Temporal trends in mammal responses to fire reveals the complex effects of fire regime attributes

Abstract

Fire is a major ecological process in many ecosystems worldwide. We sought to identify which attributes of fire regimes affect temporal change in the presence and abundance of Australian native mammals. Our detailed study was underpinned by time series data on 11 mammal species at 97 long- term sites in southeastern Australia between 2003 and 2013. We explored how temporal aspects of fire regimes influenced the presence and conditional abundance of species. The key fire regime components examined were: (1) severity of a major fire in 2003, (2) interval between the last major fire (2003) and the fire prior to that, and (3) number of past fires. Our long- term data set enabled quantifica-tion of the interactions between survey year and each fire regime variable: an ecological relationship missing from temporally restricted studies. We found no evidence of any appreciable departures from the assumption of independence of the sites. Multiple aspects of fire regimes influenced temporal variation in the presence and abundance of mammals. The best models indicated that six of the 11 species responded to two or more fire regime variables, with two species influenced by all three fire regime attributes. Almost all species responded to time since fire, either as an interaction with survey year or as a main effect. Fire severity or its interaction with survey year was important for most terrestrial rodents. The number of fires at a site was significant for terrestrial rodents and several other spe-cies. Our findings contain evidence of the effects on native mammals of heterogeneity in fire regimes. Temporal response patterns of mammal species were influenced by multiple fire regime attributes, often in conjunction with survey year. This underscores the critical importance of long- term studies of biota that are coupled with data sets characterized by carefully documented fire history, severity, and frequency. Long- term studies are essential to predict animal responses to fires and guide management of when and where (prescribed) fire or, conversely, long- unburned vegetation is needed. The complexity of observed responses highlights the need for large reserves in which patterns of heterogeneity in fire regimes can be sustained in space and over time.