What Is Earth System Modeling and What Does It Actually Do?

Earth system modeling is the practice of representing the atmosphere, ocean, land surface, ice, and living processes as interacting components of one coupled system, then running that system forward on a computer to test how it responds to forces like rising greenhouse gases. It is the tool behind most long-range climate projections and carbon-cycle research. It is not the same as weather forecasting: weather models aim at the next days, while Earth system models aim at decades to millennia and must track how carbon, water, and energy move between components, not just where the wind blows.

The components and how they are coupled

An Earth system model (ESM) is usually built from separate component models that exchange information at set intervals.

"Coupled" means each component reads the others' outputs and writes back its own. A change in ocean temperature alters evaporation, which alters clouds, which alters incoming sunlight, which alters ocean temperature again. That loop is why the models are run as one system rather than as separate pieces.

Model types you will encounter

• General Circulation Models (GCMs): the atmospheric and oceanic core; historically climate-focused without full interactive chemistry or carbon.
• Earth System Models (ESMs): GCMs extended with interactive biogeochemistry, so CO₂ and other greenhouse gases can respond to the climate rather than being fixed inputs.
• Earth system Models of Intermediate Complexity (EMICs): reduced-resolution or simplified versions used for very long simulations, large ensembles, or paleoclimate work where full ESMs would be too costly.

What the models are used for

• Climate projection: running scenarios of future emissions to estimate temperature, precipitation, and sea-level change.
• Carbon cycle research: tracking how much emitted CO₂ stays in the atmosphere versus being taken up by land and ocean.
• Paleoclimate: simulating past climates to test whether the model reproduces known conditions.
• Scenario experiments: comparing outcomes under different policy or emission pathways.

Inputs, outputs, and honest limits

Inputs are the external forcings — greenhouse gas concentrations or emissions, solar variability, volcanic aerosols, and land-use change — plus the model's initial state. Outputs are gridded fields over time: temperature, precipitation, wind, ocean circulation, ice cover, and carbon fluxes.

The main limitations are structural, not incidental:

• Resolution: grid cells are typically tens to hundreds of kilometers wide, so processes smaller than a cell cannot be simulated directly.
• Parameterization: clouds, turbulence, and small-scale mixing must be approximated by formulas, and different approximations give different results.
• Computational cost: a full ESM run at high resolution can consume enormous computing time, which is why EMICs and ensembles exist.
• Uncertainty: because several parameterizations are plausible, models are usually run as ensembles and results are reported as ranges, not single numbers.

How this connects to studying environmental science

Earth system modeling sits inside atmospheric science, oceanography, and environmental science programs rather than standing alone. At a department like Rutgers' Department of Environmental Sciences, the relevant groups and program areas include Atmospheric Science, Physical Oceanography, Ocean Modeling, Atmospheric Chemistry, and Aerosol Mechanics — the same specialties that supply the component models described above. Students typically enter through graduate programs in Environmental Sciences or Atmospheric Science, or undergraduate programs in Environmental Sciences, Environmental Engineering, Meteorology, and Bioenvironmental Engineering, then specialize in a component or in coupling. If your goal is to work with ESMs, the practical preparation is a foundation in physics, mathematics, and programming, followed by coursework in one component discipline deep enough to understand its parameterizations.