Development of Earth System model of intermediate complexity
Ocean Sciences, Climate
Research area
Climate modeling is crucial for understanding and predicting the Earth’s climate system. While state-of-the-art Atmosphere-Ocean General Circulation Models (AOGCMs) simulate the complexity of climate processes, their high computational cost limits their use in long-term simulations. Models of Intermediate Complexity (EMICs) offer a balance by simplifying physical processes while retaining key mechanisms, enabling efficient exploration of low-frequency climate variability.
This project develops a simplified EMIC using the Modular Ocean Model (MOM) for the ocean-sea ice system and the SPEEDY model for the atmosphere. The model reduces computational demand while maintaining essential dynamics. We will evolve this AOGCM into an EMIC with sub-models for ocean biogeochemistry, land-vegetation, atmospheric chemistry, and ice sheets. The project leverages high-performance computing to study long-term climate interactions, providing insights into Earth system dynamics.
Project goals
The project concerns development and use of a coupled climate model of intermediate complexity in two stages. First, the physical model will be built, consisting of an atmosphere-ocean-sea ice general circulation model (AOGCM). The AOGCM will consist of the ocean-sea ice model MOM and the atmospheric model SPEEDY.
The model will use the three-dimensional primitive equations for both ocean and atmosphere but will be simplified from a 'state-of-the-art' coupled climate model in terms of atmospheric physics and parameterization schemes.
These simplifications will provide considerable savings in computational expense and allow mechanisms of low-frequency climate variability to be investigated more efficiently than with a full-blown model. Building on the AOGCM, an Earth system Model of Intermediate Complexity (EMIC) will be developed in a second stage, with sub-models for different components of the climate system (e.g., ocean biogeochemistry land-vegetation, atmospheric chemistry, ice sheet model etc.)
Computational approach
This project pushes the boundaries of climate modeling by developing and integrating advanced models to simulate long-term climate variability and Earth system interactions. At its core, we adopt the ACCESS-OM2 climate model, developed by the Consortium for Ocean-Sea Ice Modelling in Australia (COSIMA) and widely used as the ocean-sea ice component for the Ocean Model Intercomparison Project (OMIP) within CMIP. The first computational challenge is adapting the AOGCM to the Leonardo supercomputer. This involves porting the project from the PBS scheduler, for which it was initially designed, to the SLURM scheduler, now the standard in high-performance computing (HPC). The next step is increasing flexibility in the choice of compilers and libraries, followed by replacing the current atmospheric component with SPEEDY, an intermediate-complexity model that preserves essential dynamics while reducing computational cost. Balancing model complexity with computational efficiency is crucial. Despite simplifications, running the coupled AOGCM on long timescales relevant to low-frequency variability requires extensive computational resources. Efficient parallelization and optimization will ensure simulations remain feasible. By leveraging CINECA’s HPC infrastructure, this project will overcome computational limits and enable high-resolution simulations and long-term climate predictions.
Image
Block diagram of the Earth system Model of Intermediate Complexity (EMIC)
Natalia Tilinina
Istituto Nazionale di Oceanografia e di Geofisica Sperimentale
I am a scientist working in air-sea interaction and climate studies since the beginning of my PhD course in 2010. After completing the PhD course (thesis titled “The role of North Atlantic cyclone activity in the air-sea interaction processes over the ocean”) I was working as a post-doctoral scientist on the topics of the Agulhas Current dynamics (under the project “Regional and global importance of the Agulhas Current system, led by Prof. Dr. Arne Biastoch”) and ocean mixed layer dynamics (under the project “MEDLEY - Mixed Layer Heterogeneity”, led by Prog. Anne Marie Treguier). My work was mostly based on the output of the ocean circulation models, a combination of the models’ output with observational datasets, and exploration of the big climate datasets. In 2023 I’ve started the Master in High Performing Computing (SISSA/ICTP) course to combine my knowledge in oceanography and climate with the modern HPC approaches.