ICOS Science Conference 2020, Session 18

1. Variability in radial stem growth
– how much determination of the past is in the present response?
Roman Zweifel
Eidgenössische Forschungsanstalt für Wald, Schnee und Landschaft (WSL)
Sophia Etzold (WSL), Käthi Liechti (WSL), Anne Thimonier (WSL), Lukas Hörtnagl (ETH Zürich), Mana Gharun (ETH Zürich), Nina
Buchmann (ETH Zürich)
1
ICOS Science Conference, 17st of September 2020, Skype

2. 2
• Swiss Alps
• Norway spruce (Picea abies)
• 1640 m a.s.l.
• Average tree age 270 years
• Tree height max: around 30 m
• ICOS Class 1 vegetation site
• 22 years of continuous data
Subalpine site Davos (CH-DAV)

3. 3
Growth measurements
Water exchange
between wood
and bark
F∆Ψ XB
FTissue
∆RF∆Ψ AS
Wood (Xylem) Bark (Cambium, Phloem, Parenchyma)
Wood I Bark
Sapflow TWD
GRO[µm]
18 days
Tree water deficit
TWD
Radial stem growth (GRO)
GRO
• Automated
high-presicion
dendrometer
• 30 sec exec/10
min averages
• Growth includes wood and
bark increments
• Growth = cell expansion
(xylem and phloem)
• Detrended for water-
related variations
(Zweifel et al 2016, New Phytol.)

4. 0
5
10
15
0 100 200 300
annualGRO[mm]
●
●
●
2003
2015
2018
0
20
40
60
80
100
0 100 200 300
annualpercentageGRO[%]
●
●
●
2003
2015
2018
4
• Median radial
increment: 1.5 mm
yr-1
• 80% of growth in
Jun/Jul
• Largest C-uptake in
May
Variability in radial stem growth
01 02 03 04 05 06 07 08 09 10 11 12
020406080100
Percentmonthlycontribution(%)
month
NEE
DOY

5. 5
• Gradually changing
increments (with
one jump in 2006)
• Trend of increasing
growth (1997-2018)
Variability in annual tree growth
Rel.annualincrement(%)

6. y
6
• Trend of increasing
temperature (1997-
2018)
Variability in annual temperature
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
−2
−1
0
1
2
DeviationgsTemperature[°C]

7. 7
• Trend of decreasing
precipitation
(1997-2018)
Variability in annual precipitation
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
−200
−100
0
100
200
300
DeviationgsPrecipitation[mm]

8. 8
Variability in daily growth
Linear regression model (single variable):
Daily growth is most strongly related to RH (out of Temp, Precip, VPD, WB,
SWP, RH):
R2 = 0.19
Linear regression model:
• Atmospheric moisture
conditions mainly
determine/limit growth
• Less < 20% of variance
explained

9. 9
Explanation of variability in annual tree growth
Multi variables linear regression modelling
0
500
1000
1500
2000
0 500 1000 1500 2000
Predictedgrowth[um]
Observed growth [um]
adj.R2=0.54**
Current year meteorological variables
Rgwinter Precipwinter WBspring.min VPDspring.max
-434*** -320** -303** -236*
adj R2 = 0.54**
Best model with maximum predictors=4
• Annual and seasonal meteorological
variables:
• Temperature, Precipitation, Relative
Humidity, Radiation, VPD, Water balance
• Included:"rg.winter","precip_sum.winter
","temp.winter_min","temp.winter_max"
,"rg.winter_max",
"vpd.winter_max","vpd.winter"
,"WB.winter_min", "precip_sum.spring"
,"temp.spring_min","temp.spring_max",
"rg.spring_max","vpd.spring_max","vpd.
spring","WB.spring_min",
"precip_sum.summer","temp.summer_
min","vpd.summer","WB.summer_min",
"temp.veg_min" ,"rg.veg_max",
"vpd.veg_max","vpd.veg"

10. 0
500
1000
1500
2000
0 500 1000 1500 2000
Predictedgrowth[um]
Observed growth [um]
adj.R2=0.70***
10
Explanation of variability in annual tree growth
Multi variables linear regression modelling
0
500
1000
1500
2000
0 500 1000 1500 2000
Predictedgrowth[um]
Observed growth [um]
adj.R2=0.54**
Current year Current and past year variables
Rgwinter Precipwinter WBspring.min VPDspring.max
-434*** -320** -303** -236*
Precipautumn-1 Tempspring.min Tempspring.min-1 VPDautumn-1
-338*** 226*** -208*** -185**
adj R2 = 0.54** adj R2 = 0.70***

11. 0
500
1000
1500
2000
0 500 1000 1500 2000
Predictedgrowth[um]
Observed growth [um]
adj.R2=0.70***
11
Explanation of variability in annual tree growth
Multi variables linear regression modelling
Current year Current and past year variables
Precipautumn-1 Tempspring.min Tempspring.min-1 VPDautumn-1
-338*** 226*** -208*** -185**
adj R2 = 0.70***Conclusion …
• Past conditions considerably explain
growth
… raises new questions
• ???
• What do these delayed responses
mean?
• How can we link past conditions
into a physiolopgical model?

12. 12
Past conditions determine current response SCIENCE sciencemag.org 31 JULY 2015 • VOL 349 ISSUE 6247
What are potential
mechanisms behind
‘legacy effects’
‘ecological memory’
‘carry-over effects’?

13. 13
Explanation of variability in annual tree growth
Requirement
• Past conditions need to be stored in a tree to
affect today’s physiology
Idea
• Functional organs and reserves that are built up
over several years are the ‘ecological memory’

14. 14
Explanation of variability in annual tree growth
Ecological memory
• Crown: patchwork of needles from several years
• Sapwood: Collection of tree rings built over
many years
• Carbon storage: Accumulated and depleted over
decades
• Buds: predisposed in the year before

15. 15
Explanation of variability in annual tree growth
Today’s functionality depends on conditions of
many years back in time
• Crown: needle size/shape changes with
conditions
• Sapwood: a tree ring of a dry year is different
from a tree ring of a wet year -> different
hydraulic properties

16. 16
Explanation of variability in annual tree growth
Conclusion
• In such an approach a good/poor year keeps its
effect on future physiological responses as long
as the structure that was built in this year is kept
functional.
• KEY is lifetime/turnover rate of organs and
reserves

17. 17
Explanation of variability in annual tree growth

18. 18
Explanation of variability in annual tree growth
Expectation
• A tree with short organ/reserve
lifetimes will closely respond to
current environmental conditions
Expectation
• A tree with long organ/reserve
lifetimes will respond with delay to
current environmental conditions

19. 19
Explanation of variability in annual tree growth
Model findings
• The longer the lifetimes of organs and reserves are, the stronger is the legacy
effect of past conditions on the physiological response
• The stronger the legacy effect is, the lower is the explanatory power of current
conditions on growth

20. 20
Pine trees:
• Responded with a delayed
growth response on a
treatment change of 4 years.
• Pines’ needle turnover rate
was 4-5 years!!
Explanation of variability in annual tree growth

21. 21
Case of Davos spruce is not
solved yet:
• No experimental change
of conditions possible
• More difficult to pinpoint
influences from the past
• However, crown size may
also here play a role:
relationship between
crown transparency of
the past year is more
closely related to stem
growth than the current
year one.
●
●
●
●
● ●
●
●
●
●
●
●
●●
●
●
●
● ●
●
●
●
●
●
●
●
10
15
20
25
30
35
40
2006 2008 2010 2012 2014 2016 2018
meancrowntransparency(%)
●
●
●
●
●
●
●
●
●
●
●
● ●
●
600
800
1000
1200
1400
1600
meanGRO.yr(mm)
●
●
crown transparency
growth
R2 GRO vs current crown transparency: 0.26
R2 GRO vs past year crown transparency: 0.52
Variability in radial stem growth
– how much determination of the past is in the present response?

22. 22
Take home message
• Past conditions have an effect on the actual growth
performance of trees
• Legacy effect: trees’ current physiology is altered by
it’s history
• Past conditions are suggested to be stored in
functional structures e.g. needles, sapwood, carbon
reserves
• The longer the lifetime/turnover rate of these
structures, the more current growth is decoupled
from current conditions and the more it can be
explained by past environmental conditions.
Variability in radial stem growth
– how much determination of the past is in the present response?

23. Variability in radial stem growth
– how much determination of the past is in the present response?
Roman Zweifel
Eidgenössische Forschungsanstalt für Wald, Schnee und Landschaft (WSL)
Sophia Etzold (WSL), Käthi Liechti (WSL), Anne Thimonier (WSL), Lukas Hörtnagl (ETH Zürich), Mana Gharun (ETH Zürich), Nina
Buchmann (ETH Zürich)
23
ICOS Science Conference, 17st of September 2020, Skype