The 2014 Camp Swift Fire experiments: a FAIR, spatiotemporally integrated dataset for evaluating 3D fire models

Background A significant limitation to evaluating three-dimensional, time-dependent wildland fire models is the absence of pre-fire, active-fire and post-fire time-synchronized and quality-assured datasets. Aims The 2014 Camp Swift Fire Experiment integrated data from three controlled burns to create an openly available FAIR (Findable, Accessible, Interoperable, Reusable) dataset for model evaluation. Methods The study collected infrared (IR) and electro-optical (EO) active-fire imagery, field-based wind and fire measurements, pre- and post-fire aerial imagery and vegetation sampling. Rate of spread (ROS), fire depth and flame height were derived, spatially and temporally aligned. We assessed data quality for accuracy, precision and completeness and presented it through GIS products. Key results Despite sensor and synchronization challenges, vegetation classifications and heat flux data showed strong spatial correspondence. IR- and EO-derived ROS values generally agreed with field-based estimates though anemometer data quality varied. Fire behavior observations were linked consistently to vegetation type and wind measurements. Conclusions The Camp Swift dataset represents fully integrated spatiotemporal measurements suitable for evaluating 3D fire models, highlighting the importance of interdisciplinary planning, rigorous quality control and documentation in producing datasets for modeling applications. Implications This work provides a blueprint for future fire experiments and underscores the critical need for quality-assured, integrated datasets to advance operational and research fire modeling capabilities.