1 | INTRODUCTION
Tropical forest species are believed to be particularly sensitive to global warming, as they are adapted to conditions of limited seasonal temperature variation. Furthermore, temperatures in the lowland tropics are already the highest known to support closed-canopy forest, and distances to cooler refugial areas can be large (Wright, Muller-Landau & Schipper 2009). Tropical species will therefore need to acclimate to changing conditions for tropical forests to continue to have a mitigating effect on anthropogenic climate change.
Despite millions of years of thermal stability in the tropics, tropical trees have the capacity to acclimate to warming. For example, experimental nighttime warming results in down-regulation of respiratory carbon loss from leaves (Cheesman & Winter 2013; Slot et al. 2014; Slot & Winter 2017a, 2018), consistent with thermal acclimation (Atkin & Tjoelker 2003); under elevated growth temperature the optimum temperature for photosynthetic carbon uptake shifts towards the new, higher growth temperature (Kotisup et al. 2009; Slot & Winter 2017a); and growth rates do not necessarily decrease under either nighttime warming or daytime warming (Cheesman & Winter 2013; Scafaro et al. 2017; Slot & Winter 2018). Thus, despite differences among species, the comparative thermal stability of the tropics has not deprived tropical species of the physiological plasticity that enables them to acclimate to moderate warming.
As plants are confronted with rising temperatures, atmospheric CO2 concentrations ([CO2]atm) are also increasing, and in the short term this has both direct and indirect effects on photosynthetic carbon uptake and its response to temperature. CO2 as the substrate for photosynthesis directly stimulates rates of carbon uptake. Higher [CO2]atm also suppresses carbon loss associated with photorespiration, the result of oxygenation of Rubisco (e.g., Long Ainsworth, Rogers & Ort. 2004; Ainsworth & Rogers 2007). Photorespiration increases with temperature, so when photorespiration is suppressed by elevated [CO2]atmplants can achieve a higher optimum temperature of photosynthesis than at ambient [CO2]atm (Berry & Björkman 1980; Long 1991). The decrease in Rubisco limitation of photosynthesis under elevated CO2 conditions means that net photosynthesis is increasingly limited by the maximum photosynthetic electron transport rate (JMax), reflecting the maximum rate of ribulose 1,5-bisphosphate (RuBP) regeneration (Sage & Kubien 2007). This affects the temperature response of net photosynthesis because electron transport limited-photosynthesis has a higher temperature optimum than Rubisco limited photosynthesis (Sage & Kubien 2007; Hikosaka, Ishikawa, Borjigidai, Muller & Onoda 2006). In the long term, elevated [CO2]atm may cause a reduction in both the maximum capacity of RuBP carboxylation (VCMax) and RuBP regeneration (JMax). VCMax might decrease as a result of lower investment in Rubisco when high [CO2]atm reduces carbon limitation of photosynthesis and optimization requires proportionally greater investment in electron transport (Ainsworth & Rogers 2007). VCMax and JMax might also decrease if leaf nitrogen (N) concentrations decrease (Medlyn et al. 1999) as a result of N dilution by rapid growth (Luo, Field & Mooney 1994), or as a result of progressive soil N limitation (Luo et al. 2004; Warren, Jensen, Medlyn, Norby & Tissue 2015). Changes in leaf N in plants grown at elevated CO2 could directly affect the thermal acclimation capacity of plants, as there appears to be an important role for N allocation to Rubisco—an N-rich enzyme—during acclimation to warming (Scafaro et al. 2017).
To assess thermal acclimation of photosynthesis, the short-term temperature response of the photosynthetic parameters is compared between warmed and control plants. Berry & Björkman (1980), summarizing previous research, showed that photosynthesis peaked at higher temperatures in plants acclimated to warmer conditions than in cool grown plants. A recent global meta-analysis showed that across sites, the optimum temperature for net photosynthesis (TOpt) scaled with growth temperature, and this pattern could be explained by thermal acclimation; adaptation—expressed as inherent differences based on source populations—was of lesser importance (Kumarathunge et al. 2019), stressing the importance of physiological plasticity. A shift in the optimum temperature (TOpt) towards higher values may or may not be accompanied by increases in POpt, the rate of photosynthesis at TOpt (Berry & Björkman 1980; Way & Yamori 2014; Slot, Garcia & Winter 2016). Vegetation models are highly sensitive to the formulation of temperature responses (Booth et al. 2012). To inform such models about acclimation, temperature response parameters of VCMax and JMax are required, their activation energies, de- activation energies, entropy factors (ΔS, sensu Medlyn et al. 2002), and their temperature optima (Stinziano, Way & Bauerle 2018; Mercado et al. 2018). Because of the inherent non-linearity of temperature responses and Jensen’s inequality, implementation of acclimation parameters would be most meaningful if the parameterization reflected the true diversity of these parameter values, rather than single averages.
Acclimation processes are currently not well represented in most dynamic global vegetation models and earth system models (Smith & Dukes, 2013; Smith, Malyshev, Shevliakova, Kattge & Dukes. 2016; Lombardozzi, Bonan, Smith, Dukes & Fisher. 2015; Mercado et al. 2018), and limited experimental data is available on the combined effects of warming and elevated [CO2] that can provide mechanistic foundations for modeling acclimation (Way, Oren & Kroner 2015), particularly for tropical plants. To address thermal acclimation of photosynthesis several modeling studies have capitalized on the clear trend of a decreasing JMax/VCMax ratio with acclimation to higher growth temperatures, as synthesized by Kattge & Knorr (2007), and more recently confirmed by Smith & Dukes (2018) (e.g., Lombardozzi et al. 2015; Smith et al. 2016; Mercado et al. 2018). However, [CO2]atm may affect the JMax/VCMax ratio independent of temperature. While acclimation to warming consistently decreases the JMax/VCMax ratio, increased [CO2]atm may either increase the ratio (e.g. meta-analysis by Ainsworth & Rogers 2007), or not affect it (e.g., meta-analysis by Medlyn et al. 1999). These differences may be related to nutrient supply and /or to source-sink relationships (Arp 1991; Sage 1994), as nutrient- or sink limitation may cause proportionally greater reduction in VCMax. Fauset et al. (2019) found that the JMax/VCMax ratio of the tropical tree species Alchornea glandulosa decreased with increasing growth temperature, but increased with elevated [CO2], such that JMax/VCMax for 800 ppm CO2 was higher at 35°C than that of control plants at 30°C. The same pattern of opposing effects of warming and elevated CO2 on JMax/VCMax was found in the boreal tree species Larix laricina , and to a lesser extent in Picea mariana (Dusenge et al. 2020). The utility of JMax/VCMax changes to model acclimation may thus be limited when both temperature and CO2increase, and additional information on photosynthetic parameters is needed.
The short-term temperature response of net photosynthesis can be controlled by different factors, including the temperature sensitivities of VCMax, JMax­, and respiration in the light, and by stomatal conductance (Lin, Medlyn & Ellsworth 2012). We have shown that the temperature response of net photosynthesis of field-grown lowland tropical trees is largely controlled by decreases in stomatal conductance as the leaf-to-air vapor pressure deficit (VPD) increases with increasing measurement temperature (Slot & Winter 2017b, 2017c; Hernández, Winter & Slot2020; see also Smith et al. 2020), whereas Vårhammer et al. (2015) reported significant limitations by JMax in tropical montane species in Rwanda. Growth at elevated CO2 generally results in decreases in stomatal conductance (Saxe, Ellsworth, & Heath 1998; Ainsworth & Rogers 2007), potentially increasing stomatal control over net photosynthesis. However, upregulation of stomatal conductance at a given VPD during acclimation to elevated temperature and VPD has also been observed in some species (Marchin, Broadhead, Bostic, Dunn & Hoffmann 2016; Wu et al. 2018; Dusenge, Madhavji & Way 2020). Knowing which process limits photosynthetic carbon fixation is important to inform vegetation models and to better predict how environmental change will impact photosynthetic carbon uptake of tropical forest trees.
Here we report on an experiment with neotropical tree speciesTabebuia rosea , grown under combined warming and elevated CO2 conditions. We evaluated the capacity for acclimation of photosynthesis and respiration. Specifically, we examine leaves developed under treatment conditions (“long-term acclimation”), and of pre-existing leaves of plants transferred from control to treatment conditions and vice versa (“short-term acclimation”). We hypothesized that (i) acclimation would result in higher temperature optima for photosynthetic parameters, and in lower respiration rates at a set temperature; (ii) long-term acclimation (of newly-developed leaves) would be stronger than short-term acclimation (of pre-existing leaves); and (iii) stomatal limitation of photosynthesis would increase under treatment conditions of warming and elevated CO2.  Nonlinear models characterizing the temperature responses for each treatment were fitted using a probabilistic Bayesian approach. This approach enabled us to present the parameters of interest in the form of a probability distribution of values, to better reflect the range of potential parameter values.