How Archival Data Can Inspire New Research

A dataset mobilised through NFDI4Biodiversity, containing 5,467 historical records from 1845, is attracting considerable interest among today’s researchers – including as a valuable baseline for comparison.

5,467 historical records – 180 years old and more relevant than ever

In meticulous handwriting, forestry officers recorded where they had spotted a lynx, when they had encountered a bear, or how many otters lived in their district. The order came from above: the Bavarian forestry administration wanted a comprehensive account covering the entire kingdom. Officially, the aim was scientific: the survey was intended to document the geographical distribution of selected animal species. Yet occasional comments in the files suggest that some forestry officers were hardly neutral observers. One noted: “Unfortunately, the wildcat has not yet been completely eradicated, and every year still yields a few specimens.” Whatever the individual motives, the extensive survey produced a collection of 5,467 records documenting the presence or absence of animal species – a collection that, 181 years later, provides a valuable baseline for comparison.

Because today, the data are not only legible — they are analysable, visualisable, and comparable. This has been made possible by a collaborative project under the umbrella of NFDI4Biodiversity, involving researchers from the Chair of Computational Humanities at the University of Passau, the German Centre for Integrative Biodiversity Research (iDiv), and the General Directorate of the Bavarian State Archives, which includes the Bavarian Main State Archive where the original forestry files are held.

Through a multi-step process, the handwritten records were transcribed, georeferenced, standardised, and finally published via research platforms such as GBIF and Zenodo, making them accessible to researchers around the world. [More about the mobilisation process can be found in this earlier article.]

Old Data, New Questions

*The impact was immediate. Shortly after its release, the dataset was used in its first peer-reviewed publication in Nature Scientific Data. The study places the historical material in a scientific context and shows that the animal observations from 1845 are far more than mere lists. They offer insights into wildlife populations, habitats and range boundaries at the time—documenting, for example, where beavers had already disappeared even though their former habitat still existed. The study is also notable for its methodology: it demonstrates how disparate handwritten sources can be transformed into a structured dataset that meets modern research standards—interoperable, open and machine-readable. The article has now been cited five times.*

A second study, „From Historical Archives to Algorithms: Reconstructing Biodiversity Patterns in 19th Century Bavaria“, was published in the open-access journal Diversity in April 2025. Drawing on the historical data, it examines past species distributions, habitats and human–wildlife relationships, among other topics, and highlights the research potential of digitising historical observations.

The historical records are also being used to develop new methods of analysis. In a 2025 workshop paper researchers at the University of Passau explore how large language models can help extract information about species abundance from the historical survey texts.

Early-career researchers have also taken an interest in the digitised data. Ricarda Huter from the University of Jena used the dataset in her master’s thesis to analyse changes in biodiversity. Her work shows how historical sources can help reveal ecological change over decades – from the gradual decline of individual species to shifts in entire habitats.

"It’s wonderful to see the data mobilised in this project now actively feeding into current research."

Dr Bernhard Grau, Director General, Bavarian State Archives

The usage figures also speak for themselves. On GBIF, the dataset has featured in 9,790 downloads to date, including downloads that may contain only some of its records or data from additional sources. GBIF also links the dataset to 18 literature records, including works that reference GBIF downloads containing records from the dataset. On Zenodo, the associated repository has registered 3,931 views and 5,898 downloads across all versions. It also provides background information, references to the original sources and technical documentation. Figures as of 27 October 2026.

In recognition of its exemplary findability, accessibility, interoperability and reusability, the dataset received the Thuringian FAIRest Dataset Award in 2026. The jury particularly highlighted its comprehensive documentation and praised the interdisciplinary collaboration between several institutions.

From Scientific Journal to Daily Newspaper

The data attracted attention beyond the research community, too. NFDI4Biodiversity published a blog post about the project’s results, while the Bavarian State Archives issued an official press release and invited the public to a talk on the subject. Promotional postcards featuring historical animal illustrations and a social media campaign accompanied these efforts; the Münchner Merkur also reported on the project.

For Dr Bernhard Grau, Director General of the Bavarian State Archives, this broad public resonance is not a byproduct but an intended outcome: "It’s wonderful to see the data mobilised in this project now actively feeding into current research. That was precisely our goal — to highlight the value of archival heritage and the knowledge potential of historical, analogue sources for biodiversity research."

More Than a One-Off

Biodiversity research depends on long-term, high-quality data — not only from labs and field studies, but also from archives, museums, historical maps, and collections. Only through such sources is it possible to look back across decades or centuries — an essential foundation for understanding species decline, analysing range shifts, or planning conservation measures.

However, many of these data remain difficult to find or are not readily accessible, often because they exist only in analogue form. This is precisely where NFDI4Biodiversity comes in: to make such data — previously scattered, non-digital, or locked away in institutional silos — systematically accessible for long-term use by science, administration, and society.

The 1845 wildlife observations are therefore far more than an isolated success. They represent what becomes possible when historical treasures are unearthed — and they point to the vast potential in systematically mobilising archival data.

The information on reuse of the dataset was last updated on 27 October 2026.