Temperature-assimilation curves
To estimate the carbon assimilation parameters used in this study (i.e. Popt, Topt Tmax, and Ω), we first repeatedly measured leaf temperature and net carbon assimilation for each of our study species using a LI-6800 portable photosynthesis system (LICOR, Lincoln, NE, USA). More specifically, we randomly visited each focal plant during sunny days between June 21 and September 1, 2018, and measured carbon assimilation over a range of leaf temperatures within the canopy of each individual following the general methods of Slot and Winter (2017a). Leaf temperature was first measured on a set of randomly selected leaves within the canopy of each individual with a MT6 MiniTemp infrared thermometer (Raytek, Wilmington, NC USA). For each leaf, the LI-6800 cuvette was set to the observed temperature, and the leaf was allowed to acclimate to chamber conditions before its net assimilation was measured. During all measurements, the LI-6800 leaf chamber was maintained at saturating light levels (1000 μmol quanta m-2 s-1).
The CO2 concentration was maintained at either 400 or 405ppm in the reference chamber (differences due to operator error). Varying CO2 reference chamber concentrations can potentially bias estimates of our carbon assimilation parameters. To correct for this, we conducted a separate set of measurements for 17 of our 21 species in which we varied the reference chamber CO2 over a range of concentration to measure the effect that this could have on CO2 assimilation in the sample chamber. We modeled this effect using the ‘smooth.spline’ function in base R’s ‘stats’ package (Core 2020) to calculate the difference in leaf assimilation rates between sample chamber CO2concentrations of 400 and 405 ppm. This difference was then added to assimilation measurements taken at 400ppm to correct for any potential bias in our results. Even prior to correction, CO2concentrations within the sample chamber were uniformly distributed with a mean CO2 concentration of 386ppm and a standard deviation of 8 ppm – a level of variation that is only slightly greater than those observed in leaf chambers of other studies (sd = 6ppm; Slot & Winter 2017a), and is unlikely to have affected our results.
The sample chamber’s relative humidity was set to 50% during sampling, but was automatically varied as needed to prevent moisture condensation within the LI-6800. In order to avoid sensor drift, the LI-6800 reference and sample chambers’ infrared gas analyzers were matched any time the sample chamber’s leaf temperature was changed by ≥5˚C since the previous match. We visually assessed stabilization of leaf temperatures, assimilation rates, and stomatal conductances before recording carbon assimilation rate (μ mol m-2 s-1).
Assimilation was modeled as a function of temperature following the model presented in June, Evans & Farquhar (2004) and adapted by Slot & Winter (2017a):
\(P\left(T\right)=P_{\text{opt}}\times\ e^{-\left(\frac{T_{\text{leaf}}-T_{\text{opt}}}{\Omega}\right)^{2}}\)Eq. 1
where Tleaf is leaf temperature and Ω is defined as the difference between the temperatures above and below Toptat which assimilation (P ) is reduced by ~37% from Popt (Fig. 1a) .
We estimated Tmax and additional values of Topt and Popt following the model from Cunningham S. C. & Read J. (2003), which provides better fits for asymmetrical temperature-assimilation curves:
\(P=\left\{b(T_{\text{leaf}}-T_{\min})\times\left[1-e^{c(T_{\text{leaf}}-T_{max)}}\right]\right\}^{2}\)Eq. 2
where b and c are constants, P is the assimilation rate, Tmin is the theoretical low-temperature compensation point and Tmax is the theoretical high-temperature compensation point (Fig. 1a) .
The Popt, Topt and Ω parameters of Eq. 1, and the b , c , Tmin andTmax parameters from Eq. 2 were estimated based on the fits of logistic non-linear least squares (nls) functions in R’s base ‘stats’ package (Core 2020). We bootstrapped the parameter estimates for each model and species by randomly resampling our leaf temperature and assimilation dataset 1000 times with replacement. We present the bootstrapped means for Popt, Topt and Ω from Eq 1, and Tmax, Popt, and Topt from Eq. 2.