Marine heatwaves (MHWs), characterized by prolonged anomalously high sea surface temperatures (SSTs), have severe ecological and socioeconomic consequences. While extensively studied in other regions, Arctic MHWs have received less attention despite their potential to significantly impact Arctic ecosystems, disrupting food chains, harming fish stocks, and causing biodiversity loss. These impacts can have cascading effects on indigenous communities and fisheries. This study aims to quantitatively assess the role of greenhouse gas (GHG) forcing in the emergence and recurrence of Arctic MHWs, a crucial question given the Arctic's rapid warming—nearly four times faster than the global average—and the pronounced seasonality of this warming, with winter warming exceeding summer warming. This warming is occurring alongside significant transformations in Arctic sea ice extent, timing, and characteristics, including a regime shift since 2007 towards thinner and more uniform ice cover. Arctic sea-ice variability is primarily driven by atmospheric temperature fluctuations, with other factors explaining only 25% of the variability. While GHG increases drive Arctic land surface temperature trends, their impact is partially offset by other anthropogenic forcings, mainly aerosols. This study employs an extreme event attribution technique based on causal counterfactual theory to determine whether GHG forcing is a necessary or sufficient cause for Arctic MHWs. It uses multiple satellite observations, reanalyses, and three large ensembles of coupled general circulation models with GHG-forcing only experiments (CESM1-LE, MPI-ESM-LR, and CanESM5).

Previous research highlights the increasing frequency and intensity of marine heatwaves globally, with studies focusing on their drivers and impacts. However, Arctic MHWs have received comparatively less attention. Studies have shown the significant impacts of MHWs on marine ecosystems, including disruption of food webs and damage to sensitive species. The Arctic's rapid warming and sea ice decline have been well-documented, with studies demonstrating the central role of diminishing sea ice in Arctic temperature amplification and the increasing frequency of extreme temperature events. The link between atmospheric temperature fluctuations and sea-ice variability has been established, although other factors also play a role. Attribution studies have successfully linked extreme events, such as record minimum sea ice extents, to anthropogenic forcings. However, an event attribution analysis specifically for Arctic MHWs was lacking prior to this study. Existing research on Arctic MHWs has focused on characterizing their spatial and temporal patterns and their relationship to sea ice conditions but lacked a comprehensive analysis of the role of GHG forcing. This study builds upon this existing knowledge by conducting a quantitative attribution analysis.

This study uses multiple data sources including satellite observations (NOAA OISSTV2 for SST, NOAA/NSIDC for sea ice concentration), reanalyses (ERA5 for surface heat fluxes), and simulations from three large ensembles of coupled general circulation models (CESM1-LE, MPI-ESM-LR, and CanESM5). The analysis involves two main components: 1) characterizing Arctic MHWs and their relationship to sea ice and atmospheric forcing; 2) attributing the likelihood of MHW events to GHG forcing. MHWs were identified using a standardized definition based on SST exceeding a seasonally varying 95th percentile threshold for at least five consecutive days. The study region focused on the Arctic marginal seas, characterized by shallow mixed-layer depths and predominantly first-year ice. The relationship between SST variability and net atmospheric surface fluxes (Qnet) was assessed using regression analysis. Event attribution analysis used an extreme event attribution technique based on causal counterfactual theory to estimate the probabilities of MHWs occurring with and without GHG forcing, using both observed and model-simulated data. The probabilities were used to calculate the probability of necessary causation (PN) and the probability of sufficient causation (PS) for MHW intensity, duration, and cumulative heat intensity. Long-term trend detection and attribution employed a univariate total least squares (TLS) regression analysis to project observed SST changes onto the model-simulated response to GHG forcing, accounting for uncertainties due to internal climate variability using time-evolving internal variability estimated from ALL-forcing simulations. The study also calculated the cumulative solar heat input to the ocean, considering incident solar irradiance, ice concentration, and ocean albedo, to understand the underlying mechanisms of MHWs. The open-water period was defined as the time between the last day of sea ice concentration above 15% before the annual minimum and the first day above 15% after the annual minimum. A 5-day moving average was applied to daily sea ice concentration data to reduce short-term fluctuations. Detection of GHG signal in observed SST trends was assessed by testing null hypotheses for scaling factors derived from the regression model using time-evolving internal variability from ALL-forcing simulations of MPI-ESM-LR and CanESM5.

Satellite observations revealed a shift towards more frequent and intense MHWs in the Arctic since 2007, primarily occurring in shallow marginal seas covered by first-year ice. Eleven MHW events were identified between 2007 and 2021, with the 2020 event being the most intense (4°C) and longest (103 days). Regression analysis showed that 82% of SST variability in the marginal seas could be explained by variability in net atmospheric surface fluxes (Qnet), driven by changes in shortwave and longwave radiation and turbulent heat fluxes. Event attribution analysis demonstrated that GHG forcing is virtually certainly a necessary cause for MHWs with intensities exceeding 1.5°C, meaning that these events would be exceptionally unlikely without GHG forcing. For example, the probability of necessary causation (PN) for the 2020 event was approximately 0.99, indicating a 99% chance that GHG forcing was required for its occurrence. However, GHG forcing was not a sufficient cause for these extreme events, meaning that even with GHG forcing, such events remain rare. For moderate events (0.5–1°C intensity), GHG forcing emerged as a likely sufficient cause (66–99% probability), suggesting that continued GHG emissions will lead to persistent moderate MHWs. Analysis of the underlying mechanisms revealed that MHWs are primarily triggered by abrupt early-summer sea ice retreat, coinciding with the peak of downward radiative fluxes. The rate of sea ice melt showed a substantial increase (38% in 25 years). The 2020 MHW was stronger than the 2012 MHW despite less total sea-ice loss in 2020 due to earlier sea ice retreat near the peak of atmospheric heat flux. In autumn, extra heat stored in the ocean is released back into the atmosphere, with outgoing fluxes increasing substantially (5.6 W/m²/decade). Long-term SST trend detection showed a clear GHG signal in observed warming (1.2 °C/decade from 1996-2021), confirming the MHW attribution results.

The findings strongly support the role of GHG forcing in driving the observed increase in Arctic MHWs. The strong correlation between SST variability and net atmospheric surface fluxes, combined with the event attribution analysis showing GHG forcing as a necessary cause for extreme events, clearly links anthropogenic climate change to these events. The identification of abrupt sea ice retreat as a trigger mechanism further highlights the interplay between warming temperatures, sea ice loss, and MHWs. The conclusion that GHG forcing is a likely sufficient cause for moderate MHWs underscores the potential for continued and even more frequent occurrences of these events in the future under ongoing emissions. The results also emphasize the importance of the timing of sea-ice retreat relative to atmospheric heat input in determining the intensity of MHWs, indicating a complex interaction between different forcing mechanisms. The long-term SST trend detection reinforces the importance of GHG forcing in shaping the baseline conditions for MHWs. The findings have important implications for understanding and predicting the future of Arctic ecosystems and highlight the need for mitigation efforts to reduce GHG emissions. The shallow mixed layer depth in Arctic marginal seas enhances the susceptibility to MHWs by facilitating rapid absorption of solar energy in areas of open water.

This study demonstrates a clear link between GHG forcing and the increasing frequency and intensity of Arctic MHWs. GHG forcing is a necessary, though not always sufficient, cause for extreme MHW events. Continued GHG emissions will very likely result in persistent moderate MHWs, exacerbating climate change impacts in the Arctic and accelerating sea ice loss. Future research could focus on improving the understanding of regional variations in MHWs and their impacts, exploring the role of other climate drivers besides GHG forcing, and further refining the attribution methodologies to account for uncertainties and complex interactions within the Arctic climate system.

The study relies on specific model ensembles and datasets, limiting the generalizability of some findings. The event attribution analysis uses a causal counterfactual framework, which involves assumptions and limitations inherent in such approaches. The analysis focuses on a specific region and time period, which may not fully represent the Arctic Ocean as a whole. Uncertainties exist related to internal climate variability, model biases, and data quality. The use of reanalysis data for atmospheric heat fluxes introduces potential biases that could affect the results. Further research is needed to refine and verify these findings. The choice of sea ice concentration threshold for defining the open-water period could influence the results. However, the general findings are robust and do not depend on the choice of threshold.