# Km-scale coupled simulation and model–observation SST trend discrepancy

**Authors:** Sarah M. Kang, Dian A. Putrasahan, Noel G. Brizuela, Helmuth Haak, Jürgen Kröger, Jochem Marotzke, Bjorn Stevens, Jin-Song von Storch

PMC · DOI: 10.1073/pnas.2522161123 · Proceedings of the National Academy of Sciences of the United States of America · 2026-02-19

## TL;DR

A high-resolution climate model successfully reproduces observed cooling in the Southern Ocean and tropical Pacific, which most models fail to capture.

## Contribution

The study demonstrates that km-scale resolution improves simulation of key climate features previously missed by coarser models.

## Key findings

- A 5 km ocean and 10 km atmosphere model reproduces observed cooling in the Southern Ocean and southeastern tropical Pacific.
- Fine resolution enables realistic representation of eddy heat transport and stratocumulus cloud feedbacks.
- Improved resolution may reduce model–observation discrepancies and enhance climate projections.

## Abstract

Accurately simulating historical sea surface temperature (SST) trends is critical for reliable near-term climate projections. Yet, most climate models fail to capture the observed cooling in the Southern Ocean and southeastern tropical Pacific—two persistent model–observation discrepancies. Using a km-scale coupled simulation, we successfully reproduce these key features by explicitly resolving ocean eddies and simulating realistically strong stratocumulus cloud feedbacks. We argue that km-scale spatial resolution could improve representations of ocean heat uptake and extratropics-to-tropics teleconnections, potentially reducing long-standing model–observation discrepancies and increasing confidence in near-term climate projections.

The spatial pattern of sea surface temperature (SST) change in the tropical Pacific plays a central role in shaping global climate through its influence on atmospheric circulation, rainfall, and extreme weather. Accurately simulating this pattern is critical for reliable near-term climate projections, yet persistent discrepancies between models and observations remain. While global warming is unequivocal, satellite-era observations reveal pronounced cooling in the southeastern tropical Pacific and Southern Ocean, contrary to the warming simulated by most CMIP models. These discrepancies raise concerns about the reliability of current models in projecting near-term regional climate change. Here, we present the historical simulation with the ICON coupled model at a 5 km ocean and 10 km atmosphere grid-spacing. ICON successfully reproduces the observed SST trends, including cooling in both the Southern Ocean and the southeastern tropical Pacific. Its fidelity is enabled by the direct representation of eddy heat transport across the Antarctic Circumpolar fronts and realistic stratocumulus cloud feedbacks in the subtropical southeast Pacific. These findings highlight the importance of fine spatial resolution in capturing the mechanisms of heat uptake in the eddy-rich Southern Ocean. Our results suggest a pathway for resolving long-standing biases in historical simulations and improving confidence in near-term climate projections.

## Full-text entities

- **Diseases:** CMIP (MESH:D004195)
- **Chemicals:** CO2 (MESH:D002245), ice (MESH:D007053), ACC (-), ozone (MESH:D010126), ICON (MESH:C037304), water (MESH:D014867), PNAS (MESH:D020135)

## Full text

_Full body text omitted from this summary view._ Fetch the complete paper as Markdown: https://tomesphere.com/paper/PMC12933147/full.md

## Figures

6 figures with captions in the complete paper: https://tomesphere.com/paper/PMC12933147/full.md

## References

57 references — full list in the complete paper: https://tomesphere.com/paper/PMC12933147/full.md

---
Source: https://tomesphere.com/paper/PMC12933147