BGBiogeosciencesBGBiogeosciences1726-4189Copernicus GmbHGöttingen, Germany10.5194/bg-12-3695-2015Modelling the climatic drivers determining photosynthesis and carbon
allocation in evergreen Mediterranean forests using multiproxy long time
seriesGea-IzquierdoG.gea-izquierdo@cerege.frguigeiz@gmail.comGuibalF.JoffreR.OurcivalJ. M.SimioniG.GuiotJ.https://orcid.org/0000-0001-7345-4466CEREGE UMR7330, CNRS/Aix-Marseille Université, Europole de
l'Arbois, BP 80 13545, Aix-en-Provence CEDEX 4, FranceIMBE, CNRS/Aix-Marseille Université, UMR7263 Europole de l'Arbois, BP
8013545, Aix-en-Provence CEDEX 4, FranceCentre d'Ecologie Fonctionnelle et
Evolutive CEFE, UMR5175, CNRS, Université de Montpellier,
Université Paul-Valéry Montpellier, EPHE, 1919 Route de Mende,
34293 Montpellier CEDEX 5, FranceEcologie des Forêts
Méditerranéennes, INRA UR629, Domaine Saint Paul, 84914 Avignon
CEDEX 9, FranceG. Gea-Izquierdo (gea-izquierdo@cerege.fr; guigeiz@gmail.com)17June201512123695371228December20146February201510May201518May2015This work is licensed under a Creative Commons Attribution 3.0 Unported License. To view a copy of this license, visit http://creativecommons.org/licenses/by/3.0/This article is available from https://bg.copernicus.org/articles/12/3695/2015/bg-12-3695-2015.htmlThe full text article is available as a PDF file from https://bg.copernicus.org/articles/12/3695/2015/bg-12-3695-2015.pdf
Climatic drivers limit several important physiological processes involved in
ecosystem carbon dynamics including gross primary productivity (GPP) and
carbon allocation in vegetation. Climatic variability limits these two
processes differently. We developed an existing mechanistic model to analyse
photosynthesis and variability in carbon allocation in two evergreen species
at two Mediterranean forests. The model was calibrated using a combination
of eddy covariance CO2 flux data, dendrochronological time series of
secondary growth and forest inventory data. The model was modified to be
climate explicit in the key processes addressing the acclimation of
photosynthesis and the pattern of C allocation, particularly to water
stress. It succeeded in fitting both the high- and the low-frequency response of
stand GPP and carbon allocation to stem growth. This would support its
capability to address both C-source and C-sink limitations. Simulations
suggest a decrease in mean stomatal conductance in response to a recent
enhancement in water stress and an increase in mean annual intrinsic water
use efficiency (iWUE) in both species during the last 50 years. However,
this was not translated into a parallel increase in ecosystem water use
efficiency (WUE). The interannual variability in WUE closely followed that in
iWUE at both sites. Nevertheless, long-term decadal variability in WUE
followed the long-term decrease in annual GPP matching the local trend in
annual precipitation observed since the late 1970s at one site. In contrast,
at the site where long-term precipitation remained stable, GPP and WUE did
not show a negative trend and the trees buffered the climatic variability.
In our simulations these temporal changes were related to acclimation
processes at the canopy level, including modifications in LAI and
stomatal conductance, but also partly related to increasing [CO2]
because the model includes biochemical equations where photosynthesis is
directly linked to [CO2]. Long-term trends in GPP did not match those
in growth, in agreement with the C-sink hypothesis. The model has great potential for use with abundant dendrochronological data and
analyse forest performance under climate change. This would help to
understand how different interfering environmental factors produce
instability in the pattern of carbon allocation and, hence, the climatic signal
expressed in tree rings.
Introduction
Global change challenges forest performance because it can enhance forest
vulnerability (IPCC, 2013). Trees modify multiple mechanisms on different
scales to tackle environmental stress, including changes in
photosynthesis and carbon allocation within plants (Breda et al., 2006;
Niinemets, 2007; Chen et al., 2013). Many factors affect the different
physiological processes driving forest performance. Among them, the net
effect of the rising CO2 mixing ratio ([CO2]) and climate change is
meaningful when determining the forests' capacity of acclimation to enhanced
xericity (Peñuelas et al., 2011; Keenan et al., 2011; Fatichi et al., 2014). Forest process-based models have been developed to mimic these
mechanisms. They can include different levels of complexity but generally
implement calculations of leaf photosynthesis upscaled to the canopy and
carbon allocated to different plant compartments (Le Roux et al., 2001;
Schaefer et al., 2012; De Kauwe et al., 2013). Although there is evidence that
the tree performance depends to some extent on stored carbohydrates (Breda
et al., 2006; McDowell et al., 2013; Dickman et al., 2015), these models have
received some criticism when used to understand plant performance in
response to climate change. This is in part because they are C-source
oriented, therefore can exhibit certain limitations to represent the C-sink
hypothesis (i.e. that growth rates are limited by environmental factors such
as water stress, minimum temperature or nutrient availability rather than by
carbohydrate availability) and address dysfunctions related to the tree
hydraulics (Millard et al., 2007; Breshears et al., 2009; Sala et al., 2012;
Körner, 2013; McDowell et al., 2013; Fatichi et al., 2014).
Complex process-based models benefit from multiproxy calibration,
particularly when such data are applied on different spatio-temporal scales
(Peng et al., 2011). The temporal scale can be approached using time growth
series of dendrochronological data. However, the analysis of the past always
adds uncertainties related to the influence of unknown stand conditions to
properly scale productivity. Flux data, including stand productivity, can be
estimated using the eddy covariance technique (Baldocchi, 2003). These data
overcome many of the limitations of dendroecological data (e.g. intra-annual
resolution, control of stand conditions and scaling of net productivity), but
they lack their spatial and temporal coverage. Thus, CO2 flux data can
be used to implement unbiased models of canopy photosynthesis and can then be combined with dendroecological data to study how carbon is allocated to stem
growth as a function of environmental forcing (Friedlingstein et al., 1999;
Chen et al., 2013, McMurtrie and Dewar, 2013).
Mechanistic models can also be used to analyse the environmental factors
determining instability in the climate-growth response (D'Arrigo et al.,
2008). Different process-based models have been applied with dendroecological
data used either in forward or inverse mode (see Guiot et al., 2014, for a
review). Among these models, the process-based model MAIDEN (Modeling and Analysis In DENdroecology).
(Misson, 2004) was originally developed using dendroecological data.
The model explicitly includes [CO2] to calculate photosynthesis (hence
its influence on carbon allocation) and includes a carbohydrate storage
reservoir, this being one of its strengths compared to other models (Vaganov
et al., 2006; Sala et al., 2012; Guiot et al., 2014). It has been previously
employed to analyse growth variability in one temperate and two Mediterranean
species (Misson et al., 2004; Gaucherel et al., 2008) and recently, in
inverse mode (also including C and O stable isotopes), to reconstruct past
climate (Boucher et al., 2014). However, it requires further development to
ensure that it provides unbiased estimates of forest productivity and
assesses uncertainties in the response of trees to climatic variability on a
greater spatial scale at the regional level. In particular, its
parameterization would need improvement if the model is applied to assess how
climate modulates forest performance and the pattern of C allocation within
plants (Niinemets and Valladares, 2004; Fatichi et al., 2014).
In this study we use multiproxy data to develop a process-based model and
investigate how evergreen Mediterranean forests have modified stand
photosynthesis and carbon allocation in response to interacting climatic
factors and enhanced [CO2] in the recent past. The first objective was
to develop a process-based model based on MAIDEN (Misson, 2004). Within the
new version of the model, photosynthesis, carbon allocation, canopy turnover
and phenology are now calculated using climate-explicit functions with a
mechanistic basis. The model is adapted to give unbiased estimates of canopy
photosynthesis and stem growth using instrumental data. Specifically, within
the new model formulation, (1) photosynthesis is penalized by prolonged
water stress conditions through reductions in leaf area index (LAI) and
maximum photosynthetic capacity; (2) the pattern of carbon allocation is
directly determined by soil water content (i.e. water stress) and
temperature through nonlinear relationships; (3) these relationships can be
contrasting for different phenophases and affect photosynthesis and the pattern of C allocation independently. Once the model was
developed, a second objective was to analyse how [CO2] and climatic
variability affect the temporal instability in annual forest productivity,
water use efficiency and carbon allocation. We hypothesize that they will
exhibit differences in their long-term variability in relation to recent
climate change driven by different functional acclimation processes within
trees.
Material and methodsStudy sites and climatic data
The study sites were two evergreen Mediterranean monitored forests in
Southern France where CO2, water vapour and energy fluxes are
measured using the Eddy covariance technique (Baldocchi, 2003). Both sites
are included in FLUXNET (http://fluxnet.ornl.gov/). The first
site, Fontblanche (43.2∘ N, 5.7∘ E; 420 m), is
a mixed stand where Pinus halepensis Mill. dominates the open-top canopy layer reaching about
12 m; Quercus ilex L. forms a lower canopy layer, reaching about 6 m, and there is a sparse
shrub understory which includes Quercus coccifera L. (Simioni et al., 2013). The second site,
Puechabon (43.4∘ N, 3.4∘ E; 270 m), is a
dense coppice in which overstorey is dominated by Q. ilex with a density of around
6000 stems ha-1 (Rambal et al., 2004; Limousin et al., 2012). Both forests grow on
rocky and shallow soils that have a low retention capacity and are of Jurassic
limestone origin. The climate is Mediterranean, with a water stress period
in summer, cold or mild winters and most precipitation occurring between
September and May. Meteorological data were obtained from the neighbouring
stations of St. Martin de Londres (for Puechabon) and Aubagne (for
Fontblanche). According to those data Puechabon is colder and receives more
precipitation than Fontblanche (Table 1). Meteorological data showed a
decrease in total rainfall since the 1970s in Puechabon but no trend in
Fontblanche. Both sites exhibit a positive trend in temperatures more
evident for the maximum values (Fig. A1).
Characteristics of mean annual gross primary productivity,
climatic (annual means) and growth data. Standard deviations are shown
in parentheses. Precipitation: mean annual precipitation;
Tmax: annual mean of mean daily maximum temperature; Tmin: annual mean of
mean daily minimum temperature; length: chronology year replicated with
more than 5 radii; RW: mean annual ring width; Rbs: mean correlation
between series; AR: mean autocorrelation of raw series; MS: mean
sensitivity; EPS: mean expressed population signal. Rbs, AR, MS and EPS
are classical statistics to characterize growth chronologies, and they follow
Fritts (1976).
We assumed that GPP (gross primary productivity) is driven by the top pine and/or oak layers and that the
percentage of LAI related to the understory shrub layer will behave like
that of the oak species (evergreen, shrubby). For Fontblanche we considered
a maximum leaf area index (LAImax) of 2.2 m2 m-2
(3 m2 m-2 plant area index, PAI), composed of 70 % pine and 30 % oak (Simioni et al., 2013). For Puechabon we considered
an LAImax of 2.0 m2 m-2 (2.8 m2 m-2
PAI) monospecific to Q. ilex (Baldocchi et al., 2010; Limousin et al., 2012).
The specific leaf area (SLA) considered was 0.0045 m2 g-1 for
Q. ilex and 0.0037 m2 g-1 for P. halepensis (Hoff and
Rambal, 2003; Maseyk et al., 2008).
The model
We used MAIDEN (Misson, 2004), a stand productivity mechanistic model driven
by a number of functions and parameters representing different processes.
The model inputs are precipitation, maximum and minimum temperature, and
[CO2] with a daily time step. This model has been previously
implemented for monospecific forests, including two oaks and one pine
species,
using dendroecological chronologies of growth and, when available, stand
transpiration estimates from sap-flow sensors (Misson et al., 2004; Gaucherel
et al., 2008). However, the model has never been compared to actual CO2
flux data to ensure that it provides unbiased estimates of forest
productivity. In this study, the model was further developed to match
ground-based observations and generalize model use by modifying the
photosynthesis and allocation modules (including the different phenophases)
in relation to climatic drivers. To properly scale model outputs and get
unbiased estimates of stand productivity, we used CO2 eddy covariance
fluxes (Baldocchi, 2003). Different parameters were calibrated to different
data sources, including some species-dependent and some site-dependent
parameters, as follows. The transpiration rate (E) of day i is calculated
using a conductance approach: E(i)=gs(i)× VPD(i)/Patm(i), where Patm is
atmospheric pressure and gs and VPD are stomatal conductance and vapour
pressure deficit, respectively, as described below (Misson, 2004). The
other equations used to calculate micrometeorological covariates, soil
humidity and photosynthetic active radiation, as well as those functions
describing the water cycle (including soil evaporation and plant
transpiration) are explained in the original model formulation by Misson (2004). Therefore, they will not be described here. The rest of the model was
modified as follows.
Modelling the effect of climatic forcing on photosynthesis
Leaf photosynthesis (An) is calculated based on the biochemical model of
Farquhar et al. (1980). An is a function of the carboxylation
(Vc), oxygenation (Vo) and leaf dark respiration rates (Rd):
An=Vc-0.5Vo-Rd, where photosynthesis during the day i is limited by
either the rate of carboxylation when Rubisco is saturated (Wc) or when
it is limited by electron transport (Wj), i.e.
Ac=Vc-0.5Vo=min{Wc,Wj}. Rd was considered a fixed function
of Ac (0.006×Ac) because this formulation performed better than an exponential function of temperature (Sala and Tenhunen,
1996; De Pury and Farquhar, 1997; Bernacchi et al., 2001). Following De Pury
and Farquhar (1997):
Wc(i)=Vcmax(i)⋅(Ci(i)-Γ(i))Ci(i)+Kc(i)1+[O2]Ko(i),Wj(i)=Jmax(i)⋅(Ci(i)-Γ(i))4Ci(i)+8Γ(i),
where Ci is the CO2 intercellular concentration, Γ is the
[CO2] compensation point for photosynthesis in the absence of dark respiration, and Kc and
Ko are the kinetic Michaelis–Menten constants for carboxylation
and oxygenation, respectively. Vcmax and Jmax are
temperature-dependent parameters, as outlined below. Photosynthesis is known
to respond to the carbon concentration within chloroplasts Cc
rather than to Ci. Throughout the paper we retain the notation
presented here in Eqs. (1) and (2) but discuss below how mesophyll
conductance is taken into account empirically in relation to water stress
when calculating gs and acknowledge the possible limitations of
our approach (Reichstein et al., 2002; Grassi and Magnani, 2005; Flexas et
al., 2006; Sun et al., 2014).
Climate influences leaf photosynthesis calculations through the temperature
dependence of different parameters (Bernacchi et al., 2001; Nobel, 2009).
Γ, Kc and Ko were modelled using Arrhenius functions of
daily mean temperature (Tday, in ∘C) with parameters from De Pury and Farquhar (1997). We modelled Jmax as a fixed rate of
Vcmax (Jmax(i)=Jcoef⋅Vcmax(i)) after comparing it with
different temperature-dependent formulations (De Pury and Farquhar, 1997;
Maseyk et al., 2008). The model behaviour was better when the temperature
dependence of Vcmax was modelled using a logistic function
(Gea-Izquierdo et al., 2010) rather than an exponential function as in Misson (2004):
Vcmax(i)=Vmax(1+exp(Vb⋅((Tday(i)+273)-Vip)))⋅θp.Vmax, Vb and Vip are parameters to be estimated, with
Vmax being the asymptote and Vip the inflection point. θp is a soil water stress function dependent on the soil moisture
conditions of the previous year. It takes into account the downregulation of
photosynthesis in response to protracted drought through its impact on the
photosynthetic capacity of active LAI in evergreen species caused by
constraints in Vcmax, in turn produced by irreversible photoinhibition,
modifications in leaf stoichiometry and/or the aging of standing foliage through
lower leaf replacement rates in response to long-term water stress (Sala and Tenhunen, 1996; Niinemets and Valladares, 2004;
Niinemets, 2007; Vaz et al., 2010).
θp=1-exp(pstr⋅SWC180),
where pstr is a parameter to be estimated and
SWC180 is the mean soil water content (mm) from July to December of the
previous year.
Photosynthesis is coupled to the calculation of stomatal conductance, which is
estimated using a modified version of the Leuning (1995) equation:
gs(i)=g1⋅An(i)Cs(i)-Γ(i)⋅1+VPD(i)/VPD0⋅θg(i),
where g1 and VPD0 are parameters; VPD(i) is
daily vapour pressure deficit; Cs is the leaf surface [CO2];
θg is a non-linear soil water stress function calculated as
θg(i)=11+exp(soilb⋅(SWC(i)-soilip));
soilb and soilip are parameters; and SWC(i) is daily soil
water content (mm). θg accounts for variability in gas
exchange under drought conditions which cannot be taken into account through
stomatal control alone; thus, the variability can also be related to, e.g.,
mesophyll conductance or stomatal patchiness. Therefore, with this
empirical expression, we partly represent the effect of CO2
fractionation during mesophyll conductance under water stress, acknowledging
that this will likely be more complex under environmental stress (Reichstein
et al., 2002; Grassi and Magnani, 2005; Flexas et al., 2006; Sun et al.,
2014). The coupled photosynthesis-stomatal conductance system of equations
was estimated separately for sun and shade leaves. Canopy photosynthesis was
integrated using LAI, divided into its sunlit and shaded fractions (De Pury
and Farquhar 1997). Transmission and absorption of irradiance was calculated
following the Beer–Lambert law as a function of LAI, with
LAIsun=(1-exp(-LAI))⋅Kb (Kb is the beam light
extinction coefficient, which was set to 0.8) and
LAIshade=LAI-LAIsun (Misson, 2004). In the mixed stand
(Fontblanche), photosynthesis was calculated separately for Q. ilex and P. halepensis and then
integrated to obtain stand estimates of forest productivity.
Modelling the effect of climatic forcing on carbon allocation
Outline of the different phenological phases (P1 to P5) and
carbon allocation in the model within a given year. An: net daily
carbon assimilation; NSC: storage (non-structural carbohydrates);
GDD: growing degree days; GDDl= parameter determining shift from P2
to P3 (see text); C: carbon allocated either to the stem, canopy or roots;
d: day of year. Solid arrows correspond to allocation within the plant, whereas dashed arrows correspond to litterfall (canopy or roots). f3
and f4 are nonlinear functions of soil water content and temperature, determining carbon allocation to different compartments (see text for more
details).
The model allocates daily carbon assimilated either to the canopy, stem,
roots or storage of non-structural carbohydrates (NSC) to mimic intra-annual
carbohydrate dynamics (Misson, 2004; Dickman et al., 2015). Although trees can
store carbon within different above-ground and below-ground compartments
(Millard et al., 2007), carbon storage is treated as a single pool within the
model. Tree autotrophic respiration (Ra, in addition to Rd) is
modelled as a function f(i) of daily photosynthesis and maximum daily
temperature (Tmax; Sala and Tenhunen, 1996; Nobel, 2009):
Ra(i)=An(i)⋅max{0.3,f(i)},withf(i)=0.47⋅(1-exp(prespi⋅Tmax(i)),
where prespi is a parameter. Net photosynthesis is calculated for
day i as AN(i)=An(i)-Ra(i). This assumption means that
respiration would be considered zero when there is no photosynthesis; hence,
maintenance respiration would not be taken into account those days. Although
this could bias the overall carbon balance, we assume that this effect will
be very reduced in the studied forests because they present photosynthetic
activity all year round (see results). The model simulates several
phenological phases during the year (see Fig. 1):
winter period where all photosynthates assimilated daily,
AN(i), are allocated to the storage reservoir (NSCs) but there is no
accumulation of growing degree days (GDD).
winter period where all AN(i) are allocated to storage (i.e. the
same as in (P1)), but in contrast to (P1) there is active accumulation of
GDD, which defines the threshold GDD1 to trigger the next phenophase (P3) (budburst, leaf flush).
budburst, where carbon-available CT(i)=AN(i)+Cbud (Cbud is daily C storage utilized from
buds, a parameter) is either allocated to the canopy, to roots or to the
stem.
once the canopy has been completed in (P3), the next phenophase (P4) starts; in this period, daily photosynthates AN(i) are allocated either
to the stem or to storage;
the last phenophase (P5) starts when the photoperiod (parameter)
crosses a minimum threshold in fall. In this phase, root mortality occurs.
Otherwise (P5) is similar to (P1) and (P2) in the sense that all
AN(i) is used for storage until next year's (P3) starts.
The allocation of carbon to different plant compartments is complex because it
can be decoupled from photosynthetic production depending on different
factors, some of them climatic, acting on different temporal scales
(Friedlingstein et al., 1999; Sala et al., 2012; Chen et al., 2013; McMurtrie
and Dewar, 2013). In this new version of the model, we set the different
allocation relationships as nonlinear functions of temperature and soil
water content, h(i)=f1(Tmax)⋅f2(SWC), in (P3) and (P4) following the
functional relationships described in Gea-Izquierdo et al. (2013). This
means that now we take into account homeostatic acclimation processes at the
canopy level related to LAI dependence on water availability (Hoff and
Rambal, 2003; Sala and Tenhunen, 1996; Reichstein et al., 2003). LAI is
negatively related to long-term drought because litterfall is negatively
linked to water stress (Limousin et al., 2009; Misson et al., 2011) and bud
size depends on the climate influencing the period of bud formation
(Montserrat-Marti et al., 2009). Therefore, the actual carbon that can be
allocated to the canopy in (P3) of year j (AlloCcanopy(j)) was set as a function
of the previous year's moisture conditions (θLAIj) and
the maximum carbon that can be allocated to the canopy (MaxCcanopy).
MaxCcanopy is
calculated from LAImax and SLA, and
AlloCcanopy(j)=θLAI(j)×MaxCcanopy, where
θLAI(j)=(1-2⋅pLAI-SWC250pLAI),constrained toθLAI(j)∈[0.7,1.0].pLAI is a parameter to be calibrated representing the threshold over
which θLAIj=1 and SWC250 is mean soil water
content for May–December of the previous year.
Leaf turnover is variable within years and partly related to water
availability (Limousin et al., 2009, 2012). We considered a mean leaf
turnover rate of 3 years for pines and 2 for oaks. To model within-year
variability in leaf phenology (i.e. leaf growth and litterfall), we followed
Maseyk et al. (2008) and Limousin et al. (2009; Fig. 1). C allocation to
the canopy (i.e. including primary growth) in (P3) is calculated as
Ccanopy(i)=CT(i)×(1-0.2×h3_1(i))×Ratioroot/leaf-1; Ratioroot/leaf
was fixed to 1.5 for both species (Misson et al., 2004; Ourcival, unpublished
data), and
h3_1(i)=1-exp(p3moist⋅SWC(i))⋅exp-0.5⋅Tmax(i)-p3tempp3sd2,
where p3moist, p3temp and p3sd are parameters representing the scale of
the SWC and the optimum and dispersion of the Tmax functions, respectively.
The carbon allocated to the stem (Cstem) in (P3) is
Cstem(i)=CT(i)×0.2×h3_1(i)×h3_2(i), where
h3_2(i)=1-exp(st3moist⋅SWC(i))⋅exp-0.5⋅Tmax(i)-st3tempst3sd_temp2,
with h3_1(i) as in Eq. (9); st3moist, st3temp and
st3sd_temp are parameters as in h31(i).
The carbon allocated to roots in (P3) is set complementary to that of the
other compartments to close the carbon budget within the tree, i.e.
Croots(i)=CT(i)-Cstem(i)-Ccanopy(i).
Finally, in (P4) carbon-assimilated daily AN(i) is allocated either to
stem growth or to storage until changing to (P5). In (P4), the amount of carbon to be allocated to stem growth
is also set as a function of climatic forcing:
Cstem(i)=AN(i)×(1-h4(i)) and
Cstor(i)=AN(i)×h4(i), with
h4(i)=(1-exp(st4temp⋅Tmax(i)))⋅exp-0.5⋅SWC(i)st4sd_moist2,
where st4temp and st4sd_temp
are parameters.
Eddy covariance CO2 flux and dendrochronological data
The process-based model was calibrated using daily gross primary
productivity (GPP), dendrochronological data and inventory data. To develop
the model, those functions used to model daily stand
photosynthesis (i.e. Eq. 1 to 9) were first calibrated against GPP values. GPP
estimates were obtained from half-hourly net CO2 flux measurements
(NEP). GPP was obtained as the difference between measured net ecosystem
productivity and calculated ecosystem respiration (Reichstein et al., 2005).
Negative GPP values were corrected following Schaefer et al. (2012).
Half-hourly GPP data were integrated to obtain daily estimates for the
period 2001–2013 (Puechabon, methods detailed in Allard et al., 2008) and
2008–2012 (Fontblanche, Table 1).
In the second step, those functions used to model how carbon assimilated
and/or storage is allocated to the growth of the tree stem (i.e. Eqs. 10 and 11) were developed using calculated annual stem biomass increment time
series. Stem biomass increment chronologies were built combining
dendroecological data and forest inventory data collected at each site. We
built one chronology for Q. ilex in Puechabon, a second for Q. ilex in Fontblanche and a
third one for P. halepensis at Fontblanche (Fig. 2). For pines, two perpendicular cores
were extracted using an increment borer from 25 trees in fall 2013, whereas
for oaks we used cross sections. In Fontblanche, 15 oak stems were felled and
basal sections collected in spring 2014. A total of 17 oak stems from
Puechabon were logged in 2005 and 2008. The age and diameter distributions
of the studied forests are depicted in Fig. A2.
Growth (basal area increment, BAI, cm2 yr-1) and biomass allocated to the tree stem
(g C m-2 yr-1) of Q. ilex and P. halepensis at Fontblanche (growth shown in a,
biomass in b) and Q. ilex at Puechabon (growth and stem biomass shown in c). A
vertical dashed line marks the release event in Fontblanche produced by the
enhanced winter mortality in 1985 in (a). Dark lines for BAI correspond to
yearly means while grey polygons show confidence intervals (at 95 %) on
the standard errors of the mean.
All samples were processed using standard dendrochronological methods
(Fritts, 1976). Annual growth (RW) was measured using a stereomicroscope and
a moving table connected to a computer. RW cross-dating was visually and
statistically verified. RW estimates were transformed to basal area
increments (BAI, cm2 yr-1). Mean BAI chronologies
were obtained by averaging individual tree BAI time series. In Fontblanche,
BAI during the period 1987–1995 was standardized relative to the mean
calculated after excluding that period (Fig. 2). BAI data were
standardized because we did not find a climatic explanation for the abrupt
growth peak observed in Fontblanche during that period (Fig. 2). Therefore,
we assumed that it had been caused by a release event (i.e. reduction in
competition) produced by the death of neighbours as a consequence of winter
frost during 1985 and 1987 (Vennetier, personal communication, 2014). These two frosts
were reflected by the presence of characteristic frost rings in most
individuals from Fontblanche.
To scale BAI chronologies to the same units as annual stem biomass (which is
an output of the model), we used plot inventory data collected around the
flux towers at the two sites. Inventory data included stem diameter for all
trees and tree height collected for a subsample every 2 years during
2007–2011 in Fontblanche as well as annual diameter estimates for the period
1986–2011 for Puechabon. Individual annual biomass increments were estimated
by subtracting the stem biomass of one year from that of the next; then, stand stem biomass
increments (SBIs, g C m-2 yr-1) were calculated by integrating plot
data. Stem biomass was calculated using allometric functions. For pines, we
calculated stem biomass using diameter and estimated stem height assuming
that the tree bole follows a paraboloid shape (Li et al., 2014). For oaks,
stem biomass was calculated following Rambal et al. (2004). Once SBI had
been estimated for the years when we had available inventory data, BAI
chronologies were correlatively scaled to SBI units (g C m-2 yr-1). We built two mean stand SBI chronologies, one for each site,
meaning that we analysed carbon allocation within stands, not
differentiating between species in Fontblanche. These two SBI chronologies
were used to calibrate sitewise Eqs. (10) and (11).
Model development and analyses
Parameters were selected according to the ecological characteristics of the
species, exploring the model using comprehensive sensitivity analysis to
sequentially optimize groups of parameters. In a first step, a group of
common parameters was selected using GPP
data from Fontblanche (Table 2). The species-dependent parameters selected
for Q. ilex in this first step (those parameters in Table 2 which are common to the two sites) were independently validated when applied in
Puechabon. In a second step, a
subset of site-dependent parameters was calibrated against GPP and SBI data.
Four parameters from Eqs. (6) and (9) were calibrated using GPP data, and five parameters
in Eqs. (10) and (11) were calibrated using stem biomass increment data (Table 2). The local parameters were calibrated constrained to an ecologically
realistic range and using a global optimization algorithm and maximum likelihood
principles (Gaucherel et al., 2008).
Model parameters. Those parameter differing between sites
were optimized either with GPP data (photosynthesis and allocation module)
or with growth-based biomass increment chronologies (allocation module). The
rest were common parameters for both sites and selected while developing the
model in the first step for Fontblanche using GPP data (represented in the
“Cal” column by “–”). Meaning of parameters, equation number (E) and phenophase
(P) are as in the Material and Methods section. Fontb: Fontblanche;
Puech: Puechabon; Cal: local parameters to be calibrated with GPP or stem
biomass increment data (SBI); PIHA: Pinus halepensis; QUIL: Quercus ilex.
To compare model output with stem biomass chronologies as estimated from
dendroecological data, we used only the period for which we had available daily
meteorological data (1960–2013), which was also a period that did not
include juvenile years with increasing BAI (BAIs reached an asymptote after
increasing for the first 15–20 juvenile years; Fig. 2). The model does not
take into account how size differences in allometry or ontogeny affect
carbon allocation (Chen et al., 2013). We tried to keep the model as simple
as possible also because we had no such data to calibrate ontogenic effects.
Hence, the model is designed for non-juvenile stands with canopies that
reached a steady state with asymptotic LAImax. For the same reasons
it does not take into account how changes in management affect carbon
allocation. The model was analysed in terms of goodness of fit.
Additionally, for the period for which we had available daily meteorological
data, we simulated time series of GPP, ecosystem water use efficiency
(WUE=GPP/ET, with
ET being actual evapotranspiration) and intrinsic water use efficiency
of sun leaves (iWUE=AN/gs) calculated
following Beer et al. (2009).
Results
The studied evergreen forests exhibit a bimodal pattern in GPP with maxima
in spring and autumn (Fig. 3) as often observed in Mediterranean
ecosystems (e.g. Baldocchi et al., 2010). GPP was above 0 almost every day
of the year, including in winter, particularly at the milder site, Fontblanche
(Table 1). This means that there is active photosynthesis all year round in
these evergreen forests, including during both periods of climatic stress, i.e. those with low
temperature and short photoperiod in winter and with low moisture
availability in summer (Fig. 3). Mean annual GPP was 1431.4 ± 305.4 g C m-2 yr-1 and precipitation 642.7 ± 169.7 mm in
Fontblanche, whereas it was 1207.3 ± 206.7 g C m-2 yr-1 and
1002.6 ± 328.2 mm in Puechabon (see Table 1 for more details). Mean GPP
was higher at Fontblanche because carbon assimilation was greater in the low-temperature winter period but similar the rest of the year (Fig. 3). Stem
growth did not show any long-term (decadal) growth trend for any of the
species studied (Fig. 2).
Daily gross primary productivity (GPP) at Puechabon
(2001–2013, black dots, blue line) and Fontblanche (2008–2012, orange dots,
red line). DOY: day of year. Thick lines correspond to smoothers fitted to
the mean to highlight seasonal trends at the two sites.
The model accurately represented the low-frequency response of daily GPP: both the
seasonal variability in GPP within years and variability in GPP among years
(Fig. 4). The model explained over 50 % of the annual biomass growth
variance, and 46 and 59 % of daily GPP in Fontblanche and Puechabon,
respectively (Fig. 4). This means that we were able to mimic the daily,
seasonal and long-term trends in stand productivity with unbiased estimates
but also to model how carbon is allocated to stem growth throughout the year
during the different phenophases described. The model assumed species-specific
carbon allocation responses set to the different plant compartments as
nonlinear functions of temperature and soil moisture. These relationships
were biologically meaningful in the sense that photosynthesis and carbon
allocation could be decoupled to some extent as a function of climatic
variability. Once the canopy had been formed in spring, the model allocated
more carbon to the stem and less to storage when less severe climatic stress
occurs, i.e. with decreasing temperatures and more humid conditions (Fig. 5).
Model fit to stem biomass increment (a) and GPP (b) in
Fontblanche and stem biomass increment (c) and GPP (d) in Puechabon.
R2: coefficient of determination; ρ: linear correlation between
estimated and observed data; ρlow15: linear correlation between
estimated and observed data smoothed with a 15-year low-pass filter (blue
and red lines in b and c). Polygons behind the estimated values in
(a) and (c) correspond to confidence intervals of the mean: solid grey polygons
for estimated values and dashed polygons for observed stem biomass increment
values.
Modelled carbon allocation trajectory to the stem when leaf flush has
finished in phenological period (P4). We show the unitless modifier
1-h4(i) (i.e. h4(i) is the portion of carbon allocated to storage)
from Cstem(i)=AN(i)× [(1-h4(i))]
in Eq. (11). The modifier [0,1] is a function of soil water content (SWC) and maximum
temperature (Tmax); multiplied with available daily carbon, it gives the
distribution of daily carbon allocated to secondary growth and storage.
Both sites exhibited an increase in temperature particularly evident in the
maximum values, but only Puechabon suffered a decrease in annual
precipitation between 1960 and 2012 (Fig. A1). In the model, the studied
forests acclimated to changing conditions in the last decades,
coupling different physiological traits, and simulated annual GPP largely
followed the overall trends in precipitation observed. In Fontblanche, which
is milder and receives less precipitation, GPP has remained stable since the
1960s and presented no apparent long-term trend (Fig. 6). In contrast, at
the coldest and rainiest site (Puechabon), the model simulated a decrease in
GPP (Fig. 6), which was driven by the prevailing decrease in precipitation
observed since the 1970s (Figs. A1; 6). This reduction in GPP was
partly a consequence of decreased LAI in response to enhanced long-term
water stress (Fig. A3; Limousin et al., 2009; Misson et al., 2011).
Simulated long-term decadal trends in mean annual stomatal conductance were
similar and decreased at the two sites with greater water stress as a
consequence of enhanced temperatures (Fig. 6). The two species studied
showed a long-term increase in simulated iWUE (Fig. 7) following the
decrease in simulated gs (Fig. 6). The interannual variability in WUE
and iWUE were highly and positively correlated (Fig. 7). However, in the
long-term they followed a different pattern, particularly in Puechabon, where
there was a recent decline in WUE (not observed in iWUE) forced by trends in
ET and GPP (Fig. 7). This means that the recent reduction in simulated GPP
was proportionally greater than that of simulated ET (Figs. 6; A3).
DiscussionLinking photosynthetic production to carbon allocation as a function of climate
The model calculates stand productivity and carbon allocation to stem growth
in response to climate and [CO2] with realism. It is particularly well
suited to mimic the effect of water stress in plant performance by the
explicit assessment of different acclimation processes at the canopy level,
including changes in stomatal conductance and photosynthetic capacity (Sala
and Tenhunen 1996; Reichstein et al., 2003; Limousin et al., 2010; Misson et al., 2011). Additionally, the model simulates carbohydrate storage
dynamically as a function of environmental variability. Climate can affect
differently the carbon dynamics and pattern of C allocation to different
tree compartments at different phenophases. In the model the storage
reservoir is an active sink for assimilated carbon during some periods of
the year and a source in spring to be used in primary and secondary growth
(Fig. A5). Additionally stem growth is limited by climatic constraints (in
(P3) and (P4)) rather than just by the amount of available carbohydrates
(Millard et al., 2007). This means that water stress and optimum temperature
directly affect the modelled processes, assuming that cell-wall expansion in
the xylem can be related to climatic variability differently from photosynthetic
production (Sala et al., 2012). The model showed C limitation (for primary
growth) in the years when LAImax was not achieved (i.e. a limitation in LAI
is driven by limitations in the C supply in spring), e.g. all years in
Puechabon for the period shown in Fig. A5 (1995–2012) but only those years
in Fontblanche when the minimum value considered as a threshold was reached.
Therefore, both C-source (photosynthesis) and C-sink (just related to growth;
other sinks such as volatile organic compounds or root exudates are not
explicitly included in the model) limitations can be assessed in different
years within one site and even at different periods within the same year
(Millard et al., 2007; Sala et al., 2012; Chen et al., 2013; Fatichi et al., 2014). This hypothesis seems plausible as drought stress affects both
C-source (e.g. through reduced stomatal conductance) and C-sink limitations
(e.g. cell water turgor, hydraulic performance; McDowell et al., 2013).
Whether the pattern of C storage simulated is realistic is something that
needs to be validated against actual data. However, the flexible way in
which stored C is modelled has much potential to improve ecosystem models
that only view a carbon-source limitation (Sala et al., 2012; Friend et al., 2014).
Modelled total annual stand gross primary productivity (GPP) and
mean stomatal conductance of sun leaves (gs) for Fontblanche (a) and
Puechabon (b) for the period for which meteorological data were
available. To show the influence of the precipitation decline observed in
Puechabon on GPP, we run a sensitivity simulation in which precipitation was
fixed for 1980–2012 on the basis of precipitation in 1960–1979 (Fig. A1) and all other input variables (Tmin,
Tmax, [CO2]) were actual values. GPP values from this
simulation are depicted as dashed grey lines in (b).
Ecosystem WUE (integral annual) and iWUE for sun leaves
(mean daily, for PIHA and QUIL separated in Fontblanche) for (a) Fontblanche
and (b) Puechabon for the period for which we had available meteorological data.
Water stress is generally considered the greatest limitation for
Mediterranean ecosystems, driving a close relationship between precipitation
and both growth and photosynthesis (Breda et al., 2006; Pereira et al., 2007;
Baldocchi et al., 2010; Gea-Izquierdo and Cañellas, 2014). Our results
show that a long-term decrease in precipitation triggered a decrease in
simulated GPP at the rainier and coldest site. However, this decline was
not expressed in the growth trends. This means that long-term productivity
and the allocation of C to secondary growth were decoupled and did not match
(Sala et al., 2012; Chen et al., 2013; Fatichi et al., 2014). The existence of
trade-offs between carbon assimilation and allocation in relation to
environmental variability suggests exercising caution when using growth as a direct
proxy to investigate stand productivity dynamics (e.g. Piovesan et al., 2008;
Peñuelas et al., 2008; Gea-Izquierdo and Cañellas, 2014). GPP was
greater at the site receiving less precipitation, which could be related to
differences in soil retention capacity. However, both soils are calcareous,
shallow and stony, and differences in GPP were, to a large part, explained by less
limitation for carbon assimilation of low winter temperatures at the warmest
site (Fontblanche). They can also be a result of a different species
composition (oak vs. pine oak). LAI is greater at the site yielding higher
annual GPP. Nonetheless, had this factor been responsible for the observed
differences in winter photosynthesis, there would also have been differences
in spring photosynthesis, which was not the case (Fig. 3).
A better understanding of the underlying processes determining carbon
allocation will benefit process-based models (Sala et al., 2012; Fatichi et
al., 2014). Model parameters were within the range found in the literature,
bearing in mind that using a daily time step to study differential processes
or not distinguishing between leaf ages will affect the scaling of parameters
such as Jmax, Vcmax or Rd (De Pury and
Farquhar, 1997; Grassi and Magnani, 2005; Maseyk et al., 2008; Vaz et al.,
2010). Daily climatic data are readily available on a greater spatial scale
than data with a higher temporal resolution, which increases the
applicability of daily models. Model performance could be improved by
addressing respiration changes related to ontogeny, allometry and nutrient
limitations (e.g. N/P) on photosynthesis or by including more complex
upscaling of leaf-level photosynthesis (Niinemets et al., 1999; Niinemets,
2007; Chen et al., 2013; McMurtrie and Dewar, 2013). However, it is difficult
to find suitable data to calibrate such processes. Similarly, it would be
challenging to include allocation to reproductive effort in the carbon
budget. This is because, even if it is influenced by water stress in the
studied forests (Pérez-Ramos et al., 2010), there is still great
uncertainty regarding the causal factors driving multi-annual variability in
fruit production (Koenig and Knops, 2000). Addressing stand dynamics would
also help to generalize model applicability. Stand disturbances modifying
stand competition can leave an imprint on growth for more than a decade
whereas they do not seem to affect stand GPP over more than 1 or 2 years if
the disturbance is moderate (Misson et al., 2005; Granier et al., 2008). In
response to changes in competition, the trees modify the carbon allocation
pattern or keep the root : shoot ratio constant to enhance productivity on
a per-tree basis but up to an asymptotic
stand GPP. Still, the model behaviour was good compared with other studies
that addressed ontogenic changes in the carbon allocation response to
photosynthesis (Li et al., 2014) and similar or better than that of other
mechanistic approaches calibrated to standardized dendroecological data
(Misson et al., 2004; Evans et al., 2006; Gaucherel et al., 2008;
Tolwinski-Ward et al., 2011; Touchan et al., 2012).
Forest performance in response to recent climate change and [CO2] enhancement
Few studies under natural conditions have observed a net increase in growth rates
in response to enhanced [CO2] levels since the late 1800s, meaning that
other factors, such as water stress and/or N or P, were more limiting for
photosynthesis and/or allocation to growth than [CO2] (Niinemets et al., 1999; Peñuelas et al., 2011;
McMurtrie and Dewar, 2013; Lévesque et al., 2014). Yet the forests have increased their iWUE. This can be partly a
passive consequence of enhanced [CO2], but higher iWUE observed in more
water-stressed sites suggests that climate is co-responsible for an active
acclimation of physiological plant processes (Keenan et al., 2013; Leonardi
et al., 2012; Saurer et al., 2014). These processes would include a higher
stomatal control, like in our results, where, in turn, we did not observe any
increase in long-term carbon assimilation. The mean annual stomatal
conductance simulated was driven by climate but also decreased
simultaneously in time with increasing [CO2] (Fig. A4).
Furthermore, there is debate on whether there has been an increase in
ecosystem WUE in response to recent changes in [CO2] under a warming
climate (Beer et al., 2009; Reichstein et al., 2002; Keenan et al., 2013). In
our results the high frequency of WUE followed that of iWUE, but there was
some mismatch between the two traits in the low frequency. We observed no
dominant time trends in simulated annual WUE for the period 1980–2000 at the
site where precipitation remained stable, whereas a decrease in
WUE following that in GPP was particularly evident at the site experiencing a
drier climate in recent years. This trend was not observed in iWUE, which
means that reductions in GPP and gs were proportionally greater than
those in ET (Figs. 6, 7, A3).
Higher [CO2] concentrations enhance photosynthesis with the equations
used to calculate leaf photosynthesis in biochemical models (e.g. Gaucherel
et al., 2008; Keenan et al., 2011; Leonardi et al., 2013; Boucher et al., 2014).
Thus, the absence of a long-term increase in GPP and growth would not mean
that enhanced [CO2] was not beneficial for model outputs (particularly
in the case of C-source limitation) but that the net control exerted by
other factors such as climatically driven stress was more limiting than that
of [CO2] availability: growth and photosynthesis would have been lower
had we used constant [CO2] with the same model parameters. The absence
of any modification in the growth trends, even if there are changes in WUE,
would express a sink limitation mostly related to hydraulic constraints
(Peñuelas et al., 2011; Sala et al., 2012; Keenan et al., 2013). Often, the
trees show a growth decline at those sites where there is an enhancement
in long-term water stress that dominates species performance (e.g. Bigler et al., 2006; Piovesan et al., 2008; Gea-Izquierdo et al., 2014). In contrast, it
has been observed that, under certain conditions, trees have increased growth
with warming since the 1850s (Salzer et al., 2009; Gea-Izquierdo and
Cañellas, 2014). These studies suggest the existence of a positive effect
of warming, rather than of CO2 fertilization, upon growth in forests
where water stress is not the most limiting factor. Our study sites are
located at the northern limit of the Mediterranean region, meaning that
the two species studied occupy drier and warmer areas further to the south. The
two species have different functional characteristics, e.g. oaks are
anisohydric, whereas pines tend to be isohydric. This confers different
capacities of adaptation to climate change on them, which means that they should
play different roles in future stand dynamics. Our results show the
existence of trade-offs in response to climate at different phenological
periods. This is important since synergistic environmental stresses acting
at different periods can trigger tree mortality (McDowell et al., 2013;
Voltas et al., 2013). Model sensitivity analysis could be performed to
discuss the influence of specific factors, such as climate or [CO2],
causing instability in the climate-growth response (D'Arrigo et al., 2008;
Boucher et al., 2014). However, [CO2] enhancement and climate warming are
mixed in analysis performed using data from field studies, which can make
the isolation of their effect problematic. The model can be applied using
abundant dendrochronological data used to determine the site-dependent
parameters. This would provide much flexibility for investigating growth trends
and forest performance in response to global change on a larger scale.
Conclusions
By developing an original process-based model with carbon allocation
relationships explicitly expressed as functions of climate, we accurately
simulated gross primary productivity and the allocation of carbon to
secondary growth in evergreen Mediterranean forests. Different processes were
modelled as functions of environmental variability, including [CO2] and
climate. The studied forests showed trade-offs in carbon allocation to
different plant compartments in response to stress in different seasons: with
low temperatures and a short photoperiod in winter and with moisture shortage
in summer. We modelled a decreasing time trend in stomatal conductance, which
would suggest a partly active increase in iWUE in the forests studied.
Interannual variability in WUE followed closely that in iWUE. However, WUE
exhibited a decreasing trend at the site where we simulated a decrease in LAI
and GPP in response to a decrease in annual precipitation since the 1970s.
Long-term GPP has remained at similar levels in the last 50 years in just one
stand, whereas it declined in the forest suffering a reduction in
precipitation. This suggests different acclimation processes at the canopy
level and in the pattern of allocation in response to enhanced xericity and
increasing [CO2] levels; these acclimation processes could not
counterbalance the negative effect of warming only at the site where there
was a simultaneous decrease in precipitation. Tree growth was partly
decoupled from stand productivity, highlighting that it can be risky to use
growth as a direct proxy for GPP. The model is flexible enough to assess both
C-source and C-sink limitations and includes a dynamic estimation of stored
C. These features would improve ecosystem models with a fixed C-source
formulation. By calibrating a limited number of parameters related to carbon
allocation, there is great potential for using the model with abundant
dendroecological data to characterize past instability in the growth response
in relation to environmental variability and to simulate future forest
responses under different climatic scenarios.
Mean climatic time series of the last 50 years. (a) Annual
precipitation; and annual maximum (Tmax) and minimum (Tmin)
temperatures for Fontblanche (b) and Puechabon (c).
Diameter (dbh, cm) and age (years) distribution of trees
included in the chronologies. Frequencies are calculated separately by
species and site.
Simulated maximum annual leaf area index LAI
(m2 m-2) and total annual stand transpiration E
(mm yr-1) in Fontblanche (a) and Puechabon (b).
Simulated mean annual stomatal conductance (gs) as a
function of mean [CO2] (a) and mean maximum temperature (b).
Simulated non-structural carbohydrate content (NSC) in the
storage pool at both sites. The period 1995–2012 is shown to highlight
within-year variability.
Acknowledgements
G. Gea-Izquierdo was funded by the Labex OT-Med (no. ANR-11-LABEX-0061) from
the “Investissements d'Avenir” program of the French National Research
Agency through the A*MIDEX project (no. ANR-11-IDEX-0001-02).
Federation de Recherche FR3098 ECCOREV, the labex IRDHEI and OHM-BMP also
supported the study. We are grateful to Roland Huc for sharing data from
Fontblanche.
Edited by: Akihiko Ito
References
Allard, V., Ourcival, J. M., Rambal, S., Joffre, R. and Rocheteau, A.:
Seasonal and annual variation of carbon exchange in an evergreen
Mediterranean forest in southern France, Glob. Change Biol., 14,
714–725, 2008.
Baldocchi, D. D.: Assessing the eddy covariance technique for evaluating carbon
dioxide exchange rates of ecosystems: Past, present and future, Glob. Change Biol., 9, 479–492, 2003.
Baldocchi, D. D., Ma, S. Y., Rambal, S., Misson, L., Ourcival, J. M.,
Limousin, J. M., Pereira, J., and Papale, D.: On the differential advantages
of evergreenness and deciduousness in mediterranean oak woodlands: a flux
perspective, Ecol. Appl., 20, 1583–1597, 2010.Beer, C., Ciais, P., Reichstein, M., Baldocchi, D., Law, B. E., Papale, D.,
Soussana, J.-F., Ammann, C., Buchmann, N., Frank, D., Gianelle, D.,
Janssens, I. A., Knohl, a., Köstner, B., Moors, E., Roupsard, O.,
Verbeeck, H., Vesala, T., Williams, C. A., and Wohlfahrt, G.: Temporal and
among-site variability of inherent water use efficiency at the ecosystem
level, Global Biogeochem. Cy., 23, GB2018, 10.1029/2008GB003233, 2009.
Bernacchi, C. J., Singsaas, E. L., Pimentel, C., Portis, A. R., and Long, S.
P.: Improved temperature response functions for models of Rubisco-limited
photosynthesis, Plant Cell Environ., 24, 253–259, 2001.
Bigler, C., Braker, O. U., Bugmann, H., Dobbertin, M., and Rigling, A.:
Drought as an inciting mortality factor in Scots pine stands of the Valais,
Switzerland, Ecosystems, 9, 330–343, 2006.Boucher, É., Guiot, J., Hatté, C., Daux, V., Danis, P.-A., and Dussouillez, P.:
An inverse modeling approach for tree-ring-based climate reconstructions under changing atmospheric CO2
concentrations, Biogeosciences, 11, 3245–3258, 10.5194/bg-11-3245-2014, 2014.
Breda, N., Huc, R., Granier, A., and Dreyer, E.: Temperate forest trees and
stands under severe drought: a review of ecophysiological responses,
adaptation processes and long-term consequences, Ann. For. Sci., 63,
625–644, 2006.
Breshears, D. D., Myers, O. B., Meyer, C. W., Barnes, F. J., Zou, C. B.,
Allen, C. D., McDowell, N. G., and Pockman, W. T.: Tree die-off in response
to global change-type drought: mortality insights from a decade of plant
water potential measurements, Front. Ecol. Environ., 7, 185–189, 2009.
Chen, G., Yang, Y., and Robinson, D.: Allocation of gross primary production
in forest ecosystems: allometric constraints and environmental responses,
New Phytol., 200, 1176–1186, 2013.
D'Arrigo, R., Wilson, R., Liepert, B., and Cherubini, P.: On the “Divergence
Problem” in Northern Forests: A review of the tree-ring evidence and
possible causes, Global Planet. Change, 60, 289–305, 2008.
De Pury, D. G. G. and Farquhar, G. D.: Simple scaling of photosynthesis from
leaves to canopies without the errors of big-leaf models, Plant Cell
Environ., 20, 537–557, 1997.De Kauwe, M. G., Medlyn, B. E., Zaehle, S., Walker, A. P., Dietze, M. C.,
Hickler, T., Jain, A. K., Luo, Y. Q., Parton, W. J., Prentice, I. C., Smith,
B., Thornton, P. E., Wang, S. S., Wang, Y. P., Warlind, D., Weng, E. S.,
Crous, K. Y., Ellsworth, D. S., Hanson, P. J., Seok Kim, H., Warren, J. M.,
Oren, R., and Norby, R. J.: Forest water use and water use efficiency at
elevated CO2: a model-data intercomparison at two contrasting temperate
forest FACE sites, Glob. Change Biol., 19, 1759–1779, 2013.
Dickman, L. T., McDowell, N. G., Sevanto, S., Pangle, R. E., and Pockman, W.
T.: Carbohydrate dynamics and mortality in a piñon-juniper woodland
under three future precipitation scenarios, Plant. Cell Environ., 729–739, 2015.Evans, M. N., Reichert, B. K., Kaplan, A., Anchukaitis, K. J., Vaganov, E.
A., Hughes, M. K., and Cane, M. A.: A forward modeling approach to
paleoclimatic interpretation of tree-ring data, J. Geophys. Res., 111, G03008, 10.1029/2006JG000166,
2006.Farquhar, G. D., von Caemmerer, S., and Berry, J. A.: A biochemical model of
photosynthetic CO2 assimilation in leaves of C3 species, Planta,
149, 78–90, 1980.
Fatichi, S., Leuzinger, S., and Körner, C.: Moving beyond photosynthesis: from carbon source to sink-driven vegetation modeling, New Phytol., 201,
1086–1095, 2014.
Flexas, J., Bota, J., Galmes, J., Medrano, H., and Ribas-Carbo, M.: Keeping a
positive carbon balance under adverse conditions: responses of
photosynthesis and respiration to water stress, Physiol. Plant., 127,
343–352, 2006.
Friedlingstein, P., Joel, G., Field, C. B., and Fung, I. Y.: Toward an
allocation scheme for global terrestrial carbon models, Glob. Change Biol.,
5, 755–770, 1999.
Fritts, H.C,: Tree Rings and Climate, Blackburn Press, 567 pp., 1976.
Gaucherel, C., Campillo, F., Misson, L., Guiot, J., and Boreux, J. J.:
Parameterization of a process-based tree-growth model: Comparison of
optimization, MCMC and Particle Filtering algorithms, Environ. Model.
Softw., 23, 1280–1288, 2008.
Gea-Izquierdo, G., Mäkelä, A., Margolis, H., Bergeron, Y., Black, T.
A., Dunn, A., Hadley, J., Kyaw Tha Paw U, Falk, M., Wharton, S., Monson, R.,
Hollinger, D. Y., Laurila, T., Aurela, M., McCaughey, H., Bourque, C.,
Vesala, T., and Berninger, F.: Modeling acclimation of photosynthesis to
temperature in evergreen conifer forests, New Phytol., 188, 175–186,
2010.
Gea-Izquierdo, G., Fernández-de-uña, L., and Cañellas, I.: Growth
projections reveal local vulnerability of Mediterranean oaks with rising
temperatures, Forest Ecol. Manag., 305, 282–293, 2013.
Gea-Izquierdo, G. and Cañellas, I.: Long-term climate forces instability
in long-term productivity of a Mediterranean oak along climatic gradients,
Ecosystems, 17, 228–241, 2014.
Gea-Izquierdo, G., Viguera, B., Cabrera, M., and Cañellas, I.: Drought
induced decline could portend widespread oak mortality at the xeric ecotone
in managed Mediterranean pine-oak woodlands, Forest Ecol. Manag. 320, 70–82,
2014.
Granier, A., Breda, N., Longdoz, B., Gross, P., and Ngao, J.: Ten years of
fluxes and stand growth in a young beech forest at Hesse, North-eastern
France, Ann. For. Sci., 64, 704, 2008.
Grassi, G. and Magnani, F.: Stomatal, mesophyll conductance and biochemical
limitations to photosynthesis as affected by drought and leaf ontogeny in
ash and oak trees, Plant Cell Environ., 28, 834–849, 2005.
Guiot, J., Boucher, E., and Gea-Izquierdo, G.: Process models and model-data
fusion in dendroecology, Front. Ecol. Evol., 2, 1–12, 2014.
Hoff, C. and Rambal, S.: An examination of the interaction between climate,
soil and leaf area index in a Quercus ilex ecosystem, Ann. For. Sci., 60,
153–161, 2003.
IPCC: Climate Change: The Physical Science Basis, Summary for policymakers,
edited by: Stocker, T. F., Qin, D., Plattner, G.-K., Tignor, M., Allen, S. K., Boschung, J.,
Nauels, A., Xia, Y., Bex, V., and Midgley, P. M., IPCC, Geneva, Switzerland, 33 pp., 2013.Keenan, T., Maria Serra, J., Lloret, F., Ninyerola, M., and Sabate, S.:
Predicting the future of forests in the Mediterranean under climate change,
with niche- and process-based models: CO2 matters!, Glob. Change Biol.,
17, 565–579, 2011.
Keenan, T. F., Hollinger, D. Y., Bohrer, G., Dragoni, D., Munger, J. W.,
Schmid, H. P. ,and Richardson, A. D.: Increase in forest water-use efficiency
as atmospheric carbon dioxide concentrations rise, Nature, 499,
324–7, 2013.
Koenig, W. D. and Knops, M. H.: Patterns of annual seed production by
Northern Hemisphere trees: a global perspective, Am. Nat., 155, 59–69,
2000.
Körner, C.: Growth Controls Photosynthesis – Mostly, Nova
Acta Leopoldina, 283, 273–283, 2013.Leonardi, S., Gentilesca, T., Guerrieri, R., Ripullone, F., Magnani, F.,
Mencuccini, M., Noije, T. V., and Borghetti, M.: Assessing the effects of
nitrogen deposition and climate on carbon isotope discrimination and
intrinsic water-use efficiency of angiosperm and conifer trees under rising
CO2 conditions, Glob. Change Biol., 18, 2925–2944, 2012.
Le Roux, X., Lacointe, A., Escobar-Gutierrez, A., and Le Dizes, S.:
Carbon-based models of individual tree growth: A critical appraisal, Ann.
For. Sci., 58, 469–506, 2001.
Leuning, R.: A critical appraisal of a combined stomatal-photosynthesis
model for C3 plants, Plant Cell Environ., 18, 339–355, 1995.
Lévesque, M., Siegwolf, R., Saurer, M., Eilmann, B., and Rigling, A.:
Increased water-use efficiency does not lead to enhanced tree growth under
xeric and mesic conditions, New Phytol., 203, 94–109, 2014.Li, G., Harrison, S. P., Prentice, I. C., and Falster, D.: Simulation of tree-ring widths
with a model for primary production, carbon allocation, and growth, Biogeosciences, 11, 6711–6724, 10.5194/bg-11-6711-2014, 2014.Limousin, J. M., Rambal, S., Ourcival, J. M., Rocheteau, A., Joffre, R., and
Rodriguez-Cortina, R.: Long-term transpiration change with rainfall decline
in a Mediterranean Quercus ilex forest, Glob. Change Biol., 15, 2163–2175, 2009.Limousin, J. M., Longepierre, D., Huc, R., and Rambal, S.: Change in
hydraulic traits of Mediterranean Quercus ilex subjected to long-term throughfall
exclusion, Tree Physiol., 30, 1026–1036, 2010.
Limousin, J.-M., Rambal, S., Ourcival, J.-M., Rodriguez-Calcerrada, J.,
Perez-Ramos, I. M., Rodriguez-Cortina, R., Misson, L., and Joffre, R.:
Morphological and phenological shoot plasticity in a Mediterranean evergreen
oak facing long-term increased drought, Oecologia, 169, 565–577, 2012.
Maseyk, K. S., Lin, T., Rotenberg, E., Grünzweig, J. M., Schwartz, A.,
and Yakir, D.: Physiology-phenology interactions in a productive semi-arid
pine forest, New Phytol., 178, 603–16, 2008.
McDowell, N.G., Fisher, R.A., Xu, C., Domec, J., Höltta, T., Mackay,
D.S., Sperry, J. S., Boutz, A., Dickman, L., Gehres, N., Limousin, J.M.,
Macalady, A., Pangle, R. E., Rasse, D.P., Ryan, M.G., Sevanto, S., Waring,
R.H., Williams, A.P., Yepez, E. A., and Pockman, W.T.: Evaluating theories of
drought-induced vegetation mortality using a multimodel – experiment
framework, New Phytol., 200, 304–321, 2013.
McMurtrie, R. E. and Dewar, R. C.: New insights into carbon allocation by
trees from the hypothesis that annual wood production is maximized, New
Phytol., 199, 981–990, 2013.
Millard, P., Sommerkorn, M., and Grelet, G. A.: Environmental change and carbon
limitation in trees: A biochemical, ecophysiological and ecosystem
appraisal, New Phytol., 175, 11–28, 2007.
Misson, L.: MAIDEN: a model for analyzing ecosystem processes in
dendroecology, Can. J. For. Res., 34, 874–887, 2004.
Misson, L., Rathgeber, C., and Guiot, J.: Dendroecological analysis of
climatic effects on Quercus petraea and Pinus halepensis radial growth using
the process-based MAIDEN model, Can. J. For. Res., 34, 888–898, 2004.
Misson, L., Tang, J., Xu, M., Mckay, M. ,and Goldstein, A.: Influences of
recovery from clear-cut , climate variability , and thinning on the carbon
balance of a young ponderosa pine plantation, Agric. Fore, 130, 207–222,
2005.
Misson, L., Degueldre, D., Collin, C., Rodriguez, R., Rocheteau, A.,
Ourcival, J.-M., and Rambal, S.: Phenological responses to extreme droughts
in a Mediterranean forest, Glob. Change Biol., 17, 1036–1048, 2011.
Montserrat-Marti, G., Camarero, J. J., Palacio, S., Perez-Rontome, C.,
Milla, R., Albuixech, J., and Maestro, M.: Summer-drought constrains the
phenology and growth of two coexisting Mediterranean oaks with contrasting
leaf habit: implications for their persistence and reproduction,
Trees-Structure Funct., 23, 787–799, 2009.
Niinemets, U.: Photosynthesis and resource distribution through plant
canopies, Plant. Cell Environ., 30, 1052–71, 2007.
Niinemets, U. and Valladares, F.: Photosynthetic acclimation to simultaneous
and interacting environmental stresses along natural light gradients:
optimality and constraints, Plant Biol. (Stuttg), 6, 254–68, 2004.Niinemets, Ü., Tenhunen, J. D., Canta, N. R., Chaves, M. M., Faria, T.,
Pereira, J. S., and Reynolds, J. F.: Interactive effects of nitrogen and
phosphorus on the acclimation potential of foliage photosynthetic properties
of cork oak, Q. suber, to elevated atmospheric CO2 concentrations, Glob. Change Biol., 5, 455–470, 1999.
Nobel, P. S.: Physicochemical and environmental plant physiology. 4th edn.
Academic Press, Elsevier, Oxford UK, 568 pp., 2009.
Peng, C. H., Guiot, J., Wu, H. B., Jiang, H., and Luo, Y. Q.: Integrating
models with data in ecology and palaeoecology: advances towards a model-data
fusion approach, Ecol. Lett., 14, 522–536, 2011.
Peñuelas, J., Hunt, J. M., Ogaya, R., and Jump, A. S.: Twentieth century
changes of tree-ring delta C-13 at the southern range-edge of Fagus
sylvatica: increasing water-use efficiency does not avoid the growth decline
induced by warming at low altitudes, Glob. Change Biol., 14, 1076–1088,
2008.
Peñuelas, J., Canadell, J. G., and Ogaya, R.: Increased water-use
efficiency during the 20th century did not translate into enhanced tree
growth, Glob. Ecol. Biogeogr., 20, 597–608, 2011.Pereira, J. S., Mateus, J. A., Aires, L. M., Pita, G., Pio, C., David, J. S., Andrade, V.,
Banza, J., David, T. S., Paço, T. A., and Rodrigues, A.: Net ecosystem carbon exchange in
three contrasting Mediterranean ecosystems – the effect of drought, Biogeosciences, 4, 791–802, 10.5194/bg-4-791-2007, 2007.
Pérez-Ramos, I. M., Ourcival, J. M., Limousin, J. M., and Rambal, S.:
Mast seeding under increasing drought: results from a long-term data set and
from a rainfall exclusion experiment, Ecology, 91, 3057–3068, 2010.Piovesan, G., Biondi, F., Di Filippo, A., Alessandrini, A., and Maugeri, M.:
Drought-driven growth reduction in old beech (Fagus sylvatica L.) forests of the central
Apennines, Italy, Glob. Change Biol., 14, 1265–1281, 2008.Rambal, S., Joffre, R., Ourcival, J. M., Cavender-Bares, J. ,and Rocheteau,
a.: The growth respiration component in eddy CO2 flux from a Quercus ilex mediterranean
forest, Glob. Change Biol., 10, 1460–1469, 2004.Reichstein, M., Tenhunen, J. D., Roupsard, O., Ourcival, J. M., Rambal, S.,
Miglietta, F., Peressotti, A., Pecchiari, M., Tirone, G., and Valentini, R.:
Severe drought effects on ecosystem CO2 and H2O fluxes at three
Mediterranean evergreen sites: revision of current hypotheses?, Glob. Change Biol., 6, 999–1017, 2002.Reichstein, M., Tenhunen, J., Roupsard, O., Ourcival, J. M., Rambal, S.,
Miglietta, F., Peressotti, A., Pecchiari, M., Tirone, G., and Valentini, R.:
Inverse modeling of seasonal drought effects on canopy CO2/H2O exchange in
three Mediterranean ecosystems, J. Geophys. Res., 108, 4726, 10.1029/2003JD003430, 2003.
Reichstein, M., Falge, E., Baldocchi, D., Papale, D., Aubinet, M.,
Berbigier, P., Bernhofer, C., Buchmann, N., Gilmanov, T., Granier, A.,
Grünwald, T., Havránková, K., Ilvesniemi, H., Janous,
D., Knohl, A., Laurila, T., Lohila, A., Loustau, D., Matteucci, G., Meyers,
T., Miglietta, F., Ourcival, J. M., Pumpanen, J., Rambal, S., Rotenberg, E.,
Sanz, M., Tenhunen, J., Seufert, G., Vaccari, F., Vesala, T., Yakir, D., and
Valentini, R.: On the separation of net ecosystem exchange into assimilation
and ecosystem respiration: Review and improved algorithm, Glob. Change Biol., 11, 1424–1439, 2005.Sala, A. and Tenhunen, J. D.: Simulations of canopy net photosynthesis and
transpiration in Quercus ilex L. under the influence of seasonal drought, Agric. For.
Meteorol., 78 203–222, 1996.
Sala, A., Woodruff, D. R., and Meinzer, F. C.: Carbon dynamics in trees:
feast or famine?, Tree Physiol., 32, 764–75, 2012.Salzer, M. G., Hughes, M. K., Bunn, A. G., and Kipfmueller, K. F.: Recent
unprecedented tree-ring growth in bristlecone pine at the highest elevations
and possible causes, P. Natl. Acad. Sci., 106, 20348–20353, 2009.
Saurer, M., Spahni, R., Frank, D.C., Joos, F., Leuenberger, M., Loader,
N.J., McCarroll, D., Gagen, M., Poulter, B., Siegwolf, R. T. W.,
Andreu-Hayles, L., Boettger, T., Dorado, I., Fairchild, I. J., Friedrich,
M., Gutierrez, E., Haupt, M., Hilasvuori, E., Heinrich, I., Helle, G.,
Grudd, H., Jalkanen, R., Levanič, T., Linderholm, H. W., Robertson, I.,
Sonninen, E., Treydte, K., Waterhouse, J. S., Woodley, E. J., Wynn, P. M.,
and Young, G. H. F.: Spatial variability and temporal trends in water-use
efficiency of European forests, Glob. Change Biol., 20, 3700–3712, 2014.Schaefer, K., Schwalm, C. R., Williams, C., Arain, M. A., Barr, A., Chen,
J. M., Davis, K. J., Dimitrov, D., Hilton, T. W., Hollinger, D. Y., Humphreys,
E., Poulter, B., Raczka, B. M., Richardson, A. D., Sahoo, A., Thornton, P.,
Vargas, R., Verbeeck, H., Anderson, R., Baker, I., Black, T.A., Bolstad, P.,
Chen, J., Curtis, P.S., Desai, A. R., Dietze, M., Dragoni, D., Gough, C.,
Grant, R. F., Gu, L., Jain, A., Kucharik, C., Law, B., Liu, S., Lokipitiya,
E., Margolis, H.A., Matamala, R., McCaughey, J. H., Monson, R., Munger,
J.W., Oechel, W., Peng, C., Price, D. T., Ricciuto, D., Riley, W. J.,
Roulet, N., Tian, H., Tonitto, C., Torn, M., Weng, E., and Zhou, X.: A
model-data comparison of gross primary productivity: Results from the North
American Carbon Program site synthesis, J. Geophys. Res., 117, G03010,
10.1029/2012JG001960,
2012.
Simioni, G., Durand-Gillmann, M., and Huc, R.: Asymmetric competition
increases leaf inclination effect on light absorption in mixed canopies,
Ann. For. Sci., 70, 123–131, 2013.Sun, Y., Gu, L., Dickinson, R. E., Norby, R. J., Pallardy, S. G., and
Hoffman, F. M.: Impact of mesophyll diffusion on estimated global land
CO2
fertilization, P. Natl. Acad. Sci. USA, 111, 15774–15779, 2014.
Tolwinski-Ward, S. E., Evans, M. N., Hughes, M. K., and Anchukaitis, K. J.:
An efficient forward model of the climate controls on interannual variation
in tree-ring width, Clim. Dyn., 36, 2419–2439, 2011.Touchan, R., Shishov, V. V., Meko, D. M., Nouiri, I., and Grachev, A.: Process based model sheds
light on climate sensitivity of Mediterranean tree-ring width, Biogeosciences, 9, 965–972, 10.5194/bg-9-965-2012, 2012.
Vaganov, E. A., Hughes, M. K., and Shashkin, A. V.: Growth Dynamics of Conifer Tree
Rings: Images of Past and Future Environments, Springer, New York, 354 pp., 2006.Vaz, M., Pereira, J. S., Gazarini, L. C., David, T. S., David, J. S.,
Rodrigues, A., Maroco, J., and Chaves, M. M.: Drought-induced photosynthetic
inhibition and autumn recovery in two Mediterranean oak species (Quercus ilex and
Quercus suber), Tree Physiol., 30, 946–956, 2010.
Voltas, J., Camarero, J. J., Carulla, D., Aguilera, M., Ortiz, A., and
Ferrio, J. P.: A retrospective, dual-isotope approach reveals individual
predispositions to winter-drought induced tree dieback in the southernmost
distribution limit of Scots pine, Plant. Cell Environ., 36, 1435–48,
2013.