Zhao et al. (2025) Eco-evolutionary modelling of global vegetation dynamics and the impact of CO ₂ during the late Quaternary: insights from contrasting periods

Identification

• Journal: Earth System Dynamics
• Year: 2025
• Date: 2025-10-09
• Authors: Jierong Zhao, Boya Zhou, Sandy P. Harrison, I. Colin Prentice
• DOI: 10.5194/esd-16-1655-2025

Research Groups

• Department of Geography and Environmental Science, University of Reading, UK
• Department of Life Sciences, Georgina Mace Centre for the Living Planet, Imperial College London, UK

Short Summary

This study uses an eco-evolutionary optimality (EEO)-based modeling approach to quantify the relative impacts of climate fluctuations and CO2 levels on global vegetation dynamics and gross primary production (GPP) during the Last Glacial Maximum (LGM) and mid-Holocene (MH) compared to pre-industrial (PI) conditions. It reveals that low CO2 was nearly as important as climate in reducing GPP and increasing C4 plant abundance at the LGM, and it significantly offset positive climate impacts on GPP during the MH.

Objective

• To examine the impacts of climate fluctuations and CO2-induced alterations on gross primary production (GPP) and C3/C4 plant competition during two contrasting periods: the Last Glacial Maximum (LGM; 21,000 years before present) and the mid-Holocene (MH; 6,000 years before present), compared to pre-industrial (PI) conditions.
• To evaluate the unique and synergistic effects of climate, atmospheric CO2, and solar radiation (photosynthetic photon flux density, PPFD) on GPP changes using factorial experiments.

Study Configuration

• Spatial Scale: Global, with regional analysis for tropics (25° N–25° S), northern extra-tropics (> 25° N), and southern extra-tropics (> 25° S). Model outputs were down-scaled to a resolution of 0.5° longitude/latitude.
• Temporal Scale: Late Quaternary, specifically Last Glacial Maximum (LGM; 21,000 years before present), mid-Holocene (MH; 6,000 years before present), and pre-industrial (PI) conditions.

Methodology and Data

• Models used:
  • Eco-evolutionary optimality (EEO)-based modelling approach.
  • P model: A light-use efficiency model simulating GPP, based on the Farquhar–von Caemmerer–Berry photosynthesis model with "coordination" and "least-cost" EEO hypotheses.
  • C3/C4 competition model: A simple process-based scheme derived from the P model, validated against worldwide soil carbon stable isotope data.
  • LAI model: An EEO-based model predicting the seasonal cycle of leaf area index (LAI), incorporating a seasonal maximum LAI model with a water–carbon trade-off.
  • SPLASH model (Simple Process-Led Algorithms for Simulating Habitats): A generic soil water accounting model.
• Data sources:
  • Climate simulations from the Max Planck Institute Earth System Model (MPI-ESM1.2-LR, low-resolution version, T63 spectral resolution) from the Palaeoclimate Modelling Intercomparison Project (PMIP4-CMIP6) for LGM, MH, and PI.
  • Input variables: air temperature (degrees Celsius), vapour pressure deficit (pascals), air pressure (pascals), incident photosynthetic photon flux density (micromoles per square meter per second), and ambient CO2 concentration (parts per million).
  • Validation data: Eddy covariance flux tower data, satellite-derived LAI, worldwide soil carbon stable isotope data, and independent atmospheric estimates of carbon isotopes.
  • Factorial experiments using the Stein–Alpert decomposition method to isolate the effects of climate, CO2, and PPFD.

Main Results

• Global GPP at the LGM was 83.9 PgC/year, a substantial reduction compared to the simulated pre-industrial (PI) value of 109.6 PgC/year. The Northern Hemisphere extra-tropics experienced over a 50% reduction in GPP.
• Global GPP in the MH was 110.3 PgC/year, similar to PI and greater than LGM. The northern extra-tropics showed a 4% increase in GPP compared to PI in the MH.
• C4 plants constituted 40% of the vegetation fraction at the LGM (compared to 23% in PI and 25% in MH) and were responsible for 56% of total GPP at the LGM (compared to 21% in PI and 25% in MH).
• Factorial experiments for LGM (relative to PI):
  • Climate change had a negative effect on GPP (-14.8 PgC/year).
  • CO2 reduction (from 284.3 ppm to 190 ppm) had a negative effect on GPP (-12.2 PgC/year), nearly as important as climate.
  • Light (PPFD) had a small positive effect.
  • The two-way synergy between climate and CO2 was positive, indicating a less-than-additive combined negative impact.
• Factorial experiments for MH (relative to PI):
  • Climate changes had a positive effect on GPP.
  • PPFD changes had a positive effect on plant growth.
  • CO2 reduction (from 284.3 ppm to 264.4 ppm) resulted in a larger overall reduction in GPP than the enhancements due to climate or PPFD changes, effectively offsetting positive climate impacts.

Contributions

• First application of a globally uniform, eco-evolutionary optimality (EEO)-based approach (independent of plant functional types, PFTs) to simulate LGM and recent vegetation function, significantly reducing uncertainties associated with PFT parameterization.
• Provides robust quantitative estimates of global GPP changes and the relative contributions of C3 and C4 plants during glacial and interglacial periods, consistent with palaeovegetation data.
• Quantifies the individual and synergistic importance of climate, CO2, and light availability on GPP changes, highlighting the critical role of CO2 levels in shaping past vegetation dynamics.
• Demonstrates that the direct physiological impacts of CO2 on plant growth were comparably important to climate in reducing GPP at the LGM and were sufficient to annul the positive impacts of climate on GPP during the MH.
• Utilizes independently validated, parameter-sparse EEO models, suggesting higher realism and reduced uncertainty in GPP estimates compared to previous studies.

Funding

• Natural Environment Research Council (NERC) (grant no. NE/W00075X/1) for the project "When and Why does it Rain in the Desert: Utilising unique stalagmite and dust records on the northern edge of the Sahara".
• Schmidt Sciences, LLC (grant no. 355) for the Land Ecosystem Models based On New Theory, obseRvations and ExperimEnts (LEMONTREE) project.
• European Research Council (EU HORIZON EUROPE) (grant no. 787203) for the project Re-inventing Ecosystem And Land-surface Models (REALM).

Citation

@article{Zhao2025Ecoevolutionary,
  author = {Zhao, Jierong and Zhou, Boya and Harrison, Sandy P. and Prentice, I. Colin},
  title = {Eco-evolutionary modelling of global vegetation dynamics and the impact of CO <sub>2</sub> during the late Quaternary: insights from contrasting periods},
  journal = {Earth System Dynamics},
  year = {2025},
  doi = {10.5194/esd-16-1655-2025},
  url = {https://doi.org/10.5194/esd-16-1655-2025}
}