We present FDS and AERMOD simulations of the Alaska Clean Seas (ACS) oil burn experiments to improve dispersion modeling of large, open-air fires relevant to industrial settings. We propose a method in which FDS smoke fields with available ground measurements are used to empirically calibrate AERMOD configured using volume sources to represent the fire source. FDS is first run for the three ACS experiments at high resolutions (~10 m) and verified against ground monitoring to provide detailed three-dimensional smoke fields. The fractional allocation of total fire emissions (weights, w_i) is then empirically specified for each volume source i so AERMOD smoke predictions fit both the ground level measurements and FDS simulations to acceptable accuracy. Runs for volumes at the surface (i = 1), 100 m AGL (i = 2) and 300 AGL (i = 3) and w_i = [0.01, 0.09, 0.9]–[0.04, 0.36, 0.6] accurately represent these data, suggesting this range as suitable for fire heat fluxes (~800–3000 kW/m²), wind speeds (5–10 m/s) and PBL depths (300–500 m with and without capping temperature inversions) of the three ACS experimental burns. Further work exploring the applicability of this AERMOD setup to a broader range of conditions is ongoing.