Pushing the Frontiers: High-Resolution Modeling and the Future of Climate Synthesis.
The CMIP26 Workshop, held in the historic city of Kyoto, Japan, provided a timely platform to discuss the evolution of Earth System Models (ESMs). Specifically, Session 3.2, “Advancing climate model science with high resolution simulations, downscaling, next generation model evaluation and observations,” showcased a vibrant collection of research aimed at refining our understanding of the climate system. The session brought together global experts to discuss how high-resolution simulations and novel evaluation methods can bridge the gap between global projections and regional impacts. This post highlights key takeaways from the poster session that the CMIP community can leverage as we transition toward CMIP7.
The Power of Resolution: From Global Trends to Local Extremes
While the broad strokes of global warming are well understood, the “multiverse” of models often diverges when it comes to regional phenomena. A recurring theme in the session was how increasing resolution—moving toward kilometer-scale (km-scale) simulations—fundamentally changes our representation of the Earth. Research using the AWI-ESM3, NICAM and ICON models demonstrated that higher resolution isn’t just about “sharper pictures”; it’s about capturing the correct dynamics of the climate system. For instance, studies on global extreme wind speeds and the Mediterranean Sea showed that resolving non-hydrostatic atmospheric processes and variable-resolution grids generates structural and intensity differences that emphasize the need of coupled models resolving all the different variables of the climate system, such as the ocean sea surface temperature and currents.
Next-Gen Parameterization: Machine Learning and Physics
Even with increasing computational power, we cannot resolve everything. The session highlighted a shift toward “Next-Generation” evaluation via integration of Deep Learning for parameterizing atmospheric gravity waves and the exploration of Quantum Machine Learning. These tools are beginning to enhance traditional sub-grid scales, offering a way to account for complex feedbacks like cloud-aerosol interactions. By learning from high-resolution data, these models are becoming better equipped to handle the internal variability that often leads to uncertainty in decadal projections without the cost of global full high-resolution simulations.
Practical Insights: Cyclones, ENSO, and Scale Interactions
The poster presentations provided concrete examples of how resolution dependency impacts our climate outlook:
- Tropical Cyclones: Comparisons between observations and HighResMIP models revealed that significant challenges remain in capturing the precise historical tendencies.
- Ocean Modes of variability: Updates on the TaiESM and NorESM models showed how El Niño-Southern Oscillation representation shifts as we move to eddy-permitting ocean configurations. Similarly, research involving the North Atlantic and the Northwest Pacific highlighted the importance of scale-dependent energy budgets in understanding ocean-scale interactions.
- Land-Water Coupling: The move from fixed boundaries to dynamic lakes and river-land interactions is proving vital for simulating the seasonal-to-subseasonal variability more accurately.
Bridging Downscaling and Observations
A significant portion of the session focused on CORDEX and dynamical downscaling. Whether looking at the Australian national projections or benchmarking models over South America, the consensus was: regional downscaling remains a bridge between global ESMs and local decision-making.
Concluding Thoughts: Towards a Collaborative CMIP7
Small-scale processes and the precise simulation of long term variability are the building blocks of the climate system. The research presented in Kyoto underscores that the future of climate science lies in the synergy between high-resolution modeling, interactive coupling, and innovative evaluation frameworks like Shannon’s Entropy for decadal variability.
Many thanks to the conveners and all the presenters in Session 3.2 for the fruitful discussions. The insights gathered here in Kyoto will undoubtedly help the broader community as we strive to reduce uncertainty and provide more robust climate information to everyone.
