Effects of macro vs. micro initialization and ocean initial-condition memory on the evolution of ensemble spread in...

Area of Science:

• Climate Science
• Earth System Modeling
• Oceanography

Background:

• Single Model initial-condition Large Ensembles (SMILEs) are crucial for understanding internal climate variability and forced climate change.
• Initialization procedures in SMILEs impact initial-condition memory, affecting ensemble-mean and spread interpretation.
• The Community Earth System Model version 2 (CESM2) SMILE provides a unique dataset for analyzing these effects.

Purpose of the Study:

• To investigate how initialization methods (micro- vs. macro-perturbations) influence ensemble spread in CESM2 SMILE.
• To examine the duration and characteristics of ocean initial-condition memory, particularly related to the Atlantic Meridional Overturning Circulation (AMOC).
• To provide recommendations for future SMILE experimental design.

Main Methods:

• Analysis of a 100-member CESM2 SMILE dataset.
• Investigation of ensemble spread dependence on micro- vs. macro-perturbation initialization.
• Assessment of initial-condition memory duration and spatial patterns, focusing on AMOC states.

Main Results:

• Ensemble spread evolution is largely insensitive to initialization method after two decades, except in the tropical Indo-Pacific during the 4th decade with macro-initialization.
• AMOC initial-condition memory manifests in two stages: North Atlantic (4-5 decades) and Southern Ocean (delayed, 35 years post-initialization, lasting 3-4 decades).
• The delayed Southern Ocean memory's dependence on AMOC and its mechanisms require further research.

Conclusions:

• Initialization method choice has a limited long-term impact on ensemble spread, with notable regional exceptions.
• AMOC significantly influences climate variability across different ocean basins over extended periods.
• Understanding these memory effects is critical for designing effective future climate modeling experiments.