Antarctic sea ice polynya and polynya ice production based on high resolution SSP585 simulation

Data for "Projected change of Antarctic polynya under SSP5-8.5 forcing in a high-resolution coupled climate model" This archive holds the sea ice data behind the paper. Everything comes from a single AWI-CM3 simulation that runs from 1920 to 2100, following CMIP6 historical forcing until 2014 and SSP5-8.5 afterwards. The ocean component FESOM2 reaches about 5 km around the Antarctic margin. Daily sea surface fields were archived for three periods, 2004 to 2013, 2050 to 2059, and 2090 to 2100. All gridded files below live on the same 6.25 km Antarctic polar stereographic grid with 1264 by 1328 pixels, and every file carries the lon and lat arrays that belong to it. sic_daily.zip One NetCDF file per year, 31 files in total. The variable a_ice is daily sea ice concentration in percent. To keep the archive small we stored it as 16 bit integers with a scale factor of 0.1. Any normal NetCDF reader such as xarray, netCDF4, ncview or cdo applies the scaling for you, so in practice you just read the variable and get percent. The packing error is at most 0.05 percent, far below anything the analysis is sensitive to. sip_daily.zip One NetCDF file per year, covering April to October. The variable polynya_ice_growth is the daily thermodynamic ice growth in metres of ice per day, kept only inside the detected polynyas. Pixels outside the polynya masks hold the fill value -999. Multiply by the pixel area of 39.0625 km2 and sum over a region to obtain volume production. The seasonal totals in the paper were computed exactly this way. polynya_masks.zip Three tar archives, one per period, with the daily output of the DEEP polynya detection. In each daily file the variable Map labels every pixel with the ID of the polynya it belongs to, and zero means open pack ice or ocean. The variable CoastalPolynyaFlag tells you whether an ID refers to a coastal or an open ocean polynya. LonLat.nc inside each archive gives the coordinates of the grid. derived_diagnostics.zip Three small csv tables with the regional means used in the figures. regional_ice_production.csv holds monthly polynya ice production per region in km3. regional_daily_sic.csv holds the climatological daily sea ice concentration per region and period. regional_timeseries_sst_2t.csv holds annual April to October means of sea surface temperature and 2 m air temperature per region from 1920 onwards. One warning about this last table. The column named era5_2t contains the 2 m air temperature of the model atmosphere and not ERA5. The name survived from an early version of our scripts.Polynyas were detected with the DEEP algorithm of Lin et al. (2024). The matching observational dataset DEEP-AA is available at doi 10.5281/zenodo.11379148. Raw output of the AWI-CM3 simulation is far too large for this archive and is available on request from the authors.