High spatial variability has been observed in simple fuelbeds (Keane et al.2012, Hiers et al. 2009) which accentuates the need to describe the actual fuel arrangement and composition of fuels in a prescribed or wildfire operation setting as opposed to abstract representations, regardless of model application and domain size. Working with high resolution 3D datasets, I will apply TLS scan information of highly instrumented burn plots in Pebble Hill and Eglin Air Force Base to create voxelized data arrays containing bulk density and fuel height for surface fuels <1m on a horizontal resolution of 0.5 m2. This will be coupled with a common method to populate the environment with trees described in Linn et al. (2005), where fuel is assigned to voxels depending on tree inputs including tree location, height, height to live crown, crown radius, and crown concavity. Starting at the resolution of the fire grid on QUIC-Fire (2m), fuels will be aggregated, and characteristics will be calculated with resolutions of 2, 4, 6, 8, 16, and 32 m, as well as using a single fuel characteristic for the entire plot. I will then use QUIC-Fire under a full range of wind and dead fuel moisture conditions common to prescribed and wildland fires to represent potential ranges of fire intensity, since it is possible that fuel resolution is only relevant for fire intensities under a particular threshold of wind speed, moisture content, or some combination of each. These thresholds will be investigated using a suite of different metrics that have been developed to fully describe the fire environment. Some of these include the bulk rate of spread (analogous to perimeter growth rate), the reduced area, and the normalized canopy consumption, as well as more traditional metrics such as surface fuel consumption and max downwind spread rate. Using QUIC-Fire in an ensemble fashion will allow us to calculate a margin of error that’s considered acceptable for every run, and sensitivity to the fuel heterogeneity can then be tested
