Skip to main content
U.S. flag

An official website of the United States government

Toward Efficient Coupled Earth System Model Initialization for E3SM Version 3

Presentation Date
Thursday, August 8, 2024 at 1:05pm - Thursday, August 8, 2024 at 1:11pm
Authors

Lead Presenter

Co-Author

Abstract

A well-initialized and equilibrated state for coupled atmosphere-land-ocean components is essential to skillfully predict and study the influence of anthropogenic forcing with Coupled Earth System Models (ESMs) including DOE’s E3SM. However, spinning up such a coupled system to equilibration is scientifically and computationally challenging. For the standard configuration of E3SM Version 3 (E3SMv3), it took over 2000 years of spin-up to reach a quasi-equilibrium state. This becomes a major bottleneck for high-resolution models for which we cannot afford thousands of simulation years. Motivated by previous studies, this work explores the potential application of an alternating B-/G-Cases approach to improve the efficiency of coupled model initialization in E3SMv3, using its nominal low-resolution configuration for demonstration purposes. The experimental spin-up procedure alternates between a computationally expensive coupled configuration (B-Case) and a computationally cheaper configuration (G-Case) where a data model replaces the active atmospheric component. Compared to the traditional approach with long-term simulations through continuous B-Case, the alternating B-/G-Cases approach acts in a computationally tractable manner with the capability of achieving faster and computationally cheaper spin-up of coupled ESMs by optimizing the length of alternating simulations for B- and G-Cases. Results from our preliminary simulations have indicated that, when set up properly, the approach with alternating B-/G-Cases can reasonably track the evolution of atmospheric, land, and ocean model states predicted by the continuous B-Case simulation. We will report the details of applying the proposed alternating approach to low-resolution coupled E3SM model spin-up simulation for pre-industrial climate, as well as demonstrate the applicability and value of the method to support the high-resolution E3SMv3 model development.

Presentation File(s)
Category
Methods in Model Integration, Hierarchical Modeling, Model Complexity
Energy, Water, and Land System Transition
Strengthening EESM Integrated Modeling Framework – Towards a Digital Earth
Funding Program Area(s)
Additional Resources:
ALCC (ASCR Leadership Computing Challenge)
NERSC (National Energy Research Scientific Computing Center)