Padronização e automação de dados da coleção acarológica do centro de coleções taxonômicas da UFMG: Um enfoque na gestão de metadados e modelagem de distribuição de espécies

Abstract

The digitization and standardization of biological data are essential for the efficient management of scientific collections, enabling their integration into global repositories and expanding their potential for ecological and biogeographical research. This dissertation sought to standardize and automate data from the UFMG-AC Acarological Collection to address challenges concerning data fragmentation and inconsistency. The study implemented computational methodologies for data conversion and organization, ensuring compliance with the Darwin Core (DwC) standard. Scripts and pipelines were developed to structure metadata, correct taxonomic and geospatial inconsistencies, and integrate records into international platforms such as the Global Biodiversity Information Facility (GBIF) and the Brazilian Biodiversity Information System (SiBBr). Concurrently, Species Distribution Modeling (SDM) was applied to infer biogeographical patterns and predict potential occurrence areas for cataloged species. The EcoDistrib library was used to optimize environmental variable selection and machine learning based model execution, while cross-validation techniques and statistical metrics ensured result robustness. Comparative analysis of data before and after standardization revealed significant improvements in quality, consistency, and accessibility, including taxonomic redundancy correction and enhanced spatial accuracy. The integration of enhanced data into GBIF and SiBBr increased the collection’s visibility, facilitating its use in biogeographical and conservation studies. Distribution models identified relevant spatial patterns, highlighting gaps in records and suggesting new areas for future sampling. The research demonstrates that combining computational techniques with traditional curation approaches can significantly improve biological collection management, emphasizing the importance of automation and standardization for preserving, sharing, and applying biodiversity data. The findings reinforce the relevance of digitization and computational tools in modernizing scientific collections, advancing biogeography, ecology, and biodiversity conservation.