Preamble

Sticking to a project’s Data Policy should not be seen as a chore. Its purpose is to help you to create, maintain and safeguard high-quality data, to share data and gain access to data. These guidelines exists to help you to document your data so that it is useful to others (and to you too, as you will soon forget the details). Sharing data from the start with close colleagues and data specialists will help you to detect and eradicate errors and improve calibration, as well as to safeguard your hard-won data. Later, well-documented data from your project can be compared with other projects, leading to improved understanding and more publications.
The Data Policy is a guide to good working practice, which will in the end reduce the time taken to manage your data and maximise the research you get out of it. For detailed recipes to guide you step by step, consult the IMBER Data Management Cookbook.

Involve Data Professionals

A successful, efficient project involves end-to-end data management, which requires involvement and dialogue right from the start between researchers and Data Professionals (people whose primary task is management of data rather than research, for example the staff of national Data Centres). Data professionals are familiar with the technical requirements (and jargon) that are needed; researchers usually are not.
A project seeking IMBER endorsement must show that data professionals have assisted in drawing up the project’s data management plan and that their involvement is on-going.
In many cases, a National Data Centre will be appointed/funded as the Project Data Assembly Centre (DAC) to advise a project and to receive data and metadata from that project.
If the experiment has international participation, the project’s data management plan must show data centre responsibilities and identify the Project DAC.
If a country does not have a national data centre, the plan must show what arrangements have been made to create a Project DAC and to submit data to a recognised Data Centre. (See http://www.imber.info/DM_resources_links.html for a list of recognised National Data Centres and World Data Centres).

Appoint Data Scientist

IMBER observational projects are frequently large and multidisciplinary. Managing data for such a project is thus a substantial task, and it is cost-effective, both financially and in research efficiency, to appoint or assign an individual (the Data Scientist) to be responsible, part- or full-time for data management.
To receive IMBER endorsement, a Project must show, as part of its Data Management plan, what steps are being taken to appoint a Data Scientist.
The Data Scientist does not have to be a data professional, indeed he/she may well be a post-doc (for example) with a research role in the project, who will learn data and people management skills, and will be responsible for data management within the project (Cookbook).

Data sharing within a project

The generally multi-disciplinary nature of IMBER projects means that all data from a project need to be accessible to all project participants to aid interpretation.
Project Participants should agree from the start that there be open access to each other’s data, but that it should not be passed on to non-participants for an agreed period (usually two years).
Only the data originator may pass on data (or authorize data to be passed on) to non-participants at his/her discretion.

Metadata submission

As soon as a project (or a cruise* as part of a project) is funded, project (or cruise) metadata should be submitted as a DIF (Directory Interchange Format) to the GCMD (Global Change Master Directory) (see Cookbook for explanation).
As soon as any cruise ends, a CSR (Cruise Summary Report) must be submitted to the Project DAC and the IMBER IPO. This is the responsibility of the Cruise Principal Scientist, though likely to be greatly assisted by the Data Scientist. The DIF for the project/cruise should be updated to reference the CSR. (It is hoped that we will soon provide translation software so that a full post-cruise DIF can be automatically created from a CSR. In the interim, it is important that the main DIF be updated so that IMBER data are fully recorded in the IMBER Data Portal.)
As soon as the cruise ends, the Event Log (Cookbook) must also be submitted to the Project DAC.
Within 6 months after the end of a cruise, the Cruise Principal Scientist should submit a detailed cruise report to the Project DAC and the IMBER IPO.
Individual researchers are encouraged to create and submit DIFs to the GCMD on their own data sets, linked to their project or cruise.
All metadata will be publically accessible via the GCMD Web site as soon as it has been submitted to the GCMD (after checking by the IMBER Data Liaison Officer).
*When IMBER fieldwork is not conducted on cruises, “cruise” should be more generally interpreted as “fieldwork”, and metadata equivalent to the CSR, etc., submitted on the same time scales.

Data submission

Initial versions of data files should be duplicated as soon as they are created and copied to the Project DAC at the end of the cruise.
This provides data security. The data will not be released beyond the project participants, and will only be released to project participants with the approval of the originator (since early versions of data sets may be unsuitable for release even to participants).
Data originators should interact with the Data Scientist and Project DAC to provide updated and final versions of data sets as they become available

Public Access to data

Data collected as part of IMBER will be made publicly available no later than 2 years from end of cruise or project.
Exceptions to this may be allowed by the SSC, for example where the policy is overridden by national constraints on data access.