Best Practices for Publishing Biodiversity Data from Environmental Impact Assessments

Once an organization agrees to share EIA data, its staff must establish provisions ensuring that stakeholders in each stage of data collection, curation and management agree to the terms by which data publishing takes place and are properly acknowledged in their role. These provisions should include agreeing to the GBIF Data Publisher Agreement (the English version is valid for legal purposes) and understanding the conditions of the GBIF Data User Agreement that users of GBIF-mediated data must abide when reusing data.

Ensuring the standardization of data capture at point of collection will

GBIF primarily relies on the Darwin Core (DwC) and Ecological Metadata Language (EML) standards, which set out the structure and format of published datasets (learn more about applicable standards).

GBIF provides pre-configured Excel spreadsheets that can serve as templates for capturing checklist, occurrence and sampling-event data. These spreadsheets are simple tools aimed at providing a common format and standard for collecting data. They use consistent terminology and can be completed with additional DwC terms to fit the data collection purpose. This standardized approach makes it easier to exchange data between users, compare it across sites, and integrate it into national and global biodiversity databases. No metadata template is provided as publishers can use the built-in metadata editor in GBIF’s data publication tool–the Integrated Publishing Toolkit (IPT) (see Step 3 below)–to populate the metadata. The IPT ensures that the data and metadata are in a valid XML format.

GBIF.org does not itself host data. The system relies on each data publisher maintaining their own datasets and making them available online in a GBIF-supported format. It also relies on organizations registering datasets and providing GBIF with a stable endpoint for finding and indexing the data. GBIF recommends using the Integrated Publishing Toolkit (IPT) to do this. Highly skilled publishers can also use an API to register datasets programmatically (contact the GBIF help desk for more details).

Organizations may install the IPT if they have the capacity to host and maintain data on servers that always remain online, ensuring that the data that they share will have a persistent, stable point of access. An organization that either does not have this capacity or does not wish to maintain its own installation can choose one of the following options for data hosting (more details available here):

The first two options provide a range of helpdesk services to potential publishers, while the final option provides very limited support to publishers. Potential publishers can request guidance from the GBIF help desk on the most suitable option. Regardless of the hosting option selected, data publishers retain full control of the data, including the ability to correct and update datasets at any time. Data citations will always acknowledge the data publisher, irrespective of how or where the datasets are hosted.

The IPT is the most commonly-used tool and is maintained and developed by the GBIF Secretariat. IPTs can generate a Darwin Core Archive (DwC-A), the preferred exchange format, for each dataset and register them on GBIF. To use the IPT, data must already be digitized. Acceptable file types include delimited text files (e.g. text files using comma or tab-separated values) or Microsoft Excel. Database connections can also be made. If the IPT is to be hosted within the publishing institution, upon installation of the IPT, the publishing organization should register as the host. If the IPT is hosted elsewhere, the IPT administrator can add the publishing organization to the IPT using an IPT token that is issued upon endorsement of the publisher.

To share data through GBIF.org, publishers must collate or transform and describe existing datasets into a standardized format. This work may require additional processing, content editing and mapping the content of a dataset into one of the available formats. Publishers thus play an essential role not simply in sharing datasets, but also in managing their quality, completeness and usefulness as well as ensuring their integration and value within GBIF’s global knowledge base. GBIF provides guidance on the data quality requirements and recommendations. The GBIF Data Validator is a tool that lets publishers check datasets prior to publication and makes recommendations on how datasets can be improved and cleaned by flagging, for example, duplicate identifiers, incomplete fields and recognized inconsistencies in formatting.

Publishers should use a precautionary approach and seek input from specialists on the publishing of precise locations of sensitive species, for example threatened or valuable species, when there are concerns that doing so could enable poaching or other threats to the species population. For a thorough discussion of this topic, see Current Best Practices for Generalizing Sensitive Species Occurrence Data.

The GBIF IPT supports automatic registration in the GBIF network (see the IPT manual). If publishers are using an IPT, GBIF registers datasets when publishers click the ‘register’ button. Data should be published as soon as possible following the EIA. However, if there are concerns about commercial confidentiality or other time-sensitive issues, publication may be delayed or embargoed until the completion of a licensing process.

To publish data to GBIF, publishers must assign one of three Creative Commons licences to a dataset:

Note that the CC-BY-NC license has a significant effect on the reusability of data, and that GBIF does not consider non-commercial use restrictions to be enforceable. GBIF encourages data publishers to choose the most open option possible.

Once datasets are registered, GBIF indexes them to facilitate access to the data by users. Each dataset has its own page (example) and can be found using the search function on the website and on the publisher’s page (example). The indexing process allows the search and discovery of records from all published datasets, showing, for example, all records of a particular species or groups of species in a given geographical area.

Because search results mix records from different datasets, the GBIF Data User Agreement requires appropriate citation of data regardless of the licence applied to any individual dataset. Through the use of Digital Object Identifiers (DOIs), GBIF tracks data reuse and provides publishers with key metrics for downloads, which appear the ‘activity’ tab of each dataset page (example) and for documented citations in other research and assessments, linked from both dataset and publisher pages. Publishers can use this information to demonstrate the value of their contribution to science and society through sharing data from EIAs.