CLMS LAI/FAPAR/FCOVER¶

Copernicus Land Monitoring Service global vegetation biophysical variables LAI, FAPAR and FCover (300 m V2.0 / 1 km V2, 10-daily) — formerly Copernicus Global Land Service (CGLS) products

Available

At a glance¶

Note on period: The 300 m V2.0 runs from January 2014 to the present, using PROBA-V for 2014–2018 and Sentinel-3 OLCI from January 2019 (in CDSE file names the last PROBAV product is 2018-12-31 and the first OLCI product is 2019-01-10). The 1 km V2 (SPOT-VGT, then PROBA-V) is an archive from January 1999 to June 2020; the earlier 300 m V1 series (V1.0 PROBA-V 2014-01 to 2020-06, V1.1 Sentinel-3 2020-07 to 2026-02) also remain on CDSE. On 2026-10-13 the newest V2.0 file on CDSE was the RT0 for the nominal date 2026-10-10.

Overview¶

Global vegetation variables from the EU Copernicus Land Monitoring Service (CLMS): leaf area index (LAI), fraction of absorbed photosynthetically active radiation (FAPAR) and fraction of green vegetation cover (FCover), every 10 days. Neural networks trained per vegetation type turn satellite surface reflectances into instantaneous estimates, which are then fitted in time and gap-filled to give dekadal values. The flagship is the 300 m V2.0, which joins PROBA-V (2014–2018) and Sentinel-3 OLCI (2019 onwards) in a single version; the 1999–2020 period is covered by the 1 km V2 (GEOV2) record. These products used to be distributed by VITO under the 'Copernicus Global Land Service (CGLS)' name and are now obtained from the CLMS website and CDSE.

What changes with the successor¶

CLMS LAI, FAPAR, FCover 300 m V1 (V1.0 PROBA-V / V1.1 Sentinel-3) → CLMS LAI, FAPAR, FCover 300 m V2.0¶

Transition: 2026-03

V2.0 rebuilds the separate V1.0 (PROBA-V) and V1.1 (Sentinel-3) records as one version. The Sentinel-3 period uses the same neural networks, so values barely change, but the PROBA-V period (2014–2018) uses new input reflectances and FCover and, to a lesser extent, FAPAR were adjusted. NRT operations began in January 2026 and V1.1 ended at the end of February 2026.

Sources: Version 2.0 of LAI, FAPAR and FCover 300 m in NRT operations (CLMS, 2026-01-16) · CLMS PUM LAI, FAPAR, FCover 300 m V2.0 (I1.20, 2026-07-16) · CLMS Validation Report LAI, FAPAR, FCover 300 m V2.0 (2026-03-09)

CLMS LAI, FAPAR, FCover 1 km V2 (GEOV2, SPOT-VGT and PROBA-V) → CLMS LAI, FAPAR, FCover 300 m series¶

Transition: 2020-06

The 1 km V2 record, running since 1999, stopped when the PROBA-V mission ended in June 2020 and is now an archive. Only the 300 m series (currently V2.0) is updated since then. The two series differ in resolution, inputs and compositing windows and cannot be joined as one time series.

Sources: CLMS product page — Leaf Area Index 1999-2020 (1 km) V2 · CDSE documentation — lai_global_1km_10daily_v2 bands · Verger et al. (2023), GEOV2

Available variables¶

netCDF4 (CF-1.6) layers from the 300 m V2.0 Product User Manual (PUM). LAI, FAPAR and FCover come in separate files (LAI300, FAPAR300, FCOVER300) carrying the same auxiliary layers. Everything is stored as 8-bit integers: physical value = stored value × scale_factor (LAI 1/30, FAPAR and FCover 1/250, others 1), with 255 as missing. Water pixels are not processed and are set to 255. The CDSE STAC metadata list the same layers in lower case (lai, rmse, ...), so check the names in an actual file.

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

300 m V2.0 vegetation variables (estimates for the nominal date)¶

300 m V2.0 auxiliary layers¶

1 km V2 (1999-01 to 2020-06, archive)¶

The full variable list is in the official documentation.

How to get the data¶

• CLMS product page — LAI 300 m V2.0 — Product description, DOI and access routes; CLMS Data viewer
  • FAPAR (fraction-of-absorbed-photosynthetically-active-radiation-v2-0-300m) and FCover (fraction-of-green-vegetation-cover-v2-0-300m) have their own pages. CDSE download options are recommended for large volumes.
• Copernicus Data Space Ecosystem — OData, S3, STAC — OData API / S3 (EODATA) / STAC / Copernicus Browser (free CDSE account)
  • The OData datasetIdentifier values are lai_global_300m_10daily_v2, fapar_global_300m_10daily_v2 and fcover_global_300m_10daily_v2 (and lai_global_1km_10daily_v2 etc. for 1 km); the S3 path is /eodata/CLMS/bio-geophysical/vegetation_properties/lai_global_300m_10daily_v2/YYYY/MM/DD/. Both netCDF (_nc) and COG (_cog) products are available.
• CDSE openEO / Sentinel Hub — openEO (CLMS_LAI_GLOBAL_300M_10DAILY_V2 etc.) / Sentinel Hub BYOC collections (separate RT0, RT1, RT2, RT6)
  • Each global file is several GB, so these are handy for cutting out an area of interest. CDSE openEO applies scale and offset automatically.
• WEkEO (unified CLMS dataset) — HDA API / web viewer
  • WEkEO says it serves CLMS data moved to CDSE through a unified dataset (EO:JRC:DAT:CLMS), but the help article checked on 2026-10-13 does not name LAI, FAPAR or FCover. CDSE is the primary distribution point.

Example code¶

Finds and downloads the final (RT6, netCDF) LAI 300 m V2.0 product for the last dekad of July 2024 through the CDSE OData API. Put a free CDSE account in the CDSE_USERNAME and CDSE_PASSWORD environment variables. One global netCDF is about 2.8 GB, so for a small area the COG (_cog) products or openEO are better choices.

import os
import requests
import xarray as xr

token = requests.post(
    "https://identity.dataspace.copernicus.eu/auth/realms/CDSE/protocol/openid-connect/token",
    data={"client_id": "cdse-public", "grant_type": "password",
          "username": os.environ["CDSE_USERNAME"], "password": os.environ["CDSE_PASSWORD"]},
).json()["access_token"]

odata = "https://catalogue.dataspace.copernicus.eu/odata/v1"
flt = ("Collection/Name eq 'CLMS' and "
       "Attributes/OData.CSC.StringAttribute/any(att:att/Name eq 'datasetIdentifier' and "
       "att/OData.CSC.StringAttribute/Value eq 'lai_global_300m_10daily_v2') and "
       "contains(Name,'RT6_20240731') and endswith(Name,'_nc')")
prod = requests.get(f"{odata}/Products", params={"$filter": flt}).json()["value"][0]
name = prod["Name"]                       # c_gls_LAI300-RT6_202407310000_GLOBE_OLCI_V2.0.1_nc
ncfile = name[:-3] + ".nc"                # actual file inside the product folder

s = requests.Session()
s.headers["Authorization"] = f"Bearer {token}"
url = f"{odata}/Products({prod['Id']})/Nodes({name})/Nodes({ncfile})/$value"
r = s.get(url, allow_redirects=False)
while r.status_code in (301, 302, 303, 307):   # keep the auth header across the redirect to the download server
    url = r.headers["Location"]
    r = s.get(url, allow_redirects=False)
with s.get(url, stream=True) as resp, open(ncfile, "wb") as f:
    for chunk in resp.iter_content(1 << 20):
        f.write(chunk)

ds = xr.open_dataset(ncfile)              # xarray applies scale_factor and the 255 fill value
lai = ds["LAI"].sel(lat=slice(39, 33), lon=slice(124, 132))   # latitude runs north to south
print(float(lai.mean()))

Practical tips¶

• RT0, RT1, RT2 and RT6 give different values for the same nominal date. Use RT6 only for research time series and say so if the last few dekads were filled with RT0–RT2. RT0 is known to show spatial noise such as stripes and unrealistically high values.
• Do not mix V1 (V1.0, V1.1) with V2.0 in one time series. The 2014–2018 PROBA-V values in particular changed in V2.0, so re-download the whole record as V2.0.
• V2.0 still shows a moderate level shift (OLCI slightly higher) across the PROBA-V to Sentinel-3 OLCI transition in 2018–2019. Check this boundary when computing trends from 2014.
• Multiply LAI by 1/30 and FAPAR or FCover by 1/250 to get physical values. xarray does this automatically, but reading the COGs with GDAL or rasterio returns the stored values (0–250). 255 means missing or water.
• Nominal dates fall on the 10th, 20th and last day, so the third dekad is 8–11 days long. Consider weighting by dekad length rather than taking a plain mean when building monthly values.
• Unrealistic values can occur over wetlands and near water, and evergreen broadleaf forest (QFLAG bit 2) uses a simplified treatment (mean of 10 instantaneous values, or the previous dekad when too few are available).
• The C3S Climate Data Store dataset 'satellite-lai-fapar' has a similar name but is a different line (produced for C3S, with effective LAI from V2 onwards). Cite the exact product name and DOI you used.

Notices and change history¶

References for your paper¶

Dataset citation¶

LAI 300 m V2.0 (CLMS) — DOI 10.2909/a696fab4-c3dc-4dcc-b1d8-6fcd9a2ce1eb

@misc{clms_lai300_v2,
  author    = {{Copernicus Land Monitoring Service}},
  title     = {Leaf Area Index 2014-present (raster 300 m), global, 10-daily - version 2},
  year      = {2026},
  publisher = {Copernicus Land Monitoring Service},
  doi       = {10.2909/a696fab4-c3dc-4dcc-b1d8-6fcd9a2ce1eb},
  note      = {European Union's Copernicus Land Monitoring Service information}
}

FAPAR 300 m V2.0 (CLMS) — DOI 10.2909/c828c1b2-6ea8-40f3-897d-c92080ba1daf

@misc{clms_fapar300_v2,
  author    = {{Copernicus Land Monitoring Service}},
  title     = {Fraction of Absorbed Photosynthetically Active Radiation 2014-present (raster 300 m), global, 10-daily - version 2},
  year      = {2026},
  publisher = {Copernicus Land Monitoring Service},
  doi       = {10.2909/c828c1b2-6ea8-40f3-897d-c92080ba1daf},
  note      = {European Union's Copernicus Land Monitoring Service information}
}

FCover 300 m V2.0 (CLMS) — DOI 10.2909/e7b99351-7a89-42c7-8bba-1d9fc0463946

@misc{clms_fcover300_v2,
  author    = {{Copernicus Land Monitoring Service}},
  title     = {Fraction of Vegetation Cover 2014-present (raster 300 m), global, 10-daily - version 2},
  year      = {2026},
  publisher = {Copernicus Land Monitoring Service},
  doi       = {10.2909/e7b99351-7a89-42c7-8bba-1d9fc0463946},
  note      = {European Union's Copernicus Land Monitoring Service information}
}

LAI 1 km V2 (CLMS, 1999–2020) — DOI 10.2909/d5fdc595-2e03-4cbe-a39e-5f006f9cef07

@misc{clms_lai1km_v2,
  author    = {{Copernicus Land Monitoring Service}},
  title     = {Leaf Area Index 1999-2020 (raster 1 km), global, 10-daily - version 2},
  year      = {2017},
  publisher = {Copernicus Land Monitoring Service},
  doi       = {10.2909/d5fdc595-2e03-4cbe-a39e-5f006f9cef07},
  note      = {European Union's Copernicus Land Monitoring Service information}
}

Reference papers¶

Quality assessment of the 300 m PROBA-V products — DOI 10.3390/rs12061017

@article{fuster2020cgls300,
  author  = {Fuster, Beatriz and S{\'a}nchez-Zapero, Jorge and Camacho, Fernando and Garc{\'i}a-Santos, Vicente and Verger, Aleixandre and Lacaze, Roselyne and Weiss, Marie and Baret, Fr{\'e}d{\'e}ric and Smets, Bruno},
  title   = {Quality Assessment of {PROBA-V} {LAI}, {fAPAR} and {fCOVER} Collection 300 m Products of {Copernicus} Global Land Service},
  journal = {Remote Sensing},
  volume  = {12},
  number  = {6},
  pages   = {1017},
  year    = {2020},
  doi     = {10.3390/rs12061017}
}

1 km V2 (GEOV2) algorithm and validation — DOI 10.1016/j.jag.2023.103479

@article{verger2023geov2,
  author  = {Verger, Aleixandre and S{\'a}nchez-Zapero, Jorge and Weiss, Marie and Descals, Adri{\`a} and Camacho, Fernando and Lacaze, Roselyne and Baret, Fr{\'e}d{\'e}ric},
  title   = {{GEOV2}: Improved smoothed and gap filled time series of {LAI}, {FAPAR} and {FCover} 1 km {Copernicus} Global Land products},
  journal = {International Journal of Applied Earth Observation and Geoinformation},
  volume  = {123},
  pages   = {103479},
  year    = {2023},
  doi     = {10.1016/j.jag.2023.103479}
}

Acknowledgements¶

The CLMS data policy asks publications to include "European Union's Copernicus Land Monitoring Service information" and the relevant dataset DOIs, and dataset citations to give the product title, that phrase, the name and URL of the access point, the access date and the DOI. Look up the DOIs of the FAPAR and FCover 1 km V2 and the 300 m V1 products on their CLMS product pages.

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¶

• CLMS product page — LAI 300 m V2.0
• CLMS product page — FAPAR 300 m V2.0
• CLMS product page — FCover 300 m V2.0
• CLMS product page — LAI 300 m V1 (2014–2026, superseded)
• 300 m V2.0 Product User Manual (PUM)
• 300 m V2.0 Validation Report
• CDSE documentation — accessing CLMS data (OData, S3, STAC)
• CLMS data policy (citation wording)
• Old Global Land page address mapping