Taxonomic, geographic and thematic structure of research on annual killifishes in Brazil: data and reproducible analysis

This dataset supports a scoping review and descriptive scientometric analysis of research on annual or seasonal killifishes recorded in Brazil. The study characterises the temporal, editorial, taxonomic, geographic, ecological, thematic, methodological and conservation-related structure of the available scientific literature.Systematic searches were conducted in Scopus, Web of Science Core Collection and SciELO on 25 September 2026, without language or initial publication-year restrictions. The searches identified 712 records: 321 from Scopus, 349 from Web of Science and 42 from SciELO. After removing 319 duplicates, 393 unique records were screened and 243 studies were included in the final evidence matrix.The deposited package contains the final harmonised evidence matrix, a data dictionary, complete search strategies, portable R scripts, derived tables, figures in English and Portuguese, taxonomic and geographic audit files, and the IBGE biome boundary layer used for spatial attribution. The package also includes licences, software requirements, citation metadata and SHA-256 integrity manifests.Taxonomic names, synonyms, geographic information, habitats, thematic dimensions, study designs, field and laboratory methods, analytical approaches, reported threats and conservation implications were harmonised while preserving the information originally reported by the sources. The principal biome analysis comprised 139 studies assigned with high or moderate confidence; 14 low-confidence studies were retained for sensitivity analysis, and 90 studies without an applicable Brazilian territorial unit remained in the general corpus but were excluded from biome comparisons.The original database exports and full-text publications are not redistributed because they may be subject to platform or publisher restrictions. Their filenames, formats, record counts and cryptographic hashes are documented for audit purposes. OpenAlex, Crossref and Semantic Scholar were used only to verify or complete metadata and locate associated resources; they did not add records to the closed screening flow.The dataset describes the structure of published documentation. Frequencies and documentary gaps must not be interpreted as biological abundance, complete species distributions, threat intensity or evidence of species absence.