Earth System Model Evaluation and Diagnostics

Earth system models (ESMs) have evolved into comprehensive simulators of the coupled atmosphere, ocean, land surface and cryosphere, integrating physical, chemical and biological processes. Rigorous evaluation and diagnostic analyses are vital for assessing model fidelity, quantifying uncertainty and guiding development. Model evaluation involves systematic comparison of simulated fields—such as temperature, precipitation, circulation patterns and carbon fluxes—to observational and reanalysis datasets using standardised performance metrics that quantify biases, variability and trend reproduction. Diagnostics delve deeper into process-level behaviour, examining phenomena such as monsoon dynamics, ocean heat uptake, cloud feedbacks and carbon-cycle interactions. Community-developed frameworks and interoperable data conventions have enabled automated, near-operational evaluation workflows embedded within global data infrastructures, supporting rapid monitoring of multi-model ensembles. By identifying common biases, attributing errors to parameterisations and benchmarking model spread, this diagnostic culture accelerates model improvement and underpins robust climate projections for research, adaptation planning and policy.

Research from Nature Portfolio

No recent Nature Portfolio content available.

Research from all publishers

Recent work has argued for a clear separation between the operational service demands of model intercomparison projects and their scientific research objectives. Proposals highlight transitioning intercomparison efforts into sustained climate-information services while refocusing scientific agendas on methodological advances and next-generation model evaluation.

An extended community evaluation toolkit now offers automated preprocessing, a wide suite of large-scale diagnostics and performance metrics for atmospheric, oceanic and terrestrial variables. Seamless integration with a distributed data archive enables quasi-operational evaluation of model output shortly after publication, reducing latency in model benchmarking and facilitating rapid dissemination of diagnostic results.

A foundational overview of the latest intercomparison phase outlines a federated experimental design that includes control, historical and process-oriented experiments. Standardised data formats, metadata conventions and coordinated diagnostic tiers support consistent evaluation across diverse models, addressing key questions of climate response to forcing, systematic bias origins and uncertainty quantification.

Earth System Model Evaluation and Diagnostics publication trend

The graph below shows the total number of articles in earth system model evaluation and diagnostics across all publications each year (not limited to Nature Index journals).

Technical terms

Earth system model (ESM): A numerical representation of the coupled atmosphere, ocean, land and cryosphere used to simulate past, present and future climate.

Coupled Model Intercomparison Project (CMIP): A coordinated international initiative organising standard experiments and data protocols to compare Earth system models.

Diagnostic: A targeted analysis probing specific processes or phenomena within model simulations to evaluate their physical realism.

Performance metric: A quantitative measure used to assess model fidelity in reproducing observed climate statistics, variability and trends.

ESMValTool: A community software suite providing automated preprocessing, diagnostics and performance metrics for Earth system model evaluation.

Earth System Grid Federation (ESGF): The distributed data infrastructure facilitating storage, access and routine evaluation of model simulations and observational datasets.

References

1. A Perspective on the Future of CMIP. AGU Advances (2024).
2. Overview of the Coupled Model Intercomparison Project Phase 6 (CMIP6) experimental design and organization. Geoscientific Model Development (2016).
3. ESMValTool (v1.0) – a community diagnostic and performance metrics tool for routine evaluation of Earth system models in CMIP. Geoscientific Model Development (2016).
4. Earth System Model Evaluation Tool (ESMValTool) v2.0 – an extended set of large-scale diagnostics for quasi-operational and comprehensive evaluation of Earth system models in CMIP. Geoscientific Model Development (2020).
5. Requirements for a global data infrastructure in support of CMIP6. Geoscientific Model Development (2018).