Fire creates complex patterns of burnt, scorched and unburnt vegetation, commonly referred to as fire severity. This study examines the spatial complexities of fire severity and explores the relationships between common methods of measuring fire severity and their utility in predicting the effects of fire severity on plant ecological processes. Experimental burns were undertaken in dry sclerophyll vegetation in Gibraltar Range National Park, NSW, Australia. Following fires, fire severity was assessed using three common methods (scorch height, burn tip diameter on shrubs, leaf scorch on graminoids). In addition, the relationship between on-ground measurements of fire severity and ecological processes known to be influenced by fire severity/intensity were also examined (emergence of seedlings of serotinous species, emergence of seedlings of hard-seeded species and resprouting ability of shrubs). Our results indicate that fire severity at the canopy level, understorey level and below-ground were not strongly correlated with each other. Furthermore, fire severity measurements in closest proximity to the ecological process being measured were found to have the strongest relationships. Finally, these results indicate that fire severity should be measured at all strata (below-ground, understorey level and canopy level) in order to examine the overall fire severity at a site. © Springer Scienc+Business Media Dordrecht 2016.