Limitations, assumptions, and considerations
Methods
- This work does not establish a change in fire risk between ‘TIME A’ and ‘TIME B’ using incident data. It establishes a baseline that facilitates assessment of a ‘TIME A’ and ‘TIME B’ risk change for the next review.
- The analysis of MVAs was introduced late in the analysis. Therefore, only kernel density estimation (KDE) analysis was undertaken on the MVAs, rather than them being part of the entire incident dataset, on which the various spatial techniques were undertaken.
- This methodology is not static. It will continue to be enhanced and adjusted as more research is undertaken and more data becomes available.
Data
- Not all incidents serviced by the CFA and FRV are considered by this report. Only incidents classified as fire and explosion data or MVAs are considered.
- False alarm data was examined to understand the demand pattern for such incidents; however, the result from that analysis was not directly used in the review process.
- Because the ABS variables were only available in SA1 scale, the base unit for fire risk analysis was conducted on the SA1 scale. Future reviews may employ a grid-based analysis or SRA-based analysis.
- SDS is different for FRV and CFA. To implement a consistent approach to the analysis of response time, response time data was only used to ascertain whether the SDS was met or failed for each incident. As a result, there was no requirement for the Emergency Services Telecommunications Authority (ESTA) data.
- FRV and CFA have different terms to describe the area that individual stations respond to. As a result, the term ‘station response area’ or SRA has been used in this report.
- Hydrant data was not used in this analysis but remains a planning consideration if boundaries are to be changed.
- Only human impact, in terms of casualty47 as a result of fire, is considered in this work. Economic impact is not considered. Property loss estimates recorded at the time of incident by the fire services are not sufficient for analysis, and insurance estimate data are not available at the SA1 scale.
- When providing population estimates for suburbs from the ABS Census, the suburb boundaries do not align with the SRA boundaries. As a result, the population statistics do not exactly represent the population demand of the SRA.
- This review used population projections for 2036 at the SA2 scale. Population projection for 2036 was captured for each SRA based on the proportion of SRA falling within the SA2s.
- Coronial data, which could provide an insight to deaths that did not occur at the time of the fire but after a period of time, was not considered in this review.
- Street network data (HERE data) was not utilised as part of this review. Access to this data is recommended for future reviews to enable enhanced spatial accessibility and coverage analysis.
- Regression modelling results conducted at SA1 scale was not included in this review. Future reviews may consider a different approach in regression modelling (for example, a point-based regression modelling).
Footnotes
[47] Casualty = Injury + death
Updated