Tropical rainforest species have larger increases in temperature optima with warming than warm-temperate rainforest trees.

While trees can acclimate to warming, there is concern that tropical rainforest species may be less able to acclimate because they have adapted to a relatively stable thermal environment. Here we tested whether the physiological adjustments to warming differed among Australian tropical, subtropical and warm-temperate rainforest trees. Photosynthesis and respiration temperature responses were quantified in six Australian rainforest seedlings of tropical, subtropical and warm-temperate climates grown across four growth temperatures in a glasshouse. Temperature-response models were fitted to identify mechanisms underpinning the response to warming. Tropical and subtropical species had higher temperature optima for photosynthesis (ToptA ) than temperate species. There was acclimation of ToptA to warmer growth temperatures. The rate of acclimation (0.35-0.78°C °C-1 ) was higher in tropical and subtropical than in warm-temperate trees and attributed to differences in underlying biochemical parameters, particularly increased temperature optima of Vcmax25 and Jmax25 . The temperature sensitivity of respiration (Q10 ) was 24% lower in tropical and subtropical compared with warm-temperate species. Overall, tropical and subtropical species had a similar capacity to acclimate to changes in growth temperature as warm-temperate species, despite being grown at higher temperatures. Quantifying the physiological acclimation in rainforests can improve accuracy of future climate predictions and assess their potential vulnerability to warming.

Introduction

The effect of climate warming on rainforest trees is uncertain and under debate (e.g. Huntingford et al., 2013; Mercado et al., 2018; Smith et al., 2020). Some modelling studies have suggested that tropical Amazonian forests are seriously threatened by climate warming (Malhi et al., 2008; Cook et al., 2012; Fu et al., 2013; Bastin et al., 2019) and project a large reduction in vegetation biomass by 2100 (Cox et al., 2004). Moreover, drought stress and associated tree mortality are also predicted to increase in tropical rainforests (Rowland et al., 2015) leading to reduced carbon stocks (Brando et al., 2019). Such an extensive loss of tropical rainforest would have major feedbacks to the global climate (Lewis et al., 2009; Fu et al., 2013) and pronounced effects on the global carbon budget (Malhi, 2012). However, these predictions depend in part on the representation of physiological responses to warming, which is one of the major uncertainties in current global vegetation models (Lombardozzi et al., 2015; Huntingford et al., 2017). Many models do not account for the possibility that tree species can acclimate to rising temperatures. Several studies have highlighted the need to incorporate thermal acclimation of photosynthetic and respiratory processes to improve the accuracy of future climate projections (Atkin et al., 2008; Smith & Dukes, 2013; Mercado et al., 2018). The potential adjustment of the photosynthetic capacity in tropical species could reduce their predicted vulnerability to warming (e.g. Mercado et al., 2018), so it is important to quantify how much rainforest species can adjust to warming.

Generally, plants can maintain or increase carbon gain via increased photosynthetic capacity (Kattge & Knorr, 2007; Sage & Kubien, 2007) and/or an increase in the temperature optimum of photosynthesis (T optA) (Cowling & Sage, 1998; Gunderson et al., 2010; Yamori et al., 2014; Scafaro et al., 2017; Crous et al., 2018) in response to warmer growth temperatures. Given that tropical species are adapted to stable climatic conditions with high growth temperatures within a narrow temperature range (Janzen, 1967; Wright et al., 2009; Perez et al., 2016), we might expect that their physiological capacity to adjust to warming would be more limited compared with temperate species. Previous studies have found that tropical species were more susceptible to growth declines at increased temperatures, whereas temperate species had enhanced growth with warming (Way & Oren, 2010; Drake et al., 2015), suggesting that tropical species may have constrained capacity to adjust to climate warming (Perez et al., 2016; Crous et al., 2018).

The limited studies investigating thermal acclimation of photosynthesis in rainforest species have generally reported reduced, not increased, photosynthetic rates with warming (Cunningham & Read, 2003; Slot & Winter, 2016; Scafaro et al., 2017; Dusenge et al., 2021). Slot & Winter (2017) found that tropical seedlings can acclimate to moderate warming via an increase in T optA, but photosynthesis declined under a high level of warming.

T optA may be related to species' climate of origin (Slatyer, 1978; Robakowski et al., 2012), but other studies reported a lack of T optA response to a species' climate of origin (Gunderson et al., 2000; Kumarathunge et al., 2019; Dusenge et al., 2021). A change in growth temperatures was found to be the most common driver of T optA (Kumarathunge et al., 2019). Our study compared the response of T optA using a range of growth temperatures rather than just two levels to understand how T optA adjusted across temperatures, whether there were limits to this adjustment (range of warming up to 10°C) and how rainforest species from different climates adjusted their T optA across this temperature range. We also aimed to identify the mechanisms underpinning the physiological responses of temperate and tropical species to warming.

Several underlying physiological processes contribute to the temperature response of net photosynthesis (A net), including stomatal conductance (Lin et al., 2012), biochemical processes, particularly changes in the carboxylation of ribulose‐1,5‐bisphosphate (RuBP) activity (V cmax) and regeneration of RuBP (i.e. the maximum rate of electron transport, J max) (Farquhar et al., 1980), each of which have their own temperature dependency. The direct dependence of g s on temperature has not been consistent across studies (Sage & Sharkey, 1987; von Caemmerer & Evans, 1991), but low g s in response to increased vapour pressure deficit (VPD) can reduce the temperature optimum of photosynthesis (Lin et al., 2012; Slot & Winter, 2016). In addition, changes in the biochemical component processes of the photosynthesis temperature response may be involved in T optA adjustments. Some studies have found increased T optA to be associated with an increase in the activation energy of V cmax, E aV, together with a decrease in the J max : V cmax ratio with increasing growth temperatures (Hikosaka et al., 2006). Other studies have found that the adjustment of T optA was associated with a decline in V cmax25 or J max25 (Medlyn et al., 2002b; Yamori et al., 2005; Sage & Kubien, 2007; Scafaro et al., 2017). Identifying which of these parameters changed across several growth temperatures will give us better insights into the principal processes responsible for photosynthetic temperature acclimation and whether any relationships underlie these processes.

Similar to photosynthesis, leaf respiration responds to short‐ and long‐term changes in growth temperatures. Leaf respiration typically shows an exponential increase with short‐term changes in temperature (Atkin & Tjoelker, 2003; Heskel et al., 2016), but can adjust rapidly to warmer temperatures via thermal acclimation in a manner that promotes homeostasis in metabolic function (Lee et al., 2005; Crous et al., 2011; Atkin et al., 2015; Aspinwall et al., 2016) via reduced rates of respiration at a standard temperature or via reduced temperature sensitivity of respiration (Atkin & Tjoelker, 2003). However, at higher growth temperatures, temperature acclimation of respiration may be constrained and not achieve homeostasis (Drake et al., 2017). A meta‐analysis including 103 species from different biomes found a general pattern of acclimation of respiration to warming, via reduced respiration rates at a set temperature (Slot & Kitajima, 2015). In addition, respiration rates at a set temperature can vary geographically, with higher rates of dark respiration at a standard temperature (25°C) in temperate species compared with tropical species (Atkin et al., 2015). Based on these reports, we would expect that respiration would be reduced with long‐term warming in all species to minimise carbon loss.

The magnitude of physiological adjustments in rainforest tree species and whether these physiological adjustments differed in tropical vs temperate tree species are currently not well known. This study aimed to determine the key components responsible for the photosynthesis and respiration temperature responses in Australian woody rainforest species from different climates. We also tested whether there was a difference between tropical and temperate species in their capacity to acclimate photosynthesis and leaf respiration to a range of warmer growth temperatures. As growth temperature is an important driver of thermal acclimation (Kumarathunge et al., 2019), using a range of growth temperatures in our experimental design enabled us to develop relationships across growth temperatures, while comparing how rainforest species from different climate zones differed across this temperature gradient. We addressed the following hypotheses: (1) the temperature optima of photosynthesis (T optA), V cmax (T optV) and J max (T optJ) will increase with increasing growth temperatures; (2) physiological processes at a common temperature including net photosynthesis at 25°C (A net25), V cmax25 and J max25 will be downregulated with increasing T growth; (3) respiration rate at 25°C (R 25) will be reduced with warming; and (4) adjustments in T optA, A opt and R 25 will be larger in temperate than tropical species. Using growth temperatures experienced within the native range, we aimed to imitate some of the growth conditions in the field. However, disentangling these detailed mechanisms would be hard to realise in the field. Our study in a controlled environment across similar soils focuses on the mechanistic differences in acclimation responses of rainforest species to warming across a large geographical scale.

Materials and Methods

Plant materials and experimental design

We grew seedlings of six Australian rainforest tree species at a range of growth temperatures. All species' distributions were located along the east‐coast margin of Australia between 12 and 40°S (Table 1). We selected common rainforest species with minimally overlapping distribution ranges. The species included two tropical species (Atractocarpus fitzalanii (F. Muell.) Puttock and Xanthostemon chrysanthus (F. Muell.) Benth.), two subtropical species (Backhousia citriodora F. Muell. and Flindersia australis R. Brown), and two warm‐temperate species (Cryptocarya laevigata Blume and Tristaniopsis laurina Sm.). All plant species are evergreen angiosperms with similar leaf traits (i.e. size, thickness) and none is classified as a pioneer species. Seedlings of the six species were obtained from two commercial nurseries (Burringbar Rainforest nursery, NSW and Yuruga nursery, QLD) with seeds of each species locally obtained from one seed source (Table 1). While rainforest species have a similar natural history, selecting species with similar traits from different climates enabled us to test whether acclimation capacity would differ depending on climate origin while also reflecting a temperature range that they currently experience. Species distributions were obtained from Atlas of Living Australia, and WorldClim climatology data (WorldClim 1.4; Hijmans et al., 2005) were used to calculate the average summer temperature (December–February) of the southernmost and northernmost latitude of each species' native occurrence (Table 1).

Table 1

Six rainforest species from three climates including their distributional range and the corresponding average summer temperatures at the southernmost and northernmost latitude occurrence based on WorldClim climatology data (WorldClim1.4; Hijmans et al., 2005).

Rainforest group	Species	Family	Latitude range of native distribution	Average summer temperature (°C)	Coordinates of seed collection
Warm‐temperate	Cryptocarya laevigata	Lauraceae	24.6–31°S	22.9–25.0°C	28°08′56″S, 153°25′05″E (Tallabugera, QLD)
	Tristaniopsis laurina	Myrtaceae	24.6–40°S	17.4–25.4°C	28°31′16″S, 153°32′28″E (Big Scrub, NSW)
Subtropical	Backhousia citriodora	Myrtaceae	17.1–33.9°S	22.0–23.4°C	27°28′13″S, 153°01′28″E (Brisbane, QLD)
	Flindersia australis	Rutaceae	17.2–35.2°S	19.7–23.4°C	28°23′16″S, 153°33′29″E (Pottsville, NSW)
Tropical	Atractocarpus fitzalanii	Rubiaceae	14.4–27.6°S	24.8–28.3°C	17°15′58″S, 145°29′09″E (Atherthon, QLD)
	Xanthostemon chrysanthus	Myrtaceae	12.4–19.2°S	27.1–27.2°C	17°07′40″S, 145°25′40″E (Walkamin, QLD)

Collection coordinates (latitude, longitude) are included in the last column.

To assess the effects of warming, seedlings of each species were grown under four growth temperatures. Six mean diel temperature treatments ranging from 17 to 34.5°C in 3.5°C increments were implemented in six adjacent, natural sunlit glasshouse bays. All growth temperature regimes were implemented with a diurnal range of 10°C. Therefore, temperatures during the day were warmer than the average diel temperatures reported above, with the target daily maximum temperatures ranging from 23 to 40°C in the coolest to hottest bays, respectively. The temperature treatments for each rainforest group included three temperature regimes that spanned the average summer temperatures in their native range and one temperature regime that was c. 3.5°C warmer than average summer temperatures currently experienced. We included a growth temperature of 24°C for tropical species (cooler than their average summer temperature) to obtain four growth temperatures for all species. Warm‐temperate species were grown under mean diel temperatures of 17, 20.5, 24 or 27.5°C, while subtropical and tropical species were grown at 24, 27.5, 31 or 34.5°C. Relative humidity in the respective glasshouse bays (17–34.5°C) were on average 81%, 84%, 88%, 72%, 76% and 72%, respectively, over 24 h (Carel Humidisk 65 humidifier, Sydney, Australia).

The experiment ran at Western Sydney University (Richmond, NSW, Australia) during the Austral summer of 2017–2018 (November–February). Seedlings were transplanted individually into 7 l pots of loamy sand soil. Ten seedlings of each of the six species were randomly assigned to each of the four temperature treatments on 6 November 2017. While some species grew faster than others, all seedlings within a given species started with similar heights. Seedling height at the beginning of the experiment ranged from 10 to 32 cm across species. After transplanting and before measurements, plants were allowed to establish for at least 4 wk to develop new leaves under the experimental conditions. After the establishment period, most species developed two to four new leaves every 2 wk with the exception of Backhousia citriodora and Tristaniopsis laurina which developed more than five leaves per fortnight. Seedlings were generally growing well following an exponential growth curve. Throughout the experiment, plants were kept well watered with an automated irrigation system and were fertilised weekly with a commercial fertiliser (50 ml at 2 g l−1; 25% N, 5% P, 8.8% K, 0.004% Zn, 0.005% Cu, 0.001% Mo, 0.01% Mn, 0.18% Fe, 0.005% B; Thrive soluble, Yates, Padstow, NSW, Australia).

Leaf photosynthesis and A net–C curves

Three plant replicates per species were randomly selected within each temperature treatment (6 species × 4 growth temperatures × 3 replicates = 72 plants) for leaf physiological measurements. Gas‐exchange measurements were conducted using several portable open gas‐exchange systems using the 2 × 3 cm leaf chamber with red and blue lamps (LI‐6400XT; Li‐Cor, Lincoln, NE, USA). One newly developed, fully expanded leaf was marked and measured on each plant replicate across 1–3 d to accommodate all leaf temperatures but avoid measurement ‘fatigue’. Leaves were measured at one point in time for light‐saturated photosynthesis (A net) between 08:30 h and 16:00 h, local time. Initial A net measurements were conducted at saturating light (1800 μmol m–2 s–1) and ambient CO2 concentration (c. 415 μmol mol–1) using a flow rate of 300 μmol s−1, followed by an A net–C response curve using a sequence of CO2 concentration levels (40, 150, 235, 330, 415, 700, 1200, 1500, 1800 μmol mol–1).

To establish the temperature responses of the apparent maximum carboxylation rate, V cmax and the apparent maximum electron transport rate, J max, these A net–C response curves were measured at five leaf temperatures (17, 25, 30, 35 and 40°C) on the same leaf (6 species × 4 growth temperatures × 3 replicates × 5 leaf temperatures = 360 A net–C curves). The temperature‐response curves were started at the respective growth temperature of each replicate, after which the rest of the temperatures were measured from low to high. Leaf temperature was controlled to within ±1°C of the target leaf temperature by manually adjusting the temperature of the chamber block. To achieve good temperature control, plants were temporarily moved to different bays for each of the five target measurement temperatures and measured after at least a 1‐h adjustment period. The relative humidity in the leaf cuvette was controlled between 50% and 70%. The leaf‐to‐air VPD during these measurements increased consistently with leaf temperature among the target measurement temperatures (from c. 1 to 4 kPa; Supporting Information Fig. S1).

Net photosynthesis temperature responses

Based on the initial A net measurements at each of five leaf temperature temperatures, temperature responses of photosynthesis were fitted using the following parabolic equation (Gunderson et al., 2010): where A net is the light‐saturated net photosynthetic rate (μmol m–2 s–1) at a given leaf temperature (T in °C); T opt, the temperature optimum for photosynthesis (°C). A opt is the light‐saturated net photosynthetic rate at T opt, and the parameter b describes the broadness of the curvature of the parabola.

Net photosynthesis at a common C (A 300)

We also examined A 300, the net photosynthesis rate at the mean C (300 μmol mol–1), which was obtained from each A net–C curve by interpolating the curve using the Farquhar model (Farquhar et al., 1980) with parameters fitted to that curve. When the photosynthetic rate is scaled to a common C , it eliminates the effect of variation in stomatal conductance on C , therefore isolating the temperature effects on photosynthetic biochemistry (Kumarathunge et al., 2019). The temperature optimum for photosynthesis at the mean C (T optA300) was estimated for each species by fitting Eqn 1. By comparing T optA300 and T optA, we estimated the effect of variation in stomatal conductance on the temperature optimum for photosynthesis.

Stomatal limitation

The stomatal limitation (Sl) of net photosynthesis was calculated by comparing photosynthesis from the fitted A net–C curve at the measured ambient C (A measuredC ) and photosynthesis from the fitted curves at the mean C (A 300) using the following equation: Stomatal limitation (Sl) was compared among species at each T growth in relation to leaf temperature.

Temperature dependence of photosynthetic biochemistry parameters

Each A net–C curve was fitted to the Farquhar et al. (1980) photosynthesis model using the ‘fitacis’ function in the plantecophys package in R (Duursma, 2015). We used the standardised kinetics parameters using the parameterisation given by Bernacchi et al. (2001). The reported rates of photosynthetic capacity are apparent J max and apparent V cmax values were based on data from intracellular CO2 (C ) concentrations rather than CO2 concentrations at the site of carboxylation. The temperature dependencies of apparent V cmax and J max were fitted using the modified version of the Arrhenius equation to reflect a peaked function (Medlyn et al., 2002a): where: k 25, value of V cmax or J max at 25°C; R, universal gas constant (8.314 J mol–1 K–1; T k, leaf temperature in K; E a (J mol–1), activation energy. E a describes the exponential rise of enzyme activity with increasing temperature. H d (J mol–1), deactivation energy and ΔS is the entropy term (J K−1). H d and ΔS together describe the rate of decrease in the function above the optimum. To avoid overparameterisation, H d was held at a constant of 200 kJ mol–1 for all species (Medlyn et al., 2002a; Kattge & Knorr, 2007).

The optimum temperatures (T opt) of V cmax and J max were calculated from the following equation (Medlyn et al., 2002a): where the variable abbreviations are explained above and loge represents the natural log.

Temperature dependence of dark respiration

The short‐term temperature dependence of leaf dark respiration (R T) was measured on three plants of each species in two of the four growth temperatures. Different pairs of growth temperatures were measured for temperate (17 and 20.5°C), subtropical (24 and 27.5°C) and tropical (27.5 and 31°C) species to assess the effect of +3.5°C warming on leaf respiration in each group. Leaves were kept in darkness for at least 30 min before measurements by covering them with aluminium foil. Measurements on excised leaves were conducted during the day using a flow rate of 300 μmol s–1 and a reference CO2 concentration of 400 μmol mol–1. Dark respiration was measured over a temperature range from 14 to 60°C, using a large gas‐exchange chamber (3010‐GWK1; HeinzWalz GmbH, Effeltrich, Germany) connected to an infrared gas analyser (IRGA, LI‐6400XT; Li‐Cor). The leaf temperature was raised at a rate of 1°C min−1 to obtain high‐resolution temperature‐response curves of dark respiration (O'Sullivan et al., 2017). The following equation was fitted to the data between 15 and 45°C: where: R T, respiration rate measured at a given temperature; T, leaf temperature, and the parameters R 25 and Q 10 characterise the respiration rate at 25°C and the proportional increase in respiration with a 10°C increase in temperature, respectively.

Data analysis

All graphs and statistical analysis were conducted in R v.3.3.2 (R Development Core Team, 2018). Data were checked for homogeneity and normality. Temperature‐response curves (Eqns 1, 4) were fitted to the data using the ‘nls’ function within the nlstools package (Baty et al., 2015). We used ANCOVA to distinguish effects of growth temperatures vs rainforest groups, including an interaction term. Regressions against T growth were compared across rainforest groups using ANCOVA; the slopes and intercepts of the groups were obtained using lstrends within lsmeans package. As R 25 and Q 10 were only measured at two growth temperatures, an ANOVA was used with growth temperature and rainforest group as categorical variables.

Results

Temperature responses of net photosynthesis

The optimum temperature for leaf net photosynthesis (T optA300 and T optA) was higher in tropical and subtropical species than in warm‐temperate species (Figs 1, 2). The optimum temperature of net photosynthesis at mean C , T optA300 in tropical and subtropical species ranged from 30.1 to 38.9°C, whereas in warm‐temperate species T optA300 ranged from 25.9 to 28.3°C (Table 2). Similar differences among rainforest groups were observed for both T optA and T optA300, indicating a minor role of stomatal conductance in causing the difference in optimum temperature.

Fig. 1

Temperature response of photosynthesis of six rainforest tree species at an intercellular CO2 concentration of 300 μmol mol–1 (A 300) at four different growth temperatures (means ± 1 SE): (a, b) warm‐temperate species, (c, d) subtropical species (Flindersia australis could not be fit at 34.5°C, data not shown) and (e, f) tropical species. Symbols representing growth temperatures are maintained in relevant subsequent graphs. The temperature response fits at each growth temperature is represented by different line types and colours: dashed blue lines, 17°C; dashed cyan lines 20.5°C; : solid black lines, 24°C; solid grey lines, 27.5°C; : long‐dashed yellow lines, 31°C; dashed red lines, 34.5°C.

Fig. 2

Means and ± SE of the temperature optimum for (a) leaf net photosynthesis (T optA) and (b) net photosynthesis at an intercellular CO2 concentration of 300 μmol mol–1 (T optA300) at different growth temperatures (T growth) for each rainforest group: warm‐temperate species in blue, subtropical species in orange and tropical species in red.

Table 2

Mean coefficients (± 1 SE) from the photosynthesis temperature response fits at mean C (300 μmol mol–1) and net photosynthesis at 25°C (A net25) for six rainforest tree species at four different growth temperatures.

Rainforest group	Species	Growth temperature	A opt300 (µmol m–2 s–1)	T optA300 (°C)	b (broadness)	A net25 (µmol m–2 s–1)
Warm temperate	Cryptocarya laevigata	17.0°C	6.62 ± 0.65	28.3 ± 1.8	0.020 ± 0.008	5.71 ± 0.81
		20.5°C	6.12 ± 0.66	27.5 ± 1.8	0.018 ± 0.007	5.48 ± 0.50
		24.0°C	6.24 ± 0.69	27.1 ± 1.5	0.023 ± 0.008	4.19 ± 1.21
		27.5°C	5.37 ± 0.60	27.0 ± 2.4	0.013 ± 0.006	3.59 ± 0.75
	Tristaniopsis laurina	17.0°C	11.01 ± 0.44	27.4 ± 0.5	0.040 ± 0.005	11.23 ± 0.33
		20.5°C	9.17 ± 0.66	26.3 ± 0.8	0.04 ± 0.007	10.00 ± 0.35
		24.0°C	13.10 ± 0.55	25.9 ± 0.9	0.038 ± 0.007	11.34 ± 0.63
		27.5°C	12.24 ± 0.94	27.9 ± 1.6	0.043 ± 0.015	10.58 ± 1.46
Subtropical	Backhousia citriodora	24.0°C	13.33 ± 0.59	31.3 ± 0.6	0.044 ± 0.007	12.01 ± 0.77
		27.5°C	13.71 ± 0.61	37.7 ± 4.3	0.016 ± 0.008	14.50 ± 0.76
		31.0°C	10.84 ± 0.61	38.9 ± 6.0	0.011 ± 0.006	9.93 ± 0.17
		34.5°C	10.32 ± 1.12	38.6 ± 10.6	0.011 ± 0.011	8.34 ± 1.21
	Flindersia australis	24.0°C	9.58 ± 0.67	31.4 ± 1.2	0.027 ± 0.008	6.68 ± 1.24
		27.5°C	12.19 ± 0.80	33.2 ± 2.1	0.026 ± 0.009	10.89 ± 0.58
		31.0°C	10.58 ± 0.85	38.3 ± 10.7	0.009 ± 0.009	5.66 ± 0.43
		34.5°C	–	–	–	2.78 ± 0.30
Tropical	Atractocarpus fitzalanii	24.0°C	8.54 ± 1.01	30.1 ± 2.0	0.022 ± 0.012	7.00 ± 0.39
		27.5°C	11.38 ± 0.93	31.1 ± 1.6	0.028 ± 0.011	8.94 ± 0.63
		31.0°C	8.25 ± 1.24	33.4 ± 5.4	0.014 ± 0.014	8.24 ± 2.092
		34.5°C	2.36 ± 0.48	30.3 ± 6.9	0.003 ± 0.005	1.76 ± 0.70
	Xanthostemon chrysanthus	24.0°C	14.54 ± 1.22	30.0 ± 1.1	0.055 ± 0.016	10.73 ± 0.99
		27.5°C	13.62 ± 1.00	33.3 ± 3.3	0.022 ± 0.014	9.37 ± 2.06
		31.0°C	10.58 ± 0.77	37.4 ± 8.4	0.009 ± 0.008	9.72 ± 1.73
		34.5°C	9.25 ± 0.47	35.8 ± 2.9	0.016 ± 0.006	7.98 ± 0.88

A parabolic function was used to fit according to Eqn 1 in the Materials and Methods section. Flindersia australis fits at 34.5°C had T optA300 outside the measurement range.

We observed acclimation of both T optA and T optA300 to growth temperature (Table 3). T optA increased with growth temperature and the rate of increase was similar across groups (Fig. 2a). When normalised for differences in stomatal conductance, T optA300 displayed relationships with T growth that differed among rainforest groups P = 0.02; Table 3; Fig. 2b). The sensitivity (i.e. slope) of T optA300 to T growth was higher in the subtropical (0.78 ± 0.20°C °C–1 increase in T growth) and tropical species (0.35 ± 0.17°C °C–1 increase in T growth) than in warm‐temperate species that showed little increase in T optA300 (P = 0.020; ANCOVA; Table 3; Fig. 2).

Table 3

Analysis of covariance F‐statistic and P‐values of the regressions of photosynthetic parameters against T growth across rainforest groups (Group).

Parameter	T growth df = 1		Group df = 2		T growth × Group df = 2
	F‐statistic	P‐value	F‐statistic	P‐value	F‐statistic	P‐value	Resid df
T optA	25.10	0.00016	19.52	< 0.0001	1.54	0.25	15
T optA300	55.58	< 0.0001	15.44	0.00015	5.00	0.020	17
A opt	0.23	0.64	4.27	0.034	0.82	0.46	15
A opt300	0.07	0.79	3.00	0.077	1.60	0.23	17
A net25	1.67	0.21	2.69	0.081	0.74	0.54	16
V cmax25	0.09	0.77	1.37	0.28	0.52	0.60	17
J max25	8.77	0.0088	1.48	0.25	5.47	0.015	17
J max25 : V cmax25	2.33	0.14	0.74	0.49	1.48	0.25	18
T optV	15.48	0.0012	1.69	0.22	1.34	0.29	16
T optJ	22.18	0.0002	5.95	0.012	0.78	0.47	16
E aV	0.46	0.50	1.79	0.20	1.73	0.21	17
E aJ	0.15	0.70	1.40	0.27	1.98	0.17	17
ΔS V	8.28	0.011	0.38	0.69	0.19	0.82	16
ΔS J	15.05	0.0013	3.94	0.040	0.32	0.73	16

Parameters are: the temperature optima of net photosynthesis (T optA) and of photosynthesis at [CO2] = 300 μmol mol–1 (T optA300), and the maximum photosynthesis rates at the temperature optima (A opt and A opt300), the maximum carboxylation rate at 25°C (V cmax25) and its activation energy (E aV), the maximum electron transport rate at 25°C (J max25) and its associated activation energy (E aJ), while its deactivation energy (H d) was kept constant at 200 kJ mol–1. T optV and T optJ represent the temperature optima of V cmax and J max fits, respectively. ∆S represents the entropy factor in the model (∆S V and ∆S J for V cmax and J max fits, respectively). df stands for degrees of freedom, including the residual df in the last column. Bold values are the significant relationships at P < 0.05; italic values are significant at P < 0.1.

Consistent with the higher T optA, the photosynthetic rate at the temperature optimum (A opt) was significantly higher in tropical and subtropical species than in temperate species (P = 0.03; Table 3; Fig. S2). However, there was no evidence that either A opt or A opt300 acclimated to growth temperature (P > 0.64; Table 3). Net photosynthesis at 25°C (A net25) also showed no difference with growth temperatures but tended to be decreased more in tropical and subtropical species compared with temperate species (P = 0.08; Table 3), especially at higher growth temperatures (Table 2; Fig. S2).

Stomatal component of the photosynthetic response to temperature

Temperature responses of photosynthesis can potentially be affected by stomatal closure at high VPD associated with high leaf temperatures (Fig. S3), which was the case for warm‐temperate T. laurina and subtropical B. citriodora (Fig. S3) while only B. citriodora significantly decreased its conductance at higher leaf temperatures (Fig. 3). However, the g s responses to T leaf did not differ at different T growth in any species (Fig. S4), indicating that changes in T optA with T growth were not caused by changes in stomatal response (Fig. 3). Moreover, stomatal limitation was not significantly related to leaf temperature (P > 0.05) for most species at any growth temperature, except in three out of 24 instances for C. laevigata grown at 17°C, B. citriodora grown at 31°C and F. australis grown at 27.5°C (Fig. S5).

Fig. 3

Relationships between leaf temperature (T leaf) and stomatal conductance (g s) at different growth temperatures (different symbols) in six rainforest tree species: (a, b) warm‐temperate species in blue, (c, d) subtropical in orange, and (e, f) tropical species in red. The relationship was only significant for B. citriodora (g s = −0.0109, T leaf + 0.70, R 2 = 0.77, P = 0.01).

Biochemical component of the photosynthetic response to temperature

The observed acclimation of T optA was mainly driven by adjustments of photosynthetic biochemistry, which showed significant differences in response to growth temperatures in most parameters (Table 3). Temperature responses of V cmax and J max are shown for each species and growth temperature in Figs S6 and S7, while the parameters extracted by fitting Eqn 3 to these curves are shown in Fig. 4 and Table 4.

Fig. 4

Means and ± SE of parameter values of individually fitted functions to characterise the temperature dependence of maximum carboxylation rate (V cmax) (a, c, e, g) and maximum electron transport rate (J max) (b, d, f, h) at 25°C across growth temperatures. (a, b) Standard value of V cmax and J max at 25°C, (c, d) optimum temperature (T opt) of V cmax and J max, (e, f) activation energy (E a) for V cmax and J max, (g, h) entropy term (∆S) for V cmax and J max fits. Warm‐temperate species are indicated in blue, subtropical species are orange and tropical species are red. ANCOVA of linear regressions across growth temperatures and rainforest groups are reported in Table 3. Significant linear regressions (and no significant differences between neither slopes nor intercepts) were found: (c) T optV and T growth (T optV = 0.49, T growth + 27.12; R 2 = 0.38, P = 0.001); (g) ∆S V and T growth (∆S V = −1.00, T growth + 659.86; R 2 = 0.29, P = 0.005).

Table 4

Table of coefficients ± 1 SE from temperature response fits for two warm‐temperate, two subtropical and two tropical rainforest species across four growth temperatures reflecting warming within and beyond their native range.

Rainforest group	Species	T growth	Rubisco carboxylation					Electron transport
			V cmax25 (µmol m–2 s–1)	E aV (kJ mol–1)	∆S V (J mol–1 K–1)	T optV (°C)	R 2	J max25 (µmol m–2 s–1)	E aJ (kJ mol–1)	∆S J (J mol–1 K–1)	T optJ (°C)	R 2
Warm temperate	Cryptocarya laevigata	17°C	30.4 ± 8.0	57.3 ± 23.2	640 ± 20	37.36	0.37	69.6 ± 6.3	46.4 ± 15.8	640 ± 10	34.90	0.72
		20.5°C	25.4 ± 5.0	101.3 ± 48.8	650 ± 10	36.40	0.70	63.1 ± 7.3	54.1 ± 21.0	641 ± 10	34.80	0.52
		24°C	27.3 ± 8.0	96.9 ± 75.6	650 ± 20	35.00	0.55	56.9 ± 5.0	43.4 ± 14.8	637 ± 4	33.58	0.59
		27.5°C	19.0 ± 4.0	72.4 ± 29.0	630 ± 10	41.66	0.65	55.2 ± 6.0	39.3 ± 14.1	632 ± 8	37.45	0.54
	Tristaniopsis laurina	17°C	47.8 ± 3.3	61.3 ± 11.8	638 ± 4	36.11	0.77	86.3 ± 4.2	41.9 ± 10.2	640 ± 4	32.26	0.77
		20.5°C	44.0 ± 4.6	75.4 ± 36.6	652 ± 8	31.57	0.75	77.4 ± 8.2	48.0 ± 24.9	646 ± 8	31.84	0.60
		24°C	57.2 ± 3.0	66.4 ± 8.6	637 ± 3	37.89	0.92	108.3 ± 5.7	48.1 ± 10.9	644 ± 4	33.07	0.72
		27.5°C	57.5 ± 4.8	35.0 ± 11.1	623 ± 10	40.58	0.68	89.6 ± 6.2	34.9 ± 10.4	633 ± 5	36.60	0.68
Subtropical	Backhousia citriodora	24°C	55.7 ± 5.4	80.9 ± 13.5	640 ± 3	39.34	0.93	91.8 ± 5.4	29.3 ± 8.0	620 ± 10	42.27	0.80
		27.5°C	68.3 ± 10.4	51.0 ± 18.8	610 ± 40	47.81	0.78	74.9 ± 5.5	38.6 ± 4.7	*		0.88
		31°C	53.0 ± 5.3	71.0 ± 14.7	630 ± 10	42.50	0.93	68.4 ± 6.0	38.4 ± 12.9	620 ± 10	41.06	0.73
		34.5°C	38.7 ± 7.6	77.4 ± 28.3	640 ± 10	38.71	0.73	56.2 ± 3.6	38.1 ± 8.7	630 ± 10	40.00	0.81
	Flindersia australis	24°C	32.8 ± 6.1	95.1 ± 26.6	640 ± 10	41.12	0.86	63.4 ± 5.4	43.2 ± 12.4	630 ± 10	40.65	0.73
		27.5°C	48.6 ± 10.0	70.0 ± 25.3	630 ± 20	43.70	0.78	74.3 ± 6.7	50.8 ± 11.7	620 ± 10	43.90	0.89
		31°C	35.1 ± 7.0	93.7 ± 23.2	630 ± 10	43.19	0.89	77.8 ± 6.4	39.1 ± 9.2	610 ± 30	46.68	0.82
		34.5°C	25.5 ± 7.0	85.2 ± 23.2	620 ± 40	40.58	0.68	47.8 ± 6.2	44.3 ± 17.3	620 ± 30	44.40	0.74
Tropical	Atractocarpus fitzalanii	24°C	32.5 ± 7.5	113.6 ± 45.5	640 ± 10	38.50	0.83	90.0 ± 7.4	42.9 ± 16.1	640 ± 5	35.30	0.71
		27.5°C	47.0 ± 7.5	75.2 ± 21.8	630 ± 10	40.36	0.78	88.3 ± 8.8	52.9 ± 14.8	630 ± 10	37.86	0.67
		31°C	38.9 ± 10.2	70.0 ± 38.0	630 ± 20	41.42	0.76	63.6 ± 12.3	69.1 ± 34.2	640 ± 10	37.82	0.75
		34.5°C	17.6 ± 4.8	8.7 ± 35.1	620 ± 40	42.46	0.21	21.6 ± 5.4	58.7 ± 37.5	640 ± 10	38.20	0.47
	Xanthostemon chrysanthus	24°C	51.7 ± 12.6	98.7 ± 46.8	640 ± 10	37.65	0.68	103.6 ± 13.5	32.2 ± 19.7	630 ± 10	37.54	0.33
		27.5°C	48.2 ± 8.4	87.4 ± 23.3	630 ± 10	42.00	0.79	95.4 ± 8.8	31.6 ± 13.4	620 ± 3	43.95	0.52
		31°C	47.3 ± 4.7	65.4 ± 12.7	620 ± 20	46.76	0.96	65.7 ± 7.1	48.5 ± 18.4	630 ± 10	40.40	0.82
		34.5°C	33.1 ± 4.4	80.2 ± 7.3	*		0.95	47.0 ± 5.2	39.9 ± 14.2	610 ± 50	47.30	0.74

Each fit was performed across three replicates per treatment using a peaked Arrhenius function, except two fits for which a standard Arrhenius was used (indicated with *) because they did not show a peak (V cmax of X. chrysanthus at 34.5°C and J max of B. citriodora at 27.5°C). Coefficients are the maximum carboxylation rate at 25°C (V cmax25), the maximum electron transport at 25°C (J max25), the activation energy (E aV and E aJ for V cmax and J max, respectively), and the entropy (∆S V and ∆S J, respectively). The temperature optimum (T opt) of each fit was calculated. R 2 represents the overall model fit.

In general, J max25 declined with increasing growth temperature (P = 0.009; Table 3) with an interaction between growth temperature and rainforest group (P = 0.015). The interaction indicated significantly higher reduction of J max25 in pooled subtropical and tropical species (−6.13 ± 1.33 for each °C increase in T growth) than warm‐temperate species (−0.12 ± 1.33 for each °C increase in T growth) (Fig. 4b). Several species exhibited large J max25 reductions (Table 4): tropical A. fitzalanii displayed more than 75% reduction of J max25 at the highest T growth compared with 27.5°C, while subtropical B. citriodora reduced J max25 by c. 39% at 34.5°C compared with 24°C. By contrast, V cmax measured at 25°C (V cmax25) showed no difference in response to growth temperatures (P = 0.77; Fig. 4a; Table 3) and there were also no differences among rainforest groups (P = 0.28). While the ratio of J max25 to V cmax25 (J max25 : V cmax25) did not decline with higher growth temperatures (P = 0.14) and was similar among rainforest groups (P = 0.49; Fig. S8a), J max : V cmax did reduce significantly across all growth temperatures (P = 0.003).

Similar to T optA, the optimum temperatures of both V cmax (T optV) and J max (T optJ) increased significantly with increasing growth temperatures across species (P < 0.0012; Fig. 4c,d; Table 3). The slopes of linear regressions of T optV and T optJ, with growth temperature were 0.49 ± 0.13°C °C–1 and 0.56 ± 0.14°C °C–1, respectively (Fig. 4c,d) and slopes were not different among groups, indicating a similar degree of acclimation for all rainforest groups (P > 0.05; Table 4). The change in T optA was not associated with a change in activation energy (E a) of either V cmax or J max; both E aJ and E aV were independent of growth temperature (P > 0.5; Fig. 4e,f; Table 3). Across growth temperatures, the increases in T optV and T optJ were associated with a significant decline in ∆S V and ∆S J (P < 0.011; Fig. 4g,h; Table 3). Therefore, several biochemical parameters (J max25, T optV, T optJ, ∆S V, ∆S J) responded to higher growth temperatures across species, suggesting a strong influence of growth temperatures on the underlying biochemical components of the temperature response of photosynthesis. Similarly to J max25, the associated parameters T optJ and ΔS J varied among rainforest groups (P < 0.041; Table 3) resulting in higher T optJ and lower ∆S J in subtropical and tropical species compared with warm‐temperate species (different intercepts in Fig. 4d,h).

There were strong relationships between several biochemical parameters and the temperature optimum of photosynthesis, T optA300 (Fig. 5). We found a positive relationship between T optA300 and the optimum temperatures of V cmax (T optV) and of J max (T optJ) (Fig. 5a,b), indicating that higher T opt in V cmax and J max were correlated with a higher temperature optimum of photosynthesis. A low J max25 : V cmax25 ratio was also related to high temperature optima of photosynthesis (Fig. 5c). While there was no significant relationship between T optA300 and the activation energy (either E aV or E aJ, data not shown), a strong negative relationship was found between ΔS J and T optA300 (R 2 = 0.43; Fig. 5d) and between ΔS V and T optA300 (R 2 = 0.17, data not shown), in line with the reduction of ΔS V and ΔS J with increasing growth temperature across species (Fig. 4g,h; Table 3).

Fig. 5

Relationships of means and ± 1 SE between several biochemical parameters and the temperature optimum of photosynthesis at a common C of 300 µmol mol–1 (T optA300) among six rainforest tree species. Warm‐temperate species are indicated in blue, subtropical species are orange and tropical species are red: (a) T optA300 and T optV (T optA300 = 0.72, T optV + 2.72; R 2 = 0.20, P = 0.02), (b) T optA300 and T optJ (T optA300 = 0.92 T optJ − 3.85; R 2 = 0.52, P < 0.01), (c) T optA300 and J max25/V cmax25 (T optA300 = −5.57, J max25/V cmax25 + 41.71; R 2 = 0.38, P = 0.001), (d) T optA300 and ∆S J (T optA300 = −0.41, ΔS J + 287.43; R 2 = 0.43, P < 0.01).

Temperature response of respiration

The response of dark respiration to leaf temperature acclimated differently to warming in warm‐temperate vs subtropical and tropical species (Figs 6, S9), reflected by adjustments in both respiration rates at a common temperature, R 25 and the temperature sensitivity of respiration, Q 10. While R 25 rates tended to increase with warming in temperate species, and decrease in tropical species (Fig. 6a), there were no differences in response to 3.5°C warming (P = 0.88). There were differences among rainforest groups in R 25 (P = 0.045) with subtropical species having the highest R 25 (Fig. 6a) while tropical and warm‐temperate species exhibited similar rates.

Fig. 6

Means and ± SE of (a) dark respiration rate at 25°C (R 25) across six rainforest tree species, (b) Q 10 of respiration (calculated across 15–45°C), and (c) dark respiration each at two T growth (R Tgrowth). Warm‐temperate species are indicated in blue, subtropical species are orange and tropical species are red.

Similarly, the decrease in the temperature sensitivity of respiration (Q 10 between 15 to 45°C) was significantly different among rainforest groups (P = 0.027) but not different with categorical growth temperatures (P = 0.90). On average, Q 10 was reduced by 24% in subtropical and tropical species, compared with the average Q 10 of warm‐temperate species (P < 0.025). These adjustments resulted in homeostasis in the respiration rate at T growth (R Tgrowth) among subtropical and tropical species, while warm‐temperate species had lower R Tgrowth (Fig. 6c; P = 0.003).

Discussion

We found that Australian rainforest species adjusted both photosynthesis and respiration to warming. Tropical and subtropical species showed as much, if not more, capacity to acclimate as did warm‐temperate species, despite being grown at a higher range of temperatures. The optimum temperature for photosynthesis increased with growth temperature and the rate of increase was higher in tropical and subtropical species than in warm‐temperate species. We showed that adjustments of biochemical processes, particularly changes in J max25, ΔS and the optimal temperatures for RuBP regeneration and Rubisco carboxylation, were the principal mechanisms underlying the shift in temperature optimum of photosynthesis in response to increased growth temperatures. Tropical species showed greater acclimation of respiration rates via both a reduction in R 25 and Q 10 whereas warm‐temperate species did not reduce R 25 at higher growth temperatures. These adjustments resulted in similar respiration rates at higher growth temperatures in tropical species, rather than enhanced R Tgrowth by warming in temperate species.

The photosynthetic response to growth temperature

The rate of increase in T optA300 was 0.35–0.78°C °C–1 increase in T growth, a shift similar in magnitude to that observed in earlier studies (e.g. Berry & Björkman, 1980; Kumarathunge et al., 2019). Tropical and subtropical species showed a higher increase of T optA300 (c. 8°C increase) with warming than warm‐temperate species (c. 3°C increase), similar to Cunningham & Read (2002). The larger shift in T optA300 in our study in tropical species compared with temperate species could, in part, be related to the fact the tropical species were exposed to higher growth temperatures compared with temperate species.

The increase in T optA300 with higher growth temperatures was underpinned by an increase in T opt of V cmax25 and T opt of J max25 (Fig. 4) supporting our hypothesis that the temperature optima of photosynthesis (T optA), V cmax (T optV), and J max (T optJ) would increase with increasing growth temperatures. The magnitude of acclimation of T optV (0.49°C °C–1) was similar to Dreyer et al. (2001), in a study of seven European tree species, and Kattge & Knorr (2007) (0.44°C °C–1), in a meta‐analysis of 36 species. There was a strong relationship between T optA300 and T optV and between T optA300 and T optJ (Fig. 5a,b) suggesting that biochemical components strongly influenced the temperature optimum of photosynthesis.

A strong negative correlation between ∆S J and T optA300 (Fig. 5d) indicated that the reduction in ∆S also contributed to an increased temperature optimum of photosynthesis in response to warming. Using a peaked Arrhenius function, the change in T optV and T optJ is generally explained by the sensitivity of ∆S to temperature (Kattge & Knorr, 2007; Kumarathunge et al., 2019). We observed a decline of ∆S V and ∆S J in response to growth temperature, and the magnitude of this decline (−1.00 ± 0.32 and −1.10 ± 0.31 J mol–1 K–1 °C–1, for ∆S V and ∆S J, respectively) was very similar to Kattge & Knorr (2007) and to Kumarathunge et al. (2019). There was no positive relationship observed between either activation energy (E aV or E aJ) and T growth. While some studies have found a positive relationship between activation energy and growth temperature (Hikosaka et al., 1999, 2006; Onoda et al., 2005), these studies typically did not use a peaked temperature‐response function in which the ΔS parameter is quantified. Kattge & Knorr (2007) also found lack of E aV and E aJ responses to T growth in 36 rainforest species. Overall, the selected rainforest species clearly adjusted the underlying biochemical components of photosynthesis to warmer growth temperatures, mainly via reduced J max25, reduced ∆S and increased temperature optima (T optV and T optJ).

The reduction in J max25 with increased growth temperatures supported our hypothesis of downregulation of photosynthetic capacity at warmer T growth via the decline of J max25, but not V cmax25 (Fig. 4). By contrast, Scafaro et al. (2017) associated the decline in net photosynthetic rate with a decline in V cmax25 as T growth increased across six temperate and four tropical rainforest species. However, other studies have reported increased V cmax25 with warming (Hikosaka et al., 1999; Onoda et al., 2005; Ghannoum et al., 2010; Smith & Dukes, 2017; Lamba et al., 2018). While photosynthetic capacity can be adjusted by either a change in V cmax25 of J max25 or both, many studies have reported a reduced J max25 : V cmax25 ratio with increasing growth temperatures across many species (Atkin et al., 2006; Hikosaka et al., 2006; Kattge & Knorr, 2007; Sage & Kubien, 2007; Lin et al., 2012; Crous et al., 2013; Dusenge et al., 2020). The reduction of J max/V cmax is likely to be related to the greater dependence of J max upon membrane stability than V cmax at higher temperatures (Hikosaka et al., 2006; Sage & Kubien, 2007). Moreover, a lower J max25 : V cmax25 ratio was related to a higher T optA300 (Fig. 5; Kumarathunge et al., 2019; Dusenge et al., 2020). As J max and V cmax have different temperature optima, the optimum temperature of photosynthesis is determined by the most limiting component process of photosynthesis. At lower J max25 : V cmax25, there is less Rubisco limitation compared with a higher J max25 : V cmax25, an adjustment that leads to a higher T optA at higher growth temperatures (Hikosaka et al., 2006).

Our study was designed to distinguish the acclimation capacity among rainforest species from different climates in a controlled environment. Most biochemical parameters describing the temperature response of photosynthesis responded to a change in growth temperatures without a difference among rainforest groups (Table 3), suggesting similar acclimation capacity among rainforest groups. Similar acclimation to experimental warming has also been reported in Slot & Kitajima (2015), Slot & Winter (2017) and Crous et al. (2022). One novel result here was that J max25 not only reduced with increased growth temperatures but the reduction was stronger in tropical and subtropical species compared with warm‐temperate rainforest species. The reduction in J max25 was underpinned by lower ΔS J in tropical and subtropical species. While a reduction in J max25 could lead to impaired photosynthesis rates with warming, this reduction was counteracted by increased temperature optima with warming, resulting in similar photosynthesis rates (A net25 and A opt300) across rainforest groups (P = 0.08; Table 3). Similar photosynthesis rates across a range of growth temperatures were also reported for a tropical montane species, Syzygium guineense (Dusenge et al., 2021), although rates did decrease above 30°C. Other studies have found reduced rates of photosynthesis in tropical species with warming especially at growth temperatures above 30°C (Cunningham & Read, 2003; Slot & Winter, 2016; Scafaro et al., 2017; Crous et al., 2018; Dusenge et al., 2021). While not statistically significant, photosynthesis tended to decline at the highest growth temperatures in our study (Table 2) suggesting that some species may decline photosynthesis at high growth temperatures despite thermal acclimation. There is clearly a need to understand how acclimation capacity differs among species from different latitudes or biogeographic regions, which our study addressed in Australian rainforest species. Recently, Crous et al. (2022) reviewed the acclimation responses in evergreen species from boreal to tropical latitudes and found more negative responses of photosynthesis to warming in species experiencing higher growth temperatures.

Another factor potentially contributing to reduced photosynthesis at higher growth temperatures could be reduced stomatal conductance. In contrast to most biochemical components, the stomatal conductance response to growth temperatures remained mostly unchanged and stomatal limitation was generally not related to leaf temperature in our study. This implies that g s was not a major component driving a shift in the temperature response of photosynthesis in these rainforest species, which is likely due to the well watered conditions in this experiment. Gunderson et al. (2010), Crous et al. (2018) and Kumarathunge et al. (2019) also demonstrated that stomatal conductance was not a major component driving the adjustments of the photosynthesis temperature response. However, water limitations can reduce the temperature optima for photosynthesis in trees (Lin et al., 2012; Kumarathunge et al., 2020). A recent study found that the stomatal response to high VPD could be a major driver of the decline of photosynthesis at higher temperatures in tropical forests (Smith et al., 2020), especially when VPD is not manipulated, which is in contrast with most controlled environments in which VPD is minimised via high humidity. The decrease of g s with increasing temperatures is not uncommon and has been observed in several studies (e.g. Slot et al., 2016; Carter et al., 2021) suggesting that photosynthesis rates could additionally be constrained by stomatal conductance at higher growth temperatures.

Response of leaf dark respiration to growth temperature

Since photosynthesis and respiration are coupled (Reich et al., 1998; Whitehead et al., 2004; Dusenge et al., 2019), plants with strong photosynthetic acclimation might also show strong thermal respiratory acclimation. R Tgrowth tended to increase with warming in temperate rainforest species (Fig. 6c) while tropical and subtropical species displayed similar respiration rates regardless of growth temperature. This homeostatic response of R Tgrowth can be explained via the adjustments of both a reduced temperature sensitivity (Q 10) and reduced rates of R 25 in tropical compared with temperate rainforest species (Fig. 6).

Subtropical and tropical species showed 24% lower temperature sensitivity (Q 10) compared with warm‐temperate species. Changes in Q 10 enabled respiration to adjust dynamically to changes in growth temperature (Atkin & Tjoelker, 2003), probably via changes in substrate and/or adenylate control (Armstrong et al., 2008). This reduction in Q 10 in subtropical and tropical species was likely related to the higher growth temperatures they experienced compared with temperate species, and enabled tropical and subtropical species to reduce carbon loss via reduced respiration. Therefore tropical species had similar R 25 as found for temperate species, regardless of warming combined with a lower Q 10. While we could not attribute this response to warming, reduced respiration rates in response to long‐term warming are commonly observed in many species (Atkin et al., 2005; Crous et al., 2011; Aspinwall et al., 2016; Drake et al., 2015, 2017), including in tropical species (Slot & Kitajima, 2015). Adjustments in R 25 and Q 10 resulted in homeostasis in the respiration rate at T growth (R Tgrowth) among subtropical and tropical species, regardless of warming, while warm‐temperate species had lower R Tgrowth (Fig. 6c).

Because tropical species reduced both R 25 and Q 10, respiration rates in tropical species adjusted more than respiration rates in temperate species, refuting our fourth hypothesis. Therefore tropical and subtropical species have a large capacity to acclimate respiration in response to moderate warming (Slot & Kitajima, 2015). The adjustment of leaf dark respiration in tropical and subtropical species partly offsets the decrease of photosynthetic capacity (J max25), although acclimation of stem or root respiration to warming may be more limited (Drake et al., 2017; Noh et al., 2020). In terms of leaf‐level responses, tropical species may adjust to warming via a combined effect of an increased T optA and reduced respiration, but it is unclear whether these responses are enough to prevent reduced photosynthesis with future warming.

Mercado et al. (2018) showed that thermal acclimation of photosynthetic capacity reduced the vulnerability of temperate and tropical species to warming in an Earth system model, although acclimation of respiration was not accounted for. Moreover, recent field studies on tropical tree species have observed limited thermal acclimation (Carter et al., 2021; Dusenge et al., 2021). Our findings extends this work by investigating the mechanisms controlling the photosynthesis temperature response across a range of growth temperatures to determine differences in acclimation capacity in rainforest trees from different climates. While acclimation capacity of photosynthesis may be more limited with higher growth temperatures, including in the tropics (Carter et al., 2021; Crous et al., 2022), there is a need to investigate the limits of thermal acclimation across latitudes, including at extreme temperatures (Zhu et al., 2018). Understanding how thermal acclimation varies geographically and what factors are controlling this response would contribute to reduce uncertainties regarding physiological responses to warming and improve projections of future carbon uptake in models.

Contrasting responses between tropical and warm‐temperate rainforest trees

According to our hypothesis, the shift in the optimum temperature of photosynthesis (T optA) would be larger in warm‐temperate than subtropical and tropical species due to the larger seasonal temperature variation temperate species experience. In contrast with this hypothesis, we found larger adjustments of T optA300 in tropical and subtropical species compared with warm‐temperate species related to an increase in T optJ and T optV. While other studies have generally found a higher T opt of photosynthesis in tropical species compared with temperate species (Cunningham & Read, 2003; Scafaro et al., 2017; Mau et al., 2018; Crous et al., 2022), our study also found that the adjustment in T optA was larger in tropical species (8°C range) compared with warm‐temperate species (4.5°C range).

Tropical species also exhibited stronger reductions in J max25 compared with temperate species in response to warmer growth temperatures. Other studies also reported reduced rates of J max25 in response to warming in several broadleaf evergreen species (Aspinwall et al., 2016; Scafaro et al., 2017; Smith & Dukes, 2017; Crous et al., 2018; Carter et al., 2020) especially in species exposed to higher growth temperatures. Reduced J max rates are likely to be related to reduced photosystem II (PSII) electron flow at moderately high temperatures (Havaux, 1996; Pastenes & Horton, 1996; Sharkey, 2005). The shift away from a linear electron flow by PSII towards increased cyclic electron flow can help to maintain a pH gradient (and proton motive force) across the thylakoid membrane in response to high temperatures and counteracts increased membrane leakiness of the thylakoid membrane (Bukhov et al., 1999; Schrader et al., 2004; Sharkey & Zhang, 2010). This mechanism, together with other structural changes in the thylakoid membrane, can help to avoid thermal damage and support thermostability in response to higher growth temperatures. Other adjustments to cope with high temperatures can be increased content of saturated fatty acids (Zhu et al., 2018), induction of heat‐shock proteins (Vierling, 1991) and changes in osmotic potential (Hüve et al., 2006).

While warming may have negative effects on carbon storage, tropical species have some thermal resilience with moderate warming. Tropical and subtropical species have a large capacity to adjust to warmer temperatures via increases in the temperature optimum of photosynthesis and reduced respiration rates. Sullivan et al. (2020) reported that high daytime temperatures (> 32°C) contributed most to reduced growth rates in the tropics. Therefore, it remains important to further investigate the responses to warming in tropical forests (Cavaleri et al., 2015) both in the short term and over longer time periods (Sullivan et al., 2020) and evaluate how resilient tropical species are to future warming. This is especially true for mature trees rather than seedlings. Mature individuals may not exhibit the same responses as seedlings and much fewer studies have involved large trees (Doughty & Goulden, 2008; Crous et al., 2013; Slot et al., 2014; Aspinwall et al., 2016). Based on these and other studies conducted on large trees in the field, our evidence of thermal acclimation of photosynthesis and respiration is similar in magnitude and direction. Moreover, other evidence points to convergent acclimation of photosynthesis and respiration to both seasonal and experimental warming (Vanderwel et al., 2015; Reich et al., 2016). While this is the case for the processes of carbon uptake and loss, other processes such as growth and water transport regulation are likely to be different between seedlings and mature trees.

Overall, this study provided insights into the mechanisms controlling the photosynthesis temperature response in rainforest tree species and their capacity to adjust to warming. Subtropical and tropical species showed greater adjustments of T optA300, and leaf respiration compared with warm‐temperate species. Photosynthetic capacity was reduced via a stronger reduction in J max25 in tropical compared with warm‐temperate species. Therefore, our study found that tropical and subtropical rainforest tree species substantially acclimated to higher temperatures while exhibiting reductions in photosynthetic capacity at higher growth temperatures (> 32°C), indicating that tropical species are likely to reduce carbon uptake at higher growth temperatures. Photosynthetic biochemistry, but not stomatal limitation, was the main driver of the shift in the temperature response of photosynthesis and the increase of the temperature optimum of photosynthesis.

Both the temperature‐driven shifts in photosynthesis and respiration should be considered together when forecasting future warming impacts at larger scales (Smith & Dukes, 2013; Mercado et al., 2018). In addition, there is a need to understand how the responses to experimental warming are different from responses to extreme temperatures. Our findings indicated that the magnitude of acclimation can differ depending on climate zone, and that warming responses at 3.5°C warming may not be the same compared with larger temperature changes. These differences in acclimation capacity can be modified in combination with other climate change factors, for example limited water availability, a common co‐limitation with warming in the field (Kumarathunge et al., 2020). These are important questions to be addressed in future research to reflect more realistic climate change scenarios when predicting carbon exchange in land surface models.

Author contributions

KYC designed the study with input of ZC and MGT. ZC led the data collection with help of AW‐K, AB and NPB. ZC performed the data analyses with the guidance of KYC, BEM and MGT. Manuscript writing was led by ZC and KYC with input from all co‐authors.

Fig. S1 Relationships between leaf temperature (T leaf) and leaf‐to‐air vapour pressure deficit (VPD) at different growth temperatures.

Fig. S2 Temperature responses of photosynthesis at the mean Ci (300 ppm) and different growth temperatures in six rainforest species from three different rainforest groups (means ± 1 SE, n = 3) (warm‐temperate species, subtropical species and tropical species).

Fig. S3 Relationships between stomatal conductance (g s) and vapour pressure deficit (VPD) at different growth temperatures in warm‐temperate species, subtropical species and tropical species).

Fig. S4 Means and SE between stomatal conductance measured at 25°C (g s25) and different growth temperatures (T growth) across six rainforest species.

Fig. S5 Relationships between A measuredC –A 300 and leaf temperature (T leaf) at different growth temperatures in warm‐temperate species, subtropical species and tropical species at each growth temperature.

Fig. S6 Temperature responses of the apparent maximum carboxylation (V cmax) for each of six rainforest species measured in their respective growth temperatures at five leaf temperatures (T leaf) in three replicates per growth temperature (means ± 1 SE).

Fig. S7 Temperature response of the apparent maximum electron transport rate (J max) for each of six rainforest species measured in their respective growth temperatures at five leaf temperatures (T leaf) in three replicates per growth temperature (means ±1 SE).

Fig. S8 Mean and SE of fitted values (3 replicates) of J max : V cmax at a common leaf temperature (25°C) as a function of six growth temperatures across three rainforest groups: J max25/V cmax25 = −0.04 × T growth + 3.03 (P = 0.01; R 2 = 0.20).

Fig. S9 The short‐term temperature response of mitochondrial leaf respiration (R dark) as a function of leaf temperature (T leaf) measured at two different growth temperatures (different symbols) to assess the responses to +3.5°C warming among rainforest groups.

Please note: Wiley Blackwell are not responsible for the content or functionality of any Supporting Information supplied by the authors. Any queries (other than missing material) should be directed to the New Phytologist Central Office.

Click here for additional data file.