Macroscopic fire parameters such as fuel regression rate, flame height and flame tilt are critical to the development of detailed fire models and empirical tools for hazard analysis [1–3]. As a result, these characteristics have been investigated by many researchers using various measurement methods in studies of liquid fuelled pool fires of different diameters and fuel types, under a range of crosswind conditions. In investigations related to transportation accidents, fire scenarios have been complicated further through interactions between the fire and upwind or downwind objects [1,2]. Of particular interest is the determination of fuel regression rate, an important parameter but one that is generally difficult to characterize accurately. Many techniques have been reported for measurement of fuel regression rate. These include load cells [2,4,5], differential pressure systems [2,5–7], sight glass and float-type level meters [6–8] and thermocouple rakes [1]. In general, load cells have been employed most successfully for measurements in smaller scale fires [2,4], while researchers have turned to differential pressure and thermocouple type systems for measurements in fires above 5 m diameter [2,6,7]. All the techniques have been used with varying levels of success to measure fuel regression rate under quiescent conditions. Under crosswind conditions and in cases with an object present, however, inherent wandering of the fire plume and dynamic wind loading on the pool can be of additional concern as they affect the accuracy and repeatability of the measurements [1,2,6,7]. In several excellent reviews, available results have been summarized and used to derive empirical correlations relating overall fire characteristics to fire diameter, fuel type and/or wind velocity [3,9–11].Copyright © 2003 by ASME