Parameter sensitivity in photosynthetic temperature response functions in terrestrial biosphere models: Evaluating the effects of fixed vs. estimated deactivation energies

Abstract

Photosynthesis is a central process in the terrestrial carbon cycle and a key determinant of predicting plant responses to climate change. Terrestrial biosphere models (TBMs) commonly simulate the temperature response of C₃ photosynthesis using the peaked Arrhenius function, which relies on parameters including Vcmax (maximum rate of Rubisco activity), Jmax (potential electron transport rate), K25 (process rate at 25 °C), Ea (activation energy), ΔS (entropy), and Hd (deactivation energy). To minimize overfitting, many models assume a fixed Hd value, though the consequences of this assumption remain underexplored. We analysed a global dataset of plant photosynthetic temperature responses to evaluate differences in parameter estimates and photosynthesis predictions between models with fixed and estimated Hd. Maximum rate of Rubisco activity at 25°C (K25) and Ea were generally higher with fixed Hd models, with ΔS showing higher or similar values and strong correlations. Potential electron transport rate at 25°C (K25) and Ea were also higher with fixed Hd, though Ea varied more, while ΔS showed higher values with weaker alignment, indicating greater inconsistency, suggesting Jmax is less sensitive to model structure compared to Vcmax, which exhibits consistent trends across parameters with fixed Hd assumptions. Percentage decline in modelled photosynthesis between fixed and estimated Hd assumptions was evaluated at optimum temperature (Topt) and ±5 °C from the optimum. Models with fixed Hd consistently estimated higher Topt values and lower declines, suggesting that fixed Hd models may overestimate thermal tolerance. These results highlight that while fixed Hd assumptions may be acceptable for estimating K25, they risk misrepresenting temperature responses of Ea, ΔS, and photosynthetic performance under warming. Incorporating variable Hd improves model flexibility and enhances the realism of TBM projections under climate change.