Tempo a y educ ions o s em CO2 e lux du ing ain all e en s ac oss ee species
1
2
E a Da eno a
3
Global Change Resea ch Ins i u e CAS, . .i., Belidla 4a, 603 00 B no, Czech Republic
4
da eno a[email p o ec ed]
5
6
7
8
Abs ac
9
S em espi a ion is an impo an componen o o es ca bon cycling, ye i s accu a e quan i ica ion
10
emains challenging. While s em espi a ion is commonly measu ed as CO2 e lux om he s em
11
su ace (EA), se e al p ocesses can decouple measu ed luxes om ac ual espi a o y ac i i y. In
12
his s udy, we in es iga ed he in luence o ain all e en s on EA ac oss empe a e ee species and
13
assessed he po en ial consequences o seasonal ca bon budge es ima es. Con inuous au oma ed
14
chambe measu emen s e ealed dis inc sho - e m dec eases in EA coinciding wi h ain e en s.
15
Time-lapse came a eco ds con i med ha s em su aces we e equen ly co e ed by a wa e ilm
16
du ing ain all, sugges ing ha we ba k empo a ily es ic ed CO2 di usion o he a mosphe e.
17
Rain- ela ed dec eases in EA accoun ed o up o 21% o he a ailable seasonal da a (May–Oc obe )
18
and caused up o 12% unde es ima ion o seasonal sums o measu ed EA. The la ges impac was
19
obse ed in beech, ollowed by ho nbeam and sp uce, whe eas oak and ash showed he smalles
20
unde es ima ion. Ou esul s demons a e ha ain all can sys ema ically bias EA measu emen s
21
and lead o unde es ima ion o seasonal and annual ca bon luxes om ee s ems. Recognizing
22
and accoun ing o hese empo a y bu epea ed e en s is he e o e essen ial o imp o ing he
23
accu acy o o es ca bon budge assessmen s unde bo h cu en and u u e p ecipi a ion egimes.
24
25
Keywo ds
26
Ba k we ing, ca bon lux, di usion esis ance, o es , ain all, s em espi a ion
27
28
29
1 In oduc ion
30
Fo es espi a ion is a key componen o he global ca bon cycle, as i ep esen s a majo lux o
31
ca bon dioxide (CO2) om ecosys ems o he a mosphe e. Al hough s em espi a ion gene ally
32
accoun s o only abou 5–25% o o al o es ecosys em espi a ion (Acos a e al., 2008; Khomik
33
e al., 2010; Kim e al., 2021; Rod íguez-Calce ada e al., 2014; Song e al., 2023; Yang e al.,
34
2016; Zha e al., 2007), i plays a non-negligible ole.
35
S em espi a ion is mos commonly measu ed as CO2 e lux (EA) om he s em su ace using
36
chambe sys ems. This me hod p o ides aluable insigh s in o bo h sho - e m dynamics and long-
37
e m seasonal ends. Howe e , EA does no necessa ily equal he ac ual espi a ion o s em issues
38
a he measu emen poin . Se e al p ocesses can al e he ela ionship, including anspo o
39
dissol ed CO2 by he xylem sap (Teskey e al. 2008; Ta ainen e al. 2018), pa ial CO2 e ixa ion
40
by pho osyn he ically ac i e cells in he ba k (Duka e al., 2024), and di usional limi a ions wi hin
41
woody issues (S eppe e al., 2007). Ne e heless, EA emains he mos widely applied and
42
echnically easible indica o o s em espi a ion.
43
Tempo al EA dynamics is p ima ily d i en by empe a u e (Da eno a e al., 2018; Kim e al., 2021).
44
Du ing he pe iod o new wood o ma ion, EA is enhanced due o ex a ene gy demands (Da eno a
45
e al., 2020; Rod íguez-calce ada e al., 2019). F om ou expe ience and p e ious s udies, he
46
diu nal cou ses o EA ollows he empe a u e luc ua ions. We, howe e , obse ed unexpec ed
47
empo al dec eases in ou EA da ase s om au oma ed con inuous measu emen sys ems. These
48
da a we e no lagged by he da a quali y indica o , he e o e, hey we e no assessed as e o
49
measu emen s. We did no ind any eason s em espi a ion o d op suddenly. The dec eases,
50
howe e , coincided wi h ain e en s. Time-laps came a obse a ions con i med ha he s em
51
su ace was co e ed by a wa e ilm (Fig. S1). We assumed ha such su ace we ing empo a ily
52
es ic ed EA by limi ing CO2 di usion o he a mosphe e. This phenomenon has so a ecei ed
53
li le a en ion. S eppe e al. (2007) and Salomón e al. (2016) in es iga ed he ole o woody issue
54
and ba k wa e s a us in con olling CO2 di usion esis ance. They ound ha esis ance o adial
55
CO2 di usion was highe a nigh , when woody issue and ba k wa e ese oi s we e e illed in
56
he absence o anspi a ion, han du ing he day, when anspi a ion lowe ed wa e con en . This
57
pa e n e lec s he much slowe di usion o CO2 in wa e compa ed wi h ai (Nobel, 2009).
58
Rain- ela ed educ ions in measu ed EA may sys ema ically bias seasonal and annual o es ca bon
59
budge es ima es. Such disc epancies a e likely in luenced by in e annual and si e-speci ic a ia ion
60
in p ecipi a ion egimes and may also di e among ee species. The e o e, he main objec i e o
61
his s udy was o quan i y he e ec o ain all e en s on s em CO₂ e lux measu emen s in di e en
62
ee species. Speci ically, we aimed i) o iden i y and quan i y sho - e m dec eases in EA du ing
63
and immedia ely a e ain e en s, ii) o es ima e he con ibu ion o hese dec eases o o al
64
seasonal s em CO₂ e lux, and iii) o compa e he magni ude o he e ec among ee species wi h
65
con as ing ba k cha ac e is ics and c own a chi ec u e. The s udy included No way sp uce as a
66
ep esen a i e coni e , along wi h b oadlea ed species—beech and ho nbeam wi h smoo h ba k,
67
and oak and ash wi h ough ba k. Ou indings p o ide new insigh s in o he p ocesses ha al e
68
measu emen s o s em espi a ion and suppo mo e accu a e es ima es o o es ca bon luxes.
69
70
2 Ma e ials and Me hods
71
72
2.1 Si es
73
Measu emen s we e ca ied ou a h ee si es: a beech o es (Eu opean beech, Fagus syl a ica L.),
74
a sp uce o es (No way sp uce, Picea abies L.), and a b oadlea ed mixed o es domina ed by
75
common ho nbeam (Ca pinus be ulus L.), English oak (Que cus obu L.), and na ow-lea ed ash
76
(F axinus angus i olia Vahl). Si e and s and cha ac e is ics a e summa ised in Table S1.
77
78
2.2 CO2 e lux measu emen s
79
We analysed da a om six chambe s ins alled on beech and sp uce ees, and ou chambe s on
80
ho nbeam, oak, and ash ees. The chambe s we e a pa o he 12-chmabe au oma ed measu emen
81
sys ems. The con ol uni consis ed o an in a ed gas analyse (Li-840, LI-COR, Lincoln, NE,
82
USA), and a pe sonal compu e equipped wi h con ol so wa e, as well as addi ional analogue
83
inpu and digi al ou pu ha dwa e. The chambe closing was enabled by p essu e ai supplied by a
84
comp esso and egula ed h ough a sys em o al es, and pneuma ic pis ons ixed on he chambe s.
85
Each chambe (Fig. S2) comp ised a ame and a chambe head. The ames we e ec angula wi h
86
inne dimensions 7x12 cm on sp uce and 9x15 cm on he o he ee species, and hey had a neop ene
87
seals agains he ee unk. A g oo e a ew millime es wide, illed wi h ubbe , secu ed he seal
88
be ween he ame and chambe head. The chambe heads we e made o s ainless s eel and had a
89
hal -cylinde shape. Chambe s we e a ached o he ee unk a b eas heigh on he no he n side
90
using wo bel s.
91
Fo beech, ho nbeam, and sp uce (all wi h ela i ely smoo h ba k), he neop ene seal was su icien .
92
In con as , oak and ash had hick, ough ou e ba k ha needed ca e ul emo al wi h a chisel o
93
a oid damaging li ing issues. Ba k was emo ed only along he ame ci cum e ence whe e he
94
neop ene seal was applied; he a ea inside he ame emained un ouched.
95
S em CO2 e lux (EA) was measu ed sequen ially in all chambe s, wi h da a collec ed e e y 2 h a
96
each posi ion. A e chambe closu e, ai was ci cula ed be ween he chambe and he analyze ,
97
ollowed by a 60-second equilib a ion pe iod. The sys em hen eco ded 20 CO2 concen a ion
98
alues a 10-second in e als o e lux calcula ion, using he ollowing equa ion:
99
𝐸𝐴=𝑃∙𝑉∙(𝑐2−𝑐1)
𝑅∙𝑇∙𝑆∙𝑡 ,
100
whe e P is ai p essu e (Pa), V is he olume o he sys em (m3), c1 and c2 a e consecu i e CO2
101
concen a ions (mol mol–1), R is he mola gas cons an , T is sample ai empe a u e (K), S is he
102
s em su ace a ea enclosed by he ame (m2) and is measu emen in e al (10 s).
103
Fo each 200 s measu emen , mean EA and s anda d de ia ion (SD) we e calcula ed. SD se ed as
104
an indica o o da a quali y; i SD exceeded he e lux alue, he da a we e excluded. Ac oss 2020–
105
2024, 64–96% o seasonal da a we e a ailable o analyses (Table 1)
106
107
2.3 Me eo ological measu emen s
108
S em empe a u e senso s (PT 100, Sensi , Rožno pod Radhoš ěm, Czech Republic) we e
109
in eg a ed in o he s em CO2 e lux sys ems and measu ed empe a u e a he s em su ace benea h
110
he chambe s. P ecipi a ion was con inuously moni o ed wi h a ain gauge (Me One 386, Me One
111
Ins umen s, Inc., G an s Pass, OR, USA) ins alled on owe s abo e he o es s ands.
112
113
2.4 Da a p ocessing
114
The e ec o ain on s em CO₂ e lux was de e mined using he ollowing app oach (see also he
115
schema ic in Fig. S3):
116
1. De ec ion o ain- ela ed dec eases in EA
117
We iden i ied pe iods when measu ed s em EA showed unexpec ed declines ha coincided
118
wi h ain all e en s. These declines we e ecognized by disc epancies be ween he obse ed
119
EA and he alues expec ed om he empe a u e–e lux ela ionship.
120
2. Modelling he expec ed EA
121
To ob ain he e e ence e lux (Em), we i ed an exponen ial empe a u e–e lux model
122
using a 10-day mo ing window. Fo model i ing, we excluded da a a ec ed by he ain-
123
ela ed dec eases. This p o ided a smoo h, empe a u e-d i en p edic ion o EA unde
124
una ec ed condi ions.
125
3. Quan i ying he e ec o dec eases
126
Fo each ain- ela ed dec ease, we calcula ed he a io o he measu ed EA o he modelled
127
Em a he same ime. This a io ep esen s he ela i e educ ion o EA due o he p esence
128
o su ace wa e ilms du ing ain e en s.
129
4. Seasonal impac assessmen
130
We eplaced he pe iods o ain- ela ed dec eases wi h he co esponding modelled alues
131
(Em).
132
Using hese co ec ed ime se ies, we calcula ed seasonal sums o CO₂ e lux bo h wi h and
133
wi hou he ain- ela ed EA dec eases. The con ibu ion o he dec eases o seasonal o als
134
was de e mined by compa ing hese sums. Calcula ions o sums we e based only on
135
a ailable da a, wi hou gap- illing o missing alues
136
137
To de e mine he EA- empe a u e ela ionship, Q10 was calcula ed as:
138
𝑄10 = 𝑒10∙𝛼,
139
whe e α is a unc ion g ow h pa ame e de i ed om he exponen ial ela ionship be ween EA and
140
s em empe a u e.
141
A one-way epea ed measu es ANOVA was used o es di e ences be ween ee species in he
142
p opo ional impac o ain- ela ed EA dec eases on seasonal EA sums. S a is ical analyses we e
143
pe o med using SigmaPlo e sion 15.0 (Sys a So wa e Inc., San Jose, CA, USA).
144
145
146
147
3 Resul s
148
149
3.1 Mic ome eo ology
150
P ecipi a ion du ing he expe imen al seasons (May – Oc obe ) was gene ally highes in sp uce
151
o es and lowes in he b oad-lea ed mixed o es . Seasonal o als anged om 345 o 624 mm in
152
he beech o es , om 602 o 1066 mm in he sp uce o es , and om 304 o 517 mm in he b oad-
153
lea e mixed o es . A summa y o p ecipi a ion and he numbe o ainy days o each season is
154
p o ided in Table 1.
155
156
3.2 Rain- ela ed dec eases in EA
157
Rain- ela ed dec eases in EA accoun ed o up o 21% o he a ailable seasonal da a (May–Oc obe )
158
(Table 1). The la ges p opo ion was obse ed in beech (13 ± 5%), ollowed by ho nbeam (9 ±
159
2%), sp uce (8 ± 3%), ash (3 ± 1%), and oak (2 ± 2%). The bigges dec eases occu ed in beech,
160
whe e mos educ ions we e concen a ed be ween 40–60% (Fig. 1). In ho nbeam, dec eases
161
peaked a 20–30%, while in sp uce hey we e mo e e enly dis ibu ed, anging om 10 o 60%.
162
The smalles educ ions we e obse ed in oak and ash, gene ally be ween 10 and 30%.
163
The ain- ela ed dec eases in EA caused up o 12% unde es ima ion o seasonal sums o measu ed
164
EA (Fig. 2). The bigges impac was obse ed in beech wi h mean unde es ima ion o 5.3 ± 2.8%,
165
ollowed by ho nbeam (3.3 ± 1.6%) and sp uce (2.2 ± 2.4%). The smalles unde es ima ion was in
166
oak (0.3 ± 0.5%) and ash (1.1 ± 1.4%).
167
168
3.3 A e - ain all pe iod
169
To es whe he he empo a y inhibi ion o CO2 elease om s ems du ing ain e en s leads o an
170
inc ease in EA a e he ain all ends and he ba k d ies, we analyzed he di e ence be ween
171
measu ed and modeled EA in six hou s ollowing ain (mo e p ecisely h ee measu emen s a e
172
measu ed EA e u ned modeled alues). Wi hin six hou s a e ain all (co esponding o h ee EA
173
measu emen s), EA ended o inc ease in beech, sp uce, and ho nbeam, mos equen ly by 0–10%
174
(Fig. 1). In con as , no clea pos - ain inc eases we e obse ed in oak o ash.
175
176
3.4 E ec o ain- ela ed EA dec eases on Q10
177
To illus a e he impac o ain- ela ed EA dec eases on EA in e p e a ion, we selec ed one chambe
178
on beech ( he species wi h he s onges ain all e ec on EA) in 2024. We compa ed Q10 alues
179
calcula ed om he whole measu ed da ase and om a da ase in which he ain- ela ed dec eases
180
we e emo ed. Using a 10-day mo ing window, Q10 in he co ec ed da ase luc ua ed consis en ly
181
be ween 1.5 and 2.0 h oughou he season (Fig. 3-A). In con as , when based on all measu ed
182
da a, Q10 exhibi ed se e al la ge de ia ions, eaching alues as high as 15.4. A a mon hly ime
183
s ep, he disc epancy be ween he wo da ase s was educed bu emained subs an ial (Fig. 3-B).
184
185
4 Discussion
186
In ou s udy, we obse ed a empo al dec ease in EA ha was so s eep i could no be explained by
187
he ypical empe a u e d op ha o en accompanies ainy wea he . Mo eo e , he e is no
188
physiological eason o such a p onounced decline in li ing cell espi a ion, which is he main
189
sou ce o CO2 p oduc ion. On he con a y, (Salomón e al., 2016) obse ed an inc ease in s em
190
CO2 concen a ions a e ain. The explana ion mus he e o e be sough in ac o s a ec ing CO2
191
e lux om he s em o he a mosphe e.
192
A hin wa e ilm co e ing ee s ems du ing and sho ly a e ain e en s, o ba k sa u a ed wi h
193
wa e , can ep esen a empo a y ba ie o CO2 di usion om he s em o he a mosphe e. This
194
makes i mo e di icul o es ima e he ac ual amoun o CO2 espi ed by li ing s em cells and
195
eleased o he a mosphe e. The e ec o ba k wa e con en on adial CO2 di usion esis ance has
196
been con i med by (Salomón e al., 2019). Gene ally, he a e o CO2 di usion is app oxima ely
197
10⁴ imes slowe in wa e han in ai (Nobel, 2009). This undamen al di e ence in di usion a es
198
highligh s why e en a hin wa e ilm may subs an ially al e CO2 e lux dynamics.
199
O e all, EA was mos a ec ed by ain in beech, ollowed by ho nbeam and sp uce. The smalles
200
e ec was obse ed in ash and oak. T ee species and s and s uc u e s ongly in luence bo h he
201
amoun o p ecipi a ion eaching he s em and he in ensi y o s em low. In gene al, s em low is
202
highe in b oadlea es han in coni e s due o canopy mo phology (Ba bie e al., 2009; Kan o and
203
Šach, 2009). This likely explains he smalle ain e ec on EA in sp uce compa ed wi h beech and
204
ho nbeam, despi e highe seasonal p ecipi a ion a he sp uce si e. Di e ences among deciduous
205
species can be explained by ba k oughness (No osado á e al., 2023). The smoo h ba k o beech
206
and ho nbeam has a smalle soaking a ea and as e wa e low down he unk, whe eas he ough
207
ba k o oak and ash inc eases soaking capaci y and slows wa e uno . These ba k- ela ed
208
di e ences in wa e e en ion may he e o e con ol bo h he in ensi y and du a ion o ain- ela ed
209
EA educ ions. A u he me hodological in luence canno be excluded: in oak and ash, a ce ain
210
laye o ou e ba k was emo ed o achie e a igh seal o he chambe , which may ha e in oduced
211
an a i icial ba ie o s em low a ound he ame.
212
In beech, sp uce, and ho nbeam, he s onges ain e ec s we e obse ed in 2020, bo h in e m o
213
he p opo ion o da a classi ied as ain- ela ed EA dec eases and he magni ude o seasonal EA
214
unde es ima ion (Table 1). No ably, 2020 expe ienced he highes seasonal p ecipi a ion sum,
215
accompanied by a la ge numbe o ainy days compa ed wi h o he yea s. This esul highligh s
216
he ole o p ecipi a ion egime in a ec ing s em CO2 luxes.
217
The da a om ain- ela ed dec ease pe iods a e o good echnical quali y, bu hey do no ep esen
218
he ue empo al pa e n o CO2 p oduc ion in s ems. This aises he ques ion o whe he such da a
219
should be included in da ase s used o es ima ing ee CO2 budge s. We may expec ha CO2 no
220
eleased du ing we pe iods could la e be emi ed in a di e en loca ion o a a di e en ime.
221
Highe s em CO2 concen a ions ela i e o he a mosphe e due o adial di usion ba ie s a e well
222
known (S eppe e al., 2007), and empo a y s eng hening o hese ba ie s du ing ain can u he
223
be eleased, po en ially causing a ansien accele a ion o EA. Indeed, we obse ed inc eased EA
224
wi hin six hou s a e ain-supp essed EA pe iods. This pos - ain inc eases ypically anged om 0
225
o 20%, hough i was no consis en ly obse ed. While he EA d op du ing ain was apid, eco e y
226
o expec ed alues was o en mo e g adual, p obably depending on he speed o su ace d ying.
227
Mo eo e , he s em su ace inside he chambe ame did no d y uni o mly, which complica ed
228
he de e mina ion o he p ecise end o ain- ela ed we condi ions. Howe e , he pos - ain
229
inc eases we e gene ally smalle and sho e han he dec eases and did no appea o compensa e
230
o he p esumed CO2 accumula ion wi hin s ems. Thus, he ul ima e a e o he CO2 espi ed du ing
231
ain e en s emains unce ain and may in ol e i s dissolu ion in xylem sap and subsequen upwa d
232
anspo . CO2 can be hen eleased o he a mosphe e o e-assimila ed by he pho osyn he ically
233
ac i e cells la e in uppe s em pa s, b anches o lea es (Salomón e al., 2021; Teskey e al., 2008).
234
This would shi he spa ial and empo al pa e n o CO2 e lux beyond wha s em chambe s de ec .
235
Al hough he ain- ela ed dec eases accoun ed o only a ew pe cen o seasonal EA, including
236
hem in da ase s can s ongly bias models based on he EA– empe a u e ela ionship. Du ing ain,
237
ai empe a u e ypically d ops. I such pe iods a e included, he combina ion o lowe EA and lowe
238
empe a u e p oduces an a i icially s eep eg ession slope be ween EA and empe a u e. This can
239
subs an ially o e es ima e he empe a u e sensi i i y pa ame e Q10. Using a 10-day mo ing
240
window o Q10 calcula ion, inclusion o ain- ela ed dec eases can yield alues se e al imes
241
highe han he ue sensi i i y (Fig. 3). Longe ime windows educed his o e es ima ion bu did
242
no elimina e i . Howe e , longe pe iods a e also in luenced by addi ional ac o s, especially
243
di e ences in s em g ow h s a us (Da eno a e al., 2018). The e o e, ca e ul da a il e ing is
244
essen ial be o e applying EA– empe a u e models o a oid in oducing sys ema ic e o s in o
245
es ima es o o es ca bon dynamics.
246
247
5 Conclusions
248
Reduced CO2 e lux caused by we ba k gene ally led o only a sligh unde es ima ion – ypically
249
a ew pe cen – o seasonal CO2 emissions, which can be conside ed negligible. Howe e , ou
250
esul s demons a e ha in egions wi h high p ecipi a ion and in beech o es s, o in s ands
251
domina ed by ee species wi h ine ba k, his unde es ima ion can become subs an ial.
252
Fu he mo e, including ain- ela ed s em CO2 e lux da a wi hou co ec ion can signi ican ly bias
253
models ha ely on he ela ionship be ween CO2 e lux and empe a u e.
254
