Uncertainty Dominance Delays the Emergence of Marine Heatwave Signals in CMIP6 Projections

Based on simulations from the Coupled Model Intercomparison Project Phase 6 (CMIP6), projections of marine heatwave (MHW) annual accumulated days (AAC_day) and intensity (AAC_intensity) remain highly uncertain, even in regions where anthropogenic signals are expected to emerge. Total projection uncertainty is decomposed into contributions from intermodel differences, internal variability, and emission scenarios. In the near term, internal variability dominates uncertainty in climate-mode-influenced oceans, while intermodel uncertainty prevails elsewhere. From the mid- to long-term, uncertainties associated with both internal variability and intermodel differences decrease nearly globally. Scenario uncertainty remains negligible until becoming evident over tropical oceans in the long term. Anthropogenic signals in AAC_day (AAC_intensity) emerge over only 2.2% (1.9%), 16.5% (1.8%), and 43.1% (2.0%) of the global ocean in the near-, mid-, and long-term, respectively, but expand to 32.4% (11.2%), 63.5% (18.4%), and 79.9% (20.7%) when intermodel differences are removed. These results demonstrate that internal variability and model uncertainty substantially delay the detectability of MHW changes, highlighting the importance of reducing model spread to improve future projections of MHW risks.