The assumption of homogeneous fuel beds that underlies most fire spread models fundamentally limits their operational utility and future advancements in fire science, and imposes a significant disconnect between real fuels, which are highly heterogeneous, the inferences made from such models, and their subsequent applications in management. In contrast, physics-based fire models, which can model fire in heterogeneous fuels, are increasingly recognized as critically contributing to our understanding of wildland fire. In recent years new advances have been made in remote-sensing based technologies characterizing 3D vegetation structure in heterogeneous fuel beds, but these approaches must be linked to traditional fuel measures to translate point cloud information into quantitative inputs needed for fire modeling such as fuel mass, bulk density and surface area to volume. We propose multi-scale characterization of three-dimensional (3D) vegetation and fuel structure from 3D point cloud data collected using airborne and terrestrial lidar and Structure for Motion (SfM) methods applied to digital stereo photography. Integrated with traditional and innovative new field measurement methods, this suite of advanced remote sensing techniques can capture the complex 3D distribution of pre-fire vegetation and fuel elements and overcome limitations of traditional fuel sampling methods for characterizing overstory, understory, and fuel bed structure hierarchically, at high resolution, and in 3Das are the requirements for spatially explicit, heterogeneous 3D fuel data inputs at different scales commensurate with different physics-based fire models. Traditional fuel measures will be collected within these fused point clouds, in near proximity to fire power and energy flux instrumentation, to allow empirical translation of derived point cloud metrics into 3D maps of fuel loads; pre- and post-fire point cloud collections will permit mapping of fuel consumption, also in mass units (e.g., kg/m2). These 3D fuel load and consumption maps will provide unprecedented precision as heterogeneous fuel inputs into mechanistic fire and smoke models and for linking fuel patterns with fire effects. Our fuel products will advance fire science by promoting a physically based understanding of the relationships between fuels and fire energy, consumption and smoke, and other fire effects.