GLEAM4¶

Global Land Evaporation Amsterdam Model (GLEAM) 4.3 (satellite-based global land evaporation and soil moisture, 0.1°, daily)

Available

At a glance¶

Note on period: 4.3a covers 1980-01-01 to 2024-12-31 (44 years). 4.3b starts on 2003-01-01; the news item (2026-09-21) says it runs to the end of 2025, while the dataset description says 2024-12-31. Records are normally extended to the previous year around April, but on 2026-10-13 4.3a still ended in 2024.

Overview¶

GLEAM is a set of algorithms — and the resulting datasets — that estimate land evaporation and its components (transpiration, interception loss, bare-soil evaporation, sublimation, open-water evaporation) from satellite observations of radiation, precipitation, temperature, soil moisture and vegetation. The fourth generation, GLEAM4, moves to 0.1° and adds evaporative stress learned from eddy-covariance and sap-flow data, Penman-based potential evaporation and plant access to groundwater. Two archives with different forcing are provided — 'a' (1980 onward, satellite and reanalysis) and 'b' (2003 onward, mostly satellite).

What changes with the successor¶

GLEAM v3 (e.g. v3.8a) → GLEAM4 (v4.1–4.2)¶

Transition: 2024-07

The fourth generation raises the resolution from 0.25° to 0.1°, reworks potential evaporation, interception and evaporative stress, and adds plant access to groundwater. Component partitioning changed substantially (global transpiration fraction 74% → 62%), so v3 results must not be compared directly with, or appended to, GLEAM4.

Sources: Miralles et al. (2025), GLEAM4, Scientific Data · Martens et al. (2017), GLEAM v3, Geoscientific Model Development · GLEAM website

GLEAM4.2 (4.2a and 4.2b) → GLEAM4.3 (4.3a and 4.3b)¶

Within the GLEAM4 framework, 4.3 changes how evaporative stress is learned, fixes the soil-layer depths and represents irrigation explicitly. 4.2 was removed from the server when 4.3 was released.

Sources: GLEAM website (Datasets and News)

Available variables¶

The 12 variables listed in the GLEAM4 paper (Data Records) and on gleam.eu. There is one file per variable per year, named _GLEAM.nc (monthly and yearly files add _MO or _YR). From 4.2a the aerodynamic and radiative components of potential evaporation are also provided, but their variable names could not be confirmed and are not listed.

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

Evaporation components (daily totals)¶

Stress and heat flux (daily)¶

Soil moisture (daily model state)¶

The full variable list is in the official documentation.

How to get the data¶

• GLEAM website (registration, then SFTP) — Register on the web → connect by SFTP (port 2225) with the credentials sent by email
  • Enter your first name, last name and email under Downloads, and login details are emailed to you (if nothing arrives within an hour, check spam and then contact info@gleam.eu). Connect with SFTP, not plain FTP, and make sure your firewall does not block port 2225. A detailed README is on the server.
• Older versions (on request) — Email request
  • Old versions are removed from the server when a new one is released. If you need one (for example to reproduce a paper), it may be provided upon reasonable request.

Example code¶

Fill in the SFTP host, username and password from the registration email. The directory layout follows the v4.2a example in the paper (v4.2a/daily/2023/E_2023_GLEAM_v4.2a.nc); check the README on the server for the actual paths.

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

HOST, PORT = "<host from the email>", 2225
USER, PASSWORD = "<username>", "<password>"

transport = paramiko.Transport((HOST, PORT))
transport.connect(username=USER, password=PASSWORD)
sftp = paramiko.SFTPClient.from_transport(transport)
print(sftp.listdir("."))                      # check the version folder names first
remote = "v4.3a/daily/2024/E_2024_GLEAM_v4.3a.nc"   # example path
sftp.get(remote, "E_2024_GLEAM_v4.3a.nc")
sftp.close(); transport.close()

ds = xr.open_dataset("E_2024_GLEAM_v4.3a.nc")
e = ds["E"].sel(lat=slice(39, 33), lon=slice(124, 132))   # latitude runs north to south
print(e.sum("time").mean().values)            # area-mean 2024 annual evaporation (mm)

Practical tips¶

• Archives 'a' and 'b' share the method but use different forcing (per the GLEAM4 paper, a: MSWX meteorology and MSWEP precipitation; b: IMERG precipitation, AIRS temperature, CERES radiation). Use 'a' for long climate records and 'b' for a shorter, more observation-based record, and do not mix them.
• When a new version is released the old one is removed from the server and past values change. Record the version you downloaded (e.g. v4.3a) and state it in your paper.
• Evaporation is not measured directly from space; GLEAM values are model estimates. When comparing with eddy-covariance towers, the team notes footprint mismatch (~1 km vs ~10 km pixels), energy-balance non-closure and unreliable measurements during rain, and recommends masking rain and interception periods.
• Component partitioning differs strongly between versions. In GLEAM4 transpiration makes up about 62% of global evaporation, down from about 74% in v3.
• No uncertainty (error) estimates are provided with the data.
• GLEAM4 assimilates ESA CCI surface soil moisture, so GLEAM soil moisture and ESA CCI SM are not independent when compared.
• Forcing data and static parameters are not hosted on the GLEAM server; get them from their original portals (see the README and Miralles et al. 2025).

Notices and change history¶

References for your paper¶

Dataset citation¶

GLEAM4 technical notes and data record (Zenodo, the data citation given in the paper) — DOI 10.5281/zenodo.14056079

@misc{gleam4_zenodo,
  author    = {Miralles, Diego G. and Bonte, Olivier and Koppa, Akash and Baez-Villanueva, Oscar M. and Tronquo, Emma and Zhong, Feng and Beck, Hylke E. and Hulsman, Petra and Dorigo, Wouter and Verhoest, Niko E. C. and Haghdoost, Shekofeeh},
  title     = {{GLEAM4}},
  version   = {v4.2},
  year      = {2025},
  publisher = {Zenodo},
  doi       = {10.5281/zenodo.14056079}
}

Reference papers¶

GLEAM4 data paper — DOI 10.1038/s41597-025-04610-y

@article{miralles2025gleam4,
  author  = {Miralles, Diego G. and Bonte, Olivier and Koppa, Akash and Baez-Villanueva, Oscar M. and Tronquo, Emma and Zhong, Feng and Beck, Hylke E. and Hulsman, Petra and Dorigo, Wouter and Verhoest, Niko E. C. and Haghdoost, Shekoofeh},
  title   = {{GLEAM4}: global land evaporation and soil moisture dataset at 0.1{\textdegree} resolution from 1980 to near present},
  journal = {Scientific Data},
  volume  = {12},
  pages   = {416},
  year    = {2025},
  doi     = {10.1038/s41597-025-04610-y}
}

Original GLEAM paper — DOI 10.5194/hess-15-453-2011

@article{miralles2011gleam,
  author  = {Miralles, D. G. and Holmes, T. R. H. and De Jeu, R. A. M. and Gash, J. H. and Meesters, A. G. C. A. and Dolman, A. J.},
  title   = {Global land-surface evaporation estimated from satellite-based observations},
  journal = {Hydrology and Earth System Sciences},
  volume  = {15},
  number  = {2},
  pages   = {453--469},
  year    = {2011},
  doi     = {10.5194/hess-15-453-2011}
}

Evaporative stress module (hybrid deep learning) — DOI 10.1038/s41467-022-29543-7

@article{koppa2022hybrid,
  author  = {Koppa, Akash and Rains, Dominik and Hulsman, Petra and Poyatos, Rafael and Miralles, Diego G.},
  title   = {A deep learning-based hybrid model of global terrestrial evaporation},
  journal = {Nature Communications},
  volume  = {13},
  pages   = {1912},
  year    = {2022},
  doi     = {10.1038/s41467-022-29543-7}
}

Soil and groundwater module — DOI 10.1029/2022WR033731

@article{hulsman2023groundwater,
  author  = {Hulsman, P. and Keune, J. and Koppa, A. and Schellekens, J. and Miralles, D. G.},
  title   = {Incorporating Plant Access to Groundwater in Existing Global, Satellite-Based Evaporation Estimates},
  journal = {Water Resources Research},
  volume  = {59},
  number  = {8},
  pages   = {e2022WR033731},
  year    = {2023},
  doi     = {10.1029/2022WR033731}
}

Acknowledgements¶

gleam.eu asks users of GLEAM4 data to cite Miralles et al. (2011) and Miralles et al. (2025), together with the interception module (Zhong et al., 2022, HESS), the soil module (Hulsman et al., 2023) and the stress module (Koppa et al., 2022). Also give the version you used (e.g. GLEAM4.3a) and the access date. The Zenodo technical note (10.5281/zenodo.14056079) documents v4.2.

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¶

• GLEAM website
• Miralles et al. (2025) GLEAM4 paper
• GLEAM4 technical notes (Zenodo)
• Martens et al. (2017) GLEAM v3 paper