An open-source python software for standardized processing, sharing, and analysis of disdrometer data

Slack | Documentation

• About DISDRODB
• Key Features
• Installation
• Quick Start
• Documentation
• Contributing
• Contributors
• Citation
• License

DISDRODB is an international collaborative initiative to index, collect, and homogenize drop size distribution (DSD) data from disdrometers worldwide. Our mission is to establish a global standard for sharing disdrometer observations, making precipitation microphysics data accessible and interoperable.

Built on FAIR data principles (Findable, Accessible, Interoperable, Reusable) and adhering to Climate & Forecast (CF) conventions, DISDRODB provides:

• 🌐 A decentralized data archive for raw disdrometer data
• 📊 Standardized NetCDF products for seamless analysis
• 🔬 Quality-controlled datasets ready for scientific research
• 🤝 An open community for collaboration and knowledge sharing

• Download raw disdrometer data from the DISDRODB Decentralized Data Archive
• Upload your own disdrometer station data to contribute to the global archive
• Explore metadata from stations worldwide through standardized formats

• L0 Product: Convert raw data into standardized NetCDF format
• L1 Product: Generate quality-checked, homogenized measurements at multiple time resolutions
• L2 Product: Compute DSD parameters and derive radar variables (reflectivity, differential reflectivity, etc.)

• Lazy loading: Efficiently work with large datasets using Dask/Xarray
• Event detection: Automatically identify and analyze precipitation events
• Visualization: Built-in plotting functions for DSD quick-looks and data exploration
• xarray accessor: Extended functionality for disdrometer-specific operations

• Open-source and community-maintained
• Active Slack workspace for support and collaboration
• Regular updates and new features based on user feedback

DISDRODB can be installed via conda on Linux, macOS, and Windows:

conda install -c conda-forge disdrodb

If conda-forge is not set up for your system, see the conda-forge installation guide.

Alternatively, install via pip:

pip install disdrodb

To install the latest development version, see the documentation.

Get started with DISDRODB in three simple steps: download metadata, configure paths, and start analyzing data.

The Metadata Archive contains information about all disdrometer stations in DISDRODB (location, sensor type, data availability, etc.).

Option A: Clone with Git (recommended for staying up-to-date):

git clone https://github.com/ltelab/DISDRODB-METADATA.git

Option B: Download a static snapshot:

disdrodb_download_metadata_archive <path/to/DISDRODB-METADATA>

Configure DISDRODB by specifying two directories:

• metadata_archive_dir: Path to your local DISDRODB Metadata Archive (the cloned repository)
• data_archive_dir: Path where DISDRODB will store downloaded raw data and processing products

Note: Paths must end with \DISDRODB (Windows) or /DISDRODB (macOS/Linux).

import disdrodb
# Define your local paths
metadata_archive_dir = "<path_to>/DISDRODB-METADATA/DISDRODB"
data_archive_dir = "<path_to>/DISDRODB"
# Create configuration file
disdrodb.define_configs(
 metadata_archive_dir=metadata_archive_dir, data_archive_dir=data_archive_dir
)

This creates a .config_disdrodb.yml file in your home directory.

Verify your configuration:

import disdrodb
print("Metadata Archive:", disdrodb.get_metadata_archive_dir())
print("Data Archive:", disdrodb.get_data_archive_dir())

Or via command line:

disdrodb_metadata_archive_directory
disdrodb_data_archive_directory

Download all available data:

disdrodb_download_archive

Download from a specific data source:

disdrodb_download_archive --data_sources EPFL

Download a specific station:

disdrodb_download_station EPFL EPFL_2009 10

💡 Tip: Use disdrodb_download_archive --help for all available options.

Process raw data into standardized NetCDF products (L0, L1, L2) for a specific station:

disdrodb_run_station EPFL EPFL_2009 10 --parallel True --force True

💡 Tip: Use disdrodb_run_station --help to explore processing options.

The L0C product contains raw disdrometer data in standardized NetCDF format. Use open_dataset() to efficiently load data with lazy evaluation (data is only loaded into memory when needed):

import disdrodb
ds = disdrodb.open_dataset(
 product="L0C",
 data_source="EPFL",
 campaign_name="HYMEX_LTE_SOP3",
 station_name="10",
)
ds

The L1 product provides quality-controlled measurements at multiple temporal resolutions (1MIN, 5MIN, 10MIN, etc.), including hydrometeor classification and quality flags. This is the recommended product for precipitation analysis.

import disdrodb
import matplotlib.pyplot as plt
# Load L1 product at 1-minute resolution
ds = disdrodb.open_dataset(
 product="L1",
 data_source="EPFL",
 campaign_name="EPFL_2009",
 station_name="10",
 temporal_resolution="1MIN",
)
# Compute particle counts for event detection
ds["n_particles"] = ds["n_particles"].compute()
# Identify and visualize precipitation events
for ds_event in ds.disdrodb.split_into_events(
 variable="n_particles",
 threshold=10,
 neighbor_min_size=2,
 neighbor_time_interval="5MIN",
 event_max_time_gap="2H",
 event_min_duration="20MIN",
 event_min_size=5,
 sortby=lambda ds_event: ds_event["n_particles"].sum(dim="time").max(),
 sortby_order="decreasing",
):
 # Generate DSD quick-look plots
 ds_event.disdrodb.plot_dsd_quicklook(
 hours_per_slice=3,
 max_rows=6,
 )
 plt.show()

You should see quick-look plots of the PSD for the identified precipitation events, similar to this: alt text

📖 Learn more: See the products documentation for detailed information.

Open the metadata archive directory:

disdrodb_open_metadata_archive

Load all station metadata into a pandas DataFrame:

import disdrodb
df = disdrodb.read_metadata_archive()
df.head()

This README provides a quick overview. For comprehensive information, visit our documentation:

📚 https://disdrodb.readthedocs.io/en/latest/

• 📊 Detailed product specifications (L0, L1, L2)
• 🔧 Advanced processing options and customization
• 📤 Guide to contributing your own data
• 💻 API reference and code examples
• 📓 Jupyter notebook tutorials
• ❓ FAQs and troubleshooting

We welcome contributions and feedback from the community! Here's how to get involved:

• Slack Workspace : Join discussions, ask questions, and collaborate
• GitHub Discussions : Share ideas and start conversations
• GitHub Issues : Report bugs or request features

• 📊 Share your data: Contribute disdrometer observations to the archive
• 💻 Improve code: Submit bug fixes or new features via pull requests
• 📖 Enhance documentation: Help improve guides and examples
• 🧪 Develop algorithms: Propose new analysis methods or quality control procedures
• 🌍 Spread the word: Tell others about DISDRODB

See CONTRIBUTING.rst for detailed guidelines.

• Gionata Ghiggi
• Kim Candolfi
• Régis Longchamp
• Charlotte Gisèle Weil
• Jacopo Grazioli
• Alexis Berne

If you use DISDRODB in your research, please cite:

Gionata Ghiggi, Kim Candolfi, Régis Longchamp, Charlotte Weil, Alexis Berne (2023). ltelab/disdrodb. Zenodo. https://doi.org/10.5281/zenodo.7680581

For version-specific citations, visit the Zenodo record.

This project is licensed under the GPL 3.0 License.