Spatial variability and temporal trends in water-use efficiency of European forests

TLDR: It is demonstrated that the combined effects of increasing CO2 and climate change leading to soil drying have resulted in an accelerated increase in iWUE, which will help to reduce uncertainties in the land surface schemes of global climate models, where vegetation-climate feedbacks are currently still poorly constrained by observational data.

Abstract: The increasing carbon dioxide (CO2) concentration in the atmosphere in combination with climatic changes throughout the last century are likely to have had a profound effect on the physiology of trees: altering the carbon and water fluxes passing through the stomatal pores. However, the magnitude and spatial patterns of such changes in natural forests remain highly uncertain. Here, stable carbon isotope ratios from a network of 35 tree-ring sites located across Europe are investigated to determine the intrinsic water-use efficiency (iWUE), the ratio of photosynthesis to stomatal conductance from 1901 to 2000. The results were compared with simulations of a dynamic vegetation model (LPX-Bern 1.0) that integrates numerous ecosystem and land– atmosphere exchange processes in a theoretical framework. The spatial pattern of tree-ring derived iWUE of the investigated coniferous and deciduous species and the model results agreed significantly with a clear south-to-north gradient, as well as a general increase in iWUE over the 20th century. The magnitude of the iWUE increase was not spatially uniform, with the strongest increase observed and modelled for temperate forests in Central Europe, a region where summer soil-water availability decreased over the last century. We were able to demonstrate that the combined effects of increasing CO2 and climate change leading to soil drying have resulted in an accelerated increase in iWUE. These findings will help to reduce uncertainties in the land surface schemes of global climate models, where vegetation–climate feedbacks are currently still poorly constrained by observational data.

Full text

1
Spatial variability and temporal trends in water-use
efficiency of European forests
Running head: Water-use efficiency trends in Europe
Matthias Saurer
1
, Renato Spahni
2,4
, David C. Frank
3,4
, Fortunat Joos
2,4
, Markus Leuenberger
2
,
Neil J. Loader
5
, Danny McCarroll
5
, Mary Gagen
5
, Ben Poulter
6
, Rolf T.W. Siegwolf
1
, Laia
Andreu-Hayles
7
, Tatjana Boettger
8
, Isabel Dorado Liñán
9, 10
, Ian J. Fairchild
11
, Michael
Friedrich
12
, Emilia Gutierrez
9
, Marika Haupt
8
, Emmi Hilasvuori
13, 14
, Ingo Heinrich
15
, Gerd
Helle
15
, Håkan Grudd
16
, Risto Jalkanen
17
, Tom Levanič
18
, Hans W. Linderholm
19
, Iain
Robertson
5
, Eloni Sonninen
13
, Kerstin Treydte
3
, John S. Waterhouse
20
, Ewan J. Woodley
5, 21
,
Peter M. Wynn
22
, Giles H.F. Young
5
1
Paul Scherrer Institut, Villigen, Switzerland
2
University of Bern, Bern, Switzerland
3
Swiss Federal Research Institute WSL, Birmensdorf, Switzerland
4
Oeschger Centre for Climate Change Research, University of Bern, Bern,
Switzerland
5
Department of Geography, Swansea University, Swansea, UK
6
Laboratoire des Sciences du Climat et de L'Environment, Gif-sur Yvette, France
7
Tree-Ring Laboratory, Lamont-Doherty Earth Observatory of Columbia University,
Palisades, NY 10964, USA
8
Department of Catchment Hydrology, Helmholtz Centre for Environmental Research – UFZ,
Germany
9
Dept d’Ecologia, Facultat de Biologia, Universitat de Barcelona, Barcelona, Spain
10
Chair of Ecoclimatology, Technische Universität München, Freising, Germany
11
School of Geography, Earth & Environmental Sciences, University of Birmingham, B15
2TT, UK
12
Institute of Botany, Hohenheim University, Stuttgart, Germany

2
13
Laboratory of Chronology, Finnish Museum of Natural History, P.O.Box 64, 00014
University of Helsinki, Helsinki, Finland
14
Finnish Environment Institute, P.O.Box 140, 00251 Helsinki, Finland
15
GFZ German Centre for GeoSciences, Climate Dynamics and Landscape Evolution, 14473
Potsdam, Germany
16
Bolin Centre for Climate Research, Department of Physical Geography and Quaternary
Geology, Stockholm University, Stockholm, Sweden
17
Finnish Forest Research Institute, Northern Regional Unit, Rovaniemi, Finland
18
Department of Yield and Silviculture, Slovenian Forestry Institute, Večna pot 2, 1000
Ljubljana, Slovenia
19
University of Gothenburg, Gothenburg, Sweden
20
Department of Life Sciences, Anglia Ruskin University, East Road, Cambridge CB1 1PT,
UK
21
Geography, College of Life and Environmental Sciences, University of Exeter, Rennes
Drive, Exeter, EX4 4RJ, UK
22
Lancaster Environment Centre, University of Lancaster, LA1 4YQ, UK
Accepted by Global Change Biology
Corresponding author: Matthias Saurer, matthias.saurer@psi.ch, Tel 0041 563102749
Keywords
Tree rings, carbon isotope discrimination, climate change, dynamic vegetation model
Type of paper
Primary research article

3
Abstract
The increasing carbon dioxide (CO
2
) concentration in the atmosphere in combination with
climatic changes throughout the last century are likely to have had a profound effect on the
physiology of trees: altering the carbon and water fluxes passing through the stomatal pores.
However, the magnitude and spatial patterns of such changes in natural forests remain highly
uncertain. Here, stable carbon isotope ratios from a network of 35 tree-ring sites located
across Europe are investigated to determine the intrinsic water-use efficiency (iWUE), the
ratio of photosynthesis to stomatal conductance from 1901–2000. The results were compared
with simulations of a dynamic vegetation model (LPX-Bern 1.0) that integrates numerous
ecosystem and land–atmosphere exchange processes in a theoretical framework. The spatial
pattern of tree-ring derived iWUE of the investigated coniferous and deciduous species and
the model results agreed significantly with a clear south-to-north gradient, as well as a general
increase in iWUE over the 20
th
century. The magnitude of the iWUE increase was not
spatially uniform, with the strongest increase observed and modelled for temperate forests in
Central Europe, a region where summer soil-water availability decreased over the last
century. We were able to demonstrate that the combined effects of increasing CO
2
and climate
change leading to soil drying have resulted in an accelerated increase of iWUE. These
findings will help to reduce uncertainties in the land surface schemes of global climate
models, where vegetation–climate feedbacks are currently still poorly constrained by
observational data.

4
Introduction
Interactions among direct CO
2
(fertilization) effects on plants and climatic conditions such as
drought are of particular interest for understanding past and for predicting future forest growth
and carbon sequestration. The continually increasing atmospheric CO
2
concentration and
concurrent climatic change are both likely to strongly affect the physiology of forests
ecosystems and alter productivity, species distribution and vegetation–climate feedbacks. Yet,
both the magnitude and mechanisms of forest response are unclear and furthermore are
expected to be spatially very heterogeneous, depending at least upon both the local growth
limitations (Babst et al., 2013) and the trajectory of future climate (IPCC, 2013). For many
regions, ecosystem transitions are already underway (Allen et al., 2010, Parmesan & Yohe,
2003). The effects of climate change and increasing CO
2
concentrations could stimulate tree
growth in some regions via enhanced photosynthesis through CO
2
fertilization (Ainsworth &
Long, 2005). However, adverse effects from an increase in drought severity, for example,
could result in reduced growth and increased stress and mortality (Zhao & Running, 2010).
Such changes will inevitably also modify biospheric CO
2
and water fluxes and the
relationships between them (Keenan et al., 2013, Schimel et al., 2001).
The slow adaption of late-successional forests, however, is difficult to assess either by
observations or by experiments. Therefore it is not well known how the physiology of natural
forests has already changed due to the increase of atmospheric CO
2
concentration in the ca.
150 years since major global industrialisation. Increases in net ecosystem productivity
inferred from small-scale CO
2
fumigation or depletion experiments may overestimate the
CO
2
-response in natural forests (Norby et al., 2010). Down-regulation of photosynthesis
under elevated CO
2
was observed and also the reduction in transpiration was found to be
relatively small in a mature mixed deciduous forest (Leuzinger & Körner, 2007). Accurate
quantification of changes in water and CO
2
fluxes over the last century, which are spatially

5
highly variable, would be important because of their relationship with the carbon
sequestration potential of the forests (Pan et al., 2011, Schimel et al., 2001), and evaporation–
temperature feedbacks, i.e. possibly additional or reduced warming due to changes in
evapotranspiration (Betts et al., 1997).
Stable carbon isotope ratios (
13
C) of tree-rings represent a valuable tool to improve
understanding of forest response to the combined influence of climate and CO
2
over time
(McCarroll & Loader, 2004). 
13
C in plant organic matter is related to the ratio of net
photosynthesis (A) to stomatal conductance to water (g), which is the intrinsic water-use
efficiency (iWUE = A/g), and therefore provides a measure for the relative water loss per
molecule carbon acquired at the leaf level (Farquhar et al., 1982). Plant water-use efficiency
(WUE) defined as the ratio of carbon uptake to actual water loss at the plant level is an
essential element of the survival and productivity of plants. Actual WUE depends on the
evaporative demand, as transpiration is determined by the product of stomatal conductance
and vapour pressure deficit, while iWUE may be considered as a potential WUE and does not
consider this variable environmental constraint and respiratory losses (Seibt et al., 2008). This
limitation of the isotope approach may be overcome by using plant physiological models to
assist data interpretation. For assessments that strive to consider both spatial and temporal
variability in plant-climatic-CO
2
interactions, a dynamic vegetation model such as the Lund–
Potsdam–Jena (LPJ) model that combines process-based vegetation dynamics with land–
atmosphere carbon and water exchange is particularly useful (Sitch et al., 2003). Such models
have been widely applied to study the influence of increasing CO
2
on net primary productivity
changes (Hickler et al., 2008). Few studies have discussed possible changes in WUE under
variable climate and CO
2
concentrations, and those that do generally find increasing WUE
over time (De Kauwe et al., 2013, Tian et al., 2010); however, no long-term verification of

Frequently asked questions

Q1. What have the authors contributed in "Spatial variability and temporal trends in water-use efficiency of european forests" ?

The authors were able to demonstrate that the combined effects of increasing CO2 and climate change leading to soil drying have resulted in an accelerated increase of iWUE. 

Q2. What are the plant functional types used in iWUE simulations?

The plant functional types used in simulations of iWUE were ’temperate and boreal needle evergreen’, ’temperate broadleaf summergreen’ and ’boreal needle summergreen’ in LPX,from here on referred to as conifers, oak and larch, respectively. 

Q3. What is the effect of other pollutants on tree physiological properties?

Other air pollutants, e.g. sulphur dioxide or ozone, may also influence tree physiological properties such as stomatal conductance and photosynthesis (Fairchild et al., 2009, Rinne et al., 2010). 

Q4. What was the method used for the analysis of carbon isotopes?

Carbon isotope analysis was conducted on CO2 obtained from combustion of the samples in an elemental-analyser and measurement in an isotope-ratiomass-spectrometer (McCarroll & Loader, 2004). 

Q5. How many trees were sampled at each site?

At least four trees were sampled at all sites (usually 2 cores per tree), sufficient to develop an isotope site record representative of the population (Leavitt & Long, 1984). 

Q6. What is the significance of the tree-ring isotopic approach?

tree-ring stable isotopes provide an absolutely dated, replicable historical perspective on past carbon isotopic fractionation that may be coupled within modelling experiments to help evaluate and improve the model performance and resolve the nature/significance of key feedback mechanisms. 

Q7. What is the effect of constant CO2 on the iWUE increase?

Also simulations with constant climate were performed: the resulting iWUE-increase when added to the results from the constant CO2-run was almost equal, but not identical (slightly higher) to the total iWUE-increase from the standard run, indicating some non-linearity in the model. 

Q8. What is the definition of isotope discrimination between plant material and atmospheric CO2?

The isotope discrimination between plant material ( 13 Cplant) and atmospheric CO2 ( 13 Catm) is defined as = ( 13 Catm – 13 Cplant)/(1+ 13 Cplant/1000). 

Q9. What is the effect of increasing CO2 on the iWUE in the conifers?

Their results therefore suggest that the trees (mainly conifers) underwent an accelerated increase in iWUE in this region, as the sum of the effects of increasing CO2 and soil drying, resulting in an effective reduction in stomatal conductance. 

Q10. How does the physiology of natural forests change?

Therefore it is not well known how the physiology of natural forests has already changed due to the increase of atmospheric CO2 concentration in the ca. 150 years since major global industrialisation. 

Q11. Why did the authors use the model settings from previous investigations?

1. The authors decided to use model settings from previous investigations (Spahni et al., 2013, Stocker et al., 2013), i.e. standard parameters, and not to optimize for best agreement between model and data, because this would involve some circular reasoning in testing modelled trends. 

Q12. What was the correlation analysis for all European sites?

A correlation analysis between model and tree-ring derived iWUE was also conducted for all individual sites, yielding generally high correlations with only a few exceptions (Fig. S1). 

Q13. What is the effect of increasing atmospheric CO2 concentration on iWUE?

Some studies indicated a passive response of the plants to increasing atmospheric CO2 concentration (ca), where the intercellular CO2 concentration (ci) increased by the same amount as ca (implying ca–ci = constant), which resulted in no iWUE improvement (Marshall & Monserud, 1996). 

Q14. What is the effect of a large-scale eddy-flux assessment?

a large-scale eddy-flux assessment showed indeed wide-spread forest WUE increase at least for the last 10–15 years(Keenan et al., 2013).