Versioned data access with Icechunk

Summary: Demonstrates how to read the IS2SMGPSIT-V1 dataset from its Icechunk repository on S3, including how to pin an immutable release tag for reproducibility, inspect the version history, and make a few key figures directly from the store.

Version history: Version 1 (08/2026)

Note on environments: Icechunk requires Python >=3.12 and Zarr v3, so this notebook does not run in the standard book environment. Create the dedicated environment and kernel first (from the parent directory of icesat2-book):

uv venv --python 3.12 is2book_icechunk_env
source is2book_icechunk_env/bin/activate
cd icesat2-book
uv pip install -r requirements-icechunk.txt
python -m ipykernel install --user --name is2book_icechunk_env --display-name "Icechunk (py3.12)"

Then select the Icechunk (py3.12) kernel to run this notebook.

Why Icechunk?

A plain Zarr store on S3 is a single mutable snapshot: when a new winter of data is added, anyone reading the store gets different numbers than before, and there is no record of what changed. Icechunk adds git-like version control on top of the Zarr data model:

• every update is a commit with a message and timestamp,

• releases are immutable tags (e.g. v1.0 = the peer-reviewed record),

• the main branch always points at the latest complete record,

• chunks shared between versions are stored once, so appending a new season only stores the new data.

If you are reproducing published results, pin the tag cited in the paper. If you just want the latest data, read main.

### Import notebook dependencies

import icechunk as ic
import xarray as xr
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import cartopy.crs as ccrs
import cartopy.feature as cfeature
import warnings

warnings.filterwarnings('ignore')

print('icechunk:', ic.__version__, '| xarray:', xr.__version__)

icechunk: 2.0.6 | xarray: 2026.4.0

Open the repository (anonymous read)

The repository lives on the same public S3 bucket as the plain Zarr store. No AWS credentials are needed for read access.

BUCKET = 'icesat-2-sea-ice-us-west-2'
PREFIX = 'is2smsitgp/v0.1/icechunk'
TAG = 'v0.1'  # immutable release tag to pin below

storage = ic.s3_storage(bucket=BUCKET, prefix=PREFIX, region='us-west-2', anonymous=True)
repo = ic.Repository.open(storage)

print('branches:', repo.list_branches())
print('tags:', repo.list_tags())

branches: {'main'}
tags: {'v0.1'}

Inspect the version history

Each commit records what changed and when. As new winters are appended (one commit + one tag per release), this history becomes the provenance record for the dataset.

for snapshot in repo.ancestry(branch='main'):
    print(f'{snapshot.written_at:%Y-%m-%d %H:%M}  {snapshot.id}  {snapshot.message}')

2026-08-09 00:41  F5VZQ9D22KKJ4369WRS0  IS2SMGPSIT v0.1: import from plain Zarr (s3://icesat-2-sea-ice-us-west-2/is2smsitgp/dev_v7/zarr/GPSat_multivar_20181101-20250430.zarr)
2026-06-10 11:43  1CECHNKREP0F1RSTCMT0  Repository initialized

Read the latest data (main)

Opening a session on main gives the most recent complete record as a regular xarray.Dataset, lazily loaded.

session_latest = repo.readonly_session(branch='main')
IS2_SMOS_SMAP = xr.open_zarr(session_latest.store, consolidated=False)

print(f"time range: {pd.Timestamp(IS2_SMOS_SMAP.time.values.min()):%Y-%m-%d} to "
      f"{pd.Timestamp(IS2_SMOS_SMAP.time.values.max()):%Y-%m-%d} "
      f"({IS2_SMOS_SMAP.sizes['time']} days)")
IS2_SMOS_SMAP

time range: 2018-11-01 to 2025-04-30 (1635 days)

<xarray.Dataset> Size: 10GB
Dimensions:                             (time: 1635, y: 448, x: 304)
Coordinates:
  * time                                (time) datetime64[ns] 13kB 2018-11-01...
  * y                                   (y) float32 2kB 5.838e+06 ... -5.338e+06
  * x                                   (x) float32 1kB -3.838e+06 ... 3.738e+06
Data variables: (12/20)
    crs                                 (time) int32 7kB dask.array<chunksize=(1,), meta=np.ndarray>
    ice_thickness_unc                   (time, y, x) float32 891MB dask.array<chunksize=(1, 448, 304), meta=np.ndarray>
    freeboard                           (time, y, x) float32 891MB dask.array<chunksize=(1, 448, 304), meta=np.ndarray>
    longitude                           (time, y, x) float32 891MB dask.array<chunksize=(1, 448, 304), meta=np.ndarray>
    latitude                            (time, y, x) float32 891MB dask.array<chunksize=(1, 448, 304), meta=np.ndarray>
    freeboard_unc                       (time, y, x) float32 891MB dask.array<chunksize=(1, 448, 304), meta=np.ndarray>
    ...                                  ...
    mean_snow_depth_regions_1_5         (time) float32 7kB dask.array<chunksize=(1,), meta=np.ndarray>
    sea_ice_conc                        (time, y, x) float32 891MB dask.array<chunksize=(1, 448, 304), meta=np.ndarray>
    snow_depth_unc                      (time, y, x) float32 891MB dask.array<chunksize=(1, 448, 304), meta=np.ndarray>
    sea_ice_volume                      (time, y, x) float32 891MB dask.array<chunksize=(1, 448, 304), meta=np.ndarray>
    total_sea_ice_volume                (time) float32 7kB dask.array<chunksize=(1,), meta=np.ndarray>
    total_sea_ice_volume_regions_1_5    (time) float32 7kB dask.array<chunksize=(1,), meta=np.ndarray>
Attributes: (12/14)
    caa_uncertainty_caution:    Canadian Arctic Archipelago (CAA; NSIDC-0780 ...
    contact:                    Alek Petty (akpetty@umd.edu)
    description:                Monthly gridded Arctic sea ice thickness (fus...
    history:                    Created 2026-07-29
    inner_arctic_regions_note:  Inner-Arctic scalars (mean_*_regions_1_5, tot...
    n_timesteps:                1635
    ...                         ...
    time_coverage_start:        2018-11-01
    training_season_caution:    GP training is limited to a seasonal window (...
    product:                    IS2SMGPSIT
    product_version:            v0.1
    release_status:             pre-release (not peer reviewed)
    source_zarr:                s3://icesat-2-sea-ice-us-west-2/is2smsitgp/de...

xarray.Dataset

• Dimensions:
  • time: 1635
  • y: 448
  • x: 304
• Coordinates: (3)
  • time
    (time)
    datetime64[ns]
    2018-11-01 ... 2025-04-30

    long_name :

    time

    standard_name :

    time

    array(['2018-11-01T00:00:00.000000000', '2018-11-02T00:00:00.000000000',
           '2018-11-03T00:00:00.000000000', ..., '2025-04-28T00:00:00.000000000',
           '2025-04-29T00:00:00.000000000', '2025-04-30T00:00:00.000000000'],
          shape=(1635,), dtype='datetime64[ns]')

  • y
    (y)
    float32
    5.838e+06 5.812e+06 ... -5.338e+06

    long_name :

    projection y coordinate

    units :

    meters

    array([ 5.8375e+06,  5.8125e+06,  5.7875e+06, ..., -5.2875e+06, -5.3125e+06,
           -5.3375e+06], shape=(448,), dtype=float32)

  • x
    (x)
    float32
    -3.838e+06 -3.812e+06 ... 3.738e+06

    long_name :

    projection x coordinate

    units :

    meters

    array([-3.8375e+06, -3.8125e+06, -3.7875e+06, ...,  3.6875e+06,  3.7125e+06,
            3.7375e+06], shape=(304,), dtype=float32)

• Data variables: (20)
  • crs
    (time)
    int32
    dask.array<chunksize=(1,), meta=np.ndarray>

    GeoTransform :

    -3850000.0 25000.0 0 5850000.0 0 -25000.0

    false_easting :

    0.0

    false_northing :

    0.0

    grid_mapping_name :

    polar_stereographic

    latitude_of_projection_origin :

    90.0

    long_name :

    NSIDC Sea Ice Polar Stereographic North

    proj4text :

    +proj=stere +lat_0=90 +lat_ts=70 +lon_0=-45 +k=1 +x_0=0 +y_0=0 +a=6378273 +b=6356889.449 +units=m +no_defs

    srid :

    urn:ogc:def:crs:EPSG::3411

    standard_parallel :

    70.0

    straight_vertical_longitude_from_pole :

    -45.0

    units :

    meters

    Array Chunk Bytes 6.39 kiB 4 B Shape (1635,) (1,) Dask graph 1635 chunks in 2 graph layers Data type int32 numpy.ndarray

  • ice_thickness_unc
    (time, y, x)
    float32
    dask.array<chunksize=(1, 448, 304), meta=np.ndarray>

    description :

    Standard deviation of thickness (sqrt of predictive variance from GP). Thickness from ICESat-2/SMAP fusion.

    grid_mapping :

    crs

    long_name :

    sea ice thickness uncertainty (standard deviation)

    units :

    meters

    Array Chunk Bytes 849.43 MiB 532.00 kiB Shape (1635, 448, 304) (1, 448, 304) Dask graph 1635 chunks in 2 graph layers Data type float32 numpy.ndarray

  • freeboard
    (time, y, x)
    float32
    dask.array<chunksize=(1, 448, 304), meta=np.ndarray>

    description :

    Mean freeboard across the grid cell (m). Uses ICESat-2 total freeboard and CDR sea ice concentration (open water) in the solution; gridded via GPSat interpolation.

    grid_mapping :

    crs

    long_name :

    sea ice freeboard (ICESat-2 total freeboard)

    units :

    meters

    Array Chunk Bytes 849.43 MiB 532.00 kiB Shape (1635, 448, 304) (1, 448, 304) Dask graph 1635 chunks in 2 graph layers Data type float32 numpy.ndarray

  • longitude
    (time, y, x)
    float32
    dask.array<chunksize=(1, 448, 304), meta=np.ndarray>

    grid_mapping :

    crs

    long_name :

    longitude

    units :

    degrees_east

    Array Chunk Bytes 849.43 MiB 532.00 kiB Shape (1635, 448, 304) (1, 448, 304) Dask graph 1635 chunks in 2 graph layers Data type float32 numpy.ndarray

  • latitude
    (time, y, x)
    float32
    dask.array<chunksize=(1, 448, 304), meta=np.ndarray>

    grid_mapping :

    crs

    long_name :

    latitude

    units :

    degrees_north

    Array Chunk Bytes 849.43 MiB 532.00 kiB Shape (1635, 448, 304) (1, 448, 304) Dask graph 1635 chunks in 2 graph layers Data type float32 numpy.ndarray

  • freeboard_unc
    (time, y, x)
    float32
    dask.array<chunksize=(1, 448, 304), meta=np.ndarray>

    description :

    Standard deviation of freeboard (sqrt of predictive variance from GP). Freeboard uses ICESat-2 and CDR SIC with GPSat interpolation.

    grid_mapping :

    crs

    long_name :

    sea ice freeboard uncertainty (standard deviation)

    units :

    meters

    Array Chunk Bytes 849.43 MiB 532.00 kiB Shape (1635, 448, 304) (1, 448, 304) Dask graph 1635 chunks in 2 graph layers Data type float32 numpy.ndarray

  • ice_thickness
    (time, y, x)
    float32
    dask.array<chunksize=(1, 448, 304), meta=np.ndarray>

    description :

    Mean sea ice thickness across the grid cell (m). From fusion of ICESat-2 and SMAP (thin ice) via GP interpolation (petty).

    grid_mapping :

    crs

    long_name :

    sea ice thickness

    units :

    meters

    Array Chunk Bytes 849.43 MiB 532.00 kiB Shape (1635, 448, 304) (1, 448, 304) Dask graph 1635 chunks in 2 graph layers Data type float32 numpy.ndarray

  • ice_volume_per_grid_cell_area
    (time, y, x)
    float32
    dask.array<chunksize=(1, 448, 304), meta=np.ndarray>

    description :

    Sea ice thickness / sea_ice_conc, in meters. Masked where sea ice concentration < 0.15.

    grid_mapping :

    crs

    long_name :

    sea ice volume per grid-cell area

    units :

    m

    Array Chunk Bytes 849.43 MiB 532.00 kiB Shape (1635, 448, 304) (1, 448, 304) Dask graph 1635 chunks in 2 graph layers Data type float32 numpy.ndarray

  • mean_freeboard_regions_1_5
    (time)
    float32
    dask.array<chunksize=(1,), meta=np.ndarray>

    description :

    Area-weighted mean freeboard over grid cells in regions 1–5. Freeboard uses ICESat-2 and CDR SIC with GPSat interpolation.

    grid_mapping :

    crs

    long_name :

    mean total freeboard (area-weighted, NSIDC regions 1–5 only)

    units :

    m

    Array Chunk Bytes 6.39 kiB 4 B Shape (1635,) (1,) Dask graph 1635 chunks in 2 graph layers Data type float32 numpy.ndarray

  • grid_cell_area
    (y, x)
    float32
    dask.array<chunksize=(448, 304), meta=np.ndarray>

    description :

    Grid cell area from NSIDC Polar Stereographic North (EPSG:3411).

    grid_mapping :

    crs

    long_name :

    grid cell area

    source :

    https://doi.org/10.5067/N6INPBT8Y104

    units :

    m2

    Array Chunk Bytes 532.00 kiB 532.00 kiB Shape (448, 304) (448, 304) Dask graph 1 chunks in 2 graph layers Data type float32 numpy.ndarray

  • mean_sea_ice_conc_regions_1_5
    (time)
    float32
    dask.array<chunksize=(1,), meta=np.ndarray>

    description :

    Mean sea ice concentration over grid cells in regions 1–5.

    grid_mapping :

    crs

    long_name :

    mean sea ice concentration (NSIDC regions 1–5 only)

    units :

    1

    Array Chunk Bytes 6.39 kiB 4 B Shape (1635,) (1,) Dask graph 1635 chunks in 2 graph layers Data type float32 numpy.ndarray

  • mean_sea_ice_thickness_regions_1_5
    (time)
    float32
    dask.array<chunksize=(1,), meta=np.ndarray>

    description :

    Area-weighted mean ice thickness over grid cells in regions 1–5: sum(ice_thickness × grid_cell_area) / sum(grid_cell_area). Thickness from fusion of ICESat-2 and SMAP (thin ice).

    grid_mapping :

    crs

    long_name :

    mean sea ice thickness (area-weighted, NSIDC regions 1–5 only)

    units :

    m

    Array Chunk Bytes 6.39 kiB 4 B Shape (1635,) (1,) Dask graph 1635 chunks in 2 graph layers Data type float32 numpy.ndarray

  • snow_depth
    (time, y, x)
    float32
    dask.array<chunksize=(1, 448, 304), meta=np.ndarray>

    description :

    Mean snow depth across the grid cell (m). Uses CDR sea ice concentration in the solution; from NESOSIM v1.1 (IS-2 sub-sampled + GPSat interpolation).

    grid_mapping :

    crs

    long_name :

    snow depth on sea ice

    units :

    meters

    Array Chunk Bytes 849.43 MiB 532.00 kiB Shape (1635, 448, 304) (1, 448, 304) Dask graph 1635 chunks in 2 graph layers Data type float32 numpy.ndarray

  • region_mask
    (y, x)
    float32
    dask.array<chunksize=(448, 304), meta=np.ndarray>

    description :

    NSIDC-0780 sea ice region mask (sea_ice_region_surface_mask). CAA (Canadian Archipelago, value 12) is retained in gridded fields; see global attribute caa_uncertainty_caution.

    flag_values :

    0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 30, 31, 32

    grid_mapping :

    crs

    long_name :

    Northern Hemisphere region mask

    source :

    https://doi.org/10.5067/CYW3O8ZUNIWC

    Array Chunk Bytes 532.00 kiB 532.00 kiB Shape (448, 304) (448, 304) Dask graph 1 chunks in 2 graph layers Data type float32 numpy.ndarray

  • mean_snow_depth_regions_1_5
    (time)
    float32
    dask.array<chunksize=(1,), meta=np.ndarray>

    description :

    Area-weighted mean snow depth over grid cells in regions 1–5. Snow depth uses CDR SIC in the solution; from NESOSIM v1.1.

    grid_mapping :

    crs

    long_name :

    mean snow depth (area-weighted, NSIDC regions 1–5 only)

    units :

    m

    Array Chunk Bytes 6.39 kiB 4 B Shape (1635,) (1,) Dask graph 1635 chunks in 2 graph layers Data type float32 numpy.ndarray

  • sea_ice_conc
    (time, y, x)
    float32
    dask.array<chunksize=(1, 448, 304), meta=np.ndarray>

    description :

    Sea ice concentration for the 15th of the month from the NOAA/NSIDC Climate Data Record of Passive Microwave Sea Ice Concentration, daily product v6.

    grid_mapping :

    crs

    long_name :

    sea ice concentration

    source :

    NOAA/NSIDC Climate Data Record of Passive Microwave Sea Ice Concentration, daily v6. https://nsidc.org/data/G02202/

    units :

    1

    Array Chunk Bytes 849.43 MiB 532.00 kiB Shape (1635, 448, 304) (1, 448, 304) Dask graph 1635 chunks in 2 graph layers Data type float32 numpy.ndarray

  • snow_depth_unc
    (time, y, x)
    float32
    dask.array<chunksize=(1, 448, 304), meta=np.ndarray>

    description :

    Standard deviation of snow depth (sqrt of predictive variance from GP). Snow depth uses CDR SIC in the solution; from NESOSIM v1.1.

    grid_mapping :

    crs

    long_name :

    snow depth uncertainty (standard deviation)

    units :

    meters

    Array Chunk Bytes 849.43 MiB 532.00 kiB Shape (1635, 448, 304) (1, 448, 304) Dask graph 1635 chunks in 2 graph layers Data type float32 numpy.ndarray

  • sea_ice_volume
    (time, y, x)
    float32
    dask.array<chunksize=(1, 448, 304), meta=np.ndarray>

    description :

    Total amount of sea ice in the grid cell: thickness × grid_cell_area, in km³.

    grid_mapping :

    crs

    long_name :

    sea ice volume per grid cell

    units :

    km3

    Array Chunk Bytes 849.43 MiB 532.00 kiB Shape (1635, 448, 304) (1, 448, 304) Dask graph 1635 chunks in 2 graph layers Data type float32 numpy.ndarray

  • total_sea_ice_volume
    (time)
    float32
    dask.array<chunksize=(1,), meta=np.ndarray>

    description :

    Sum of sea_ice_volume over all grid cells (km³).

    grid_mapping :

    crs

    long_name :

    total sea ice volume

    units :

    km3

    Array Chunk Bytes 6.39 kiB 4 B Shape (1635,) (1,) Dask graph 1635 chunks in 2 graph layers Data type float32 numpy.ndarray

  • total_sea_ice_volume_regions_1_5
    (time)
    float32
    dask.array<chunksize=(1,), meta=np.ndarray>

    description :

    Sum of sea_ice_volume over grid cells in NSIDC-0780 regions 1–5 (Central Arctic through Laptev Sea), in km³.

    grid_mapping :

    crs

    long_name :

    total sea ice volume (NSIDC regions 1–5 only)

    units :

    km3

    Array Chunk Bytes 6.39 kiB 4 B Shape (1635,) (1,) Dask graph 1635 chunks in 2 graph layers Data type float32 numpy.ndarray

• Attributes: (14)

  caa_uncertainty_caution :

  Canadian Arctic Archipelago (CAA; NSIDC-0780 sea_ice_region_surface_mask value 12) is included in gridded fields but has elevated uncertainty relative to the open Arctic Ocean: complex coastlines, land-fast ice, and limited altimetry/auxiliary coverage. Prefer open-ocean regions for quantitative applications, and treat CAA values with extra caution.

  contact :

  Alek Petty (akpetty@umd.edu)

  description :

  Monthly gridded Arctic sea ice thickness (fusion of ICESat-2 and SMAP thin ice), freeboard, and snow depth from GP interpolation, with grid cell area and CDR sea ice concentration. NSIDC Polar Stereographic North (EPSG:3411).

  history :

  Created 2026-07-29

  inner_arctic_regions_note :

  Inner-Arctic scalars (mean_*_regions_1_5, total_sea_ice_volume_regions_1_5) use NSIDC-0780 regions 1–5 only (Central Arctic through Laptev Sea). Thickness/freeboard/snow means are area-weighted. total_sea_ice_volume is the pan-Arctic sum over all valid grid cells.

  n_timesteps :

  1635

  source :

  Combined from monthly IS2_interp_test_petty files (thickness/freeboard/snow_depth); grid area NSIDC0771; SIC CDR.

  time_coverage_end :

  2025-04-30

  time_coverage_start :

  2018-11-01

  training_season_caution :

  GP training is limited to a seasonal window (e.g. 1 Sep–30 Apr); there is no training data before 1 September or after 30 April in that setup. September and April sit at the edges of that window, so they have more limited in-window context than mid-season months—treat September and April fields with extra caution.

  product :

  IS2SMGPSIT

  product_version :

  v0.1

  release_status :

  pre-release (not peer reviewed)

  source_zarr :

  s3://icesat-2-sea-ice-us-west-2/is2smsitgp/dev_v7/zarr/GPSat_multivar_20181101-20250430.zarr

Pin an immutable release for reproducibility

Analyses supporting a publication should read from a tag, not main: the data at a tag never changes, even after new winters are appended. This is the version to cite.

session_release = repo.readonly_session(tag=TAG)
IS2_SMOS_SMAP_release = xr.open_zarr(session_release.store, consolidated=False)

print(f"tag {TAG!r}: {pd.Timestamp(IS2_SMOS_SMAP_release.time.values.min()):%Y-%m-%d} to "
      f"{pd.Timestamp(IS2_SMOS_SMAP_release.time.values.max()):%Y-%m-%d} "
      f"({IS2_SMOS_SMAP_release.sizes['time']} days)")

tag 'v0.1': 2018-11-01 to 2025-04-30 (1635 days)

Map one month

Select one calendar month, average its daily fields, and map the result. Change MAP_MONTH to any available month in YYYY-MM format. Only that month’s thickness chunks are read from the store.

MAP_MONTH = '2025-04'

thickness_month = (
    IS2_SMOS_SMAP.ice_thickness
    .sel(time=MAP_MONTH)
    .mean('time', keep_attrs=True)
)

longitude = IS2_SMOS_SMAP.longitude
latitude = IS2_SMOS_SMAP.latitude
if 'time' in longitude.dims:
    longitude = longitude.isel(time=0, drop=True)
if 'time' in latitude.dims:
    latitude = latitude.isel(time=0, drop=True)

fig, ax = plt.subplots(
    figsize=(5, 5),
    subplot_kw={'projection': ccrs.NorthPolarStereo(central_longitude=-45)},
)
im = ax.pcolormesh(
    longitude, latitude, thickness_month,
    transform=ccrs.PlateCarree(), cmap='viridis', vmin=0, vmax=5,
)
ax.add_feature(cfeature.LAND, color='0.92', zorder=5)
ax.coastlines(linewidth=0.3, zorder=3)
ax.set_extent([-179, 179, 54, 90], crs=ccrs.PlateCarree())
ax.set_title(pd.Timestamp(MAP_MONTH).strftime('%B %Y'), fontsize=10)
cbar = fig.colorbar(im, ax=ax, orientation='horizontal', shrink=0.65, pad=0.04, extend='max')
cbar.set_label('Sea ice thickness (m)')
plt.show()

Full daily time series

The store includes precomputed daily aggregates, so these plots need almost no data transfer: inner-Arctic-Ocean (NSIDC regions 1–5) mean thickness and pan-Arctic volume. Growth seasons only (May–August omitted) are shown with small markers so summer gaps do not connect across seasons.

thickness_daily = IS2_SMOS_SMAP.mean_sea_ice_thickness_regions_1_5
volume_daily = IS2_SMOS_SMAP.total_sea_ice_volume / 1e3

# Reindex to a continuous daily axis and drop May–August so seasons do not connect
t0 = pd.Timestamp(thickness_daily.time.values.min())
t1 = pd.Timestamp(thickness_daily.time.values.max())
all_days = pd.date_range(t0, t1, freq='D')
thickness_plot = thickness_daily.reindex(time=all_days)
thickness_plot = thickness_plot.where(~thickness_plot.time.dt.month.isin([5, 6, 7, 8]))
volume_plot = volume_daily.reindex(time=all_days)
volume_plot = volume_plot.where(~volume_plot.time.dt.month.isin([5, 6, 7, 8]))

fig, (ax0, ax1) = plt.subplots(2, 1, figsize=(9, 5.5), sharex=True)

thickness_plot.plot(
    ax=ax0, color='C1', marker='.', markersize=2, linestyle='None',
)
ax0.set_ylabel('Mean thickness (m)')
ax0.set_title('Inner Arctic Ocean mean sea ice thickness (daily)', fontsize=10)
ax0.set_xlabel('')

volume_plot.plot(
    ax=ax1, color='C0', marker='.', markersize=2, linestyle='None',
)
ax1.set_ylabel('Volume (1000 km$^3$)')
ax1.set_title('Pan-Arctic sea ice volume (daily, CAA excluded)', fontsize=10)
ax1.set_xlabel('')

for ax in (ax0, ax1):
    ax.grid(axis='y', alpha=0.4)
    ax.spines['top'].set_visible(False)
    ax.spines['right'].set_visible(False)

plt.tight_layout()
plt.show()

Seven-winter thickness maps

Resample the daily thickness fields to calendar-month means, average November–April for each growth season, and use the same staticArcticMaps_equal_panels layout as notebook 11e. This reads the gridded thickness record, so it is the most data-intensive example in the notebook.

from utils.plotting_utils import (
    staticArcticMaps_equal_panels,
    compute_gridcell_winter_means,
)

# Monthly means on month-start timestamps, matching notebook 11e.
thickness_monthly = (
    IS2_SMOS_SMAP.ice_thickness
    .resample(time='1MS')
    .mean(keep_attrs=True)
)

years = list(range(2018, 2025))
thickness_winter_means = compute_gridcell_winter_means(
    thickness_monthly, years=years,
)

# Fill missing open-water cells with zero, but keep land masked.
region_mask = IS2_SMOS_SMAP.region_mask
if 'time' in region_mask.dims:
    region_mask = region_mask.isel(time=0, drop=True)
ocean = region_mask.isin(list(range(1, 19))).compute()
thickness_winter_means = thickness_winter_means.fillna(0).where(ocean)

print('winters:', list(thickness_winter_means.time.values))

winters: [np.str_('Nov 2018 - Apr 2019'), np.str_('Nov 2019 - Apr 2020'), np.str_('Nov 2020 - Apr 2021'), np.str_('Nov 2021 - Apr 2022'), np.str_('Nov 2022 - Apr 2023'), np.str_('Nov 2023 - Apr 2024'), np.str_('Nov 2024 - Apr 2025')]

fig = staticArcticMaps_equal_panels(
    thickness_winter_means,
    dates=thickness_winter_means.time.values,
    title='',
    set_cbarlabel='Sea ice thickness (m)',
    cmap='viridis',
    vmin=0,
    vmax=5,
    ocean_mask=ocean.values,
    savefig=False,
)
fig

findfont: Failed to find font weight medium, now using 400.

Notes

• Anonymous (unauthenticated) reads work for all public prefixes; only dataset publishing requires AWS credentials.

• New winters are appended to main with scripts/icechunk_append_season.py (one commit + one new tag per release); previously tagged releases are never modified.

• The single-month map reads only one month of daily gridded thickness; the full time series use compact precomputed aggregates. The seven-winter map reads the complete gridded thickness record.

• The plain (non-versioned) Zarr snapshot of each release remains available alongside the Icechunk repository for tools that cannot read Icechunk — see the data access notes in notebook 11a.