AGRIBALYSE is a French LCI database for agricultural and food products consumed in France, developed on behlaf of ADEME, the French agency for ecological transition. It specifically focusses on food LCA datasets, for France. Ther new version released in fall 2022 is now available for openLCA users through Nexus (link at the end of the post). The database comes with the openLCA LCIA method package 2.1.5 and with an interim EF3.1 method developed by the data providers, also adapted to work for openLCA.
Figure 1 – AGRIBALYSE v3.1 in openLCA v1.11
What is new
Updates are documented under the following link: “Rapport des changements Agribalyse 3.0.1/3.1” https://doc.agribalyse.fr/documentation/documentation-complete . This version is currently being translated and will be available in English soon.
Implementation into openLCA
The database was provided to GreenDelta in SimaPro CSV format. Both SimaPro and openLCA software work with different reference elementary flow system, meaning that flows have to be mapped and replaced as a first step. The EF3.1 Method interim for AGRIBALYSE was also adapted to work with openLCA. Results were cross checked to ensure correct implementation.
As opposed to richer data formats such as JSON-LD or ILCD, SimaPro CSV is quite limited in the data fields it can communicate information in, most of them delivered as text. Some further work had to be done to adapt the database to openLCA:
- Change location information from the name field to location field
- Name correction to not include information such as “U” for unit process – this is directly portrayed in openLCA as a different icon image, and process in nature.
Figure 2 – Names of directly imported SimaPro CSV processes have information like location only in the name
- The metadata with information such as data generator, publication, etc. doesn’t come in different data fields from the SimaPro CSV data format but rather all together as text. This information is left in the description field for each process.
Firgure 3 – Example of agribalyse process in openLCA, metadata is only visible as text on the “description field” even though that openLCA supports specific fields for metadata
To conclude, the database data quality system is also only available in the general description text field as it originally is found there, and it unfortunatelly comes in different patterns from dataset to dataset, therefore hard to extract.
Background, data provider information
AGRIBALYSE 3.1 is the French LCI database for the agriculture and food sector provided by ADEME. The latest version is published in 2022, but it is built upon previous versions started since 2013.
It comprises LCIs for 2 600 agricultural and food products produced and/or consumed in France, combining a production-based approach and a consumption-based approach. The database has contributions from a network of partners including INRAE, the agriculture and agro-food french technical institutes (ACTA-ACTIA), as well as some consultancies (Gingko 21, Sayari, Blonk, Koch Consulting, Evea). For imported products, Agribalyse 3.1 relies on Ecoinvent and WFLDB data.
AGRIBALYSE uses a transparent methodology that is co-developed by an extensive partnership within the scientific consortium “Revalim”. Methodology principles follow the key international guidelines as much as possible (ISO, LEAP, PEF). The development program is ongoing, and more updates will follow.
Finally, this work relies on an important partnership between Agribalyse and Ecoinvent, including cooperation on data and modeling methods.
More information in www.agribalyse.fr
How to obtain it?
You can download it for openLCA via the database market place Nexus under this link: https://nexus.openlca.org/database/Agribalyse
When downloading, you agree to the terms and conditions of ETALAB. Users also need to have a valid license for ecoinvent 3.4 or beyond.
Interested to know more? A webinar in English language will be held on the 10th January, where ADEME will present the database and later there will be a demonstration for openLCA. Register here.