The significance of plant hydraulic parameters for modeling carbon and water fluxes across European climate zones and PFTs with CLM5

Plants are increasingly exposed to water stress under climate change, posing significant challenges for accurate simulation of carbon and water fluxes in terrestrial ecosystems. Most land surface models simulate the regulation of water and carbon fluxes in response to soil moisture stress through empirical soil hydraulic schemes. However, these schemes often introduce significant uncertainties in water and carbon simulations. To address this, the Community Land Model version 5 (CLM5) introduced a plant hydraulic stress routine that explicitly models water transport through vegetation via a hydraulic framework, improving the representation of vegetation water potential, root water uptake, and plant water stress¹. However, including plant hydraulics introduces additional parameters that are difficult to constrain due to limited field data and high variability. Understanding the influence of these plant hydraulic parameters on water and carbon flux modeling is crucial for model improvement and prediction accuracy.

In this study, we used a parameter perturbation approach to investigate the role of plant hydraulic parameters at 14 experimental sites in Europe, representing diverse plant functional types (PFTs) and climate zones. Using CLM5, we performed 128 ensemble simulations per site, systematically varying three key hydraulic parameters: plant- and root-segment maximum conductance (k_max and k_rmax) and water potential at 50% loss of segment conductance (psi₅₀). The perturbation ranges were informed by previous parameter perturbation experiments^2,3. We evaluated: (i) how the model represented plant hydraulic dynamics (i.e., vegetation water status and plant-segment conductances), (ii) the sensitivity of carbon and water fluxes—gross primary production (GPP) and evapotranspiration (ET)—to parameter variation, and (iii) model performance compared to in-situ observations.

The results showed that the model successfully captured seasonal variations in plant-segment conductance and vegetation water potential, which were reflected in the seasonal dynamics of GPP and ET. However, at drought-prone sites, the model overestimated ET reductions during summer compared to observations, due to a steep decline in root-segment conductance and stomatal closure. This highlights the need for improved parameterization of psi₅₀ and k_rmax to better represent plant responses to extreme drought. In addition, ensemble simulations revealed substantial sensitivity of GPP and ET to parameter perturbations, with variations up to 50% in GPP and 30% in ET depending on PFT and climate zone. These results underscore the importance of considering the variability in plant hydraulic properties, particularly k_max and k_rmax, which span several orders of magnitude.

To address these uncertainties, the next steps of this work will focus on refining the parameterization by integrating data on plant hydraulic traits from existing databases^4,5. This approach will help constrain parameter ranges across ecosystems and climate zones, particularly for drought-prone sites. Improving the representation of plant hydraulic traits will enhance predictions of ecosystem responses to water stress and the reliability of land surface models under current and future climate scenarios.

References

• ¹Kennedy et al. (2019). 10.1029/2018MS001500
• ²Kennedy et al. (2024). 10.22541/essoar.172745082.24089296/v1
• ³Eloundou et al. (2024). 10.5194/egusphere-egu24-16086
• ⁴Kattge et al. (2020). 10.22541/10.1111/gcb.14904
• ⁵Baca Cabrera et al. (2024). 10.1002/pld3.582