Journal Article

Distinct Roles of Surface Flux Changes in Driving Model Spread of Dynamic Sea‐Level Projections in Different Regions

Wu, Quran, Gregory, Jonathan

2026

Overview

Dynamic sea‐level change (ΔDSL)$({\\Delta }\\mathrm{D}\\mathrm{S}\\mathrm{L})$is a key process in shaping the pattern of future sea‐level rise. CMIP6 models predict a range of ΔDSL${\\Delta }\\mathrm{D}\\mathrm{S}\\mathrm{L}$under 1% increase of CO2${\\text{CO}}_{2}$per year. We analyze this CMIP6 spread into contributions from: (a) surface flux change (dF)$(\\mathrm{d}\\mathrm{F})$and (b) model sensitivity to it (Φ)$({\\Phi })$ . Specifically, we force the pre‐industrial simulation of an ocean model with space‐ and time‐varying dF$\\mathrm{d}\\mathrm{F}$diagnosed from different CMIP6 models (one at a time). The CMIP6 spread is thus decomposed into a flux‐driven spread and a residual; the latter is linked to model spread of Φ${\\Phi }$ . We improve upon previous studies by: (a) deriving the perturbed forcing ensemble using an ocean‐only setup and (b) comparing it with the CMIP6 ensemble for both variance and correlation. This reveals distinct roles of surface forcing in driving the CMIP6 spread in different regions. In the North Pacific, differences in windstress forcing primarily explain the CMIP6 spread, while in the North Atlantic, differences in model sensitivity are more important. For the latter region, although buoyancy forcing drives a ΔDSL${\\Delta }\\mathrm{D}\\mathrm{S}\\mathrm{L}$spread there, it correlates poorly with the CMIP6 spread. In the Southern Ocean, differences in both surface forcing and model sensitivity are important for explaining the CMIP6 spread. The surface forcing affects the spread along 40°S via windstress and the spread around the Antarctic via buoyancy flux. Plain Language Summary Climate model simulations provide important information to support planning for future sea‐level rise. Contemporary climate models exhibit large differences in simulated regional sea‐level change under strong CO2${\\text{CO}}_{2}$emission. These model differences can be analyzed in terms of: (a) model differences in simulated surface flux changes (heat, freshwater and wind) and (b) model differences in simulated ocean response to a given surface flux change. We find that model differences in surface flux changes explain most of model diversity in sea‐level change in the North Pacific and part of that in the Southern Ocean, but little of that in the North Atlantic. These results pave the way for reducing sea‐level projection uncertainties in future research. Key Points CMIP6 spread of dynamic sea‐level projections results from both surface forcing and model sensitivity to it Windstress forcing explains the CMIP6 spread in the North Pacific, while model sensitivity is more important in the North Atlantic In the Southern Ocean, both surface forcing and model sensitivity are important for explaining the CMIP6 spread

• Publisher: John Wiley & Sons, Inc,American Geophysical Union (AGU)
• Subject: Buoyancy / Buoyancy flux / Carbon dioxide / Climate change / CMIP6 / Decomposition / dynamic sea‐level change / Experiments / Fluctuations / Global warming / inter‐model spread / model sensitivity / Ocean models / Oceans / Sea level changes / Sea level rise / surface flux change