ENSO’s changing grip on Bering Sea ice: The emerging control of the North Pacific meridional mode
Abstract
Bering Sea winter sea ice (BSWI), vital for regional climate, ecosystem, and livelihoods, is significantly influenced by El Niño-Southern Oscillation (ENSO). Here we identify a shift in the ENSO-BSWI relationship. Pre-mid-1990s, traditional eastern Pacific ENSO dominated, driving a positive ENSO-BSWI linkage; Post-mid-1990s, more frequent central Pacific (CP) ENSO events reversed this relationship to negative. Observations and model experiments show CP El Niño coupled with positive-phase North Pacific Meridional Mode (NPMM) redirects poleward-propagating Rossby wave trains, enhancing Bering Sea southerlies. This suppresses ice advection, intensifies warm air intrusion, and reduces sea ice cover (SIC). Strengthened CP ENSO-NPMM coupling and heightened NPMM variability amplify this teleconnection, increasing its influence on winter SIC by 38.9% versus NPMM alone. Our findings underscore the growing role of subtropical and tropical Pacific climate interactions in subarctic sea ice variability and highlight the need for improved climate models that capture ENSO diversity, NPMM dynamics, and subtropical-subarctic teleconnections.
Dataset DOI: 10.5061/dryad.1rn8pk16q
Description of the data and file structure
To examine the impact of anomalous surface winds on SIC, three ocean general circulation model (OGCM) experiments were conducted using the Parallel Ocean Program version 2 (POP2) within the Community Earth System Model version 1.2.2 (CESM1.2.2). These experiments consist of one control simulation (CTRL) and two sensitivity simulations (southerly and northerly wind experiments). The CTRL simulation is forced by global monthly climatological 10-meter winds, implemented with the G_NORMAL_YEAR component sets. The sensitivity experiments are identical to the CTRL, except that observed southerly or northerly wind anomalies are prescribed within the region spanning 160°E–160°W and 50°N–70°N from January to December. The surface wind anomalies are applied based on the partial regression pattern of JFM1 surface winds over the Bering Sea onto the normalized JJA0 Niño4 index during 1995-2021, without amplitude scaling. All OGCM experiments were integrated over 35 years, with the ensemble mean of the last 30 years used for analysis in this study.
Files and variables
File: data.zip
File: data.zip contains two folders (pop_bs_south, pop_bs_north). Each folder includes *.ctl and .dat files (comp_aice_30case_320x384_JFM, comp_shf_30case_320x384_JFM, comp_transixy_30yr_320x384_JFM). The **.ctl files serve as descriptor files for the binary *.dat files. The dataset has a horizontal grid resolution of 320 points in the zonal direction (x) and 384 points in the meridional direction (y). The temporal average is the JFM (January, February, March) seasonal mean.
Description:
pop_bs_south: Ensemble mean differences in SIC, surface heat flux (sensible heat flux (SHF), latent heat flux (LHF), net shortwave radiation (NSWR), net longwave radiation (NLWR), and net surface heat flux (Qnet)) and sea ice mass transport between the southerly winds experiment and CTRL.
pop_bs_north: Ensemble mean differences in SIC, surface heat flux (sensible heat flux (SHF), latent heat flux (LHF), net shortwave radiation (NSWR), net longwave radiation (NLWR), and net surface heat flux (Qnet)) and sea ice mass transport between the northerly winds experiment and CTRL.
1. "The file pair comp_aice_30case_320x384_JFM.[ctl|dat] " represents:
Ensemble mean difference: The difference in sea ice concentration (SIC, variable aice) between the northerly/southerly winds experiment and the control (CTRL).
Statistical significance: The corresponding values from a significant t-test (variable ttest)."
2. "The file pair comp_shf_30case_320x384_JFM.[ctl|dat] " represents:
Ensemble mean difference: The difference in surface heat flux components (variable shf) between the northerly/southerly winds experiment and the control (CTRL). The five tdef values correspond to the following fluxes:
(1) Latent Heat Flux (LHF)
(2) Sensible Heat Flux (SHF)
(3) Net Shortwave Radiation (NSWR)
(4) Net Longwave Radiation (NLWR)
(5) Net Surface Heat Flux (Qnet)
Statistical significance: The corresponding values from a significant t-test (variable ttest).
3. "comp_transixy_30yr_320x384_JFM.[ctl|dat]" represents:
Ensemble mean difference: The difference in monthly means of ice mass transport on East side (variable transix) and monthly means of ice mass transport on North side (variable transiy) between the northerly/southerly winds experiment and the control (CTRL).
Statistical significance: The corresponding values from a significant t-test (variable ttest). Values where ttest > 0 signify that the ensemble mean difference is statistically significant at the ≥90% confidence level.
Methods
This dataset is the post-processed CESM model outputs of the northerly/southerly winds experiment and the control (CTRL). All the results were processed by the GrADS （Grid Analysis and Display System）.
Usage notes
The data descriptor file, commonly referred to as a control file and identified by its .ctl extension, is a plain-text file that contains all the necessary metadata to interpret its companion raw binary data file (.dat). This metadata includes the path to the data file, grid definition, time axis details, horizontal coordinates, vertical levels, and a list of variables with their units, effectively allowing the .ctl file to act as the header for the separate binary dataset, similar to a self-describing format like NetCDF. To open and view this data, the free and open-source software GrADS (Grid Analysis and Display System) can be used; it reads the .ctl file to load the corresponding .dat file, enabling users to interactively inspect, subset, and visualize the data.