Implementation of stable carbon isotopes into JULES model: A novel approach for evaluating the coupled carbon and water cycles as represented in UKESM

Abstract:

Stable carbon isotopic composition (δ¹³C) of terrestrial plant materials can offer valuable insights into plant stomatal and photosynthetic adjustments to changes in environmental conditions. However, the potential of using δ¹³C data to investigate plant responses to environmental changes and to evaluate land surface models has not been fully recognized. Despite recommendations from the Coupled Model Intercomparison Project Phase 6 (CMIP6), only one model in CMIP6 (i.e. CESM2 using CLM5 vegetation model) explicitly predicts stocks of carbon isotopes on land and ocean reservoirs and their exchange fluxes with the atmosphere.

Here we present a new carbon isotope modelling capability in JULES enabling novel evaluation of the coupled carbon and water cycles as represented in UKESM. Our aims are threefold: (1) highlight the impact of model assumptions about the environmental dependencies of stomatal functions on the simulated δ¹³C and related water-use efficiency (WUE), (2) provide recommendations on the model formulations to use for future studies and (3) investigate temporal and spatial variations of δ¹³C and WUE across the globe. We test four different submodels of leaf-gas exchange in JULES (i.e. Jacobs, Leuning, Medlyn and Prentice) by comparing modelled δ¹³C against a large global network of δ¹³C data from leaves and tree rings of woody C₃ plants over 1979-2016. We also examine and evaluate model simulations over 1980-2010 at different AmeriFlux eddy-covariance flux stations where both leaf and tree-ring isotopic data are available. We finally analyse mean values and long-term trends in WUE predicted by the well-validated version of JULES. We found that the optimal Prentice model has the strongest predictive skills and that incorporating soil water stress on the stomatal limitation of photosynthesis improve the predictions of the interannual variations and trends of δ¹³C. The new configuration of the model resulted in more consistent predicted changes in iWUE with the eddy-covariance measurements compared to the original model, although other model improvements are still required. Overall, our results have strong implications for the prediction of the magnitude of future changes in WUE of plants and ecosystems that are expected under a warmer and drier climate.