Critical timings of the Arctic sea ice decline in three generations of CMIP models from a parametric approximation approach

Rapid decline of the Arctic sea ice area (SIA) is one of the major manifestations of global warming with far‑reaching socioeconomic and ecological consequences. Climate models project further sea ice decrease with a possibility of seasonally ice-free Arctic in the coming decades. Timings of reaching this qualitatively new climate state largely vary among the models, CMIP generations, and depend on methodology used to filter interannual to decadal SIA variability. Here, we introduce a novel analytical framework based on a hyperbolic tangent function to describe long‑term SIA evolution from 1900 to 2100 in CMIP3, CMIP5, and CMIP6 under high‑emission scenarios. This approach provides a robust approximation of SIA dynamics, and enables a consistent assessment of critical timings. The results show a shift toward earlier and more synchronized timings from CMIP3 to CMIP5 and CMIP6 with only minor differences between the last two generations. The median ice-free year occurs about a decade earlier in CMIP5 and CMIP6 compared to CMIP3. The median year of maximum melting rate shifts more than two decades earlier in the newer model's generations. The findings reveal a growing model consensus on the anthropogenically forced decline of Arctic SIA and an earlier transition toward a seasonally ice-free Arctic for newer CMIP generations.
The earlier-than-expected transition to a seasonally ice‑free Arctic threatens traditional livelihoods, marine ecosystems, and coastal infrastructure, while simultaneously opening new shipping routes and resource extraction opportunities. Our results provide quantitative benchmarks for adaptation planning - from updating coastal infrastructure design and seasonal forecasting systems to informing maritime operations and risk assessments for Arctic communities.
This work was supported by Russian Science Foundation Project No. 25-77-31009