MERRA-2¶

Modern-Era Retrospective analysis for Research and Applications, Version 2

Available

At a glance¶

Note on period: Processed one month at a time and usually released around the 15th–20th of the following month (about a three-week lag); observation or computing issues can cause delays.

Overview¶

MERRA-2 is NASA GMAO's global reanalysis from 1980 onward, produced with the GEOS data assimilation system; it replaces the original MERRA. It assimilates newer hyperspectral infrared and microwave radiances and GPS radio occultation data, and is the first long-term global reanalysis to assimilate space-based observations of aerosols, so aerosol concentrations and optical depths are provided alongside the meteorology. Files are organised in collections (e.g. M2T1NXSLV), each with its own citable DOI.

Available variables¶

Main variables from the GES DISC collection pages and the File Specification. Names are the netCDF variable names; the collection is given in each group name.

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

Single-level diagnostics (M2T1NXSLV, hourly average)¶

Surface flux diagnostics (M2T1NXFLX, hourly average)¶

Pressure levels (M2I3NPASM, 3-hourly instantaneous, 42 levels)¶

Aerosol diagnostics (M2T1NXAER, hourly average)¶

The full variable list is in the official documentation.

How to get the data¶

• NASA GES DISC — MERRA-2 datasets — Web download / Subset / OPeNDAP
  • Requires a NASA Earthdata Login account with the "NASA GESDISC DATA ARCHIVE" application approved in the account profile. On-premises servers are being shut down from September 30, 2026; use Earthdata Cloud URLs (data.gesdisc.earthdata.nasa.gov).
• NASA Earthdata — MERRA-2 project page — Collection information / Earthdata Search / earthaccess
  • Collects per-collection descriptions, citation formats and reprocessing records. File lists are best obtained through CMR searches or earthaccess.
• Earthdata Cloud (AWS us-west-2 S3) — Direct S3 access
  • Best for computing inside the AWS us-west-2 region. Temporary S3 credentials are obtained with Earthdata Login.
• GMAO — MERRA-2 data access — Information page (links to tools such as the FLUID visualisation)

Example code¶

earthaccess prompts for Earthdata Login credentials or reads ~/.netrc. Specify both the collection short name and the version (5.12.4).

# pip install earthaccess xarray netCDF4
import earthaccess
import xarray as xr

earthaccess.login()  # Earthdata Login
results = earthaccess.search_data(
    short_name="M2T1NXSLV",
    version="5.12.4",
    temporal=("2024-01-01", "2024-01-02"),
)
files = earthaccess.download(results, "./merra2")
ds = xr.open_mfdataset(files)
print(ds["T2M"])

Practical tips¶

• The file-name prefix (MERRA2_100/200/300/400) is the production stream (1980–1991, 1992–2000, 2001–2010, 2011 onward). Reprocessed months get a higher number (e.g. September 2020 is MERRA2_401), so scripts that build file names from a fixed pattern can break.
• Precipitation comes in two versions: PRECTOT from the atmospheric model physics and the observation-corrected PRECTOTCORR. The corrected precipitation is what forces the land surface model; choose the one that fits your purpose and state which you used.
• Pressure levels that lie below the surface (e.g. 1000 or 850 hPa over high terrain) are undefined. Area means that include mountains can differ from other reanalyses depending on how these missing values are handled, so check.
• GES DISC on-premises THREDDS was discontinued on July 31, 2026, on-premises OPeNDAP was turned off between August 20 and September 30, and on-premises HTTPS is being retired after September 30. Move scripts that rely on goldsmr4/goldsmr5 URLs or recursive wget of whole directories to earthaccess, CMR queries or opendap.earthdata.nasa.gov.

Notices and change history¶

References for your paper¶

Dataset citation¶

Hourly single-level diagnostics (M2T1NXSLV) — DOI 10.5067/VJAFPLI1CSIV

@misc{merra2_m2t1nxslv,
  author    = {{Global Modeling and Assimilation Office (GMAO)}},
  title     = {{MERRA-2} tavg1\_2d\_slv\_Nx: 2d,1-Hourly,Time-Averaged,Single-Level,Assimilation,Single-Level Diagnostics V5.12.4},
  year      = {2015},
  publisher = {Goddard Earth Sciences Data and Information Services Center (GES DISC)},
  address   = {Greenbelt, MD, USA},
  doi       = {10.5067/VJAFPLI1CSIV},
  note      = {Accessed: [Data Access Date]}
}

3-hourly pressure-level assimilated fields (M2I3NPASM) — DOI 10.5067/QBZ6MG944HW0

Monthly single-level diagnostics (M2TMNXSLV) — DOI 10.5067/AP1B0BA5PD2K

Reference papers¶

MERRA-2 overview paper — DOI 10.1175/JCLI-D-16-0758.1

@article{gelaro2017merra2,
  author  = {Gelaro, Ronald and McCarty, Will and Su{\'a}rez, Max J. and Todling, Ricardo and Molod, Andrea and Takacs, Lawrence and Randles, Cynthia A. and Darmenov, Anton and Bosilovich, Michael G. and Reichle, Rolf and others},
  title   = {The Modern-Era Retrospective Analysis for Research and Applications, Version 2 ({MERRA-2})},
  journal = {Journal of Climate},
  volume  = {30},
  number  = {14},
  pages   = {5419--5454},
  year    = {2017},
  doi     = {10.1175/JCLI-D-16-0758.1}
}

Acknowledgements¶

GMAO notes that each MERRA-2 file collection has its own citable DOI and asks that the collections you used be cited in your reference list. The three above are examples; find the DOIs of other collections in the table on GMAO's "Citing MERRA-2 Data" page or under the Data Citation tab of each GES DISC dataset page.

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¶

• GMAO — MERRA-2
• MERRA-2 File Specification (GMAO Office Note No. 9)
• Citing MERRA-2 Data (DOI table by collection)
• MERRA-2 FAQ
• GMAO Production Status
• MERRA-21C (building on MERRA-2, ~25 km, from 1998)
• NCAR Climate Data Guide — MERRA-2