---

Overview

CMIP6-Downloader is a Python toolkit for downloading, subsetting, and processing CMIP6 (Coupled Model Intercomparison Project Phase 6) climate model data. It supports OPeNDAP subsetting, multi-server failover, spatial/temporal extraction, and bias correction — covering 61 GCMs across 197 countries.

Built for climate scientists, water resources engineers, and impact modelers who need reproducible, efficient access to global climate projections.

---

Features

• 61 GCMs Supported — All major CMIP6 models from ESGF nodes
• 197 Country Boundaries — Automatic spatial extraction by country or custom bbox
• OPeNDAP Subsetting — Server-side slicing to minimize data transfer
• Multi-Server Failover — Automatic retry across LLNL, DKRZ, CEDA, NCI nodes
• Bias Correction — Quantile mapping and delta change methods
• Variable Support — Temperature, precipitation, humidity, wind, radiation, pressure
• Scenario Coverage — SSP1-2.6, SSP2-4.5, SSP3-7.0, SSP5-8.5, historical
• Export Formats — NetCDF, CSV, GeoTIFF, Zarr
• Parallel Downloads — Concurrent retrieval with progress tracking

---

Installation

pip install cmip6-downloader

Or from source:

git clone https://github.com/rushmarshall/CMIP6-Downloader.git
cd CMIP6-Downloader
pip install -e ".[dev]"

---

Quick Start

from cmip6dl import CMIP6Downloader

dl = CMIP6Downloader()

# Download monthly precipitation for Jamaica (SSP2-4.5, 2030-2060)
data = dl.download(
    variable="pr",
    model="GFDL-ESM4",
    scenario="ssp245",
    country="Jamaica",
    time_range=("2030-01-01", "2060-12-31"),
    frequency="monthly",
)

# Apply bias correction against observed data
corrected = dl.bias_correct(
    data,
    method="quantile_mapping",
    reference="observed_precip.nc",
)

# Export
corrected.to_netcdf("jamaica_pr_ssp245_2030-2060_corrected.nc")
corrected.to_csv("jamaica_pr_ssp245_2030-2060.csv")

---

Supported Models (61 GCMs)

View full model list

Model	Institution	Resolution
ACCESS-CM2	CSIRO-ARCCSS	250 km
ACCESS-ESM1-5	CSIRO	250 km
BCC-CSM2-MR	BCC-CMA	100 km
CanESM5	CCCma	500 km
CESM2	NCAR	100 km
CNRM-CM6-1	CNRM-CERFACS	250 km
EC-Earth3	EC-Earth Consortium	100 km
GFDL-ESM4	NOAA-GFDL	100 km
INM-CM5-0	INM	200 km
IPSL-CM6A-LR	IPSL	250 km
MIROC6	MIROC	250 km
MPI-ESM1-2-HR	MPI-M	100 km
MRI-ESM2-0	MRI	100 km
NorESM2-LM	NCC	250 km
UKESM1-0-LL	MOHC-NERC	250 km
...	...	...

61 models total across all major modeling centers

---

Scenarios

Scenario	Description	Radiative Forcing
historical	Observed forcings (1850-2014)	N/A
ssp126	Sustainability	2.6 W/m2
ssp245	Middle of the road	4.5 W/m2
ssp370	Regional rivalry	7.0 W/m2
ssp585	Fossil-fueled development	8.5 W/m2

---

Architecture

cmip6dl/
├── downloader.py       # Main download orchestration
├── catalog.py          # ESGF catalog search & model registry
├── providers/          # Data source backends
│   ├── opendap.py          OPeNDAP with server-side subsetting
│   ├── http.py             Direct HTTP/HTTPS download
│   └── failover.py         Multi-server failover logic
├── processing/         # Post-processing
│   ├── spatial.py          Country extraction, regridding
│   ├── temporal.py         Aggregation, resampling
│   └── bias.py             Bias correction methods
└── countries.py        # ISO 3166 country boundary lookup

---

Configuration

providers:
  primary: "https://esgf-node.llnl.gov/esg-search"
  fallback:
    - "https://esgf-data.dkrz.de/esg-search"
    - "https://esgf-index1.ceda.ac.uk/esg-search"

download:
  n_workers: 4
  retry_attempts: 3
  timeout: 120
  chunk_size: 50_000_000

output:
  directory: "./cmip6_data"
  format: "netcdf"

---

Contributing

Contributions welcome. Please open an issue to discuss proposed changes before submitting a pull request.

---

_{Developed at Hydrosense Lab, University of Virginia
Part of the Global Hydrology and Water Resources research group}