xcube-icosdp 0.1.1

A data store to access datasets from ICOS Data Portal.

xcube-icosdp

xcube-icosdp is a Python package and xcube plugin that provides a data store for accessing data from the ICOS Data Portal.

---

Supported Datasets

The plugin currently supports the FLUXCOM-X-BASE products for carbon and water fluxes. For details on FLUXCOM-X-BASE data products, see:

Nelson and Walther et al. (2024): https://doi.org/10.5194/egusphere-2024-165

X-BASE contains four flux variables:

• FLUXCOM-X-BASE_NEE — Net Ecosystem Exchange
• FLUXCOM-X-BASE_GPP — Gross Primary Productivity
• FLUXCOM-X-BASE_ET — Evapotranspiration
• FLUXCOM-X-BASE_ET_T — Transpiration

To improve usability, the data is available through:

✅ Native resolution via cloud-optimized access (spatial/temporal subsetting) ✅ Pre-computed spatial & temporal aggregations

Native Full-Resolution Dataset

• 005_hourly — 0.05° spatial and hourly temporal resolution for 2001–2021 (~3 TB)

Each dataset has dimensions (time, hour, lat, lon) and includes a land_fraction variable.

The data is stored in Zarr format on a publicly accessible object storage hosted by DKRZ. It is cloud-optimized, allowing efficient spatial and temporal subsetting. No authentication is required to access the dataset.

Using the xcube-icosdp Data Store to access the full-resolution dataset

Example: open the NEE flux for 2020–2021 over a custom bounding box:

from xcube.core.store import new_data_store

store = new_data_store("icosdp")
ds = store.open_data(
    data_id="FLUXCOM-X-BASE_NEE",
    time_range=("2020-01-01", "2021-12-31"),
    bbox=[5, 45, 10, 50]  # lon_min, lat_min, lon_max, lat_max
)

🌐 Public data — no authentication required at this time. 📖 Example notebook

Aggregated Products

All aggregated datasets are distributed through the ICOS ERIC Carbon Portal (CC BY-4.0 license) in NetCDF format.

Current aggregations:

Dataset	Spatial Resolution	Temporal Resolution	Size per Year	Dimensions
050_monthly	0.5°	Monthly	~5 MB	(time, lat, lon)
025_monthlycycle	0.25°	Monthly mean diurnal cycle	~270 MB	(time, hour, lat, lon)
025_daily	0.25°	Daily	~380 MB	(time, lat, lon)
005_monthly	0.05°	Monthly	~450 MB	(time, lat, lon)

All files include land_fraction.

The aggregated products are available exclusively through the ICOS Data Portal. To access them, users must create an account and provide their registered email address and password to the data store. Note that the authentication process does not support lazy loading of the dataset. Therefore, the data must be preloaded, which involves downloading the global annual datasets and constructing a unified data cube from them by stacking along the time axis.

Example: access the NEE flux from the moonthly aggregate for 2015 to 2021 over a custom bounding box:

from xcube.core.store import new_data_store

store = new_data_store("icosdp", email="xxx", password="xxx")
cache_store = store.preload_data(
    "FLUXCOM-X-BASE_NEE",
    agg_mode="050_monthly",
    time_range=("2015-01-01", "2021-12-31"),
    bbox=[5, 45, 10, 50],
)
ds = cache_store.open_data("FLUXCOM-X-BASE_NEE_monthly_2015_2021.zarr")

🌐 Public data — authentication via ICOS account required. 📖 Example notebook

---

Installing the xcube-icosdp plugin

This section describes three alternative methods you can use to install the xcube-icosdp plugin.

For installation of conda packages, we recommend mamba. It is also possible to use conda, but note that installation may be significantly slower with conda than with mamba. If using conda rather than mamba, replace the mamba command with conda in the installation commands given below.

Installation into a new environment with pip

This method creates a new environment and installs the latest PyPi release of xcube-icosdp, along with all its required dependencies.

To do so, execute the following commands:

pip install xcube-icosdp

Installation from the repository

If you want to install xcube-icosdp directly from the git repository (for example in order to use an unreleased version or to modify the code), you can do so as follows:

mamba create -f environment.yml
mamba activate xcube-icosdp
git clone https://github.com/xcube-dev/xcube-icosdp.git
python -m pip install --no-deps --editable xcube-icosdp/

This installs all the dependencies of xcube-icosdp into a fresh conda environment, then installs xcube-icosdp in editable mode into this environment from the repository.

Testing

To run the unit test suite:

pytest

To analyze test coverage:

pytest --cov=xcube_icosdp

To produce an HTML coverage report:

pytest --cov-report html --cov=xcube_icosdp

Some notes on the strategy of unit-testing

The unit test suite uses pytest-recording to mock https requests via the Python library requests. During development an actual HTTP request is performed and the responses are saved in cassettes/**.yaml files. During testing, only the cassettes/**.yaml files are used without an actual HTTP request. During development, to save the responses to cassettes/**.yaml, run

pytest -v -s --record-mode new_episodes

Note that --record-mode new_episodes overwrites all cassettes. If one only wants to write cassettes which are not saved already, --record-mode once can be used. pytest-recording supports all records modes given by VCR.py. After recording the cassettes, testing can be then performed as usual.