CHIRPS¶

Climate Hazards Center Infrared Precipitation with Stations, Version 3

Available

At a glance¶

Note on period: v3 Final runs from 1981-01 to the present and v3 Prelim from 2025-01. Among the daily products, the IMERG-based (sat) archive starts in 2001 and the ERA5-based (rnl) archive in 1981, according to the Sci Data paper. v2.0 also starts in 1981; when checked on 2026-10-13 the v2 daily collection in Earth Engine ran through 2026-08-31.

Overview¶

CHIRPS is a 0.05° quasi-global precipitation dataset that calibrates satellite precipitation estimated from geostationary IR cold cloud duration (CHIRP) to a high-resolution climatology (CHPclim) and blends in station precipitation from many sources. Developed with USGS and FEWS NET for drought monitoring and seasonal rainfall trend analysis in data-sparse developing regions, it runs from 1981. v3.0 uses nearly four times as many gauge sources as v2, introduces gauge-undercatch correction, a new climatology (CHPclim v2) and a new IR algorithm (CHIRP3), and extends coverage to 60°.

What changes with the successor¶

CHIRPS v2.0 → CHIRPS v3.0¶

Transition: 2025-01

v3.0 was released in January 2025 and replaces v2.0. The IR algorithm, climatology, station count, gauge correction and spatial coverage all changed; v2.0 continues in parallel only through the end of 2026.

Sources: CHC CHIRPS v3 page · Funk et al. (2026), Scientific Data · CHC CHIRPS v2 page (end of v2 production)

Available variables¶

Precipitation is the only published variable, distributed as GeoTIFF, NetCDF, BIL and COG. The names below are Google Earth Engine band names; the variable name inside CHC NetCDF files could not be confirmed from official documents. Missing or erroneous values are -9999 (including missing IR data from 2023 onward).

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

Pentad and monthly (base products)¶

Daily (rnl and sat)¶

Diagnostics¶

The full variable list is in the official documentation.

How to get the data¶

• CHC data repository (CHIRPS v3.0) — HTTPS download (no account)
  • Folders for pentads, dekads, daily, monthly and multi-month aggregates, split into Africa, Latin America and global domains. The daily product is described in daily/readme.txt and station sources are listed in README-CHIRPSv3.0.txt.
• Google Earth Engine — Cloud analysis (UCSB-CHC/CHIRPS/V3/PENTAD, DAILY_RNL, DAILY_SAT)
  • v2 is UCSB-CHG/CHIRPS/DAILY and PENTAD. When checked on 2026-10-13, v3 pentads ran through 2026-08-26 and v3 daily data through 2026-08-31.
• USGS and CHC Early Warning Explorer (EWX) — Web maps and time series
• CHC data repository (CHIRPS v2.0, through 2026) — HTTPS download
  • The 50°S–50°N version. Production ends after December 2026, so start new work with v3.

Example code¶

Uses the Google Earth Engine Python API to compute the July 2023 precipitation total over the Korean Peninsula from CHIRPS v3 daily data (IMERG-split version). An Earth Engine account and a cloud project are required; run ee.Authenticate() the first time. Replace the project name with your own.

import ee

ee.Initialize(project="my-gee-project")   # your cloud project name

col = (ee.ImageCollection("UCSB-CHC/CHIRPS/V3/DAILY_SAT")   # ERA5-split version: DAILY_RNL
       .filterDate("2023-07-01", "2023-08-01")
       .select("precipitation"))                            # mm/d
total = col.sum()                                           # July total (mm)

korea = ee.Geometry.Rectangle([124, 33, 132, 39])
stat = total.reduceRegion(ee.Reducer.mean(), korea, scale=5566)
print(stat.getInfo())   # values over the sea are satellite-only estimates without stations

Practical tips¶

• Because of the gauge-undercatch correction (Legates-Willmott), v3 is modestly wetter than v2 in most areas. Do not splice v2 and v3 into one series, and when comparing with stations, use gauge-corrected station data as CHC recommends.
• CHIRPS is fundamentally a pentad and monthly product. Daily values borrow ERA5 or IMERG daily distributions to split pentads, which limits daily extremes and wet-day statistics.
• IR gaps in 1981–2022 were filled with ERA5 precipitation (about 20% of land in the 1980s), while missing IR from 2023 onward is left as -9999. Check the fill maps to see where filling occurred.
• The paper reports ERA5-related biases of up to about 30% in Central Africa and Central Asia, making decadal variations and trends there uncertain. Station networks are also declining over time in all archives.
• Station data are blended over land only. Ocean values are satellite-only estimates and are not suitable for ocean precipitation studies.
• Being based on IR cold cloud duration, it may handle orographic rain, warm-cloud rain and winter snowfall poorly. Where networks are dense, as in Korea, compare first with KMA gridded analyses.
• Mixing Prelim and Final means recent months will change later. Use Final in papers, and say so if you used Prelim.

Notices and change history¶

References for your paper¶

Dataset citation¶

CHIRPS v3.0 dataset (CHC, DataCite) — DOI 10.15780/G2JQ0P

@misc{chirps_v3,
  author    = {{Climate Hazards Center}},
  title     = {{CHIRPS}: Rainfall Estimates from Rain Gauge and Satellite Observations, Version 3.0},
  year      = {2025},
  publisher = {UCSB},
  doi       = {10.15780/G2JQ0P}
}

CHIRPS v2.0 dataset (if you used v2) — DOI 10.15780/G2RP4Q

@misc{chirps_v2,
  author    = {{Climate Hazards Center}},
  title     = {Climate Hazards Center InfraRed Precipitation with Station ({CHIRPSv2.0})},
  year      = {2015},
  publisher = {Climate Hazards Group},
  doi       = {10.15780/G2RP4Q}
}

Reference papers¶

CHIRPS v3 data descriptor (Funk et al., 2026) — DOI 10.1038/s41597-026-07096-4

@article{funk2026chirps3,
  author  = {Funk, Chris and Peterson, Pete and Harrison, Laura and Saldivar, Robert and Landsfeld, Martin and Pedreros, Diego and Shukla, Shraddhanand and Fink, Andreas H. and others},
  title   = {The Climate Hazards Center Infrared Precipitation with Stations, Version 3},
  journal = {Scientific Data},
  volume  = {13},
  pages   = {718},
  year    = {2026},
  doi     = {10.1038/s41597-026-07096-4}
}

CHIRPS v2 reference paper (Funk et al., 2015) — DOI 10.1038/sdata.2015.66

@article{funk2015chirps,
  author  = {Funk, Chris and Peterson, Pete and Landsfeld, Martin and Pedreros, Diego and Verdin, James and Shukla, Shraddhanand and Husak, Gregory and Rowland, James and Harrison, Laura and Hoell, Andrew and Michaelsen, Joel},
  title   = {The climate hazards infrared precipitation with stations---a new environmental record for monitoring extremes},
  journal = {Scientific Data},
  volume  = {2},
  pages   = {150066},
  year    = {2015},
  doi     = {10.1038/sdata.2015.66}
}

Acknowledgements¶

CHC asks users to cite the v3 data repository DOI (10.15780/G2JQ0P) with the access date; for the method, cite Funk et al. (2026) (check the full author list in the Scientific Data article). If you used v2.0, cite Funk et al. (2015) and the v2 DOI (10.15780/G2RP4Q), and state which version (Prelim/Final) and which daily product (rnl/sat) you used.

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¶

• CHC CHIRPS v3 page
• CHC CHIRPS v2 page
• CHIRPS FAQ (pentad definition, Prelim and Final)
• Funk et al. (2026) — CHIRPS v3 data descriptor
• CHIRPS v3 daily README
• CHIRPS v3 diagnostics (fill maps, correction factors)
• Earth Engine — CHIRPS v3 daily (IMERG split)
• Earth Engine — CHIRPS v3 daily (ERA5 split)