An understanding of fire-response traits is essential for predicting how fire regimes structure plant communities and for informing fire management strategies for biodiversity conservation. Quantification of these traits is complex, encompassing several levels of data abstraction scaling up from field observations of individuals, to general categories of species responses. We developed the Fire Ecology Database to accommodate this complexity. Its conceptual framework is underpinned by a flexible data pipeline enabling links between fire-related trait data and event information at individual, population, and community levels. Key features include: (a) concise and documented trait and method vocabularies
(b) documented uncertainty in observations and aggregation
and (c) documented origin of data including field observations, laboratory experiments, and expert elicitation. We demonstrated application of our framework using data from new field surveys and existing data sets in New South Wales, Australia. The database includes 14 traits for 6,287 plant species derived from 8,936 field work records from 2007 to 2018, 7,054 field records from surveys after 2019, and 48,306 records from 301 existing sources.