From Historical Archives to Algorithms: Reconstructing Biodiversity Patterns in 19th Century Bavaria

This study introduces Computational Historical Ecology, a framework that applies GIS mapping, text analysis, and AI tools to historical biodiversity data. The framework not only enhances the utility of archival sources but also highlights the potential of integrating computational techniques with historical and environmental humanities. The study works with the 1845 Bavarian Animal Observation Dataset: a historical survey documenting vertebrate species across 119 forestry districts in pre-industrial Bavaria, which offers invaluable insights into species distribution, habitat changes, and the ecological impact of human activities during the 19th century. By digitising, annotating, and analysing over 5,400 archival records, the research bridges historical ecology and computational methods to reconstruct past biodiversity patterns. Employing a data-centric methodology, the exemplary analyses reveal significant shifts in species presence, driven by land development, deforestation, and evolving agricultural practices. Notable examples include the decline of the Eurasian otter and the extinction of the Eurasian beaver in Bavaria. The interdisciplinary study demonstrates how historical records reflect both environmental transformations and the perspectives of the people who documented these changes. Findings underscore the importance of historical datasets as benchmarks for contemporary assessments and as contributions to ongoing debates in conservation science, restoration ecology, and environmental policy-making. The paper advocates for greater investment in digitisation and interdisciplinary collaboration, recognising the critical role of archival sources in shaping future biodiversity strategies.