Principal Investigators

At a Glance

The development of Earth System Models (ESMs) at major climate science centers around the world is intended to improve our capability to project climate changes caused by anthropogenic greenhouse gas emissions (GHGs). These models include interactions between the atmosphere, ocean, sea ice, and land. In addition, ESMs allow us to project how changes in the carbon uptake by the biosphere and terrestrial and marine sources may affect atmospheric concentrations of carbon dioxide (CO2) and other GHGs. The overall goal is to better constrain mitigation pathways that would stabilize the climate.


Research Highlight

The initial phase of evaluation of a new Earth System Model (ESM) usually occurs in pre-industrial (PI) configurations. This refers to climate conditions corresponding to the 1850s, before the onset of the Industrial Revolution and the rapid increase in anthropogenic GHG emissions. It usually takes several hundreds to thousands of simulated years in PI configurations to evaluate and revise an ESM code to establish model fidelity and produce a stable, non-drifting physical climate and carbon cycle. In the last Coupled Climate Model Intercomparison project, phase 6 (CMIP6), the stability of the simulated carbon cycle required a long-term CO2 flux between land, ocean, and atmosphere in the CO2 emission-driven PI simulations to be less than 0.1 Petagram of carbon per year (Pg C/year). GFDL conducted dozens of experiments, including an historical 1850 to 2014 experiment. These were completed in coordination with other international and national centers participating in CMIP6. GFDL also shared Petabytes of data with the broader scientific community. It is important to understand that GFDL does not tune its historical simulations to observed historical records.

Last year, Shevliakova and her colleagues at NOAA/GFDL and CMI scientists concluded an in-depth evaluation of the historical trends in the land surface climate from 1850 to 2014 simulated by the NOAA/GFDL ESM4.1 model. The evaluation focused on the fidelity of its land component, Land Model version 4.1 (LM4.1). The team also explored ways to improve hydrological, ecological, and biogeochemical processes for the next generation of GFDL ESMs. Features of LM4.1 include advanced vegetation dynamics, plant hydraulics, multi-layer canopy energy and moisture exchanges, daily fire, land use representation, and dynamic atmospheric dust coupling. These features vastly improved mean climate and variability compared to previous generations of the GFDL models (Figure 12.1). Shevliakova et al. (2024) documented the results of the analysis of land surface climate in a recently accepted manuscript.

Figure 12.1.
Comparison of the reconstructed Gross Primary Production (GPP, the measure of the gross carbon uptake by vegetation) and the simulations of GPP by the two generations of the GFDL Earth System models, ESM2G (used in CMIP5) and ESM4.1 (used in CMIP6) (Shevliakova et al., 2024).

In preparation for the CMIP7 and to enable comprehensive treatment of hydrological and biogeochemical exchanges, Sergey Malyshev, in collaboration with GFDL scientists and AOS postdoctoral fellows, led several improvements of the LM4.1 representation of land-atmosphere exchanges for water, energy, and chemical tracers. These included turbulent exchanges of moisture and energy between the ground and the canopy air. It also included a parameterization enabling a more faithful representation of the interaction between the planetary surface and the atmosphere.

In addition, the AOS research scholar and CMI contributor Enrico Zorzetto, in collaboration with GFDL scientists Malyshev, Ginoux, and Shevliakova, implemented a new Global Land-Atmosphere Snow Scheme (GLASS). This model accounts for the effects of snow aging and impurities such as dust and carbon deposits from fire on snow albedo. The team also initiated two new projects focused on improving the spatial distribution of fires in the orography-aware (meaning taking topogrophy into consideration) version of LM4.1 and linking damages and tree mortality to drought conditions. This will be used to explore linkages between fire, hydrological cycles and vegetation dynamics. The ongoing coordinated improvements to hydrological and carbon cycling treatments in LM4.1 will reduce uncertainty in the prediction and projections of land carbon stocks, including their vulnerability to changing climate and extremes. Significant changes in land carbon stocks would impact the world’s remaining budget for fossil fuel emissions.



Shevliakova, E., et al., 2024. The land component LM4.1 of the GFDL Earth System Model ESM4.1: Model description and characteristics of land surface climate and carbon cycling in the historical simulation. Journal of Advances in Modeling Earth Systems. In press.