To accurately predict and project climate changes, scientists rely on a connected series of models ranging from small-scale km-level models to comprehensive Earth System Models (ESMs) (Forster et al., 2023). These models need to be part of a cohesive framework that incorporates observations of the Earth system and data assimilation techniques. This ensures that the models are effective for various applications. Additionally, there is untapped potential in utilizing satellite data for these purposes, which should be further developed.

Methods and activities

With the aim of integrating and harmonizing the work carried out by modelling Working Groups and observational panels, ESMO will focus on activities that have model improvements at their core, for example working on the reduction of systematic errors, the use of reanalysis and data assimilation, the coordination of observational activities in the context of modelling, as well as links to external partners. Targeted workshops, community papers and task teams will be collaborative spaces to advance modelling and observations infrastructures.


Top level outcomes of the activities under this objective are to 1) identify, attribute, and reduce model systematic errors across systems and to 2) enhanced skill of S2S and I2D prediction and climate projections for decades to centuries. In detail, the envisioned outcomes include:

  • Recommendations on the use of model configurations for different applications. 
  • A traceable hierarchy of weather and climate models.
  • Early warning capacity for weather and climate extremes and associated risk.
  • Improved prediction skill and reduced model biases through application of tendency bias corrections derived from assimilation increments.
  • Quantify impacts of model deficiencies on analyses and forecasts.
  • Improvements in representation small scale processes, e.g., tropical convection and its organization. 
  • Best practices and a common understanding of observational uncertainties among the observation and modelling communities.  
  • Quantification of observational needs for modelling derived in a consistent way across all parts of WCRP at a global level.