Does physiological acclimation to climate warming stabilize the …€¦ · etal., 2010). Such shifts in the temperature response of photosyn-thesis and respiration via physiological

Does physiological acclimation to climate warming stabilize theratio of canopy respiration to photosynthesis?

John E. Drake1, Mark G. Tjoelker1, Michael J. Aspinwall1, Peter B. Reich1,2, Craig V. M. Barton1, Belinda E.

Medlyn1 and Remko A. Duursma1

1Hawkesbury Institute for the Environment, Western Sydney University, Locked Bag 1797, Penrith, NSW 2751, Australia; 2Department of Forest Resources, University of Minnesota, 1530

Cleveland Ave N., St Paul, MN 55108, USA

Author for correspondence:John E. Drake

Tel: +61 02 4570 1018Email: [email protected]

Received: 17 November 2015

Accepted: 4 March 2016

New Phytologist (2016)doi: 10.1111/nph.13978

Key words: acclimation, autotrophicrespiration, carbon cycle, climate change,Eucalyptus tereticornis, photosynthesis,positive feedback, warming.

Summary

� Given the contrasting short-term temperature dependences of gross primary production

(GPP) and autotrophic respiration, the fraction of GPP respired by trees is predicted to

increase with warming, providing a positive feedback to climate change. However, physiolog-

ical acclimation may dampen or eliminate this response.� We measured the fluxes of aboveground respiration (Ra), GPP and their ratio (Ra/GPP) in

large, field-grown Eucalyptus tereticornis trees exposed to ambient or warmed air tempera-

tures (+3°C). We report continuous measurements of whole-canopy CO2 exchange, direct

temperature response curves of leaf and canopy respiration, leaf and branch wood respiration,

and diurnal photosynthetic measurements.� Warming reduced photosynthesis, whereas physiological acclimation prevented a coinci-

dent increase in Ra. Ambient and warmed trees had a common nonlinear relationship between

the fraction of GPP that was respired above ground (Ra/GPP) and the mean daily tempera-

ture. Thus, warming significantly increased Ra/GPP by moving plants to higher positions on

the shared Ra/GPP vs daily temperature relationship, but this effect was modest and only

notable during hot conditions.� Despite the physiological acclimation of autotrophic respiration to warming, increases in

temperature and the frequency of heat waves may modestly increase tree Ra/GPP, contribut-

ing to a positive feedback between climate warming and atmospheric CO2 accumulation.

Introduction

Trees exchange large quantities of carbon (C) with the atmosphereand are a major driver of the global C cycle (Bonan, 2008; Panet al., 2011). Terrestrial C uptake via gross primary production(GPP) is the largest flux in the C cycle at ~120 Gt C yr�1, approx-imately half of which is returned to the atmosphere by themetabolism of autotrophic respiration (R; Piao et al., 2013; Sch-lesinger & Bernhardt, 2013). Projections of future climate dependon how these large C fluxes are influenced by climate change(Friedlingstein et al., 2006, 2014). Of particular concern is thatwarming of the Earth’s surface may increase R relative to GPP,providing a positive feedback to climate warming (Cox et al.,2000; Frank et al., 2010; Wang et al., 2013; Zhang et al., 2014).

Photosynthesis (or ‘A’ for assimilation) and R have contrastingtemperature sensitivities, such that the ratio of R/A is predicted toincrease with temperature; we illustrate this with a leaf-scalemodel (Fig. 1; see the Materials and Methods section). In theshort term, A increases with temperature to an optimum (Topt)and subsequently declines at higher temperatures. The biochemi-cal basis for this temperature response is complex, but relativelywell understood (see reviews; Berry & Bj€orkman, 1980;

McMurtrie & Wang, 1993; Sage & Kubien, 2007; Way & Oren,2010; Yamori et al., 2014). In addition, increasing air tempera-ture is typically associated with increasing vapor pressure deficit(VPD); stomatal closure with increasing VPD alters the realizedtemperature response of photosynthesis and reduces Topt (Linet al., 2012). In contrast with photosynthesis, R increases approx-imately exponentially with increasing temperature up to a maxi-mum (50–60°C), beyond which irreversible damage to therespiratory apparatus occurs (Huve et al., 2011; O’Sullivan et al.,2013). Given these contrasting temperature dependences, theratio of R/A is predicted to increase with temperature, such thatclimate warming would increase R/A (solid curve and red arrowin Fig. 1b).

However, the shapes of the short-term temperature responsesof photosynthesis and respiration are not static, as plants have thecapacity to adjust to a changing environment via physiologicalacclimation. Here, we use the term ‘acclimation’ to describe thealteration of the temperature response of photosynthesis and/orrespiration, regardless of whether this change is perceived as bene-ficial (Way & Yamori, 2014). Following an increase in tempera-ture of several days or longer, respiration at a given temperaturetends to be reduced, such that the rate of respiration in the new

� 2016 The Authors

New Phytologist� 2016 New Phytologist Trust

New Phytologist (2016) 1www.newphytologist.com

Research

thermal environment is similar to the rate of respiration in theprevious thermal environment (i.e. homeostatic acclimation;Atkin & Tjoelker, 2003; Atkin et al., 2005a; Lee et al., 2005;Tjoelker et al., 2009; Slot & Kitajima, 2015; Reich et al., 2016).Such acclimation to warming often manifests as a downwards shiftin the temperature response curve of respiration (dashed relativeto solid line in Fig. 1a). Photosynthetic acclimation to warming isvariable in nature and magnitude (Way & Yamori, 2014; Yamoriet al., 2014). However, a common acclimation response to long-term warming is an increase in Topt of photosynthesis (dashed rel-ative to solid line in Fig. 1a; Kattge & Knorr, 2007; Gundersonet al., 2010). Such shifts in the temperature response of photosyn-thesis and respiration via physiological acclimation can dampenthe increase in R/A at high temperatures, such that climate warm-ing would not increase R/A (dashed curve and orange arrow,Fig. 1b). Some authors have even suggested R/A to be constantacross variation in temperature, if respiratory acclimation

follows directly from photosynthetic acclimation via substratelimitation of respiration (Dewar et al., 1999; Gifford, 2003).

Although considerable research has demonstrated the impor-tance of physiological acclimation for R/A ratios at the leaf scale(e.g. Gunderson et al., 2000; Atkin et al., 2006; Reich et al.,2016), the response at the canopy scale of large trees remainslargely unstudied. In part, this is because A and R are challengingto measure at the scale of entire trees, particularly with experi-mental warming. We are aware of only one study of whole-plantR/A response to experimental warming within a woody species.Tjoelker et al. (1999) measured whole-plant respiration as a pro-portion of daily photosynthesis in a growth cabinet experimentwith five boreal tree species grown at three temperature condi-tions. R/A increased at the highest temperature, but the extent ofthis increase was ameliorated by the physiological acclimation ofR. Two other studies of whole-plant R/A response to warmingfound similar results in potted wheat (Gifford, 1995) andPlantago (Atkin et al., 2007). However, acclimation of A and R toexperimental warming has not been investigated at the canopyscale in large field-grown trees.

The purpose of this study was to measure the response ofwhole-canopy GPP, aboveground autotrophic respiration (Ra)and their ratio (Ra/GPP) in field-grown trees under naturally vari-able environmental conditions with and without experimentalwarming of +3°C using the unique infrastructure of 12 whole-tree chambers (WTCs). Whole-canopy measurements in the fieldare particularly useful because they capture the complexities ofreal canopies, including all foliage types, the contribution of woodrespiration, a variable light environment, and substantial diurnaland seasonal variation in temperature. We complement thesewhole-canopy measurements with leaf-scale measurements ofphotosynthesis and respiration. We hypothesize that physiologicalacclimation of photosynthesis and respiration to experimentalwarming will prevent a warming-induced increase in Ra/GPP atthe canopy scale. We specifically predict that: Ra will acclimateapproximately homeostatically, resulting in a downwards shift inthe temperature response curve of Ra and no difference in in-siturespiration rates per unit leaf area between ambient and warmedtreatments; and leaf-level photosynthetic rates will increase withwarming, possibly as an acclimation response (e.g. increase inTopt) or simply by moving up the temperature response curve iftemperatures are below Topt. These effects will combine such thatthe temperature response of Ra/GPP will be moderated withwarming, so that Ra/GPP will not increase with experimentalwarming (orange arrow in Fig. 1b).

Materials and Methods

Illustrative modeling of the temperature dependences of Aand R

The temperature dependences of leaf photosynthesis and respira-tion (A and R, Fig. 1) were illustrated with a leaf-scale model ofphotosynthesis (Farquhar et al., 1980) coupled to an optimalstomatal conductance model (Medlyn et al., 2011) as imple-mented in the ‘PLANTECOPHYS’ R package (Duursma et al., 2014;

20 30 40

20 30 40

0

10

20

Leaf

CO

2ex

chan

ge(µ

mol

CO

2m

−2s−1

)

A R

(a)

(b)

0

0.2

0.4

0.6

RA

Tleaf (°C)Fig. 1 Demonstration of the temperature dependence of gross leafphotosynthesis and respiration (A and R, respectively; a) and the fractionof leaf respiration to photosynthesis (R/A; b). Solid lines reflect the case ofstatic temperature response curves, whereas dashed lines reflect howtemperature response curves may change with physiological acclimationto warmed temperatures. In the absence of acclimation, warmingincreases R/A (red arrow). With acclimation, warming may not increaseR/A (orange arrow).

New Phytologist (2016) � 2016 The Authors

New Phytologist� 2016 New Phytologist Trustwww.newphytologist.com

Research

NewPhytologist2

Duursma, 2015). The temperature dependence of A at saturatinglight arises from the temperature dependence of the maximumvelocity of Rubisco carboxylation (Vc,max) and electron transport(Jmax), which follow peaked Arrhenius equations (Medlyn et al.,2002), and stomatal closure in response to increasing VPD. Thetemperature dependence of R follows a simple Arrhenius equa-tion. These represent the temperature dependences of A and Rused in many ecosystem and Earth system models (Smith &Dukes, 2013).

Model simulations were performed under constant environmen-tal conditions apart from leaf temperature (Tleaf); the dew pointwas 10°C, the atmospheric CO2 concentration was400 lmol mol�1 and the photosynthetic photon flux density(PPFD) was 1500 lmolm�2 s�1. Vc,max and Jmax were 100 and150 lmol CO2m

�2 s�1 at 25°C, respectively, and the stomatalparameter (g1) was set to 4.0 (Lin et al., 2013; Gimeno et al.,2015). Physiological acclimation (dashed lines, Fig. 1) was illus-trated by increasing Topt of photosynthesis, reducing Vc,max andJmax at 25°C and reducing the pre-exponential coefficient for respi-ration. These adjustments qualitatively reflect observations of leaf-level temperature acclimation to warming (Smith &Dukes, 2013).

Site and experimental design

We implemented a warming experiment using 12 WTCs inRichmond, New South Wales (Australia) at 25 m above sea level(33°36ʹ40ʺS, 150°44ʹ26.5ʺE). The WTCs are large, approxi-mately cylindrical structures topped with a cone that enclose asingle tree rooted in soil (3.25 m in diameter, 9 m in height, vol-ume of ~53 m3; photographs in Supporting InformationFig. S1). The WTCs control Tair, relative humidity (RH) andatmospheric CO2 concentration, while measuring the net fluxesof CO2 and H2O exchange between the entire canopy and theatmosphere (Barton et al., 2010, 2012; Crous et al., 2013;Duursma et al., 2014). A suspended plastic floor was sealedaround the stem of each tree at ~45 cm in height, isolating thecanopy gas exchange from CO2 and H2O evolved from the soil.

Soils at the site are an alluvial formation of low-fertility sandyloam (Clarendon sand). A cemented layer of manganese nodulesand clay is present below 70–100 cm depth. Soil was collectedfrom an adjacent paddock and placed in the chambers in two lay-ers (0–25 cm and from 25 cm to the depth of the hard layer) on10 July 2012. A root exclusion barrier of heavy duty polyethylene(300 lm thick) extends vertically into the hard soil layer, isolat-ing the soil volume of each chamber.

Six chambers tracked ambient Tair and six chambers trackedambient Tair + 3°C warming (n = 6; ‘ambient’ and ‘warmed’);treatments started on 12 December 2012. The average dailymean temperature difference between treatments was2.9� 0.3°C (standard deviation (SD)). As recommended forrealistic warming manipulations (Amthor et al., 2010; Lin et al.,2012), RH in the warmed treatment was controlled to match theRH observed in the ambient treatment using a custom air condi-tioning system, as described previously by Barton et al. (2010).During daylight hours, RH averaged 62.5� 20% and62.4� 20% (SD) in the ambient and warmed treatments,

respectively. Given the difference in Tair, the VPD was lower inthe ambient relative to the warmed treatment (1.27� 1.1 and1.50� 1.2 kPa (SD), respectively). The [CO2] within all cham-bers was controlled at 10 lmol mol�1 above the ambient atmo-spheric concentration; the mean [CO2] within the chambersduring daylight hours was 417� 40 lmol mol�1 (SD).

Nursery seedlings of a local provenance of Eucalyptustereticornis Sm. were established in 25-l pots inside the WTCsusing the same surface soil as used to fill the WTCs. Eucalyptustereticornis was chosen because it is a widespread and abundanttree across eastern Australia from �38.5° latitude in southernVictoria to �14° latitude in northern Queensland (Drake et al.,2015). Six potted trees were placed in each chamber on 5 Decem-ber 2012; a single tree was selected based on size similarity withineach treatment and planted in the chamber centre on 12 March2013. Trees were irrigated regularly and equally every 15 d withhalf the mean monthly rainfall. This irrigation schedule resultedin ‘well-watered’ trees, as monthly leaf water potential measure-ments with a Scholander pressure chamber (PMS Corp., Corval-lis, OR, USA) indicated that the trees were not water stressed;pre-dawn and mid-day leaf water potentials were moderate anddid not differ between treatments (mean and standard error (SE):pre-dawn values of �0.26� 0.02 and �0.24� 0.02MPa andmid-day values of �1.2� 0.3 and �1.2� 0.3 MPa in ambientand warmed treatments, respectively). A water exclusion treat-ment was added to half of the trees on 12 February 2014; wereport results from the well-watered trees only (n = 6 until 12February 2014, n = 3 thereafter).

Tree size was quantified every 14 d with measurements ofheight and stem diameter at 65 cm above the soil surface. Treeswere equivalent in size when the treatments began on 12 Decem-ber 2012 (diameters of 0.24� 0.01 and 0.26� 0.01 cm andheights of 0.41� 0.01 and 0.40� 0.02 m in ambient andwarmed treatments, respectively). The warmed trees were largerthan the ambient-grown trees when the floors were installed andflux measurements began, and tree size was equivalent acrosstreatments at the end of the measurement period (Table 1).

Temperature response of leaf and whole-canopyrespiration

We measured the direct response of leaf and whole-canopy respi-ration (Rleaf and Rcanopy) to short-term temperature variation for

Table 1 Mean (SE) tree size and total canopy leaf area at the beginningand end of the flux measurement period

Date TreatmentDiameter(cm)c

Height(m)

Total canopyleaf area (m2)

13 September 2013a Ambient 2.4 (0.1) 3.0 (0.1) 3.3 (0.4)13 September 2013a Warmed 2.9 (0.2) 3.6 (0.2) 5.0 (0.6)26 May 2014b Ambient 6.7 (0.3) 8.9 (0.4) 19.1 (3.8)26 May 2014b Warmed 7.0 (0.5) 8.8 (0.4) 21.1 (4.6)

aPlastic floor installed in each chamber – flux measurements begin.bTree harvest began – flux measurements end.cMeasured at 65 cm height.


New Phytologist� 2016 New Phytologist TrustNew Phytologist (2016)

www.newphytologist.com

NewPhytologist Research 3

two purposes: these data directly related to the hypothesis (pre-diction 1), and were used to inform the partitioning of whole-canopy CO2 exchange (see later). We express these gas exchangefluxes per unit leaf area, although the ratio of particular interest(Ra/GPP) does not depend on the normalizing unit.

We measured the temperature response curves of Rleaf via anestablished protocol (O’Sullivan et al., 2013) using a large gasexchange chamber (3010-GWK1, Heinz Walz GmbH, Effel-trich, Germany) connected to an infrared gas analyser (IRGA; LI6400-XT, LiCor, Lincoln, NE, USA), which continuouslyrecorded Rleaf as Tleaf increased at 1°Cmin�1 from 18 to 40°C.Measurements were made on one mature leaf per chamber, sam-pled from the upper third of the canopy on 7 February 2014. Weexplored many functions to describe the temperature sensitivityof Rleaf (Eqns 1–5 in O’Sullivan et al., 2013). We selected theArrhenius equation as the most parsimonious description of therelationship between Rleaf and temperature, as the relationshipbetween ln(Rleaf) and the inverse of Tleaf was linear.

We directly measured the temperature response of whole-canopy respiration on one night (10 February 2014) using theWTC environmental controls to generate five Tair setpoints (18,22.5, 25, 28.5 and 32°C). The setpoint order was randomizedacross chambers such that temperature and time of night werenot confounded; trees were allowed to adjust to each temperaturesetpoint for 30 min before a measurement interval of 1 h per set-point. Each WTC was transformed into a closed system by seal-ing the air inlets and outlets, such that the rate of CO2

accumulation inside the canopy airspace could be used as a mea-sure of canopy respiration. Chambers were vented at the begin-ning of each setpoint to return the chamber [CO2] to theatmospheric concentration. Chamber [CO2] was measured everyminute with a local IRGA mounted on each WTC (ModelSBA1, PP-systems, Amesbury, MA, USA). Canopy respirationwas estimated via mass balance as in Perez-Priego et al. (2010):

dCi

dt¼ �hðCi � CaÞ þ F

VEqn 1

where h is a chamber-specific leak coefficient (m3 s�1), Ci is the[CO2] inside the chamber (lmol mol�1), Ca is the [CO2] in theatmosphere outside the chamber (lmol mol�1), F is the flux ofCO2 from canopy respiration (lmol CO2 mol�1 m3 s�1) and Vis the chamber volume (52.8 m3). The leak rate was quantifiedfor each chamber after harvest when the chambers were empty; aCO2 cylinder was used to increase Ci to ~1000 lmol mol�1, thechambers were closed and measurements over the next 2 h wereemployed to estimate h using Eqn (2) assuming F = 0. Whole-canopy respiration was calculated by multiplying F by the molarvolume of air (mol m�3); differences in tree size were accountedfor by expressing the flux per total leaf area (see later).

Given the clearly exponential relationship between Rleaf andtemperature, the whole-canopy respiration vs Tair relationshipswere also fitted as exponential relationships, although we recog-nize that linear models also fit well. An exponential fit betweenrespiration and temperature has several advantages over linear fits.First, exponential relationships are supported by extensive

research and are more mechanistically sound relative to linearrelationships (Atkin & Tjoelker, 2003; Atkin et al., 2005b; Slot& Kitajima, 2015). Second, exponential fits have the advantageof parsimony, as only three parameters are required (Ea is equiva-lent across treatments), whereas four parameters are required forthe linear fits. Finally, exponential fits enable robust treatment oftype II acclimation of respiration via adjustment of the pre-exponential coefficient (Atkin et al., 2005b), whereas the slopeand intercept of linear models would have to be adjusted simulta-neously to accomplish the same result.

Leaf and branch wood respiration

We measured leaf and branch wood respiration at a constant tem-perature of 15°C to investigate thermal acclimation via the settemperature method (Atkin et al., 2005b). Two random primarybranches were sampled at the end of the study from each tree 2 hafter dusk by cutting the branch flush at the stem insertion point.These branches had an average diameter of 6.7 mm, were >1 m inlength and held an average of 76 leaves. All leaves were detachedand respiration was measured separately for a pooled sample ofthree randomly selected leaves from each branch (Li-6400-22L;LiCor). The cuvette temperature was maintained at 15°C with aflow rate of 350 lmol s�1. Branches were cut into 10-cm-longsegments and entirely enclosed in a large gas exchange chamber(3010-GWK1, Walz) connected to an IRGA (LI 6400-XT,LiCor). The chamber temperature was maintained at 15°C with aflow rate of 600 lmol s�1. Branch and leaf mass were measureddirectly after drying at 70°C. To compare rates across leaves andbranches, these respiration rates were expressed per unit dry mass.We recognize that cut branches could have released CO2 dis-solved in the xylem (e.g. Bloemen et al., 2013), but we expect thissystematic error to affect both treatments equally.

Continuous whole-canopy CO2 exchange measurements

An automated system measured the net exchange of CO2

between the canopy and the atmosphere within each chamber at15 min resolution; see Barton et al. (2010) for a thoroughdescription. We report hourly averages, following Duursma et al.(2014). Measurements began on 13 September 2013 when plas-tic floors were installed in each chamber and sealed around thestem of each tree at ~45 cm height, when the trees were ~3 m tall.Flux measurements finished on 26 May 2014 when the canopyharvest began and the trees were nearly 9 m tall (Table 1). Wereport >70 000 hourly flux observations aggregated into >3000daily sums across 12 trees.

Investigator entry into chambers and sensor failures createdgaps in the flux record (~6% missing data). Missing data weregapfilled using neural networks (Riedmiller & Braun, 1993;Moffat et al., 2007) with two layers of two neurons each,informed by PPFD, Tair, tree size (square of diameter at 65 cmmultiplied by height), VPD and the hour of the day. Neural net-works were fitted for each chamber using the ‘NEURALNET’ Rpackage (R Development Core Team, 2012). Ten per cent of thedata were withheld for validation (~7000 observations);



Research

NewPhytologist4

measured and gapfilled CO2 fluxes were strongly correlated witha slope that approached unity (measured flux = 0.002 +0.9139 gapfilled flux, P < 0.001, r2 = 0.88), suggesting that thegapfilling procedure was robust.

The measured net exchange of CO2 between the canopy andthe atmosphere reflects the balance between GPP and above-ground respiration (Ra). During the night, we used the directmeasurements of CO2 evolution as a measure of Ra. For daytimeperiods, we partitioned the measured net flux into GPP and Ra byestimating Ra as a function of temperature, as is widely performedin eddy covariance (Reichstein et al., 2005). We used the directlymeasured temperature sensitivity (Ea, activation energy) ofwhole-canopy respiration; we assumed that Ea was constant acrosstime, but that the basal respiration rate of whole-canopy respira-tion could have acclimated seasonally or in response to treatment.Thus, Ra was estimated as a function of temperature as:

RT ¼ RT ref � eEa

rT refð Þ� 1�T refTð Þ½ � Eqn 2

where RT is the estimated whole-tree respiration rate at tempera-ture T, RTref is the basal respiration rate at a reference tempera-ture Tref, Ea is the measured whole-canopy respiration activationenergy, r is the universal gas constant, Tref is the reference tem-perature (K) and T is the measured hourly temperature (K). Dailychamber-specific Tref and RTref values were estimated from themeasured nightly flux data as the mean Tair and net CO2 fluxmeasured over the preceding three nights. This produced hourlyestimates of GPP and Ra (e.g. Fig. S2), which were summed foreach chamber by 24-h blocks of time that began at dawn andended at the following dawn.

This flux partitioning approach (Eqn 2) ignores the inhibitionof leaf mitochondrial respiration in the light (e.g. Brooks & Far-quhar, 1985; Tcherkez et al., 2009; Crous et al., 2012; Atkinet al., 2013). We explored the implications of accounting for thiseffect, as described in detail in Notes S1. Accounting for the lightinhibition of Ra modestly affected the daily estimates of GPP, Raand Ra/GPP (Figs S3, S4), but this did not alter our conclusionsregarding the temperature dependence of Ra/GPP (Fig. S5). Forthe sake of simplicity, we present the results using the flux parti-tioning method with no light inhibition (i.e. Eqn 2).

Estimation of leaf area

Canopy leaf area was estimated to express GPP and Ra on a leafarea basis using methods applied in a previous experiment at thissite (Barton et al., 2012). Leaf area was measured for each tree ontwo occasions by counting all leaves and multiplying by a tree-specific mean leaf size measured across the canopy of each treewith a handheld leaf area meter (LI-3000; LiCor; 86–102 leavesper tree). These measurements were performed in the spring (9September 2013) and summer (10 February 2014); in total86 431 leaves were counted. A third direct measurement of leafarea was calculated from the final harvest in the autumn (26 May2014). The canopy of each tree was divided into three equalheights, all branches were cut flush to the stem, grouped by layerand all leaves were separated from the branches. After mixing the

leaves within each canopy layer, a random subsample of 100leaves was measured for total leaf area (LI-3100C; LiCor), drymass and specific leaf area (SLA). Total canopy leaf area was cal-culated by multiplying the total leaf dry mass by SLA, weightedby the leaf dry mass in each canopy layer. Litterfall was collected,dried and weighed fortnightly for each tree. The standing leafarea for each tree was estimated from these data as in Barton et al.(2012):

Lðt Þ ¼ aðH ðt ÞÞb �Z t

s¼0

Z ðsÞds Eqn 3

where L(t) is the standing leaf area at time t, H(t) is the height ofthe tree at time t, a and b are tree-specific components fitted tothe three direct measurements of leaf area and Z(t) reflects the rateat which leaf area was removed by litterfall at time t. Totalcanopy leaf area was higher in the warmed relative to the ambienttreatment at the beginning of the flux measurement period, butthis difference was reduced over time, such that the treatmentshad equivalent total canopy leaf area at the end of the measure-ment period (Table 1).

Leaf-scale photosynthetic measurements

We complemented the whole-canopy flux measurements withdiurnal measurements of leaf net photosynthesis (Anet) on fivedates. Measurements began 30 min after dawn and commencedevery 2–3 h until dusk with four gas exchange systems operatingconcurrently (Li-6400 with Li-6400-02B LED light source;LiCor). On each date, a fully expanded leaf in the top third ofthe canopy was marked and used for all measurements. Immedi-ately before each measurement time point, the light sources ofthe gas exchange systems were set to the incident PPFD mea-sured by a handheld sensor in an adjacent open field (Li-210;LiCor). Thus, Anet was measured at approximately in-situ PPFD,but PPFD was held constant across all trees within each timepoint. We did not control Tleaf or chamber RH, but the observedVPD within the leaf and WTCs were strongly correlated duringmeasurements with no treatment differences (VPDleaf = 0.3 +1.199 VPDWTC, P < 0.001, r2 = 0.81, n = 192 observations).The sample cell [CO2] was controlled to 400� 5 lmol mol�1.

Data analysis

Data were analysed following a completely randomized designwith the single treatment of warming (n = 6 for 6 months, thenn = 3 for the final 3 months when water was withheld from halfof the chambers). Many analyses consist of longitudinal measure-ments over time; these were performed using the ‘lme’ functionwithin the ‘NLME’ R package, and treatment means were esti-mated after adjustment for other terms in the model (i.e. least-square means, or LS means) with the ‘LSMEANS’ package in Rv.3.2.2 (R Development Core Team, 2012; Pinheiro et al.,2013). Analyses were evaluated to test assumptions of residualnormality and variance homoscedasticity; transformations wereoften necessary. Analyses followed a repeated-measures ANOVA





with an autoregressive temporal autocorrelation structure; cham-ber was included as a random effect, temperature treatment was awhole-plot factor and time was a subplot factor. Daily measure-ments provided strong statistical power to detect the effects ofdate and date9 warming, even if these effects were small (e.g.254 numerator and 1753 denominator degrees of freedom,Table 2). Thus, we also inferred the relative importance of modelterms by calculating the explanatory power (marginal r2 value;the variance explained by fixed effects) for models with and with-out each model term and its interactions using the methodologyof Nakagawa & Schielzeth (2013). This provided an assessmentof whether statistically significant date and date9 warming inter-actions were quantitatively important.

We visualized the relationships between daily flux measure-ments (GPP, Ra and Ra/GPP) and daily integrated PPFD and 24-h average Tair. Recognizing the non-independence of these data(e.g. >3000 daily data points, but n = 12 chambers), we used gen-eralized additive models (GAMs) with a random chamber effectto summarize the relationship between fluxes and environmentaldrivers without pseudoreplication (Duursma et al., 2014). Weused the ‘MGCV’ package in R v.3.2.2 to fit the GAMs (Wood,2011). This data-driven approach makes no assumptions regard-ing the functional relationship between variables; 95% confidenceintervals were used to infer whether the nature of these relation-ships was altered by experimental warming. We emphasize thatVPD and Tair at this location strongly and positively co-varied(Fig. S6; Duursma et al., 2014), such that it was not possible toseparate their effects. The patterns related to Tair discussed herewere probably influenced by the natural coincident variation inVPD. The data and source code are freely available; the data arepublished on Figshare (doi: https://dx.doi.org/10.6084/m9.-figshare.3122104.v1) and the R analysis is published as a Gitrepository (https://github.com/jedrake/wtc3_flux_share/; doi:http://dx.doi.org/10.5281/zenodo.48156).

Results

Acclimation of respiration

Experimental warming resulted in acclimation of respiration atthe leaf and canopy scales. The rate of leaf and whole-canopy

respiration increased exponentially over the short term (minutesto hours) with increasing temperature (Fig. 2). Arrhenius analysesof these data indicated acclimation, as experimental warming sig-nificantly reduced the pre-exponential coefficient of these rela-tionships, including the respiration of individual leaves (17%decline, P < 0.001; Fig. 2a), canopy respiration expressed as theraw flux (17% decline, P < 0.01; Fig. 2b) and canopy respirationper unit canopy leaf area (43% decline, P < 0.001; Fig. 2c). Theactivation energy (Ea) was not altered by warming in any case andwas estimated to be 62.0� 1.4 kJ mol�1 for leaf respiration and57.7� 1 kJ mol�1 for both expressions of whole-canopy respira-tion, reflecting Q10 values of 2.3 and 2.2, respectively, at 25°C.

Acclimation to warming was also apparent for tissue-specificrates of leaf and branch wood respiration (Fig. 3). The rate of leafrespiration at 15°C was reduced by ~30% with warming (Fig. 3a,P = 0.07), whereas branch wood respiration at 15°C was reducedby ~23% with warming (Fig. 3b, P = 0.03).

Ra, GPP and Ra/GPP time series

Across time, there were neutral or modest reductions in in-situ Rain response to the warming treatment, coupled with a more con-sistent and larger decrease in in-situ GPP per unit leaf area. Thenet impact of these changes was to maintain Ra/GPP at similar orslightly higher levels in the warmed relative to ambient trees.

The acclimation of respiration to experimental warmingobserved in the temperature response curves (Fig. 2) and tissue-specific rates (Fig. 3) was also apparent in the time series of dailyRa sums. Despite the consistent ~3°C increase in air temperature(Fig. 4a), Ra per unit leaf area was approximately equivalent inthe warmed relative to the ambient treatment (Fig. 4b; Table 2).Although there was a significant warming9 date interaction indi-cating lower Ra in the warmed relative to ambient treatment on afew dates, this interaction was of modest importance: it had asmall F value and removing it did not affect the explanatorypower of the model (Table 2). Similarly, removing the warmingeffect and its interaction with date had little effect on the explana-tory power of the model (Table 2). Thus, Ra varied strongly overtime in a manner that was mostly equivalent across treatments,suggesting that acclimation of Ra to warming was approximatelyhomeostatic.

Table 2 Statistical results of repeated-measures ANOVAs for the daily time series data

Model term dfa

Rab GPPc Ra/GPP

d

F P Dr2e F P Dr2 F P Dr2

Warming 1, 10 0.0 0.97 0.02 0.5 0.49 0.10 0.2 0.65 0.01Date 254, 1753 61.4 <0.001 0.50 62.9 <0.001 0.80 54.1 <0.001 0.79Warming9 date 254, 1753 2.3 <0.001 0.00 2.2 <0.001 0.01 1.8 <0.001 0.0

aDegrees of freedom (numerator, denominator).bAboveground autotrophic respiration.cGross primary production.dThe fraction of GPP respired above ground.eDr2 reflects the reduction in marginal r2 when a term and its interactions were removed from the model. Full model marginal r2 values were 0.5, 0.9 and0.8 for Ra, GPP and Ra/GPP, respectively.



Research

NewPhytologist6

https://dx.doi.org/10.6084/m9.figshare.3122104.v1

https://dx.doi.org/10.6084/m9.figshare.3122104.v1

https://github.com/jedrake/wtc3_flux_share/

http://dx.doi.org/10.5281/zenodo.48156

GPP per unit leaf area was significantly affected by experimen-tal warming in a manner that varied in time (Fig. 4c; Table 2,warming by date interaction, P < 0.001). Warming reduced GPPper unit leaf area on most days by an average of ~20%, except forsome low-light days where GPP was equivalently low across thetreatments (Fig. 4c). This warming9 date interaction was ofmodest importance, but removing the interaction resulted in astrong and negative main effect of warming (not shown). Remov-ing warming from the model substantially reduced its explana-tory power (Table 2). Thus, experimental warming reduced GPPper unit leaf area on many, but not all, days in the dataset.

The fraction of daily GPP that was returned to the atmospherevia aboveground autotrophic respiration (that is, Ra/GPP) washighly variable across days from ~0.18 to 0.8 (Fig. 4d; Table 2,date main effect, P < 0.001). Experimental warming affectedRa/GPP in a time-dependent manner (Fig. 4c; Table 2, warming

by date interaction, P < 0.001), as Ra/GPP was higher in thewarmed relative to ambient treatment on some days (Fig. 4d).However, as with Ra, the main effect of warming and the warm-ing9 date interaction were of modest quantitative importancefor Ra/GPP, as these terms could be removed from the modelwithout affecting its explanatory power (Table 2). Thus, Ra/GPPwas highly variable over time in a manner that was generallyequivalent across treatments, but warming increased Ra/GPP onsome dates. Averaged across the experiment, Ra consumed27� 2% (SE) of GPP in the ambient treatment and 29� 2% ofGPP in the warmed treatment (LS means).

Climate drivers of daily Ra, GPP and Ra/GPP

The dynamic temporal patterns of Ra, GPP and Ra/GPP largelyreflected the climate drivers of daily integrated PPFD and averageTair (Fig. 5). Daily GPP increased strongly and asymptoticallywith PPFD, but daily GPP sums were reduced by warming atany given level of PPFD exceeding 10 mol d�1 (Fig. 5a). On dayswhen incident light was high (PPFD > 35 mol d�1), GPPdeclined with increasing Tair in a sigmoidal fashion (Fig. 5b).Daily Ra also increased with PPFD (Fig. 5c) and Tair (Fig. 5d),but the magnitude of change was smaller for Ra relative to GPP.As such, the proportion of daily GPP that was respired aboveground (Ra/GPP) decreased with increasing PPFD up to~25 mol d�1 (Fig. 5e). The effect of low light on Ra/GPP was alsoevident as spikes in the daily time series (Fig. 4d). Again consider-ing only high-light days (PPFD > 35 mol d�1), Ra/GPP increasedwith Tair in a nonlinear manner (Fig. 5f). At any given

Rle

af

(µm

olC

O2

m−2

s−1)

0

1

2

3

4

0

1

2

3

4(a)

(b)

(c)

Rca

nopy

(µm

olC

O2

s−1)

0

5

10

15

20

25

30

0

5

10

15

20

25

30

Rca

nopy

, are

a

(µm

olC

O2

m−2

s−1)

15 20 25 30 35 40

15 20 25 30 35 40

15 20 25 30 35 40

0

0.5

1

1.5

2

2.5

0

0.5

1

1.5

2

2.5

Temperature (°C)

Fig. 2 The short-term temperature sensitivity of the respiration ofindividual leaves relative to leaf temperature (a) and whole-canopyrespiration relative to air temperature (b, c). Whole-canopy respirationrates are shown as the raw measured flux (b) or the flux relative to totalcanopy leaf area on the night of measurement (c). Mean total canopy leafareas on the night of measurement were 13.7 and 19.7m2 in ambient andwarmed treatments, respectively. Symbols reflect the mean� 1SE ofEucalyptus tereticornis trees exposed to two temperature treatments(n = 6; black, ambient temperature; red, +3°C warming). Curves reflectArrhenius functions fitted to the data; in all cases, warming reduced thepre-exponential coefficient but did not alter the activation energy.

(a) (b)

0

1

2

3

4Leaves

Treatment

A W A W0

0.5

1

1.5

2Branch wood

TreatmentR

espi

ratio

n(n

mol

g−1s−1

)Fig. 3 Tissue-specific rates of leaf (a) and branch wood respiration (b) inEucalyptus tereticornis trees exposed to ambient (A) temperatures or +3°Cwarming (W). All measurements were made at 15°C. Note the differencein y-axis scales. The height of each bar reflects the mean rate and errorbars reflect �1SE (n = 6).





temperature exceeding ~25°C, GPP and Ra were reduced in thewarmed treatment relative to the ambient treatment (95% confi-dence intervals do not overlap; Fig. 5b,d). This correlated reduc-tion in GPP and Ra with warming at any given temperatureresulted in a relationship between Ra/GPP and Tair that did notdiffer between the treatments (Fig. 5f). Thus, Ra/GPP had astrong temperature dependence that was equivalent across treat-ments; experimental warming of +3°C did not alter Ra/GPP atany given temperature because the reduction in GPP and Ra withwarming were correlated and of a similar magnitude.

Although we focused on daily Ra/GPP, we also investigatedsimpler and less-derived metrics, including daily net C uptakeand the ratio of night-time C loss to daytime C gain (Figs S7,S8). These metrics supported the inferences regarding the effectsof experimental warming on daily GPP and Ra/GPP (Fig. 5).

The nonlinear temperature dependence of Ra/GPP (Fig. 5f)suggests that warming may have a larger impact on Ra/GPP dur-ing hot conditions; we thus evaluated whether experimentalwarming affected Ra/GPP during the warmest days in the dataset.There were 12 exceptionally hot days on record when the maxi-mum air temperature exceeded 37 and 40°C in the ambient andwarmed treatments, respectively. On this particular subset of hotdays, experimental warming significantly increased Ra/GPP by14% from 0.43� 0.02 to 0.49� 0.02 (LS means and SE,ANOVA, main effect of warming, P = 0.03). Similar increases inRa/GPP with experimental warming were also observed in othersubsets of warm days (e.g. 15% increase on the 68 d exceeding22°C mean 24-h temperature, P = 0.06). These warm days com-prised ~25% of the flux record (68 of 255 d); a 15% increase inRa/GPP over 25% of the flux record translates to a modest 4%

increase in Ra/GPP because of experimental warming over thecourse of the experiment.

Temperature response of photosynthesis

We examined the dataset of canopy GPP per unit leaf area athigher temporal resolution (hourly) in an attempt to resolve thetemperature dependence of photosynthesis (Fig. 6). There wasstrong positive covariance between PPFD and Tair, such that con-ditions with high light and low temperature did not occur. GPPper unit leaf area increased with PPFD, as expected. At any givenlevel of PPFD, rates of leaf-level GPP were highest at the lowestobserved Tair and declined with increasing Tair (Fig. 6a,b). Forexample, leaf-level GPP declined with increasing Tair for timeperiods when PPFD was between 1200 and 1500 lmol m�2 s�1

(Fig. 6c). Thus, the trees in this experiment were frequently oper-ating above the photosynthetic Topt, which prevented the estima-tion of Topt directly from these data. At any given combinationof PPFD and Tair, leaf-level GPP tended to be modestly lower inthe warmed relative to the ambient treatment (e.g. Fig. 6c), butthe 95% confidence intervals often overlapped across treatments.Thus, experimental warming may have induced an acclimationresponse that shifted the temperature response of photosynthesisdownwards by a small amount at the hourly scale (Fig. 6c), andthis was evident in the daily sums (Fig. 5b). However, warmingalso directly reduced leaf-level photosynthesis by simply increas-ing the observed Tair (i.e. values were shifted towards the rightalong a similar temperature response curve; Fig. 6c).

The reduction in leaf-level GPP as measured by the whole-canopy fluxes (Figs 4c, 5a,b, 6) was confirmed by direct leaf-level

T air

(°C)

A W

20

30

40

0204060

PPFD

(mold−1 )

PPFD

(a)

Ra

(gCm

−2d−1 )

0.51

1.52

2.5 (b)

GPP

(gCm

−2d−1 )

2

4

6(c)

RaGPP

October

November

December

January

February

March

April

May

June

00.20.40.60.81

(d)

Fig. 4 Time-series of whole-canopy fluxes.Environmental variables of mean daily airtemperature (Tair) and daily integratedphotosynthetic photon flux density (PPFD;a), daily sums of aboveground autotrophicrespiration (Ra; b) and gross primaryproduction (GPP; c), and the fraction of dailyGPP that was respired above ground (Ra/GPP; d) in Eucalyptus tereticornis treesexposed to ambient (A) temperatures or+3°C warming (W) (n = 6 until 19 February2014, then n = 3). Note that Ra and GPP areexpressed per unit leaf area, whereas Ra/GPPdoes not depend on the normalizing unit.Lines connect daily estimates of the meanacross treatments. The vertical lines on eachdate reflect �1SE; note that these are toosmall to be visible on some dates. Major tickson the x-axis reflect months and minor ticksreflect weeks.



Research

NewPhytologist8

observations across five diurnal measurements (Fig. 7). Observedin-situ rates of net photosynthesis (Anet) tended to be lower in thewarmed relative to the ambient treatment, particularly in theafternoon (Fig. 7). These leaf-scale measurements were compara-ble with the canopy-scale measurements (e.g. Fig. S9), emphasiz-ing the high quality of both measurement approaches. Anet wassimilar across treatments in July 2013 on a day with low Tair andVPD (Fig. 7a,k). However, Anet was consistently reduced bywarming on warm days with moderate to high VPD (Fig. 7b–e,P < 0.05 on 17 October 2013, 4 December 2014 and 20 Febru-ary 2014). Stomatal conductance (gs) was also lower in thewarmed relative to ambient treatment across all days excludingthe mild day in July (Fig. 7f–j, P < 0.05 on all days except 10 July

2013). Also, gs and Anet tended to be highest during themid-morning (~10:00 h) and decline during the warm afternoonperiods, despite abundant PPFD.

Discussion

We hypothesized that acclimation of photosynthesis and respira-tion to experimental warming would change the temperaturedependence of Ra/GPP, such that experimental warming wouldnot increase Ra/GPP at the canopy scale (orange arrow inFig. 1b). Our measurements provide mixed support for thishypothesis. The temperature dependence of Ra/GPP was notaltered with experimental warming (Fig. 5f). The nonlinearnature of this relationship was such that experimental warminghad no effect on Ra/GPP on most days, but experimental warm-ing increased Ra/GPP on exceptionally hot days. Photosyntheticacclimation did not increase the photosynthetic temperatureoptimum, as hypothesized, and experimental warming consis-tently reduced photosynthetic rates by increasing in-situ tempera-tures beyond the photosynthetic optimum. These results are notconsistent with our hypothesis concerning GPP. By contrast, res-piration acclimated in a homeostatic manner to experimentalwarming, as hypothesized. The common Ra/GPP vs temperaturerelationship across treatments was largely the result of physiologi-cal acclimation of Ra to warming. The nonlinear temperaturedependence of Ra/GPP and the significant stimulation of Ra/GPPby warming on hot days are consistent with a positive feedbackbetween climate warming and atmospheric CO2 accumulationvia a stimulation of Ra relative to GPP, particularly as the fre-quency of hot days is predicted to increase in the future (Yaoet al., 2013).

Physiological acclimation of leaf and branch wood respirationrates to experimental warming was substantial and moderated theresponse of Ra/GPP to warming. The downward adjustment inrespiratory rates at any given temperature in response to warmingis consistent with a large body of literature demonstratingdynamic acclimation of respiration to long-term changes in tem-perature (see reviews; Atkin & Tjoelker, 2003; Atkin et al.,2005a; Smith & Dukes, 2013; Slot & Kitajima, 2015). Ourresults showed a stronger acclimation response relative to a reviewof prior studies (Slot & Kitajima, 2015), such that in-situ respira-tion tended towards a homeostatic rate. Similarly strong acclima-tion has been reported recently in a multi-year warmingexperiment with woody juveniles (Reich et al., 2016). This tem-perature acclimation of respiration moderated the warmingeffects on Ra/GPP, particularly as warming reduced GPP. We didnot measure root respiration; given that fine root respiration canbe ~40% of total autotrophic respiration in some mature forests(e.g. Hamilton et al., 2002), our measures of Ra are underesti-mates of whole-tree autotrophic respiration. Root respiration alsoincreases with temperature (e.g. Drake et al., 2008), and soaccounting for root respiration would probably steepen thetemperature dependence of R/GPP.

GPP per unit leaf area declined with increasing temperature atthe hourly and daily scales, in contrast with our hypothesis. Thenegative temperature dependence of GPP observed at high light

GP

P(g

Cm

−2d−1

)

2

4

6

8

1

2

3

4

0

0.2

0.4

0.6

0.8

1

Ra

(gC

m−2

d−1)

Ra

GP

P

10 20 30 40 50 60 15 20 25 30

PPFD (mol d−1) Tair (°C, 24−h mean)

2

4

6

8

1

2

3

4

0

0.2

0.4

0.6

0.8

1

(a) (b)

(c) (d)

(e) (f)

Fig. 5 Daily integrated gross primary production (GPP; a, b), abovegroundautotrophic respiration (Ra; c, d) and the fraction of daily GPP respiredabove ground (Ra/GPP; e, f) in response to daily integrated photosyntheticphoton flux density (PPFD; a, c, e) and daily (24-h) average airtemperature (Tair; b, d, f). Each point reflects a Eucalyptus tereticornis treemeasured on a day (black, ambient temperature; red, +3°C warming).Lines reflect generalized additive model fits to the data. Shaded areasshow the 95% confidence intervals; treatment effects are inferred fromnon-overlapping confidence intervals. Data from days with low light(PPFD < 35mol d�1) were excluded for the relationships involving Tair.GPP and Ra are expressed per unit leaf area; note that the normalizing unitdoes not affect Ra/GPP.





(Fig. 6c) contrasts with the classic peaked relationship betweenphotosynthesis and temperature (Sage & Kubien, 2007; Way &Yamori, 2014). This disparity may be related to light availability;the peaked relationship between photosynthesis and temperatureassumes high light incident on the leaf surface (e.g. PPFD of1500 lmol m�2 s�1 in Fig. 1a). Because of self-shading, much ofthe canopy may have been operating at relatively low light, evenwith high PPFD incident at the top of the canopy (e.g. Ellsworth& Reich, 1993; Sch€afer et al., 2003). If the majority of canopy-scale C uptake was fixed under light-limited conditions, the tem-perature dependence of photosynthesis would be dominated bythe negative effect of photorespiration at higher temperatures,resulting in monotonically declining photosynthesis with increas-ing temperature (McMurtrie & Wang, 1993; Haxeltine & Pren-tice, 1996). Alternatively or additionally, the observed response(Fig. 6c) may be explained by a low Topt of photosynthesis, suchthat the trees were operating at temperatures exceeding Topt. Alow Topt of photosynthesis may arise from non-stomatal limita-tions such as Rubisco activase lability (Salvucci & Crafts-Brandner, 2004; Sage & Kubien, 2007), or stomatal limitationof photosynthesis with increasing VPD, which lowers Topt byreducing the CO2 concentration within leaf air spaces at a giventemperature (Lin et al., 2012). A low Topt of GPP at the canopyscale could also be related to a divergence of Tleaf relative to Tair,particularly at high levels of solar irradiance (Doughty &Goulden, 2008). We were unable to extract Topt estimates ofGPP to test whether Topt increased with warming, as hypothe-sized. However, GPP per unit leaf area was reduced by warmingacross most common temperatures, particularly at the dailytimescale. This suggests that any physiological change in responseto warming led to reduced, rather than increased, photosyntheticrates.

Canopy Ra and GPP were variable and responsive to the envi-ronment, such that the fraction of daily GPP respired aboveground was also highly variable and responsive to PPFD and Tair.Light availability was a strong driver of Ra/GPP at the daily scale;cloudy days had high Ra/GPP, as GPP was strongly reduced

relative to Ra. Air temperature was also an important driver; hotdays had high Ra/GPP as GPP was reduced and Ra increased, andexperimental warming exacerbated these responses. The dynamicRa/GPP observed here runs counter to some literature suggestingthat autotrophic respiration may be a fixed fraction of GPP, giventhe physiological link between photosynthesis and respiration(Waring et al., 1998; Dewar et al., 1999; Gifford, 2003),although these studies were largely concerned with longertimescales. Given the daily timescale of the measurementsreported here, we were able to confirm the expected environmen-tal responses of Ra/GPP. Some studies found leaf respiration ratesto be positively correlated with the rate of photosynthesis on thepreceding day, presumably via a substrate effect on respiratorymetabolism (Whitehead et al., 2004; Hartley et al., 2006; Wertin& Teskey, 2008); this effect would tend to stabilize or constrainRa/GPP. We found limited evidence of this effect in the currentexperiment, as nightly sums of C loss to Ra increased only mod-estly with GPP on the preceding day (Fig. S10; y = 0.24 + 0.02x,r2 = 0.07, P < 0.05). This suggests that Ra was not strongly sub-strate limited in this experiment, and the photosynthate acquiredon days with high GPP was retained, rather than immediatelyrespired. This quasi-independence of GPP and Ra resulted in anonlinear temperature dependence of Ra/GPP, as predicted byphysiological models that simulate photosynthesis and respirationas independent processes.

These results provide some support for a positive feedbackbetween surface warming and atmospheric CO2 concentrationsas simulated by Earth system models. Such models predict a posi-tive feedback between surface warming and atmospheric CO2

accumulation (Dufresne et al., 2002; Friedlingstein et al., 2014),in accordance with empirical data correlating historical patternsof surface temperatures and atmospheric [CO2] (Frank et al.,2010; Wang et al., 2013). Much of this positive feedback hasbeen attributed to enhanced mineralization of soil organic matterwith increased temperatures (Kirschbaum, 2000; Wieder et al.,2013; Karhu et al., 2014), but autotrophic respiration also con-tributes to the positive feedback (e.g. Atkin et al., 2008;

Tair (°C) Tair (°C) Tair (°C)

PP

FD (µ

mol

m–2

s–1

)

PP

FD (µ

mol

m–2

s–1

)

GP

P (µ

mol

m–2

s–1

)

(a) (b)

Ambient Warmed

GPP (c)

2000

1500

1000

500

010 20 30 40

2000

1500

1000

500

010 20

20

16

12

8

4

0

0

5

10

15

15

20

20

25

25 30 35 40 45

ambientwarmed

30 40

Fig. 6 Hourly leaf-level rates of gross primary production (GPP; lmol CO2m�2 s�1) grouped into bins of hourly average air temperature (Tair) and incident

photosynthetic photon flux density (PPFD; measured at 8-m height atop an adjacent demountable building) for Eucalyptus tereticornis trees grown underambient (a) or warmed (+3°C; b) temperatures. Warmer colours reflect higher rates of leaf-level GPP. An example slice of these data is shown in whichPPFD was within 1200 and 1500 lmol m�2 s�1 (c). Generalized additive models were fitted to these data; the black line reflects the ambient temperaturetreatment, the red line reflects the warmed treatment and the shaded areas reflect 95% confidence intervals.



Research

NewPhytologist10

Lombardozzi et al., 2015). This study suggests that acclimationof Ra to warming may prevent an increase in tree Ra/GPP at aconstant temperature, mitigating a portion of this positive feed-back. However, given the nonlinear temperature dependence ofRa/GPP, experimental warming increased Ra/GPP on hot days,

suggesting that Ra/GPP may increase together with the predictedincrease in frequency and intensity of heat waves (Yao et al.,2013; Teskey et al., 2014). This positive feedback may be rela-tively modest, however, as warming only increased Ra/GPP by asmall amount and only during uncommonly hot conditions.

0

5

10

15

20

25 AW

(a)

Ane

t

(µm

ol m

−2s−1

)

10 July 2013

(b)

26 August 2013

(c)

17 October 2013

(d)

4 December 2013

0

5

10

15

20

25

(e)

20 February 2014

0

0.05

0.1

0.15

0.2

0.25

0.3

(f)

g s(m

ol m

−2s−1

)

(g) (h) (i) 0

0.05

0.1

0.15

0.2

0.25

0.3

(j)

0

10

20

30

40TPPFDVPD

Env

ironm

enta

l var

iabl

es

0

500

1000

1500

2000

0

1

2

3

4

5

6

7

(k) (l)

Hour

(m) (n)

5 10 15 20 5 10 15 20 5 10 15 20 5 10 15 20 5 10 15 20

(o)

Fig. 7 Diurnal measurements of leaf-level net photosynthesis (Anet; a–e) and stomatal conductance (gs; f–j) of Eucalyptus tereticornis trees exposed toambient or warmed temperatures (A vsW); values reflect the mean� SE (n = 6 trees). The x-axis reflects the hour of the day (Australian Eastern StandardTime). Measurements occurred every 2–3 h between dawn and dusk; note that day length increased during the transition to the Austral summer.Associated environmental variables are also displayed (k–o). The temperature (T; black; units of °C) reflects the mean measured leaf temperature of theambient trees; leaf temperature of warmed trees was ~3°C warmer. The incident photosynthetic photon flux density (PPFD; blue, units of lmol m�2 s�1)was measured in an adjacent field and provided in the leaf cuvette. The vapor pressure deficit (VPD; red, units of kPa) reflects the leaf-to-air VPD inside themeasurement cuvette of the ambient temperature trees. The measurement dates are shown across the top.





Acknowledgements

We thank Sune Linder and the Swedish University for Agricul-tural Sciences for providing the whole-tree chambers. This exper-iment was supported by the Australian Research Council(Discovery, DP140103415), the Hawkesbury Institute for theEnvironment and Western Sydney University. We thank BurhanAmiji for his cheerful support and dedication to maintaining thesite. We also thank David Tissue, Elise Pendall and AngelicaV�arhammer for their helpful suggestions on manuscript develop-ment. Finally, we thank four anonymous reviewers for their help-ful comments and criticisms.

Author contributions

J.E.D. collected the data, led the data processing and cleaning,and led the data interpretation and writing. M.G.T. providedoverall project leadership and contributed substantially to datainterpretation and writing. M.J.A. contributed substantially todata collection and interpretation, made many conceptual contri-butions and assisted with writing. P.B.R. provided advice, helpedwith interpretation and contributed to writing. C.V.M.B. builtand maintained the system that measured WTC fluxes and con-tributed substantially to data collection and processing. B.E.M.contributed to data interpretation and assisted with writing.R.A.D. contributed to data processing, analysis and writing.

References

Amthor JS, Hanson PJ, Norby RJ, Wullschleger SD. 2010. A comment on

“Appropriate experimental ecosystem warming methods by ecosystem,

objective, and practicality” by Aronson and McNulty. Agricultural and ForestMeteorology 150: 497–498.

Atkin OK, Atkinson LJ, Fisher RA, Campbell CD, Zaragoza-Castells J,

Pitchford JW, Woodward FI, Hurry V. 2008. Using temperature-dependent

changes in leaf scaling relationships to quantitatively account for thermal

acclimation of respiration in a coupled global climate–vegetation model. GlobalChange Biology 14: 2709–2726.

Atkin OK, Bruhn D, Hurry VM, Tjoelker MG. 2005a. The hot and the cold:

unravelling the variable response of plant respiration to temperature. FunctionalPlant Biology 32: 87–105.

Atkin OK, Bruhn D, Tjoelker MG. 2005b. Response of plant respiration to

changes in temperature: mechanisms and consequences of variations in Q10

values and acclimation. In: Lambers H, Ribas-Carbo M, eds. Plant respiration.Dordrecht, the Netherlands: Springer, 95–135.

Atkin OK, Scheurwater I, Pons TL. 2006.High thermal acclimation potential of

both photosynthesis and respiration in two lowland Plantago species in contrast

to an alpine congeneric. Global Change Biology 12: 500–515.Atkin OK, Scheurwater I, Pons TL. 2007. Respiration as a percentage of daily

photosynthesis in whole plants is homeostatic at moderate, but not high,

growth temperatures. New Phytologist 174: 367–380.Atkin OK, Tjoelker MG. 2003. Thermal acclimation and the dynamic response

of plant respiration to temperature. Trends in Plant Science 8: 343–351.Atkin OK, Turnbull MH, Zaragoza-Castells J, Fyllas NM, Lloyd J, Meir P,

Griffin KL. 2013. Light inhibition of leaf respiration as soil fertility declines

along a post-glacial chronosequence in New Zealand: an analysis using the Kok

method. Plant and Soil 367: 163–182.Barton CVM, Duursma RA, Medlyn BE, Ellsworth DS, Eamus D, Tissue

DT, Adams MA, Conroy J, Crous KY, Liberloo M et al. 2012. Effects ofelevated atmospheric CO2 on instantaneous transpiration efficiency at leaf

and canopy scales in Eucalyptus saligna. Global Change Biology 18:

585–595.

Barton CVM, Ellsworth DS, Medlyn BE, Duursma RA, Tissue DT, Adams

MA, Eamus D, Conroy JP, McMurtrie RE, Parsby J et al. 2010.Whole-tree

chambers for elevated atmospheric CO2 experimentation and tree scale flux

measurements in south-eastern Australia: the Hawkesbury Forest Experiment.

Agricultural and Forest Meteorology 150: 941–951.Berry J, Bj€orkman O. 1980. Photosynthetic response and adaptation to

temperature in higher-plants. Annual Review of Plant Physiology and PlantMolecular Biology 31: 491–543.

Bloemen J, McGuire MA, Aubrey DP, Teskey RO, Steppe K. 2013. Transport

of root-respired CO2 via the transpiration stream affects aboveground carbon

assimilation and CO2 efflux in trees. New Phytologist 197: 555–565.Bonan GB. 2008. Forests and climate change: forcings, feedbacks, and the

climate benefits of forests. Science 320: 1444–1449.Brooks A, Farquhar GD. 1985. Effect of temperature on the CO2/O2 specificity

of ribulose-1,5-bisphosphate carboxylase oxygenase and the rate of respiration

in the light – estimates from gas-exchange measurements on spinach. Planta165: 397–406.

Cox PM, Betts RA, Jones CD, Spall SA, Totterdell IJ. 2000. Acceleration of

global warming due to carbon-cycle feedbacks in a coupled climate model.

Nature 408: 184–187.Crous KY, Quentin AG, Lin YS, Medlyn BE, Williams DG, Barton CVM,

Ellsworth DS. 2013. Photosynthesis of temperate Eucalyptus globulus treesoutside their native range has limited adjustment to elevated CO2 and climate

warming. Global Change Biology 19: 3790–3807.Crous KY, Zaragoza-Castells J, Ellsworth DS, Duursma RA, L€ow M, Tissue

DT, Atkin OK. 2012. Light inhibition of leaf respiration in field-grown

Eucalyptus saligna in whole-tree chambers under elevated atmospheric CO2 and

summer drought. Plant, Cell & Environment 35: 966–981.Dewar RC, Medlyn BE, McMurtrie RE. 1999. Acclimation of the respiration

photosynthesis ratio to temperature: insights from a model. Global ChangeBiology 5: 615–622.

Doughty CE, Goulden ML. 2008. Are tropical forests near a high temperature

threshold? Journal of Geophysical Research–Biogeosciences 113: G00b07.

Drake JE, Aspinwall MJ, Pfautsch S, Rymer PD, Reich PB, Smith RA,

Crous KY, Tissue DT, Ghannoum O, Tjoelker MG. 2015. The capacity

to cope with climate warming declines from temperate to tropical latitudes

in two widely distributed Eucalyptus species. Global Change Biology 21:459–472.

Drake JE, Stoy PC, Jackson RB, DeLucia EH. 2008. Fine-root respiration in a

loblolly pine (Pinus taeda L.) forest exposed to elevated CO2 and N

fertilization. Plant, Cell & Environment 31: 1663–1672.Dufresne JL, Fairhead L, Le TH, Berthelot M, Bopp L, Ciais P, Friedlingstein

P, Monfray P. 2002.On the magnitude of positive feedback between future

climate change and the carbon cycle. Geophysical Research Letters 29:43-1–43-4.

Duursma RA. 2015. Plantecophys – an R package for analysing and modelling

leaf gas exchange data. PLoS One 10: e0143346.Duursma RA, Barton CVM, Lin YS, Medlyn BE, Eamus D, Tissue DT,

Ellsworth DS, McMurtrie RE. 2014. The peaked response of transpiration

rate to vapour pressure deficit in field conditions can be explained by the

temperature optimum of photosynthesis. Agricultural and Forest Meteorology189: 2–10.

Ellsworth DS, Reich PB. 1993. Canopy structure and vertical patterns of

photosynthesis and related leaf traits in a deciduous forest. Oecologia 96:169–178.

Farquhar GD, Caemmerer SV, Berry JA. 1980. A biochemical-model of

photosynthetic CO2 assimilation in leaves of C-3 species. Planta 149: 78–90.Frank DC, Esper J, Raible CC, Buntgen U, Trouet V, Stocker B, Joos F. 2010.

Ensemble reconstruction constraints on the global carbon cycle sensitivity to

climate. Nature 463: 527–530.Friedlingstein P, Cox P, Betts R, Bopp L, Von Bloh W, Brovkin V, Cadule P,

Doney S, Eby M, Fung I et al. 2006. Climate–carbon cycle feedback analysis:

results from the C4MIP model intercomparison. Journal of Climate 19:3337–3353.

Friedlingstein P, Meinshausen M, Arora VK, Jones CD, Anav A, Liddicoat SK,

Knutti R. 2014. Uncertainties in CMIP5 climate projections due to carbon

cycle feedbacks. Journal of Climate 27: 511–526.



Research

NewPhytologist12

Gifford RM. 1995.Whole plant respiration and photosynthesis of wheat under

increased CO2 concentration and temperature: long-term vs short-term

distinctions for modelling. Global Change Biology 1: 385–396.Gifford RM. 2003. Plant respiration in productivity models: conceptualisation,

representation and issues for global terrestrial carbon-cycle research. FunctionalPlant Biology 30: 171–186.

Gimeno TE, Crous KY, Cooke J, O’Grady AP, �Osvaldsson A, Medlyn BE,

Ellsworth DS. 2015. Conserved stomatal behaviour under elevated CO2 and

varying water availability in a mature woodland. Functional Ecology. doi:10.1111/1365-2435.12532.

Gunderson CA, Norby RJ, Wullschleger SD. 2000. Acclimation of

photosynthesis and respiration to simulated climatic warming in northern and

southern populations of Acer saccharum: laboratory and field evidence. TreePhysiology 20: 87–96.

Gunderson CA, O’Hara KH, Campion CM, Walker AV, Edwards NT. 2010.

Thermal plasticity of photosynthesis: the role of acclimation in forest responses

to a warming climate. Global Change Biology 16: 2272–2286.Hamilton JG, DeLucia EH, George K, Naidu SL, Finzi AC, Schlesinger WH.

2002. Forest carbon balance under elevated CO2. Oecologia 131: 250–260.Hartley IP, Armstrong AF, Murthy R, Barron-Gafford G, Ineson P, Atkin OK.

2006. The dependence of respiration on photosynthetic substrate supply and

temperature: integrating leaf, soil and ecosystem measurements. Global ChangeBiology 12: 1954–1968.

Haxeltine A, Prentice IC. 1996. A general model for the light-use efficiency of

primary production. Functional Ecology 10: 551–561.Huve K, Bichele I, Rasulov B, Niinemets U. 2011.When it is too hot for

photosynthesis: heat-induced instability of photosynthesis in relation to

respiratory burst, cell permeability changes and H2O2 formation. Plant, Cell &Environment 34: 113–126.

Karhu K, Auffret MD, Dungait JAJ, Hopkins DW, Prosser JI, Singh BK, Subke

JA, Wookey PA, Agren GI, Sebastia MT et al. 2014. Temperature sensitivity

of soil respiration rates enhanced by microbial community response. Nature513: 81–84.

Kattge J, Knorr W. 2007. Temperature acclimation in a biochemical model of

photosynthesis: a reanalysis of data from 36 species. Plant, Cell & Environment30: 1176–1190.

KirschbaumMF. 2000.Will changes in soil organic carbon act as a positive or

negative feedback on global warming? Biogeochemistry 48: 21–51.Lee TD, Reich PB, Bolstad PV. 2005. Acclimation of leaf respiration to

temperature is rapid and related to specific leaf area, soluble sugars and leaf

nitrogen across three temperate deciduous tree species. Functional Ecology 19:640–647.

Lin YS, Medlyn BE, De Kauwe MG, Ellsworth DS. 2013. Biochemical

photosynthetic responses to temperature: how do interspecific differences

compare with seasonal shifts? Tree Physiology 33: 793–806.Lin Y-S, Medlyn BE, Ellsworth DS. 2012. Temperature responses of leaf net

photosynthesis: the role of component processes. Tree Physiology 32: 219–231.Lombardozzi DL, Bonan GB, Smith NG, Dukes JS, Fisher RA. 2015.

Temperature acclimation of photosynthesis and respiration: a key uncertainty

in the carbon cycle–climate feedback. Geophysical Research Letters 42:8624–8631.

McMurtrie RE, Wang YP. 1993.Mathematical-models of the photosynthetic

response of tree stands to rising CO2 concentrations and temperatures. Plant,Cell & Environment 16: 1–13.

Medlyn BE, Dreyer E, Ellsworth D, Forstreuter M, Harley PC, Kirschbaum

MUF, Le Roux X, Montpied P, Strassemeyer J, Walcroft A et al. 2002.Temperature response of parameters of a biochemically based model of

photosynthesis. II. A review of experimental data. Plant Cell & Environment 25:1167–1179.

Medlyn BE, Duursma RA, Eamus D, Ellsworth DS, Prentice IC, Barton CVM,

Crous KY, de Angelis P, Freeman M, Wingate L. 2011. Reconciling the

optimal and empirical approaches to modelling stomatal conductance. GlobalChange Biology 17: 2134–2144.

Moffat AM, Papale D, Reichstein M, Hollinger DY, Richardson AD, Barr AG,

Beckstein C, Braswell BH, Churkina G, Desai AR et al. 2007. Comprehensive

comparison of gap-filling techniques for eddy covariance net carbon fluxes.

Agricultural and Forest Meteorology 147: 209–232.

Nakagawa S, Schielzeth H. 2013. A general and simple method for obtaining R2

from generalized linear mixed-effects models.Methods in Ecology and Evolution4: 133–142.

O’Sullivan OS, Weerasinghe K, Evans JR, Egerton JJG, Tjoelker MG, Atkin

OK. 2013.High-resolution temperature responses of leaf respiration in snow

gum (Eucalyptus pauciflora) reveal high-temperature limits to respiratory

function. Plant, Cell & Environment 36: 1268–1284.Pan YD, Birdsey RA, Fang JY, Houghton R, Kauppi PE, Kurz WA, Phillips

OL, Shvidenko A, Lewis SL, Canadell JG et al. 2011. A large and persistent

carbon sink in the World’s forests. Science 333: 988–993.Perez-Priego O, Testi L, Orgaz F, Villalobos FJ. 2010. A large closed canopy

chamber for measuring CO2 and water vapour exchange of whole trees.

Environmental and Experimental Botany 68: 131–138.Piao S, Sitch S, Ciais P, Friedlingstein P, Peylin P, Wang X, Ahlstr€om A, Anav

A, Canadell JG, Cong N et al. 2013. Evaluation of terrestrial carbon cycle

models for their response to climate variability and to CO2 trends. GlobalChange Biology 19: 2117–2132.

Pinheiro J, Bates DM, DebRoy S, Sarkar D 2013. nlme: linear and nonlinearmixed effects models. R package version 3.1-109.

R Development Core Team. 2012. R: a language and environment for statisticalcomputing. Vienna, Austria: R Foundation for Statistical Computing.

Reich PB, Sendall KM, Stefanski A, Wei X, Rich RL, Montgomery RA. 2016.

Boreal and temperate trees show strong acclimation of respiration to warming.

Nature 531: 633–636.Reichstein M, Falge E, Baldocchi D, Papale D, Aubinet M, Berbigier P,

Bernhofer C, Buchmann N, Gilmanov T, Granier A et al. 2005.On the

separation of net ecosystem exchange into assimilation and ecosystem

respiration: review and improved algorithm. Global Change Biology 11:1424–1439.

Riedmiller M, Braun H, IEEE. 1993. A direct adaptive method for fasterbackpropagation learning – the RPROP algorithm. New York, NY, USA: IEEE.

Sage RF, Kubien DS. 2007. The temperature response of C3 and C4

photosynthesis. Plant, Cell & Environment 30: 1086–1106.Salvucci ME, Crafts-Brandner SJ. 2004. Relationship between the heat tolerance

of photosynthesis and the thermal stability of rubisco activase in plants from

contrasting thermal environments. Plant Physiology 134: 1460–1470.Sch€afer KVR, Oren R, Ellsworth DS, Lai CT, Herrick JD, Finzi AC, Richter

DD, Katul GG. 2003. Exposure to an enriched CO2 atmosphere alters carbon

assimilation and allocation in a pine forest ecosystem. Global Change Biology 9:1378–1400.

Schlesinger WH, Bernhardt ES. 2013. The global carbon cycle. Biogeochemistry –an analysis of global change. Oxford, UK: Academic Press, Elsevier, 419–444.

Slot M, Kitajima K. 2015. General patterns of acclimation of leaf respiration to

elevated temperatures across biomes and plant types. Oecologia 177: 885–900.Smith NG, Dukes JS. 2013. Plant respiration and photosynthesis in global-scale

models: incorporating acclimation to temperature and CO2. Global ChangeBiology 19: 45–63.

Tcherkez G, Mahe A, Gauthier P, Mauve C, Gout E, Bligny R, Cornic G,

Hodges M. 2009. In folio respiratory fluxomics revealed by 13C isotopic

labeling and H/D isotope effects highlight the noncyclic nature of the

tricarboxylic acid “cycle” in illuminated leaves. Plant Physiology 151: 620–630.Teskey R, Wertin T, Bauweraerts I, Ameye M, McGuire MA, Steppe K. 2014.

Responses of tree species to heat waves and extreme heat events. Plant, Cell &Environment 38: 1699–1712.

Tjoelker MG, Oleksyn J, Lorenc-Plucinska G, Reich PB. 2009. Acclimation of

respiratory temperature responses in northern and southern populations of

Pinus banksiana. New Phytologist 181: 218–229.Tjoelker MG, Oleksyn J, Reich PB. 1999. Acclimation of respiration to

temperature and CO2 in seedlings of boreal tree species in relation to plant size

and relative growth rate. Global Change Biology 5: 679–691.Wang W, Ciais P, Nemani RR, Canadell JG, Piao S, Sitch S, White MA,

Hashimoto H, Milesi C, Myneni RB. 2013. Variations in atmospheric CO2

growth rates coupled with tropical temperature. Proceedings of the NationalAcademy of Sciences, USA 110: 13061–13066.

Waring RH, Landsberg JJ, Williams M. 1998. Net primary production of

forests: a constant fraction of gross primary production? Tree Physiology 18:129–134.





http://dx.doi.org/10.1111/1365-2435.12532

Way DA, Oren R. 2010. Differential responses to changes in growth temperature

between trees from different functional groups and biomes: a review and

synthesis of data. Tree Physiology 30: 669–688.Way DA, Yamori W. 2014. Thermal acclimation of photosynthesis: on the

importance of adjusting our definitions and accounting for thermal acclimation

of respiration. Photosynthesis Research 119: 89–100.Wertin TM, Teskey RO. 2008. Close coupling of whole-plant respiration to net

photosynthesis and carbohydrates. Tree Physiology 28: 1831–1840.Whitehead D, Griffin KL, Turnbull MH, Tissue DT, Engel VC, Brown KJ,

Schuster WSF, Walcroft AS. 2004. Response of total night-time respiration to

differences in total daily photosynthesis for leaves in a Quercus rubra L. canopy: imp-

lications for modelling canopy CO2 exchange. Global Change Biology 10: 925–938.WiederWR, BonanGB, Allison SD. 2013.Global soil carbon projections are

improved bymodellingmicrobial processes.Nature Climate Change 3: 909–912.Wood SN. 2011. Fast stable restricted maximum likelihood and marginal

likelihood estimation of semiparametric generalized linear models. Journal ofthe Royal Statistical Society Series B: Statistical Methodology 73: 3–36.

Yamori W, Hikosaka K, Way DA. 2014. Temperature response of

photosynthesis in C3, C4, and CAM plants: temperature acclimation and

temperature adaptation. Photosynthesis Research 119: 101–117.Yao Y, Luo Y, Huang JB, Zhao ZC. 2013. Comparison of monthly temperature

extremes simulated by CMIP3 and CMIP5 models. Journal of Climate 26:7692–7707.

Zhang YJ, Yu GR, Yang J, Wimberly MC, Zhang XZ, Tao J, Jiang YB, Zhu JT.

2014. Climate-driven global changes in carbon use efficiency. Global Ecologyand Biogeography 23: 144–155.

Supporting Information

Additional Supporting Information may be found online in thesupporting information tab for this article:

Fig. S1 Photographs of the whole-tree chamber facility.

Fig. S2 Example partitioning of measured net CO2 flux intogross primary production (GPP) and aboveground autotrophicrespiration (Ra).

Fig. S3 Estimates of light and dark respiration (Rlight and Rdark)at the canopy scale.

Fig. S4 The effects of light-induced inhibition of Ra on grossprimary production (GPP), Ra and Ra/GPP.

Fig. S5 Light-induced inhibition of Ra has little effect on thetemperature dependence of Ra/GPP (GPP, gross primary produc-tion).

Fig. S6 Covariance between air temperature and vapor pressuredeficit.

Fig. S7 Light and temperature as drivers of net carbon (C) gain.

Fig. S8 Light and temperature as drivers of the ratio of carbon(C) loss/C gain.

Fig. S9 Example comparison of leaf- and canopy-scale measure-ments of photosynthesis.

Fig. S10 Tonight’s C loss relative to yesterday’s gross primaryproduction.

Notes S1 Investigation of light inhibition of Ra and its possibleimplications for the temperature dependence of Ra/GPP (GPP,gross primary production).

Please note: Wiley Blackwell are not responsible for the contentor functionality of any supporting information supplied by theauthors. Any queries (other than missing material) should bedirected to the New Phytologist Central Office.



Research

NewPhytologist14

Documents

Does physiological acclimation to climate warming stabilize the …€¦ · etal., 2010). Such shifts in the temperature response of photosyn-thesis and respiration via physiological