FIRE INTERVAL SEQUENCES TO AID IN SITE SELECTION FOR BIODIVERSITYSTUDIES: MAPPING THE FIRE REGIME

摘要

Determining the impact of fire regimes on biota is often limited by the lack of good knowledge about fire history: where have fires occurred, how big were they, and at what time of the year and with what intensity did they burn? On the other hand, where fire history has been well documented, the complexity of this information can be daunting. In this paper, we show how a simplification of complex fire history data into sequences of fire intervals can provide a basis for studying the impact of contemporary fire history on biodiversity. We used a retrospective approach to classify historical fire intervals into descriptive units (short, moderate and long intervals). In particular, we wanted to view the sequence of past fire intervals within the spatial framework in which it exists (i.e. where in the landscape do contrasting fire interval patterns exist?). Our study centred on an area of 50 000 ha northeast of Walpole, Western Australia, that was last burnt in the fire season of 2002/03. This provided a unique opportunity to retrospectively investigate fire interval sequences using a common, recent and widespread fire as a baseline. Our fire history dataset spanned back to 1972, providing a maximum of 30 yearsbetween the least- and most-recent fires. All fire occurrence data was held in a Geographic Information System (GIS), with spatial information for fire boundaries, and years in which fires occurred. By exporting the database file (.dbf) associated with fire occurrence into Excel, we were able to assign codes for short, moderate and long intervals between successive fire events. Then, these codes could be combined for each landscape patch with a unique fire history in space and time to produce sequences of fire intervals. The result of this technique was a map of polygons with a display of their fire interval sequence in reverse time series. This representation of the results provided a quick overview of how the pattern of fire intervals differed across the study area, and we found it useful in determining those areas that have burnt with either successive short or long fire intervals. This method is innovative, cost-effective and attempts to deal with the problems of complex multidimensional data.