ArHa: A globally standardised, relational database of arenavirus and hantavirus sampling in wild small mammal populations

Arenaviruses and hantaviruses are globally distributed zoonotic pathogens primarily hosted by small mammals. Efforts to map their ecological drivers and predict spillover risk are hindered by the fragmentation of historical surveillance data across the primary literature, compounded by shifting host nomenclature and variable geographic precision. We present ArHa, a database of 123,101 host occurrence and pathogen diagnostic records derived from a systematic review of published wild small mammal sampling. The dataset standardises host and pathogen taxonomies, resolves records of the same animals published in more than one paper, and classifies the spatial resolution and reliability of each coordinate to support bias-aware spatial epidemiology. To accurately preserve complex study designs, such as multiple diagnostic assays or genetic sequences derived from a single host, ArHa is structured as a four-tier relational database. This architecture handles many-to-one relationships and supports linkage without data duplication. The dataset is distributed as compressed Apache Parquet files for programmatic analysis and as a Darwin Core Archive, so that it can be read by standard biodiversity data tools. An accompanying interactive web application, the ArHa Database Explorer, allows users to filter, visualise, and export subsets of the database without requiring programming expertise. All data, extraction scripts, and the interactive web application are openly accessible via a permanent Zenodo repository.