Forest fires tend to cause accelerated soil erosion with some of the redistributed sediment entering the fluvial system. Where there is a reservoir downstream, enhanced sediment accumulation will tend to take place on its floor. This paper concerns the application of different techniques to sediment cores obtained from the floor of Lake Burragorang, the main water supply reservoir for Sydney and explores the possibility not only of discriminating between sediment from burnt and unburnt sources but also between sediment derived from areas subject to different fire severities. Two main soil properties have been investigated: mineral magnetism and thermal activation characteristics (TAC). Laboratory heating experiments have been conducted on long unburnt material (>30 years) to ascertain temperature thresholds at which changes to the soil characteristics occur. Owing to the dominance of the mineral magnetite in soils from the study area, only high severity fires can be detected using mineral magnetic analysis due to the high temperature at which magnetite properties transform (577 °C). Lower temperature thresholds have been explored by TAC analysis. Using this technique it is possible to detect whether temperatures in excess of 110 °C have been reached. These experiments have shown that individual soil grains have retained a ‘memory’ effect of previous fire events, so that it has been necessary to find sources of material not previously exposed to fire (unexposed bedrock) to test the validity of the assumptions of the technique. The influence of the well-documented fire history has been explored in different parts of the Lake Burragorang catchment using mineral magnetic analysis and reference to SCA records of burn events within the catchment since 1960. This has enabled assessment of the influence of (i) different landscape units, (ii) the number of fires, and (iii) the time since the last fire on various magnetic signatures derived from within the catchment, which could possibly be used to build a picture of the sources of sediment within the catchment deposited on the lake floor. The potential and limitations of applying these techniques to sediment stacks for the purpose of reconstructing fire history, severity and frequency are discussed.