CCMP V3.1¶

RSS Cross-Calibrated Multi-Platform (CCMP) Ocean Vector Wind Analysis, Version 3.1

Available

At a glance¶

Note on period: V3.x starts in January 1993 because the OSCAR ocean current product used to adjust the ERA5 winds begins in 1993. The 'Temporal Coverage' field on the RSS product page (1993-01-02 to 2019-12-30) and the text (1993-01 to 2023-07) have not been updated; when checked on 2026-10-13, files actually ran through 2026-08-31. The previous V2.0 started in July 1987.

Overview¶

A Level 4 product that inter-calibrates wind speeds from microwave radiometers (SSM/I, SSMIS, AMSR-E, AMSR2, TMI, GMI, WindSat) and wind vectors from scatterometers (QuikSCAT, ASCAT), then blends them with adjusted ERA5 10 m neutral winds as the background using a variational analysis method (VAM), giving gap-free 6-hourly 0.25° ocean vector wind fields. It was originally developed at NASA GSFC (Atlas et al., 2011) and has been produced by RSS since V2.0. V3.0 switched the background from ERA-Interim to ERA5 and improved high-wind accuracy and long-term stability; V3.1 adds AMSR2 V8.2 and ASCAT-B from July 2012 onward and allows low-latency updates. It is widely used to force ocean circulation models, to compute wind stress and stress curl, and to evaluate reanalysis winds.

What changes with the successor¶

CCMP V2.0 → CCMP V3.0¶

Transition: 2023-02

The background changed from ERA-Interim to adjusted ERA5 neutral winds, buoy assimilation was dropped, GMI was added, and radiometer winds and ERA5 were adjusted before the analysis. The record now starts in 1993 instead of 1987, with better high-wind accuracy and long-term stability.

Sources: RSS — CCMP V3.0 release announcement (2023-02-13) · Mears et al. (2022), Remote Sensing · Mears et al. (2019), JGR Oceans (V2.0 and NRT comparison) · CCMP V3.1 User Guide (version comparison table)

CCMP V3.0 → CCMP V3.1¶

Transition: 2023-08

Same processing system, with inputs changed only from July 2012 onward. Adding AMSR2 V8.2 and ASCAT-B improves winds after that date, and using near-real-time OSCAR lets the record be extended continuously with low latency.

Sources: RSS — CCMP V3.1 release announcement (2023-08-25) · RSS CCMP product page

Available variables¶

Variable names follow the CCMP V3.1 User Guide and the PO.DAAC variable list. All winds are 10 m neutral-stability winds; u and v are relative to true north and the direction follows the oceanographic convention (the direction the wind blows toward).

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

6-hourly analyses (L4, daily files)¶

Monthly means (L4.5)¶

The full variable list is in the official documentation.

How to get the data¶

• RSS data server (HTTPS) — Direct HTTPS download (netCDF-4, e.g. v03.1/Y2026/M08/CCMP_Wind_Analysis_20260801_V03.1_L4.nc; monthly CCMP_Wind_Analysis_202608_monthly_mean_V03.1_L4.nc) No login needed
  • Folders for v03.1 (current), v03.0 and v02.0 (no longer updated) and v02.1.NRT (near real time). Example reading code is in read_routines/ccmp.py. FTP access requires free registration.
• NASA PO.DAAC — CCMP V3.1 6-hourly (CCMP_WINDS_10M6HR_L4_V3.1) — Earthdata Search / earthaccess / OPeNDAP / Harmony subsetting / podaac-data-subscriber
  • The monthly product is CCMP_WINDS_10MMONTHLY_L4_V3.1. A free NASA Earthdata Login is required.
• RSS CCMP product page — Web (input satellites, version changes, citation text, preview images)

Example code¶

Downloads PO.DAAC files with earthaccess. Run earthaccess.login() once with your NASA Earthdata account first. To download from RSS without logging in, use the file paths shown under access.

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

earthaccess.login()
results = earthaccess.search_data(
    short_name="CCMP_WINDS_10M6HR_L4_V3.1",   # monthly: CCMP_WINDS_10MMONTHLY_L4_V3.1
    version="3.1",
    temporal=("2025-01-01", "2025-01-31"),
)
files = earthaccess.download(results, "./ccmp")

ds = xr.open_mfdataset(files, combine="by_coords")
ws = ds["ws"].sel(latitude=slice(30, 40), longitude=slice(120, 135))  # around the Korean Peninsula
obs_based = ws.where(ds["nobs"].sel(latitude=slice(30, 40), longitude=slice(120, 135)) > 0)
print(ws.mean("time").values.mean(), float(obs_based.count() / ws.count()))

Practical tips¶

• Cells with nobs = 0 hold only the adjusted ERA5 background, with no satellite observations. If your analysis is about what the satellites add, filter with nobs.
• Small, intense wind features such as typhoons and other tropical cyclones are poorly resolved in ERA5, and satellite retrievals are often lost to rain, so RSS says not to use CCMP to analyse these events.
• Above about 15–18 m/s, CCMP is intentionally biased high relative to buoys and ASCAT-C (reflecting high-wind calibration of scatterometers and radiometers). Keep this in mind when comparing with buoys.
• Multi-decadal trends in global or basin-mean wind speed can be comparable to the dataset's uncertainty, and RSS urges caution. Regional changes larger than a few tenths of a metre per second are more trustworthy.
• V3.1, V3.0, V2.0, V1.1 and the near-real-time CCMP-NRT differ in background, inputs and adjustments, so do not splice them into one time series. V3.1 and V3.0 are identical before 2012-07-02.
• ASCAT-C winds and moored buoy winds were withheld from the V3.x analysis for independent validation, so comparing CCMP with them is an independent check.
• Each monthly RSS update leaves out the last day of the month, which arrives with the next update. When downloading the latest month, check that the last-day file is there.
• Values exist over land (from the background), but RSS advises against using them. Satellite retrievals influenced by sea ice are excluded from the analysis, so values in ice-covered areas are also close to the background.
• The monthly w (scalar mean speed) can differ greatly from the speed computed from the vector-mean u and v. For wind stress or energy calculations, it is safer to work from the 6-hourly files.
• The dataset DOIs on the RSS page (10.56236/RSS-uv6h30, RSS-uv1m30) are labelled Version 3.0. If you used V3.1, cite the PO.DAAC V3.1 DOI as well to make the version clear.

Notices and change history¶

References for your paper¶

Dataset citation¶

CCMP V3.1 6-hourly (PO.DAAC) — DOI 10.5067/CCMP-6HW10M-L4V31

@misc{rss2024ccmp6h,
  author    = {{Remote Sensing Systems}},
  title     = {{RSS} {CCMP} 6-Hourly 10 Meter Surface Winds Level 4 Version 3.1},
  year      = {2024},
  publisher = {NASA Physical Oceanography Distributed Active Archive Center},
  version   = {3.1},
  doi       = {10.5067/CCMP-6HW10M-L4V31}
}

CCMP V3.1 monthly (PO.DAAC) — DOI 10.5067/CCMP-MW10M-L4V31

@misc{rss2024ccmpmonthly,
  author    = {{Remote Sensing Systems}},
  title     = {{RSS} {CCMP} Monthly 10 Meter Surface Winds Level 4 Version 3.1},
  year      = {2024},
  publisher = {NASA Physical Oceanography Distributed Active Archive Center},
  version   = {3.1},
  doi       = {10.5067/CCMP-MW10M-L4V31}
}

RSS recommended citation (6-hourly; DOI labelled Version 3.0) — DOI 10.56236/RSS-uv6h30

@misc{mears2022ccmpv3,
  author    = {Mears, Carl and Lee, Tong and Ricciardulli, Lucrezia and Wang, Xiaochun and Wentz, Frank},
  title     = {{RSS} Cross-Calibrated Multi-Platform ({CCMP}) 6-hourly ocean vector wind analysis on 0.25 deg grid, Version 3.0},
  year      = {2022},
  publisher = {Remote Sensing Systems},
  address   = {Santa Rosa, CA},
  doi       = {10.56236/RSS-uv6h30}
}

Reference papers¶

CCMP V3.0 improvements (main V3.x reference) — DOI 10.3390/rs14174230

@article{mears2022ccmp,
  author  = {Mears, Carl and Lee, Tong and Ricciardulli, Lucrezia and Wang, Xiaochun and Wentz, Frank},
  title   = {Improving the Accuracy of the Cross-Calibrated Multi-Platform ({CCMP}) Ocean Vector Winds},
  journal = {Remote Sensing},
  volume  = {14},
  number  = {17},
  pages   = {4230},
  year    = {2022},
  doi     = {10.3390/rs14174230}
}

Original CCMP paper — DOI 10.1175/2010BAMS2946.1

@article{atlas2011ccmp,
  author  = {Atlas, Robert and Hoffman, Ross N. and Ardizzone, Joseph and Leidner, S. Mark and Jusem, Juan Carlos and Smith, Deborah K. and Gombos, Daniel},
  title   = {A Cross-calibrated, Multiplatform Ocean Surface Wind Velocity Product for Meteorological and Oceanographic Applications},
  journal = {Bulletin of the American Meteorological Society},
  volume  = {92},
  number  = {2},
  pages   = {157--174},
  year    = {2011},
  doi     = {10.1175/2010BAMS2946.1}
}

Near-real-time CCMP-NRT — DOI 10.1029/2019JC015367

@article{mears2019ccmpnrt,
  author  = {Mears, Carl A. and Scott, Joel and Wentz, Frank J. and Ricciardulli, Lucrezia and Leidner, S. Mark and Hoffman, Ross and Atlas, Robert},
  title   = {A Near-Real-Time Version of the Cross-Calibrated Multiplatform ({CCMP}) Ocean Surface Wind Velocity Data Set},
  journal = {Journal of Geophysical Research: Oceans},
  volume  = {124},
  pages   = {6997--7010},
  year    = {2019},
  doi     = {10.1029/2019JC015367}
}

Acknowledgements¶

RSS asks for the statement "CCMP Version-3.1 vector wind analyses are produced by Remote Sensing Systems. Data are available at www.remss.com." RSS also asks users to email support@remss.com with the journal name and expected publication date of papers that use the data, since continued funding depends on this.

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¶

• RSS — CCMP product page
• PO.DAAC — CCMP V3.1 6-hourly
• PO.DAAC — CCMP V3.1 monthly
• CCMP V3.1 User Guide (PDF)
• RSS — Terms of use
• This site — Metop ASCAT ocean winds (a CCMP input scatterometer)