A dataset of provincial-level new distribution records for birds in China from 2000 to 2025

Species distribution data form the empirical foundation for biogeography, macroecology and conservation biology. However, our understanding of where species actually occur remains incomplete and inaccurate, a persistent deficit known as the Wallacean shortfall. New distribution records help reduce this shortfall by documenting additions to known ranges, yet such records are often dispersed across local-language literature, inconsistently harmonized taxonomically and underrepresented in global biodiversity infrastructures. Here we present the China Bird New Record dataset (CBNR), a peer-reviewed literature-derived, taxonomically standardized and georeferenced dataset of provincial-level new distribution records for birds in China from 2000 to 2025. Using a large language model-assisted extraction pipeline followed by manual validation, we compiled 1,029 new distribution records representing 568 bird species in 23 orders across 33 provincial-level administrative units. The CBNR comprises 25 standardized metadata fields covering taxonomy, spatiotemporal distribution, detection method and inferred discovery reason, migratory strategy, conservation status and publication information. The CBNR provides an essential data source for updating species distribution knowledge, refining biogeographic understanding, assessing sampling bias, and supporting evidence-based biodiversity monitoring and conservation planning in China.