GK-2A (GEO-KOMPSAT-2A)¶

instantaneous mean accumulated maximum minimum static — instantaneous: value at the time stamp · mean/accumulated/maximum/minimum: over the indicated interval · static: does not change with time

Clouds (instantaneous at observation time, 2 km)¶

Precipitation (instantaneous at observation time)¶

Land and ocean (clear pixels, instantaneous at observation time)¶

Radiation (instantaneous at observation time)¶

Aerosol and atmosphere¶

The full variable list is in the official documentation.

How to get the data¶

• KMA Open MET Data Portal (satellite data downloads) — Web (register and log in, then search and download)
  • The official download point for GK-2A data since the end of 2025; NMSC directs users there for satellite data downloads. Check the menu location, archive period and product list on the portal itself (it blocks automated access, so they could not be confirmed for this page). The portal is mainly in Korean.
• KMA API Hub (satellite data API) — Open API (register → issue an authentication key → call)
  • The NMSC website Open API ended in December 2025 and moved here. Existing users must also get a new key from the API Hub and change their request URLs. Check the API list, request format and call limits on the API Hub.
• NOAA Open Data Dissemination — AWS bucket noaa-gk2a-pds (level 1B) — S3 (anonymous access)
  • A copy uploaded by NOAA in coordination with KMA. It holds level-1B channel data, not level-2 products, under AMI/L1B/FD/ (full disk) and AMI/L1B/LA/ (local area) in year-month/day/hour folders from February 2023. An SNS topic announces new files.
• NMSC — product Algorithm Theoretical Basis Documents (ATBD) — Web (PDF, in Korean)
  • Per-product definitions, variables and quality information.
• NMSC — software and sample data (lat/lon grids, calibration tables, sample code) — Web
  • Latitude-longitude files (NetCDF) for converting geostationary-projection pixels, land-sea masks, L1B Python and Fortran samples, and a receive-and-display Docker package (September 2026 release).

Example code¶

Opens one full-disk level-1B infrared (10.5 µm) image from AWS without an account. Level-2 products must come from the Open MET Data Portal or the API Hub; confirm variable names in the files. Use the latitude-longitude grid files on the NMSC software page to geolocate pixels.

# pip install s3fs xarray h5netcdf
import s3fs
import xarray as xr

fs = s3fs.S3FileSystem(anon=True)
# Full-disk (FD) L1B, ir105 channel files in the 2026-10-01 00h folder
keys = fs.glob("noaa-gk2a-pds/AMI/L1B/FD/202610/01/00/gk2a_ami_le1b_ir105_fd020ge_*.nc")
print(keys[:3])

ds = xr.open_dataset(fs.open(keys[0]), engine="h5netcdf")
print(ds)                                  # variable names, dimensions, attributes
print(ds.attrs)                            # global attributes such as observation time and calibration

Practical tips¶

• Data ordering and downloads on the NMSC website (datasvc.nmsc.kma.go.kr) and its Open API ended at the end of 2025. Download scripts written for the old addresses must be rewritten for the Open MET Data Portal and the API Hub.
• Level-2 products are arrays of 2 km pixels in the geostationary (GEOS) projection. Attach coordinates with the latitude-longitude files distributed by NMSC, and remember that pixels grow large near the disk edge (high latitudes, distant longitudes) because of the large viewing angle.
• Sea and land surface temperatures exist only for clear pixels, so averages over cloudy periods (the summer monsoon, typhoons) are biased toward clear days. Always use the quality flags (DQF).
• There are no observations during routine maintenance (usually May and November). Check for missing times when building daily means from 10-minute data.
• WMO OSCAR lists GK-2A's expected end of life as 2029 or later and the follow-on GEO-KOMPSAT-5 as 2031–2041. For long-term studies, plan for a discontinuity at the satellite changeover.
• Many studies compare or combine GK-2A with Japan's Himawari (AHI), which has similar time resolution. The channel sets are similar, but calibration and retrieval algorithms differ.

Notices and change history¶

References for your paper¶

Reference papers¶

GK-2A meteorological products overview — DOI 10.1007/s13143-020-00199-x

@article{chung2020gk2a,
  author  = {Chung, Sung-Rae and Ahn, Myoung-Hwan and Han, Kyung-Soo and Lee, Kyu-Tae and Shin, Dong-Bin},
  title   = {Meteorological Products of {Geo-KOMPSAT 2A} ({GK2A}) Satellite},
  journal = {Asia-Pacific Journal of Atmospheric Sciences},
  volume  = {56},
  number  = {2},
  year    = {2020},
  doi     = {10.1007/s13143-020-00199-x}
}

AMI in-orbit tests and performance validation — DOI 10.3390/rs13071303

@article{kim2021ami,
  author  = {Kim, Dohyeong and Gu, Minju and Oh, Tae-Hyeong and Kim, Eun-Kyu and Yang, Hye-Ji},
  title   = {Introduction of the Advanced Meteorological Imager of {Geo-Kompsat-2a}: In-Orbit Tests and Performance Validation},
  journal = {Remote Sensing},
  volume  = {13},
  number  = {7},
  pages   = {1303},
  year    = {2021},
  doi     = {10.3390/rs13071303}
}

Acknowledgements¶

The data have no DOI. In publications, state that the data were provided by the KMA National Meteorological Satellite Center (NMSC), how you obtained them (Open MET Data Portal, KMA API Hub or NOAA Open Data Dissemination), and the product name and version (algorithm version in the file attributes). NMSC works require specific attribution of the source.

Information on the datasets used in this study was compiled with the help of the Korean Climate Data Navigator (KCDN) (Sim, 2026; https://doi.org/10.5281/zenodo.23300923).

Links¶

• NMSC website (English)
• GK-2A introduction (NMSC)
• GK2A Fact Sheets (PDF set)
• GK2A AMI Rapid Scan service
• WMO OSCAR — GEO-KOMPSAT-2A
• AWS Registry of Open Data — KMA GK-2A