Boreal conifers maintain carbon uptake with warming despite failure to track optimal temperatures

To improve predictions of terrestrial CO2 exchange under future warmer and CO2-enriched climates, terrestrial biosphere models (TBMs) must account for acclimation of photosynthesis to both temperature and elevated CO2. Existing temperature response functions largely come from seedlings grown under ambient CO2 in controlled environments, raising uncertainty about their applicability to mature, field-grown trees and to elevated CO2 conditions. Short-term temperature responses of photosynthesis are characterized by a thermal optimum (ToptA), which can shift upward with long-term warming (thermal acclimation), yet reported acclimation magnitudes vary widely among species and studies. It is unknown whether mature, field-grown boreal conifers can adjust ToptA sufficiently to match warming, particularly under elevated CO2 that suppresses photorespiration and can increase ToptA. This study aims to quantify the thermal acclimation of net photosynthesis and its biochemical drivers (Vcmax and Jmax) in mature tamarack (deciduous conifer) and black spruce (evergreen conifer) exposed to whole-ecosystem warming and elevated CO2, and to test whether shifts in ToptA keep pace with growth temperature and how elevated CO2 modifies these responses.

Prior work shows plants generally increase ToptA with warming, with reported acclimation ranging from ~0.16–0.78 °C per 1 °C warming. Boreal conifers show mixed results in seedling studies, with some species acclimating ToptA and others not. Most data informing TBMs come from ambient CO2-grown seedlings in growth chambers. Elevated CO2 typically stimulates photosynthesis by increasing CO2 substrate for Rubisco and suppressing photorespiration, often increasing ToptA when plants are grown and measured at elevated CO2; however, long-term CO2 stimulation can diminish due to biochemical acclimation and sink limitations. Only a few studies have examined how elevated CO2 affects thermal acclimation of photosynthesis and its biochemical parameters; one pot seedling study in boreal conifers reported little effect of elevated CO2 on the thermal acclimation of Vcmax and Jmax (optima and activation energies), while warming increased ToptA by ~0.36–0.65 °C per 1 °C regardless of CO2, and elevated CO2 shifted absolute ToptA upward by ~3.6–4 °C due to reduced photorespiration. Field studies on mature trees have been lacking, and a multi-year field warming study on broad-leaved seedlings found ToptA shifts smaller than increases in growth temperature.

Study site and design: The experiment was conducted at the SPRUCE (Spruce and Peatland Responses Under Changing Environments) site in the Marcell Experimental Forest, Minnesota, USA (47°30.476′ N; 93°27.162′ W). The ~50-year-old bog forest is dominated by Picea mariana (black spruce) with Larix laricina (tamarack). Ten large open-top enclosures (114.8 m² interior; 66.4 m² sampling area) imposed five whole-ecosystem temperature treatments in a regression design: +0 (ambient control), +2.25, +4.5, +6.75, +9 °C above ambient. Five enclosures were ambient CO2 (aCO2), five elevated CO2 (eCO2, ~+430 to +500 ppm above ambient). Warming began Aug 15, 2015; CO2 enrichment began Jun 15, 2016. Target temperatures and CO2 were largely achieved.

Plant material and measurements: Field campaigns occurred Jun 18–30 and Aug 15–30, 2017 (daytime means ~19 °C). Mature canopy trees (up to ~45 years) of black spruce and tamarack were sampled. For black spruce, 1-year needles (developed in 2016) were measured; for tamarack, fully expanded current-year foliage. Branchlets from sun-exposed canopy positions were cut pre-dawn, recut under water, and measured the same day in walk-in growth chambers to control leaf temperature. Gas exchange used seven LI-COR 6400XT systems with conifer chambers, at saturating light (1800 µmol m⁻² s⁻¹). A–Ci curves were measured at leaf temperatures 15, 25, 32.5, 40, and 45 °C and CO2 setpoints (400, 300, 200, 50, 400, 500, 600, 800, 1200, 1600, 2000 µmol mol⁻¹). To mitigate high-VPD effects at high temperatures, soda lime columns were moistened. Needle projected area was measured (ImageJ) and total leaf area corrections applied. In total, 96 A–Ci temperature-response datasets were collected.

Modeling and parameters: The Farquhar–von Caemmerer–Berry (FvCB) C3 model was fitted (plantecophys R package, bilinear method) to derive Vcmax and Jmax across temperatures, using standard temperature dependencies for Γ*, Kc, and Ko (Bernacchi et al.). Mesophyll conductance was not measured; thus, apparent Vcmax and Jmax were estimated on intercellular CO2 (Ci) basis. Temperature sensitivity parameters—thermal optima (ToptV, ToptJ) and activation energies (EaV, EaJ)—were derived using a modified Arrhenius function with deactivation term, fixing Ha at 200 kJ mol⁻¹ to avoid over-parameterization. Net photosynthesis at growth CO2 (ambient ~400 ppm; elevated ~800 ppm) was extracted from A–Ci curves and fitted with a quadratic temperature response A(T) = Aopt − b (T − ToptA)² to estimate ToptA, Aopt, and the breadth parameter b.

Growth temperature comparisons: Mean daytime air temperature (9:00–15:00) over the 10 days preceding each measurement was used to compute ΔMeanTg = Mean growth temperature − ToptA, assessing whether acclimation matched warming. Net photosynthesis at mean growth temperature (Ag) was estimated using the quadratic fit and the 10-day mean air temperature; seasonal Ag was also estimated for the 2016–2017 growing seasons.

Stomatal influence test: To evaluate stomatal effects on ToptA shifts, photosynthesis was recalculated at fixed Ci/Ca = 0.7 (A70) using parameterized Vcmax, Jmax, Rday, and TPU; Ac, Aj, and ATPU were computed and A70 taken as the minimum, then ToptA of A70 was estimated.

Statistics: Mixed-effects regression models (nlme in R) assessed effects of warming and CO2 on parameters, with fixed effects of warming and CO2 treatment and random intercept for month (June/August). Model selection followed AIC/AICc criteria. One-sided tests evaluated hypothesized increases. ANOVAs assessed treatment effects on ΔMeanTg. Plot means (n = 1–4 trees per plot) were analyzed. Additional unpaired t-tests were used where appropriate.

• Thermal optimum shifts: ToptA increased with warming but lagged behind temperature increases—by 0.26 °C per 1 °C in tamarack (p = 0.021) and 0.35 °C per 1 °C in black spruce (p = 0.0058). Elevated CO2 did not alter the slope of acclimation in either species. Absolute ToptA was ~3 °C higher in eCO2 than ambient in tamarack; no CO2 effect on ToptA in black spruce.
• Biochemical coordination: Warming-induced increases in ToptA correlated with increases in ToptV and ToptJ. Slopes per 1 °C warming were: ToptV 0.35 (tamarack) and 0.44 (black spruce); ToptJ 0.26 (tamarack) and 0.55 (black spruce). No evidence for acclimation of activation energy of Vcmax; in black spruce, EaJ declined nonlinearly with warming under eCO2 but not under ambient CO2.
• Growth temperature exceedance: Under ambient CO2, mean daytime growth temperature exceeded ToptA by >2 °C across all warming treatments for both species, indicating incomplete tracking of warming. eCO2 reduced ΔMeanTg in tamarack at +2.25 °C; effects in black spruce were weak or absent.
• Temperature response breadth (b): In tamarack, b was unaffected by warming but was 86% higher in eCO2 than ambient overall, indicating increased short-term temperature sensitivity. In black spruce, b showed no CO2 effect in controls, but marginally increased with warming under eCO2 (p = 0.067), reaching 68% higher than ambient at +9 °C.
• Photosynthetic rates: In tamarack, Aopt was constant across warming, with higher overall rates under eCO2. In black spruce, Aopt increased with warming under eCO2 (slope 0.54 µmol m⁻² s⁻¹ per °C; p = 0.029) but not under ambient (slope 0.10; p = 0.54). Ag (net photosynthesis at mean growth temperature) showed similar patterns: no significant change with warming in tamarack (both CO2 treatments); in black spruce, Ag commonly increased with warming under eCO2 (slope 0.53; p = 0.026) but not under ambient (slope 0.047; p = 0.76).
• Seasonal integration: Across the 2016–2017 growing seasons, modeled Ag was not negatively affected by warming in either species; tamarack remained constant, while black spruce often increased under eCO2 with warming. Overall, despite insufficient ToptA shifts to match warming, both species maintained or increased carbon uptake at prevailing growth temperatures, especially under elevated CO2 for black spruce.

The study demonstrates that mature boreal conifers thermally acclimate the temperature response of photosynthesis under whole-ecosystem warming, but the upward shift in ToptA (0.26–0.35 °C per 1 °C warming) does not fully track increases in air temperature. Consequently, prevailing daytime temperatures often exceed ToptA, suggesting potential vulnerability to heat stress under future climates. Nonetheless, net photosynthetic rates at prevailing growth temperatures were maintained (tamarack) or increased (black spruce under eCO2), indicating that acclimation of biochemical processes and changes in instantaneous temperature response breadth can buffer leaf-level carbon uptake in the near term when moisture is not limiting. The underlying mechanism of acclimation appears to be adjustments in the thermal optima of Vcmax and Jmax rather than changes in stomatal conductance or respiration, aligning with previous seedling and field studies. Elevated CO2 had limited impacts on the thermal sensitivity (optima and activation energies) of Vcmax and Jmax, corroborating seedling studies and suggesting TBMs can likely ignore interactive effects of elevated CO2 on these biochemical temperature response parameters. However, elevated CO2 can increase absolute ToptA in some species (tamarack) and increase the breadth parameter b (greater short-term thermal sensitivity), which TBMs should consider. While warming alone did not reduce leaf-level carbon uptake at this moist peatland site, future increases in heat extremes and drought could limit photosynthesis, and the lack of thermal acclimation in foliar respiration observed previously at this site may diminish net carbon sequestration despite maintained photosynthesis.

This field study on mature boreal conifers provides evidence that photosynthetic temperature responses acclimate to multi-year whole-ecosystem warming, but ToptA shifts are insufficient to keep pace with rising air temperatures. Despite this, carbon uptake at prevailing growth temperatures is maintained or enhanced (under elevated CO2 in black spruce), driven by coordinated biochemical adjustments. The findings inform TBMs by indicating that elevated CO2 need not alter the temperature response functions of Vcmax and Jmax, while species-specific effects of elevated CO2 on ToptA and the response breadth (b) should be incorporated. Future work should test generality across biomes, additional plant functional types, and under varying water availability, and should include direct measurements of mesophyll conductance to resolve species differences in CO2 effects on ToptA.

• Species and site specificity: Results are from two boreal conifers in a single peatland ecosystem with ample moisture; generalization to other species, functional types, and drier sites may be limited.
• Duration and treatment history: Measurements followed ~2 years of warming and ~1 year of elevated CO2; longer-term responses could differ as acclimation and nutrient limitations evolve.
• Measurement approach: Gas exchange was conducted on cut branchlets transported to growth chambers; although precautions were taken and prior work suggests minimal artifact for conifers, detachment could influence behavior.
• Assumptions in comparisons: Analyses assume close coupling of leaf and air temperatures in conifer needles when comparing growth temperatures to ToptA; deviations could affect ΔMeanTg estimates.
• Mesophyll conductance: Not measured; Vcmax and Jmax are apparent (Ci-based) and species differences in mesophyll conductance could underlie some CO2 effects, particularly the lack of ToptA shift in black spruce under eCO2.
• Sample size and statistical power: Plot-level means with n = 1–4 trees per plot and small overall sample sizes may limit detection of subtle effects (e.g., marginal p-values for b in black spruce).