Contrasting Arctic Climate Change Patterns from Storyline-Driven Regional Climate Simulations

Arctic climate projections are characterized by pronounced uncertainty, stemming mainly from structural uncertainties related to the representation of Arctic processes and feedbacks, including those associated with permafrost, cryosphere–atmosphere coupling, and sea ice. Within the PolarRES project’s framework, we apply a storyline-based approach to address parts of these uncertainties and investigate how different physically plausible Arctic futures manifest in high-resolution regional climate simulations. We analyze an ensemble of five regional climate models (RCMs) at 11 km resolution over the Arctic, each driven by two CMIP6 global climate models representing contrasting storylines of Arctic change: CNRM-ESM2-1 and NorESM2-MM, under the high-emission scenario SSP3-7.0. The two driving GCMs differ in their representation of Arctic climate change mechanisms, with NorESM2-MM exhibiting stronger lower-tropospheric Arctic amplification and CNRM-ESM2-1 showing comparatively weaker atmospheric amplification but enhanced surface warming in the Barents–Kara Sea region.

Climate change signals are assessed by comparing end-of-century conditions (2070–2099) to a present-day reference period (1985–2014) for near-surface 2 m air temperature, total precipitation, and the seasonal number of freezing (ice) days. Across all variables, the RCMs broadly reproduce the large-scale spatial patterns imposed by their driving GCMs, but also introduce pronounced regional modifications and inter-model spread, particularly over the ice covered parts of the Arctic ocean.

The strongest warming occurs in winter, exceeding 15 K over parts of the Arctic Ocean, with several RCMs amplifying the Barents–Kara Sea warming relative to the driving models. Summer warming is comparatively weak and consistent across both storylines, whereas spring and autumn exhibit enhanced inter-RCM variability, pointing to sensitivities in snow-albedo feedbacks and melt–freeze processes. Changes in freezing days reveal substantial ensemble spread in summer, despite similar mean temperature change signals, highlighting the nonlinear dependence of threshold-based metrics on absolute temperature levels.

Projected precipitation increases are largest in winter and autumn, particularly over the Arctic Ocean and the Barents–Kara region, with relative increases often exceeding 80–100%. While overall patterns resemble those of the driving GCMs, individual RCMs exhibit notable deviations, especially over sea-ice loss regions.

These results demonstrate that regional climate models add important, physically meaningful structure to Arctic climate change signals, emphasizing the role of regional processes in shaping plausible future Arctic climates within a storyline framework.