioned nea ly ho izon ally. The images o he canopy
came a co e ed pa s o he o es canopy and some gen-
e al landscape. The c own came a was ocused o indi id-
ual ees o de ec hei phenological de elopmen (e.g. bud
bu s , shoo g ow h, needle shedding) mo e closely. A he
we land si e, he came a was adjus ed in an angle o 45◦on
op o a 2 m pole. This came a mos ly obse ed he g ound
ege a ion, wi h some B. pubescens and sky also isible in
he images. All came as we e placed acing he no h o min-
imize lens la e and maximize illumina ion o he canopy.
2.3 G ey e e ence pla es
A he o es si e, g ey e e ence pla es we e employed o
moni o he s abili y o he image colou channels. The pla es
we e a ached o he came as in such a way ha hey a e isi-
ble in e e y pic u e. The idea behind he e e ence pla es was
o de ec possible day- o-day shi s in he colou balance due
o changing wea he condi ions, such as adia ion a ia ions.
The e e ence images should also no show any ob ious sea-
sonali y (Pe ach e al., 2014). The g ey colou o he pla es
was close o he “ ue g ey” in a sense ha i has an equal mix
o ed, g een and blue colou componen s. To achie e his,
he e e ence pla es we e pain ed wi h Tikku ila g ey/1948
(RGB alues: R =95, G =95, B =95).
2.4 Au oma ic image analysis
The digi al images we e analysed wi h he FMIPROT so -
wa e ha has been designed as a oolbox o image p ocess-
ing o phenological and me eo ological pu poses (Tanis and
A slan, 2016). FMIPROT calcula es he colou ac ions o
ed, g een and blue channels. In he p esen analysis we use
he GCC de ined as
GCC =PG
PR+PG+PB,(1)
And omeda poli olia
Be ula pubescens
Figu e 1. View om he we land came a. The solid lines indica e
ou egions o in e es de ined acco ding o ege a ion ypes. The
dashed lines indica e he egion o in e es ha includes all ege a-
ion ypes excep Be ula pubescens.
whe e PG, PR and PB a e he sums o g een, ed and
blue channel digi al numbe s, espec i ely, o all pixels com-
p ising an image.
Wi hin each image, i was possible o de ine limi ed sub-
a eas o egions o in e es (ROIs). The ROI ea u e o
FMIPROT makes i possible o limi he GCC calcula ion o
an a ea ha ep esen s a homogeneous ege a ion a ea. I also
p o ides an op ion o analysing se e al suba eas wi hin he
image simul aneously.
2.5 Selec ion o he egion o in e es
A he we land si e, GCC was calcula ed sepa a ely o ou
di e en , clea ly iden i iable ege a ion ypes. These eg-
e a ion ypes we e domina ed by (1) bog osema y (An-
d omeda poli olia) and o he sh ubs, (2) sedges (Ca ex spp.)
and Sphagnum mosses, (3) big-lea ed bogbean (Menyan hes
i olia e), and (4) downy bi ch (Be ula pubescens) (Fig. 1).
The i s h ee ROIs also included o he g ound ege a ion,
while he ou h ROI was limi ed o he bi ch canopy. The
GCC alues we e also analysed om a la ge a ea ha in-
cludes he h ee i s ege a ion ypes (Fig. 1).
The o es si e had wo came as, one zoomed o he c own
o a pine ee (Fig. 2) and he o he p o iding a gene al iew
o he canopy (Fig. 3). F om he gene al canopy image, h ee
sepa a e ROIs we e subjec i ely selec ed wi h an aim o de-
ine simila homogenous a eas o o es canopy (Fig. 3).
2.6 CO2 lux measu emen s
The ecosys em–a mosphe e CO2exchange was measu ed a
bo h s udy si es by he mic ome eo ological eddy co a i-
ance (EC) me hod. The EC measu emen s p o ided con inu-
ous da a on he CO2 luxes a e aged on an ecosys em scale.
The e ical CO2 lux is ob ained as he co a iance o he
www.geosci-ins um-me hod-da a-sys .ne /5/417/2016/ Geosci. Ins um. Me hod. Da a Sys ., 5, 417–426, 2016
420 M. Linkosalmi e al.: Digi al pho og aphy o linking phenology and CO2exchange
Figu e 2. View om he pine o es c own came a. The line indi-
ca es he egion o in e es .
high- equency (10Hz) luc ua ions o e ical wind speed
and CO2mixing a io (Baldocchi, 2003). A bo h si es, he
EC measu emen sys ems consis ed o a USA-1 (METEK
GmbH, Elmsho n, Ge many) h ee-axis sonic anemome-
e / he mome e and a closed-pa h LI-7000 (LI-COR, Inc.,
Lincoln, NE, USA) CO2/H2O gas analyze . The measu e-
men sys ems and he da a p ocessing p ocedu es ha e been
p esen ed in de ail by Au ela e al. (2009).
The CO2 luxes ob ained om he EC measu emen s ep-
esen he ne ecosys em exchange (NEE) o CO2, which is
he sum o g oss pho osyn he ic p oduc ion (GPP) by plan s
and a espi a ion e m ha includes bo h he au o ophic es-
pi a ion by plan s and he he e o ophic espi a ion by mi-
c obes. GPP is ypically de i ed om he NEE da a by using
a dedica ed lux pa i ioning echnique, o example based on
nonlinea eg essions wi h pho osyn he ic pho on lux den-
si y (PPFD) and ai empe a u e as p edic o s (Reichs ein e
al., 2005). Ins ead o pe o ming such an explici pa i ion-
ing, we de e mined he daily GPP in e ms o he g oss pho-
osyn hesis index (GPI); o de ails, see Au ela e al. (2001),
whe e a simila index was e med “PI”. GPI indica es he
maximal pho osyn he ical ac i i y in op imal adia ion con-
di ions. I is ob ained by calcula ing he di e ences o he
daily a e ages o he day ime (PPFD >600 µmol m−2s−1,
which limi ep esen s ligh sa u a ion o pho osyn hesis a
ou si es) and nigh - ime (PPFD <20 µmol m−2s−1)NEE.
The esul ing GPI scales well wi h he maximal GPP ob-
ained om a adi ional NEE pa i ioning, despi e he day–
nigh di e ences in espi a ion. GPI p o ides a use ul mea-
su e especially o depic ing he seasonal GPP cycle, bu as
i is obus agains missing da a, i also es ima es pho osyn-
he ic ac i i y du ing as changes due o sho - e m a ia-
ions in ai empe a u e and humidi y (Au ela e al., 2001).
A
C
B
Figu e 3. View om he pine o es canopy came a. The lines indi-
ca e h ee egions o in e es .
2.7 Me eo ological measu emen s
An ex ensi e se o suppo ing me eo ological a iables was
measu ed a bo h measu emen si es, including ai empe -
a u e and humidi y, a ious soil pa ame e s ( empe a u e,
humidi y, soil hea lux and wa e able le el) and di e -
en adia ion componen s (incoming and ou going sho wa e
(SW) adia ion, PPFD and ne adia ion). He e we used he
empe a u e da a measu ed a 3 m heigh on he we land
(Vaisala, HMP155D) and a 8 m a he o es si e (Pen onic,
PT100). F om he SW adia ion measu emen s (Kipp & Zo-
nen, CM11) we calcula ed he su ace albedo as he p o-
po ion o inciden adia ion ha is e lec ed back o he a -
mosphe e by he unde lying su ace. In addi ion, ac ional
cloud co e (CL) da a we e a ailable om he nea by obse -
a o y.
3 Resul s and discussion
3.1 Tes ing he se -up
3.1.1 E ec o en i onmen al condi ions on GCC
An accu a e GCC obse a ion equi es a su icien illumina-
ion le el, which was he e ensu ed by selec ing only mid-day
(10:00–14:00 local win e ime) pho og aphs o u he anal-
ysis. This pe iod was de e mined on he basis o he GCC
o he g ey e e ence pla e in di e en adia ion condi ions
(Supplemen , Figs. S1–S3).
The in luence o cloudiness on GCC was es ima ed om
he da a collec ed in July 2014. This pa icula mon h was
selec ed o he es because July ep esen s he peak g ow-
ing season ( o bo h adia ion le els and LAI), and in
July 2014 sunny and cloudy days we e equally equen .
Based on he obse a ions o ac ional cloud co e ( ang-
Geosci. Ins um. Me hod. Da a Sys ., 5, 417–426, 2016 www.geosci-ins um-me hod-da a-sys .ne /5/417/2016/
M. Linkosalmi e al.: Digi al pho og aphy o linking phenology and CO2exchange 421
00 03 06 09 12 15 18 21 00
GCC (we land)
0.30
0.32
0.34
0.36
0.38
0.40
Sunny
Cloudy
00 03 06 09 12 15 18 21 00
GCC ( o es c own)
0.30
0.32
0.34
0.36
0.38
0.40
Sunny
Cloudy
a)(
b)(
Time (LWT)
Figu e 4. Mean (±s anda d de ia ion shown by e o ba s) diu nal
cycle o GCC du ing sunny and cloudy condi ions obse ed wi h
(a) he ee c own and (b) he we land came as in July 2014.
ing om clea sky wi h CL=0 o comple ely cloudy con-
di ions wi h CL =8), he images we e pooled o wo con-
as ing cloudiness g oups ep esen ing sunny (CL=0–1)
and cloudy (CL =7–8) condi ions. Du ing he daily pe iod
o 10:00–14:00, he di e ences in he mean GCC be ween
sunny and cloudy condi ions we e s a is ically insigni ican
(Mann–Whi ney U es ) (Fig. 4). The mean GCC di e ence
be ween he cloudy and sunny g oups was 0.0014 and 0.0011
o he en and o es , espec i ely. Sonnen ag e al. (2012)
ound an equi alen ly small, hough in pa s a is ically sig-
ni ican , di e ence be ween he diu nal GCC cycles o sunny
and o e cas si ua ions o hei deciduous and coni e ous
o es s.
The dependence o GCC on he sola angle wi h espec o
ROI was also es ima ed om he da a o July 2014 (Fig. 4).
The di e ence be ween he minimum and maximum alues
o he hou ly GCC means wi hin he day ime window was
0.0030 and 0.0020 o sunny and cloudy cases, espec i ely.
This is less han 5 % o he seasonal ampli ude o he GCC
cu e (0.069 be ween May and July) associa ed wi h phe-
nological g eening o he en. A he o es si e, he co e-
sponding alues we e 0.0022 (sunny) and 0.0012 (cloudy)
and, despi e he lowe annual ampli ude (0.024 be ween May
2014 2015
05 06 07 08 09 10 11 12 01 02 03 04 05 06 07 08 09 10
GCC (day ime a e age)
0.30
0.32
0.34
0.36
0.38
0.40
We land
And omeda
Ca ex
Menyan hes
Be ula pubescens
Win e
Figu e 5. Mean day ime (10:00–14:00, local win e ime) GCC o
di e en egions o in e es ( ege a ion ypes) du ing he measu e-
men pe iod o May 2014 o Oc obe 2015. We land e e s o com-
bined ROI shown in Fig. 5b. The g ey ci cles indica e he win e ime
da a ha a e in luenced by an insu icien ligh le el.
and July), he di e ence was less han 10 % o he seasonal
a ia ion.
3.1.2 Sensi i i y o GCC o selec ion o he egion o
in e es
The sensi i i y o he GCC alues o he selec ion o a sub-
a ea wi hin an image, i.e. a egion o in e es , was es ed
by compa ing he GCC calcula ed o di e en ege a ion
pa ches. In pa icula , we wan ed o examine, on he one
hand, whe he he o es images a e homogeneous and hus
insensi i e o he ROI de ini ion; on he o he hand, he we -
land images may p o ide an oppo uni y o simul aneously
obse e a ious mic oecosys ems inco po a ed in o a single
image.
The GCC alues o he we land ROIs de ined acco ding
o ege a ion ypes showed signi ican di e ences in he sea-
sonal cycle, bo h in he iming o he majo changes in sp ing
and au umn and in he magni ude o he maximum GCC
(Fig. 5). Fo example, downy bi ch had he ea lies g ow h
onse , while he big-lea ed bogbean had he la ges g owing-
season maximum. While he seasonal pa e ns o he GCCs
o di e en ROIs can be compa ed, he same may no be ue
o he absolu e GCC alues, which we e a ec ed by di -
e en iewing angles and dis ances o he a ge . To gain a
be e insigh in o he quan i a i e di e ences be ween di -
e en ROIs, hese ROI-speci ic GCC da a should be in es-
iga ed in conjunc ion wi h di ec ege a ion analysis (LAI,
biomass) and small-scale (chambe -based) CO2exchange
measu emen s. Fo u he analysis he e we chose o use he
la ge ROI combining h ee ege a ion ypes (Fig. 1), which
ma ches be e he a eally in eg a ing lux measu emen s.
The daily mean GCC alues o di e en o es canopy
ROIs emained e y simila h oughou he ime se ies
(Fig. 6). The GCC alues de e mined om c own images
di e ed om hose om he came a wi h a gene al canopy
iew, mos likely because he came as had di e en iewing
www.geosci-ins um-me hod-da a-sys .ne /5/417/2016/ Geosci. Ins um. Me hod. Da a Sys ., 5, 417–426, 2016

422 M. Linkosalmi e al.: Digi al pho og aphy o linking phenology and CO2exchange
05 06 07 08 09 10 11 12 01 02 03 04 05 06 07 08 09 10
GCC (day ime a e age)
0.30
0.32
0.34
0.36
0.38
C own
Canopy A
Canopy B
Canopy C
Win e ime
2014 2015
Figu e 6. Mean day ime (10:00–14:00, local win e ime) GCC al-
ues o di e en ROIs om wo o es came as du ing he measu e-
men pe iod o May 2014 o Oc obe 2015. The g ey ci cles indica e
he win e ime da a ha a e in luenced by an insu icien ligh le el.
angles and dis ances o he objec . The con ibu ion o g ound
is mixed wi h he canopy signal, which pa ially explains
why he GCC alues in he dis an canopy came a images
we e lowe han in he c own came a images. In win e , he e
was mo e snow isible behind he canopy in he smalle -scale
ROIs. Thus, we decided on using in u he analysis only he
images om he c own came a.
3.2 Phenological de elopmen
3.2.1 We land si e
As p e iously obse ed by Peichl e al. (2015), a he we land
he g owing season is clea ly disce nable in he de elopmen
o GCC da a (Fig. 7). GCC s a ed o inc ease as soon as he
we land ege a ion s a ed o assimila e CO2. This g ow h
onse ook place in May a e he snowmel , o which he
g ound albedo p o ides a sensi i e indica o by quan i y-
ing he p opo ion o inciden sola adia ion ha is e lec ed
back o he a mosphe e. Howe e , he onse was p eceded by
a sho pe iod o educed GCC alues, which we e associa ed
wi h he mois and da k soil.
The wa m spells du ing la e May and ea ly June in 2014
induced a apid eme gence and g ow h o annual plan s. De-
spi e he la e snowmel ha yea , by mid-June he g owing
season had de eloped much u he han in 2015. This di e -
ence is clea ly isible in he GCC as well as pho osyn he ic
ac i i y (GPI) da a (Fig. 7). The cold pe iod in la e June 2014
ceased his as de elopmen , which is also well e lec ed in
he GCC da a ha show a s abiliza ion and e en a empo a y
educ ion du ing ha pe iod. GPI shows a simila pa e n,
highligh ing he cohe ence be ween he g eenness obse a-
ion and he ac ual pho osyn he ic p ocesses.
Following he ea lie onse o he g owing season in 2014,
he peak o plan de elopmen was also obse ed ea lie
(Fig. 7). Howe e , he magni ude o he GCC maxima du -
ing he wo yea s was he same (0.385). F om mid-Augus
01 02 03 04 05 06 07 08 09 10 11 12
0.39
0.38
0.37
0.36
0.35
0.34
0.33
0.32
0.31
0.4
0.3
0.2
0.1
0.0
30
20
10
0
-10
-20
1.0
0.8
0.6
0.4
0.2
0.0
2014
2015
GCC
GPI (mgCO2 m-2s-1)
Ai empe a u e (°C)
Albedo
GCC
Ai empe a u e
GPI
Albedo
Mon h
A
B
A
Figu e 7. Mean day ime (10:00–14:00, local win e ime) GCC (o
he ROI shown in Fig. 1b) oge he wi h he daily mean ai empe -
a u e, g oss pho osyn hesis index (GPI) and albedo in 2014–2015
a he we land si e. The iangles indica e he da es o snowmel (A)
and snow appea ance (B). The g ey ci cles indica e he win e ime
da a ha a e in luenced by an insu icien ligh le el.
o mid-Sep embe , he a e o GCC decline was app oxi-
ma ely he same in 2014 and 2015. In mid-Sep embe , he
sligh ly highe GCC in 2014 can be a ibu ed o a wa m pe-
iod. By he i s sub-ze o alues in daily mean empe a u es,
he GCC had dec eased o i s minimum alue, close o he
sp ing ime minimum, and by he snow all in mid-Oc obe i
had s a ed inc easing owa ds he le el obse ed o he ully
snow-co e ed condi ions in sp ing.
P e ious obse a ions sugges ha GPP is well co ela ed
wi h he GCC o we lands, especially du ing sp ing (Peichl
e al., 2015). Ou esul s suppo hese obse a ions show-
ing a s ong ela ionship be ween he daily GCC and GPI
da a (Fig. S4), wi h a co ela ion coe icien o 0.90 and 0.92
o he snow- ee pe iod in 2014 and 2015, espec i ely. Es-
pecially du ing he sp ing ime, he ma ch be ween he GCC
and GPI ime se ies was ema kably close du ing bo h yea s,
while in he au umn o 2014 GPI lagged sligh ly behind
GCC.
3.2.2 Fo es si e
Due o he closeness o he measu emen si es, he me eo-
ological condi ions in o es we e simila o hose obse ed
a he we land (Figs. 7 and 8). Howe e , he onse o pho o-
syn he ical ac i i y di e ed sligh ly a he beginning o he
g owing season: he wa m days o ea ly May 2015 we e no
obse ed a he we land as an GPI inc ease due o he absence
o annual ege a ion igh a e he snowmel , while he pho-
osyn hesis o bo eal ees is igge ed as soon as empe a u e
Geosci. Ins um. Me hod. Da a Sys ., 5, 417–426, 2016 www.geosci-ins um-me hod-da a-sys .ne /5/417/2016/
M. Linkosalmi e al.: Digi al pho og aphy o linking phenology and CO2exchange 423
01 02 03 04 05 06 07 08 09 10 11 12
0.39
0.38
0.37
0.36
0.35
0.34
0.33
0.32
0.31
0.4
0.3
0.2
0.1
0.0
30
20
10
0
-10
-20
1.0
0.8
0.6
0.4
0.2
0.0
2014
2015
GCC
GPI (mgCO2 m-2s-1)
Ai empe a u e (°C)
Albedo
GCC
Ai empe a u e
GPI
Albedo
Mon h
12
A3
4
3
4
B
2
B
Figu e 8. Mean day ime (10:00–14:00, local win e ime) GCC
(c own came a) oge he wi h he daily mean ai empe a u e (a
18 m), g oss pho osyn hesis index (GPI) and albedo in 2014–2015
a he o es si e. The iangles indica e he s a da es o isually ob-
se ed phenological phases (1 – bud bu s , 2 – bud g ow h, 3 – shoo
g ow h, 4 – old needle b owning) and snow s a us (A – snowmel ,
B – snow appea ance). The g ey ci cles indica e he win e ime da a
ha a e in luenced by an insu icien ligh le el.
eaches a su icien le el (Tanja e al., 2003). Thus he g ow-
ing season in he o es s a ed ea lie in 2015 han in 2014,
while ha was no he case a he we land. Ne e heless, he
wa m pe iod in la e May–ea ly June 2014 also enhanced he
o es g ow h, and by mid-June bo h GCC and GPI had su -
passed he co esponding le el in 2015. The cold pe iod in
la e June 2014 was again obse ed as educed CO2up ake
and e en a clea e educ ion in GCC han a he we land.
Al hough in deciduous o es and open we lands GCC is
gene ally well co ela ed wi h he g oss ecosys em pho osyn-
hesis du ing he s a o he g owing season (Peichl e al.,
2015; Toomey e al., 2015), o e e g een needlelea o es s
i has been epo ed ha such co ela ion is o en weake
(Toomey e al., 2015; Winga e e al., 2015). In ou pine o es ,
howe e , he simul aneous de elopmen o GCC and pho o-
syn hesis was e iden du ing he yea wi h sp ing da a a ail-
able (Fig. S5).
Simila ly o he we land, he maximum GCC le el a he
o es si e did no di e be ween 2014 and 2015, bu his le el
was eached sligh ly ea lie in 2014. This was p obably due
o he highe empe a u es du ing he i s pa o he g ow-
ing season. Du ing bo h yea s, GCC s a ed dec easing a he
same ime, i.e. a he end o July. This was sligh ly ea lie
han he s a o he senescence de ec ed isually (Phase 4 in
Fig. 8). Simila ly o he we land, in 2014 he e was a clea
phase di e ence be ween GCC and GPI, he la e o which
s ayed a he maximum le el un il he end o Augus . In o -
es , his may be due o he oldes needles, whose senescence
akes place in Augus , while hei pho osyn he ic capaci y
has diminished al eady ea lie (Vesala e al., 2005).
In bo h 2014 and 2015, he pho osyn he ic ac i i y con in-
ues un il he end o Augus , bu he in e annual compa ison
is no possible he e owing o he missing CO2da a in 2015.
Ne e heless, in bo h yea s GCC dec eases o he win e ime
le el a he beginning o Oc obe , a he same ime as he
daily mean empe a u e dec eases below 0 ◦C.
Ou esul s show ha he phenological de elopmen o he
pine canopy could be accu a ely moni o ed wi h he GCC
analysis, e en hough he GCC changes in o es we e sub-
le han hose obse ed o he we land ege a ion. This was
con i med by isually iden i ying he phenological s ages o
he o es om he c own came a pic u es (Fig. 8). In 2014,
he came as we e ins alled oo la e o de ec he bud bu s , bu
he GCC ime se ies was consis en wi h he obse a ion ha
he buds s a ed hei g ow h a he beginning o June and e-
mained b own un il he beginning o July, when hey s a ed
o g een.
4 Conclusions
We demons a ed he easibili y o digi al epea pho og a-
phy o assessing he link be ween ege a ion phenology and
CO2exchange o wo con as ing high-la i ude ecosys ems.
While he seasonal changes in he g eenness index GCC a e
mo e ob ious o hose ecosys ems whe e he ege a ion is
enewed e e y yea (he e an open we land), seasonal pa -
e ns can also be obse ed in he e e g een ecosys ems (he e
a coni e ous o es ).
We examined he illumina ion sensi i i y o ou digi al
came a sys em by analysing he images o a g ey e e ence
pla e, which was included in he came a iew. Limi ed so-
la adia ion es ic s he use o images du ing he win e -
ime as well as du ing he nigh - ime. A ou si es in no he n
Finland, he day ime adia ion le els we e su icien o im-
age analysis om Feb ua y o Oc obe . Du ing ha pe iod,
a diu nal window o 10:00–14:00 (local win e ime) p o-
ides s able GCC da a. Ou esul s show ha he a iabili y
in cloudiness and sola zeni h angle du ing he day ime does
no play a signi ican ole in he GCC analysis. Howe e , i
would be ele an o in es iga e he seasonal dependence o
GCC on sun ele a ion, especially o he coni e ous o es .
We obse ed a clea seasonal GCC cycle a bo h s udy
si es. A he we land, GCC co ela ed well wi h he daily
pho osyn he ic capaci y es ima ed om he ecosys em–
a mosphe e lux measu emen s. The in e annual a ia ion in
GCC was also consis en wi h he obse ed CO2exchange
and me eo ological condi ions. A he o es si e, he sea-
sonal GCC cycle co ela ed well wi h he lux da a in 2015
bu showed mo e de ia ions du ing he summe o 2014. Fo
www.geosci-ins um-me hod-da a-sys .ne /5/417/2016/ Geosci. Ins um. Me hod. Da a Sys ., 5, 417–426, 2016
424 M. Linkosalmi e al.: Digi al pho og aphy o linking phenology and CO2exchange
bo h ecosys ems, he co ela ion be ween GCC and CO2ex-
change was highes du ing he sp ing.
In addi ion o depic ing he seasonal cou se o ecosys em
unc ioning, we showed ha GCC esponds o en i onmen-
al changes on a sho e imescale. We obse ed ha a bo h
si es he inc ease o GCC and pho osyn hesis ongoing in June
was ceased du ing a 2-week-long cold and we pe iod. Fo
an unknown eason, he GCC alues e en sligh ly dec eased
du ing ha pe iod. I is possible ha such a educ ion is an
a e ac caused by we su aces, o example, a he han a e-
sponse o an ac ual dec ease in he chlo ophyll concen a ion
in lea es and needles.
Due o he low cos o he ins umen a ion in ol ed, phe-
nology moni o ing can be es ablished in a much la ge num-
be o loca ions han ecosys em–a mosphe e lux measu e-
men s, hus p o iding a wide geog aphical basis o im-
p o emen o he phenological and pho osyn hesis compo-
nen s o land su ace models ha need mo e calib a ion and
alida ion. The digi al epea images allow he de ec ion o
phenological e en s, such as shoo elonga ion and he s a o
needle g ow h ha canno be ob ained om CO2 lux mea-
su emen s alone. The e o e, hey should be u ilized o en-
hance he analysis o lux da a. Fu he mo e, as ou esul s
show, he seasonal cycle o di e en ege a ion ypes wi hin
he oo p in o he lux measu emen s can be de e mined.
This could help decompose he in eg a ed CO2 lux obse a-
ions when he dis ibu ion o he ege a ion ypes wi hin he
a ea is known.
5 Da a a ailabili y
The da a p esen ed in his s udy a e p o ided as a supplemen
o his a icle.
The Supplemen ela ed o his a icle is a ailable online
a doi:10.5194/gi-5-417-2016-supplemen .
Acknowledgemen s. This wo k was suppo ed by he EU: he
ins alla ion o he came as and he de elopmen o he image
p ocessing ool (FMIPROT) was done wi hin MONIMET P ojec
(LIFE12ENV/FI/000409), unded by EU Li e+P og amme
(2013–2017) (h p://monime . mi. i).
Edi ed by: M. Pa on
Re iewed by: h ee anonymous e e ees
Re e ences
Ah ends, H. E., E zold, S., Ku sch, W. L., S oeckli, R., B uegge ,
R., Jeanne e , F., Wanne , H., Buchmann, N., and Eugs e , W.:
T ee phenology and ca bon dioxide luxes: use o digi al pho-
og aphy o p ocess-based in e p e a ion a he ecosys em scale,
Clim. Res., 39, 261–274, doi:10.3354/c 00811, 2009.
A ne h, A., Ha ison, S. P., Zaehle, S., Tsiga idis, K., Menon, S.,
Ba lein, P. J., Feich e , J., Ko hola, A., Kulmala, M., O’Donnell,
D., Schu ge s, G., So a i, S., Vesala, T.: Te es ial biogeochem-
ical eedbacks in he clima e sys em, Na . Geosci., 3, 525–532,
2010.
Au ela, M., Tuo inen, J.-P., and Lau ila, T.: Ne CO2exchange o a
suba c ic moun ain bi ch ecosys em, Theo . Appl. Clima ol., 70,
135–148, 2001.
Au ela, M., Lohila, A., Tuo inen, J.-P., Ha akka, J., Riu a, T., and
Lau ila, T.: Ca bon dioxide exchange on a no he n bo eal en,
Bo eal En i on. Res., 14, 699–710, 2009.
Baldocchi, D.: Assessing he eddy co a iance echnique o e alu-
a ing ca bon dioxide exchange a es o ecosys ems: pas , p esen
and u u e, Glob. Change Biol., 9, 479–492, 2003.
Baldocchi, D. D., Black, T. A., Cu is, P. S., Falge, E., Fuen es, J.
D., G anie , A., Gu, L., Knohl, A., Pilegaa d, K., Schmid, H. P.,
Valen ini, R., Wilson, K., Wo sy, S., Xu, L., and Yamamo o, S.:
P edic ing he onse o ca bon up ake by deciduous o es s wi h
soil empe a u e and clima e da a: a syn hesis o FLUXNET da a,
In . J. Biome eo ol., 49, 377–387, 2005.
Ba , A. G., Black, T. A., Hogg, E. H., G i is, T. J., Mo gens e n,
K., Kljun, N., Theede, A., and Nesic, Z.: Clima ic con ols on he
ca bon and wa e balances o a bo eal aspen o es , 1994–2003,
Glob. Change Biol., 13, 561–576, 2007.
Baue le, W. L., O en, R., Way, D. A., Qian, S. S., S oy, P. C., Tho n-
on, P. E., Bowden, J. D., Ho man, F. M., and Reynolds, R.
F.: Pho ope iodic egula ion o he seasonal pa e n o pho osyn-
he ic capaci y and he implica ions o ca bon cycling, P. Na l.
Acad. Sci. USA, 109, 8612–8617, 2012.
Be ninge , F.: E ec s o d ough and phenology on GPP in Pi-
nus syl es is: a simula ion s udy along a geog aphical g adien ,
Func . Ecol., 11, 33–43, 1997.
Black, T. A., Chen, W. J., Ba , A. G., A ain, M. A., Chen, Z., Nesic,
Z., Hogg, E. H., Neumann, H. H., and Yang, P. C.: Inc eased
ca bon seques a ion by a bo eal deciduous o es in yea s wi h a
wa m sp ing, Geophys. Res. Le ., 27, 1271–1274, 2000.
B yan , R. G. and Bai d, A. J.: The spec al beha iou o Sphagnum
canopies unde a ying hyd ological condi ions, Geophys. Res.
Le ., 30, 1134–1138, 2003.
Delba , N., Pica d, G., Le Toans, T., Ke goa , L., Quegan, S.,
Woodwa d, I., Dye, D., and Fedo o a, V.: Sp ing phenology in
bo eal Eu asia o e a nea ly cen u y ime scale, Glob. Change
Biol., 14, 603–614, 2008.
Delpie e, N., Du ene, E., Soudani, K., Ul ich, E., Cecchini, S.,
Boe, J., and F ancois, C.: Modelling in e annual and spa ial a i-
abili y o lea senescence o h ee deciduous ee species in
F ance, Ag . Fo es Me eo ol., 149, 938–948, 2009.
Dunn, A. L., Ba o d, C. C., Wo sy, S. C., Goulden, M. L., and
Daube, B. C.: A long- e m eco d o ca bon exchange in a bo eal
black sp uce o es : means, esponses o in e annual a iabili y,
and decadal ends, Glob. Change Biol., 13, 577–590, 2007.
Geosci. Ins um. Me hod. Da a Sys ., 5, 417–426, 2016 www.geosci-ins um-me hod-da a-sys .ne /5/417/2016/
M. Linkosalmi e al.: Digi al pho og aphy o linking phenology and CO2exchange 425
Goulden, M. L., Munge , J. W., Fan, S. M., Daube, B. C., and Wo sy,
S. C.: Measu emen s o ca bon seques a ion by long- e m eddy
co a iance: me hods and a c i ical e alua ion o accu acy, Glob.
Change Biol., 2, 169–182, 1996.
Ide, R., Nakaji, T., Mo ohka, T., and Oguma, H.: Ad an ages o
isible-band spec al emo e sensing a bo h sa elli e and nea -
su ace scales o moni o ing he seasonal dynamics o GPP in a
Japanese la ch o es , J. Ag . Me eo ol., 67, 75–84, 2011.
Keeling, C. D., Chin, J. F. S., and Who , T. P.: Inc eased ac i i y
o no he n ege a ion in e ed om a mosphe ic CO2measu e-
men s, Na u e, 382, 146–149, 1996.
Keenan, T. F., Da by, B., Fel s, E., Sonnen ag, O., F iedl, M.,
Hu kens, K., O’Kee e, J. F., Klos e man, S., Munge , J. W.,
Toomey, M., and Richa dson, A. D.: T acking o es phenology
and seasonal physiology using digi al epea pho og aphy: a c i -
ical assessmen , Ecol. Appl., 24, 1478–1489, 2014.
Kö ne , C. and Basle , D.: Wa ming, pho ope iods, and ee phenol-
ogy esponse, Science, 329, 278–278, 2010.
Linkosalo, T., Häkkinen, R., Te hi uo, J., Tuomen i a, H., and
Ha i, P.: The ime se ies o lowe ing and lea bud bu s o bo eal
ees (1846–2005) suppo he di ec empe a u e obse a ions o
clima ic wa ming, Ag . Fo es Me eo ol., 149, 453–461, 2009.
Miglia acca, M., Gal agno, M., C emonese, E., Rossini, M.,
Me oni, M., Sonnen ag, O., Manca, G., Dio i, F., Buse o, L.,
Cesca i, A., Colombo, R., Fa a, F., Mo a di Cella, U., Pa i,
E., Siniscalco, C., and Richa dson, A.: Using digi al epea pho-
og aphy and eddy co a iance da a o model g assland phenol-
ogy and pho osyn he ic CO2up ake, Ag . Fo es Me eo ol., 151,
1325–1337, 2011.
Miglia acca, M., Sonnen ag, O., Keenan, T. F., Cesca i, A.,
O’Kee e, J., and Richa dson, A. D.: On he unce ain y o phe-
nological esponses o clima e change, and implica ions o
a e es ial biosphe e model, Biogeosciences, 9, 2063–2083,
doi:10.5194/bg-9-2063-2012, 2012.
Mo ise e, J. T., Richa dson, A. D., Knapp, A. K., Fishe , J. I.,
G aham, E. A., Aba zoglou, J., Wilson, B. E., B eshea s, D. D.,
Heneb y, G. M., Hanes, J. M., and Liang, L.: T acking he hy hm
o he seasons in he ace o global change: Phenological esea ch
in he 21s cen u y, F on . Ecol. En i on., 7, 253–260, 2009.
No dli, O., Wielgolaski, F. E., Bakken, A. K., Hjel nes, S. H., Mage,
F., Si le, A., and Sk e, O.: Regional ends o bud bu s and low-
e ing o woody plan s in No way as ela ed o clima e change,
In . J. Biome eo ol., 52, 625–639, 2008.
Peichl, M., Sonnen ag, O., and Nilsson, M. B.: B inging Colo
in o he Pic u e: Using Digi al Repea Pho og aphy o In es i-
ga e Phenology Con ols o he Ca bon Dioxide Exchange in a
Bo eal Mi e, Ecosys ems, 18, 115–131, 2015.
Pe ach, A., Toomey, M., Aub ech , D., Richa dson, A. D: Moni-
o ing ege a ion phenology using an in a ed-enabled secu i y
came a, Ag . Fo es Me eo ol., 195, 143–151, 2014.
Pi inen, P., Simola, H., Aal o, J., Kauko an a, J.-P., Ka lsson, P., and
Ruuhela, R.: Clima ological s a is ics o Finland 1981–2010, Re-
po s 2012:1, Finnish Me eo ological Ins i u e, Helsinki, 2012.
Pudas, E., Leppälä, M., Tol anen, A., Poikolainen, J., Venäläinen,
A., and Kubin, E.: T ends in phenology o Be ula pubescens
ac oss he bo eal zone in Finland, In . J. Biome eo ol., 52, 251–
259, 2008.
Reichs ein, M., Falge, E., Baldocchi, D., Papale, D., Aubine ,
M., Be bigie , P., Be nho e , C., Buchmann, N., Gilmano , T.,
G anie , A., G ünwald, T., Ha ánko á, K., Il esniemi, H.,
Janous, D., Knohl, A., Lau ila, T., Lohila, A., Lous au, D., Ma -
eucci, G., Meye s, T., Miglie a, F., Ou ci al, J.-M., Pumpanen,
J., Rambal, S., Ro enbe g, E., Sanz, M., Tenhunen, J., Seu e , G.,
Vacca i, F., Vesala, T., Yaki , D., and Valen ini, R.: On he sepa a-
ion o ne ecosys em exchange in o assimila ion and ecosys em
espi a ion: e iew and imp o ed algo i hm, Glob. Change Biol.,
11, 1424–1439, 2005.
Richa dson, A. D., Jenkins, J. P., B aswell, B. H., Hollinge , D. Y.,
Ollinge , S. V., and Smi h, M.-L.: Use o digi al webcam images
o ack sp ing g een-up in a deciduous b oadlea o es , Oecolo-
gia, 152, 323–334, 2007.
Richa dson, A. D., Hollinge , D. Y., Dail, D. B., Lee, J. T., Munge ,
J. W., and O’Kee e, J.: In luence o sp ing phenology on sea-
sonal and annual ca bon balance in wo con as ing New England
o es s, T ee Physiol., 29, 321–331, 2009.
Richa dson, A. D., Keenan, T. F., Miglia acca, M., Ryua, Y., Son-
nen ag, O., and Toomey, M.: Clima e change, phenology, and
phenological con ol o ege a ion eedbacks o he clima e sys-
em, Ag . Fo es Me eo ol., 169, 156–173, 2013.
Rosenzweig, C., Casassa, G., Ka oly, D. J., Imeson, A., Liu, C.,
Menzel, A., Rawlins, S., Roo , T. L., Seguin, B., and T y-
janowski, P.: Supplemen a y ma e ial o chap e 1: Assessmen
o obse ed changes and esponses in na u al and managed sys-
ems. Clima e Change 2007: Impac s, Adap a ion and Vulne a-
bili y. Con ibu ion o Wo king G oup II o he Fou h Assess-
men Repo o he In e go e nmen al Panel on Clima e Change,
edi ed by: Pa y, M. L., Canziani, O. F., Palu iko , J. P., an de
Linden, P. J., and Hanson, C. E., Camb idge Uni e si y P ess,
Camb idge, UK, 2007
Sonnen ag, O., De o, M., Va gas, R., Ryu, Y., Runkle, B. R. K.,
Kelly, M., and Baldocchi, D. D.: T acking he s uc u al and
unc ional de elopmen o a pe ennial peppe weed (Lepidium
la i olium L.) in es a ion using a mul i-yea a chi e o webcam
image y and eddy co a iance measu emen s, Ag . Fo es Me eo-
ol., 151, 916–926, 2011.
Sonnen ag, O., Hu kens, K., Teshe a-S e ne, C., Young, A. M.,
F iedl, M., B aswell, B. H., Milliman, T., O’Kee e, J., and
Richa dson, A. D.: Digi al epea pho og aphy o phenological
esea ch in o es ecosys ems, Ag . Fo es Me eo ol., 152, 159–
177, 2012.
Tanis, C. M. and A slan, A. N.: FMIPROT – Finnish Me eo ological
Ins i u e Image P ocessing Tool, Use manual, a ailable a : h p://
monime . mi. i/index.php?s yle=wa m&page=FMIPROT, 2016.
Tanja, S., Be ninge , F., Vesala, T., Ma kkanen, T., Ha i, P., Mäkelä,
A., Il esniemi, H., Hänninen, H., Nikinmaa, E., Hu ula, T., Lau-
ila, T., Au ela, M., G elle, A., Lind o h, A., A ne h, A., Shibis-
o a, O., and Lloyd, J.: Ai empe a u e igge s he eco e y o
e e g een bo eal o es pho osyn hesis in sp ing, Glob. Change
Biol., 9, 1410–1426, 2003.
Toomey, M., F iedl, M., F olking, S., Hu kens, K., Klos e man,
S., Sonnen ag, O., Baldocchi, D., Be nacchi, C., Bi aud, S. C.,
Boh e , G., B zos ek, E., Bu ns, S. P., Cou solle, C., Hollinge ,
D. Y., Ma golis, H. A., McCaughey, H., Monson, R. K., Munge ,
J. W., Palla dy, S., Phillips, R. P., To n, M. S., Wha on, S., Ze i,
M., and Richa dson, A. D.: G eenness indices om digi al cam-
e as p edic he iming and seasonal dynamics o canopy-scale
pho osyn hesis, Ecol. Appl., 25, 99–115, 2015.
www.geosci-ins um-me hod-da a-sys .ne /5/417/2016/ Geosci. Ins um. Me hod. Da a Sys ., 5, 417–426, 2016