The influence of recent and future climate change on spring Arctic cyclones

The study investigates how recent and projected climate change influence spring (March–April) Arctic cyclone characteristics. Arctic amplification has led to rapid declines in sea ice extent, concentration, thickness, and snow depth, alongside longer melt seasons and reduced winter growth. These surface and atmospheric changes may alter baroclinicity, turbulent fluxes, and convection, potentially affecting cyclogenesis, intensity, and cyclone behavior. Yet observational records are sparse and reanalysis-based trend assessments are inconclusive, and global climate models often lack sufficient resolution to capture cyclone processes and their local interactions with sea ice. The research questions are: (1) To what extent has recent climate change altered present-day spring Arctic cyclone characteristics? (2) How will continued Arctic amplification and sea ice loss affect spring cyclone behavior by late century? By using convection-permitting regional modeling and pseudo-global warming perturbations derived from CMIP6, the study aims to isolate storm-scale responses to large-scale thermodynamic and dynamic changes, providing process-level insight crucial for anticipating impacts on sea ice mass balance, freshwater budgets, and the Arctic surface energy balance.

Prior work documents amplified Arctic warming and substantial sea ice declines in extent, thickness, and seasonal characteristics, with feedbacks to atmospheric processes. Studies indicate increased winter and spring Eady growth rates as static stability decreases with retreating sea ice. Cyclones transport heat and moisture into the Arctic and can act as compound extreme events, affecting sea ice motion and surface energy budgets; their impacts vary by event characteristics, location, and season. Analyses of long-term changes in Arctic cyclones from observations and reanalyses show substantial spatial and temporal variability, with no consensus on trends. Global models commonly project increased summer cyclone activity by the end of the 21st century but disagree on winter changes, likely due to coarse resolution and biases. Increased land–sea temperature contrast and evolving atmospheric thermal structure may intensify baroclinicity and cyclone activity, while enhanced surface turbulent fluxes could promote convection and storm intensity. However, uncertainties remain high, underscoring the need for high-resolution, process-focused studies.

The study employs the WRF model v3.9.1.1 configured over an Arctic domain at ~4 km horizontal resolution with 51 vertical levels up to 10 hPa. Physics include RRTMG shortwave and longwave radiation, Eta similarity surface layer, Unified Noah LSM, Mellor–Yamada–Janjic PBL, and Morrison double-moment microphysics. Sea ice concentration and SSTs are updated daily at 6-hour intervals. Simulations initialize shortly after cyclone genesis (from a Melbourne University cyclone tracking scheme applied to ERA5 SLP) and cover development, intensification, and decay. Nine spring cyclones (six in March: A–C, G–I; three in April: D–F) from 2000–2021 were selected based on season, geographic diversity within the model domain, and intensity (minimum SLP ≤980 hPa with main intensification inside the domain). Three March 2020 cases (A–C) coincided with the MOSAiC expedition for high-resolution validation. Boundary and initial conditions for current climate runs use ERA5 (0.25°, 6-hourly). Historical (1885–1914) and future (2070–2099, SSP5-8.5) simulations apply pseudo-global warming (PGW) perturbations to ERA5 conditions, adding monthly CMIP6 multi-model 30-year mean deltas for geopotential height, RH, specific humidity, surface and atmospheric temperature, winds, surface and mean sea level pressure, and sea ice concentration. Historical deltas approximate removal of recent anthropogenic changes; future deltas represent late-century projections. This isolates cyclone responses to changes in the large-scale environment without altering event selection or genesis. All runs apply 3 hours of digital filter initialization to balance fields. Greenhouse gas concentrations are set per scenario (e.g., for Cyclones A–C: CC 379/1774/319 ppm/ppb/ppb; HC 297/895/280; FC 975/2595/385). Model evaluation uses MOSAiC in situ measurements of SLP and 2 m temperature; WRF SLP agrees closely with observations, and temperature evolution is reasonably captured with small mean biases. Analyses compare cyclone tracks, intensity (minimum SLP, time at maximum intensity), 10 m winds, CAPE, surface–air temperature differences, surface sensible and latent heat fluxes, precipitation (snow, rain) within a 200 km radius, and cyclone size (median radius) across historical, current, and future perturbations. Climate deltas summarize regional environmental changes (sea ice concentration, surface and 700 hPa temperatures, humidity, and upper-level winds) between periods.

- Environmental changes since 1914 (historical to current, March–April): sea ice concentration decreased by up to ~25%; surface temperatures increased by ~1.5–6.5 °C; 700 hPa temperatures increased by ~0.5–1.4 °C; relative humidity changes were mixed; upper-level steering flow changes were minimal. - Projected future changes (current to 2070–2099): sea ice concentration decreases up to ~60%; surface temperatures increase by ~4–17 °C; 700 hPa temperatures by ~3.5–4.4 °C; relative humidity increases over Greenland and Barents Seas and decreases elsewhere; steering flow weakens and shifts slightly, altering cyclone trajectories. - Recent climate change effects on cyclones are negligible: Historical vs current simulations show minimal differences in cyclone intensity, lifecycle, size, surface fluxes, and precipitation. Slightly greater sea ice and cooler surfaces in the historical scenario produce small increases in air–surface temperature differences with negligible flux changes; some cases show more weakening during decay. Track changes are minimal (a slight poleward deviation in one case). - Future climate substantially intensifies cyclones: Reduced sea ice and warmer surfaces increase vertical air–surface temperature gradients, moisture availability, and positive turbulent fluxes into the atmosphere (sensible and latent), especially when sea ice under the cyclone falls below ~75%. CAPE develops in the future climate, indicating enhanced instability and convection. Across cases, 10 m wind speeds increase by ~1.5–8 m s⁻¹ on average and up to ~17 m s⁻¹ at times; minimum SLP decreases by ~4–12 hPa on average, up to ~28 hPa; time spent near maximum intensity increases by ~1.5–31%. Changes in cyclone size are not systematic. - Precipitation and temperature responses: In March, precipitation increases substantially yet remains predominantly snowfall; summed snowfall within 200 km typically increases (e.g., for Cyclones A–C, FC–CC average increases of ~643, ~294, and ~104 mm h⁻¹, respectively), with small amounts of rain occasionally occurring as temperatures approach 0 °C. In April and/or at lower latitudes, near-surface temperatures can exceed freezing, yielding a marked phase shift with rain sums increasing markedly (e.g., cases D–F show mean rain increases on the order of ~13–846 mm h⁻¹ from near zero). Warmer temperatures and increased downwelling longwave during cyclones may promote snow/ice melt despite potential snowpack thickening from enhanced snowfall. - Trajectory changes: Slight weakening and shifts in upper-level anticyclones and cyclonic circulations in the future scenario alter steering flow, enabling more poleward, westward, or eastward excursions depending on the case, potentially expanding the geographic reach of cyclone impacts (including toward the central Arctic).

The findings indicate that spring Arctic cyclone characteristics have not yet responded appreciably to recent climate change despite documented Arctic warming and sea ice decline, likely because absolute temperatures in spring remain below freezing and large sea ice concentration changes are mostly confined to marginal seas. Minimal changes in vertical air–surface temperature gradients and surface fluxes yield muted differences between historical and current cyclone characteristics. In contrast, projected late-century conditions—characterized by strong sea ice loss and pronounced surface warming—significantly enhance vertical temperature gradients, surface sensible and latent heat fluxes, moisture availability, and atmospheric instability (CAPE). These thermodynamic changes increase and prolong cyclone intensity, raise wind speeds, and elevate precipitation rates. As temperatures cross the freezing threshold, precipitation increasingly falls as rain, which, together with warmer air and higher downwelling longwave radiation, can accelerate snow and sea-ice melt. Cyclone trajectories remain strongly governed by large-scale steering flow; thus, even modest dynamical changes projected aloft can redirect storms toward previously less-affected regions, broadening the spatial extent and compounding nature of impacts. The results underscore the sensitivity of cyclone behavior to sea ice conditions: the largest changes coincide with the greatest SIC reductions. This has implications for sea ice mass balance (through snowfall, rain-on-snow, and melt), dynamic ice redistribution via stronger winds, and feedbacks that may further amplify Arctic warming.

This study demonstrates that recent spring Arctic cyclone characteristics show negligible changes attributable to recent climate change, whereas future projected Arctic amplification and sea ice loss are likely to substantially intensify and prolong cyclones, enhance wind speeds and surface energy fluxes, increase precipitation (with a shift toward rainfall as temperatures exceed freezing), and alter cyclone trajectories. These changes may strengthen positive feedbacks among ocean–ice–atmosphere interactions, exacerbating sea ice loss and Arctic warming, and may expand the geographic reach and compound risks of extreme weather for ecosystems, communities, and economic activities. Future research should extend this event-based, convection-permitting framework to other seasons, explore alternative emissions pathways and CMIP6 deltas, examine sensitivity to microphysics and flux parameterizations, and assess combined impacts on sea ice mass balance and ocean mixing. Evaluating cyclone genesis and frequency responses, which are beyond the scope of the PGW approach used here, remains an important avenue for subsequent studies.

- The pseudo-global warming approach seeds present-day ERA5 cyclone structures into perturbed environments; it does not assess changes in cyclone genesis likelihood or frequency. - Climate deltas use multi-model 30-year CMIP6 means (SSP5-8.5); while this reduces internal variability influence, it may smooth variability and represents a high-end emissions scenario. Results may differ under other scenarios or using single-model deltas. - Analyses focus on nine spring cyclone cases within one regional domain; while multiple years are sampled, the sample size limits generalization across all Arctic cyclones and seasons. - Observational evaluation relies on MOSAiC in situ data for three cases and ERA5 for initialization; there are no comprehensive direct observational records of Arctic cyclone characteristics, and model–observation temperature differences exist due to resolution and representativeness. - The study isolates storm-scale responses to prescribed large-scale changes; uncertainties remain in projected large-scale dynamical shifts and their case-dependent impacts on trajectories.