For each dataset published in PANGAEA, the abstract provides a concise and method-oriented description of the observation or measurement, namely what, when, where, why and how the data was collected. The summary should consist of meaningful running text. The format of the dataset abstract is the same as that of paper abstracts.
For data submissions with single data tables, the abstract should be inserted into the abstract section of the submission form. The length should be ideally limited to 5000 characters. If a submission contains several data sets (tables), an individual abstract must be provided for each data set. Dataset specific abstracts can be similar to one another, but should contain at least one sentence that specifically relates to content of the respective data table.
Information that an abstract should contain
- WHAT are the observations made, what parameters were measured (especially if you publish data in files such as netCDF, or other proprietory formats), what samples were collected, what experiments were performed, what was modeled?
- WHEN was the data collected (time of the measurement not the campaign)? What is the temporal coverage? The precision of is information coverage should be made in context of the temporal extent of the dataset, e.g. geological era, centuries, millennia, decades, months, days, ...
- WHERE was the data collection done (e.g. Drake passage), what is the geographical coverage? The precision of this information should be made in context of the spatial extent of the dataset, e.g.: global, sea, ocean, continent, country, region, ...
- WHY was the data collected? What was the purpose? This will have important implications for reusability of the data, and its possible limitations for other purposes.
- HOW was the data collected? Which devices or methods were used to collect the data? During which campaign or cruise were the data collected?
- Abstract may contain necessary references. If you include a reference, provide the article DOI or a full citation (if no DOI is available). References relevant to the dataset do not need to appear in the abstract, but can also be linked to the data independently of the abstract (provide the details to the data editor).
What to avoid
- Avoid the interpretation of results.
- Avoid acronyms and abbreviations; if you use them, first spell them out.
- Do not simply copy & paste manuscript abstracts or cruise report extracts. Abstracts need to be dataset specific and unique.
- Many metadata do not have to be part of the abstract because they are already listed in the metaheader of the published dataset. Like that, they are both human and machine readable. Examples are:
- Contact information of the PI (is listed in the Parameter table of the dataset)
- Link to the cruise report, chief scientist (is part of the Event information in the dataset or can be added as a reference)
- Exact geolocation (is part of the Event information in the dataset)
- Title and citation of the related article (article is added as a reference)
Data collection abstract
Data collections require a common abstract. In addition to that, each of the individual datasets may have a separate abstract.