Technical ReportPDF Available

A review of the role of Community Fireguard in the 2009 Black Saturday bushfires

Authors:
  • University of Melbourne / Flinders University

Abstract and Figures

This study investigated the role of Community Fireguard groups in the Victorian 2009 bushfires, finding that active members of Fireguard groups had higher rates of home defence and house survival than inactive members and neighbours not in Fireguard groups. In addition, the homes of active Fireguard members had higher survival rates even when they were not defended.
Content may be subject to copyright.
Community Fireguard in 2009 Victorian fires—DRAFT INTERIM REPORT 1
a review of the role of
community fireguard
in the 2009 Victorian bushfires
final report — august 2009
Community Fireguard in 2009 Victorian fires 2
This project was undertaken by the CFA
and acknowledges the input of Alan
Rhodes, Dr. Danielle Clode, Helen
Wositzky, Ray Fritz, Liz Langford, Eli
Niall, Karen Wealands, Harry Smiles and
Adam Jenkins.
CFA
Community Safety Directorate
PO Box 710
Mt Waverley VIC 3149
Australia
www.cfa.vic.gov.au
Draft interim report: June 2009
Final report: August 2009
© CFA 2009
Community Fireguard in 2009 Victorian fires 3
Summary
This study investigated the role of Community Fireguard groups in the
Victorian 2009 bushfires. The study recognises that Community Fireguard
groups are highly diverse, community based structures with a great range of
experience, engagement and interest in fire preparedness. This study did
not attempt to measure fire preparedness by individuals or groups, but
investigated the outcomes of the fire for different groups and households in
relation to their location, level of training and level of involvement in the
Community Fireguard program.
There are currently around 1225 Community Fireguard groups in Australia
covering an estimated 13,450 households. Fireguard groups are
disproportionately located in high bushfire risk areas. Seventy-seven
fireguard groups were within the fire zone on Black Saturday, primarily in
Kilmore-Murrindindi but also the Churchill and Bunyip fires. They
experienced a wide range of fire severity from ember attack and patchy fires
to extreme fire conditions. Many groups suffered house losses and eleven
groups had members who died.
Losses were strongly related to the level of fire severity, which was closely
associated with vegetation type (primarily level of clearing). However, the
data suggests that loss is also mitigated by active defence and active
participation in a Community Fireguard group the following factors:
active defence by household
80% of defended houses were saved
46% of undefended houses were saved
membership of a Community Fireguard group
regular members of fireguard groups saved 79% of their homes
irregular members of fireguard groups saved 64% of their homes
neighbouring non-members saved 65% of their homes.
Regular members of fireguard groups are more likely to defend their homes
than irregular members, which explains much of the variation in house
survival. However, even undefended homes of regular fireguard members
were more likely to survive the fire than those of irregular members,
suggesting that their properties are better prepared and more defendable.
People who join Community Fireguard groups are likely to be more
interested, engaged and committed to bushfire preparation so these results
are likely to be influenced by self-selection. However, anecdotal self-reports
from participants suggests that the training received from the Community
Fireguard program was instrumental in saving lives and property.
Further research is required to investigate the precise way in which the
benefits of the community fireguard program are transmitted to regular and
active members.
Community Fireguard in 2009 Victorian fires 4
Issues
Lack of information
Accurate analysis of the impact of the fire on community fireguard
groups has been hampered by a lack of data from other agencies, in
particular accurate fire severity mapping and restrictions on
information in relation to fatalities.
Resourcing of Community Fireguard
The evidence of this report suggests that active participation in a
fireguard program halves the risk of house loss This represents a
substantial economic saving, from a program that operates on a
minimal budget of $400,000 per year statewide, as well as reducing
the risk of fatalities. The potential benefits of an expanded and more
actively supported program need to be considered relative to costs.
Co-ordination of Community Fireguard
Investigation of the regional differences in implementation and
evaluation of community fireguard to ensure that all communities
within regions are equitably resourced.
Evaluation of Community Fireguard
Evaluation of the efficacy of information transfer from trainers to
facilitators to fireguard members
Evaluation of regional implementation of program
Evaluation of relative benefits of different sessions of the program
Further research
Assessment of the relative merits of the social aspects of the
community fireguard as well as the educational benefits.
Specific investigation of the circumstances of death of community
fireguard members in relation to their fireguard activity, in a similar
way to the investigation of firefighter fatalities or near-misses in an
operational context.
Community Fireguard in 2009 Victorian fires 5
background
History of community fireguard
Improving safe public responses to fire through education and information is
a core element of reducing losses and fatalities during bushfires.
Traditionally, information has been provided passively, in the form of radio
and television advertising, brochures and flyers (Petris, 1995). The
limitations of these passive systems of information delivery have been
repeatedly illustrated in post incident analyses of community responses to
fires.
The Country Fire Authority of Victoria started its first Community Fireguard
program in February 1993. The Community Fireguard program provides
facilitators who “assist small neighbourhood groups to take responsibility for
their own fire safety, and to develop strategies for reducing their vulnerability
from major fires.” (Petris, 1995, p22). This program uses a model of public
education based on active participation to complement and extend passive
information delivery on bushfire safety and promote bushfire preparation and
planning.
The program was decentralised in late 1995, giving responsibility for delivery
of the program to the areas. In 1998 a review of Community Fireguard
identified the need for improved understanding of the program within various
levels of the CFA (headquarters, regions and brigades) as well as within the
community (McWaters, 1998). These recommendations led to the
development of a more formal core program for delivery to fireguard groups.
This core program was implemented in 2006.
There are seven sessions in the core program including:
Introduction to Community Fireguard
Fire Behaviour
Personal Survival
House Survival
Street Walk
Fire Equipment
Bushfire Survival Plans
Each session can be delivered as a stand alone session however they can
also be combined or integrated into other sessions. The way in which the
program is delivered may depend upon the facilitator, the requirements of
the group and the perceived level of commitment by the group.
Due to the regional implementation of the fireguard program it is difficult to
obtain a precise budget across the state. Most staff employed to co-ordinate
or deliver fireguard programs do so on a part-time basis. However the
program is estimated to cost no more than $400,000 per year to deliver
statewide.
Community Fireguard in 2009 Victorian fires 6
Uptake of community fireguard
Uptake of the program has steadily increased since its formation (see Figure
1) with an annual formation of 50-150 new groups each year (average=97.8).
Peak years for group formation were 1997-2000 and 2006-2007 (Community
Safety Summer Report 2008-2009). These spikes are probably due to major
fire events, in particular fires in the Dandenongs in 1997/98 and widespread
fires across the north-east and Gippsland (2006/07). The uptake of
Community Fireguard programs is also influenced by regional allocation of
resources to the promotion of the program. Over the sixteen years of
operation, 403 groups have been disbanded, leading to a total 1225 current
groups .
Figure 1:
Growth in
Community
Fireguard
groups 1993-
2009.
Each of these groups is thought to cover between ten and fifteen houses (average of eleven,
based on Yarra Region groups). An average of thirteen people attend the first meetings of a
Community Fireguard group (CFA Community Fireguard database). An estimated total of
13,450 households (or 35,000
individuals) are covered by
Community Fireguard in Victoria.
Community Fireguard groups are
actively promoted in high fire risk
areas. They are most frequent in
the areas of highest risk and the
areas where they have the longest
history. They appear to suit
interface neighbourhoods
particularly well. The distribution of
fireguard groups across Victoria is
shown in Figure 2.
Figure 2: Community Fireguard groups
by region
0
200
400
600
800
1000
1200
1400
1600
1800
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
Year
Number of groups
Old groups Newly formed
Community Fireguard in 2009 Victorian fires 7
Impact of community fireguard
A 1998 study of Community Fireguard groups found that members of such
groups had high levels of knowledge and information about bushfires and
bushfire safety than non-members. CFG members tended to have more
sophisticated strategies in preparation for bushfires than non-members,
although there was no difference in relation to house preparation between
the two groups (Rohrmann, 1998).
Anecdotal evidence of CFG performance during fires frequently points
towards them being well-organised and responding successfully to fires
threatening their homes (e.g. Boura, 1998 p 11-12). Many CFG facilitators
report that fireguard groups typically respond better during a fire and recover
faster after a fire. Anecdotal reports suggest that fireguard members tend to
have a stronger internal locus of control in relation to fire, recognising how
their actions and preparation affects the impact of the fire. Their responses
to fire, even when the outcomes have been negative, tend to focus on what
they would do better or differently next time. By contrast, general members
of the community often respond to fire with a sense of lost control, causing
them to seek responsibility for what happened elsewhere and external to
their own actions.
The Black Saturday fires impacted on an area with a particularly high density
of fireguard groups. Over one hundred Community Fireguard groups were
directly or indirectly impacted by fires, the majority of these groups being in
the Yarra area (region 13), Outer metro north-west area (region 14),
Gippsland area (regions 9 and 10) and North-east area (region 12). Table 1
outlines the number of groups falling within the fire zone. Many other groups
were located on the edge of the fire and came under ember attack.
Area Region Fire Number of
groups
Groups with
fatalities
Yarra 13 Kilmore 51 7
North east 12 Kilmore-Murrindindi 13 3
Outer metro north west 14 Kilmore 2 1
Gippsland 9 Bunyip 3 0
Gippsland 10 Churchill 7 0
77 11
Table 1: Overview of fire-impacted groups by region and fire.
The boundary between region 12 and 13 dissects several Kinglake
townships. As a result, many fireguard groups physically located within the
boundaries of region 12 are actually managed by region 13 in line with local
brigade management structures. In these cases, groups are discussed in
relation to their regional management, not physical location (See Appendix
A).
Community Fireguard in 2009 Victorian fires 8
methods
Two sources of data were used to investigate the impact of the community
fireguard program on fire outcomes. The first data source was the initial
welfare checks of the fireguard groups in Region 13, which documented a
wide diversity of outcomes ranging from minimal damage to high levels of
house loss and fatalities. Detailed analysis of the impacts of the fires on
Region 13 groups was conducted in relation to program completion, group
activity and fire severity.
Additional information was compiled for groups in other regions and is
presented separately at the end of the report. Maps of the location of fire
affected community fireguard groups are provided in the appendix.
The second source of data was the survey of households within each
fireguard group. This provides information on fire impact at a household
level, rather than group level and accounts for large differences in level of
engagement with the program and preparedness.
Region 13 study
Detailed analysis of the impact of the fires on Community Fireguard groups
was restricted to the region with the greatest number of groups (13) in order
to control of regional differences in program delivery.
All Community Fireguard groups in or near the fire zone in the region 13
were contacted by the regional Community Fireguard co-ordinator.
Information on losses and fatalities for each group was classified into
categories (See Table 2).
Code Self-reported losses
1 No house loss or damage
2 Some damage no house loss
3 Few houses lost (1-3, “few”)
4 Some houses lost (4-6, “some”)
5 Many houses lost (6+ “most”)
6 Many houses lost and fatalities
Code Fire severity rating (estimated from aerial photographs)
0 Unburnt - No evidence of burning, but within impact zone
1 Patchy - Some burning but canopy intact and green
2 Low - Some burning but canopy mostly intact and green
3 Moderate - Some scorching (under 50%) mostly green
4 High - Mostly scorched (more than 50%)
5 Very high - Some canopy burnt (<50% burnt)
6 Extreme - Nearly all canopy burnt - mottled ground
Code Vegetation coverage (estimated from aerial photographs)
0 Grassland/pasture/township
1 Mostly grass/pasture some bush/township with gardens and trees
2 Half pasture/half bush
3 Mostly bushland, some pasture
4 Low open woodland
5 Tall forest
Table 2: Categories of loss, fire severity and vegetation coverage
Additional records on each fireguard group was compiled from the central
fireguard database and regional records on when the group was started,
what aspects of the core program had been completed and how often the
group met. Program status was measured on which of the seven core
sessions of the core program had been delivered to the group. Since many
groups formed before the core program was standardised and not all groups
have met to receive all of the sessions of the core program, this score could
range from 0 to 7.
Community Fireguard in 2009 Victorian fires 9
Information on fire severity and vegetation type was compiled from aerial
photographs for each fireguard group (www. images.land.vic.gov.au) taken
shortly after the fires. Relative fire severity was based on the proportion of
green, scorched and completely burnt trees (see Table 2). DSE mapping of
fire severity halted mid-year due to other commitments and has not
completed or made available to date. No timeline is available for the
provision of this data.
Household study
The contact person for each group was further asked, to the best of their
knowledge, for the address of each member household, how regularly they
attended meetings, what the status of their house was after the fires and
whether or not the house had been defended during the fires. Participation in
this survey was entirely voluntary. Where information provided in surveys
differed from that provided in the welfare check, survey information was
used. Aerial photography was also used to verify information on house loss,
damage or location where data was unclear.
All returned surveys were mapped and non Community Fireguard
neighbours were selected as matched control groups for each fireguard
group. Where possible similar vegetation and fire experiences were
selected. These control households were then mapped for house loss or
survival and compared with their fireguard group.
Community Fireguard in 2009 Victorian fires 10
region 13 results
Feedback
During welfare checks, many Community Fireguard groups volunteered
information on how the program had contributed to their survival during the
fires. These responses were overwhelming positive in support of the
perceived benefits of the program. This finding is also supported by a in-
depth study of several fire-affected fireguard groups (Gibbs et al. 2009).
Only four out of sixty groups contacted in CFA region 13 indicated that no
amount of information or program could have helped them under the
circumstances they experienced. These groups suffered high fire intensity
and most lost group members and many houses. One group commented
“CFG was thrown out the window on the day because of the conditions”.
Another noted that “it was futile in their area, nothing could have been done.”
A third group commented that “Things taught in CFG are no longer true, the
rule book has changed.”
Of the sixty groups contacted, nineteen explicitly volunteered the benefits of
the Community Fireguard program and CFA education to their survival. A
number of others mentioned that their fire plans worked successfully, asked
for a facilitator or otherwise indicated that they valued CFA education
programs and support. Many of these groups were in the worst hit areas.
“All I did through the fire was exactly what CFG instruction book said,
prepared, sheltered inside. No doubt that if not had education from CFA
would not be talking to you today
“…survived because of the knowledge CFA education programs gave them
Because of what CFA had shown and taught us we did know what to do….It
is all through you guys and what you taught us that we are still alive.
Honestly because of you guys that’s why we are here.”
“The CFG group has been really essential in getting to know as many
neighbours as possible, in building the local community and in preparing us
for the risk. It has really worked and I want you to know that.”
“each of us have said that what we learned from our CFG meetings saved
our lives because we were prepared, knew what to expect in terms of fire
behaviour and did not panic.”
“CFG group & things done in CFG been most positive thing that contributed
to it not being tragic.”
“…it was because of CFG they knew what to do and had everything
prepared. CFG gives us strong, thoroughly thought through
recommendations. Absolutely helped to have CFG every year it gets
stronger & better culture of responsibility.”
“We were relatively prepared because we were part of CFG, it gave us
strategies to cope. Everyone should be in a CFG group.”
“Having CFG group helped a lot, having the right gear & linking together.
…survived because of the knowledge CFA education programs gave them.
Without having gone through CFG would not have been prepared.
There is little doubt that community members find the fireguard program to
be beneficial and useful and many attribute their survival to the advice and
support provided by the program. This research is intended to examine the
way in which the program delivers this benefit and how much of the benefit is
real or percieved.
Community Fireguard in 2009 Victorian fires 11
Nature of the fires
Across the impact areas different neighbourhoods experienced very different
types of fires. Many neighbourhoods experienced fires that were moderate
or patchy. Many households came under ember attack but did not
experience a direct flame front. For example, even within the township of
central Kinglake, fire severity ranged from patchy (significant areas unburnt)
to extreme (substantial areas of crown fire with total canopy loss). This is
illustrated in Figure 3-4.
Figure 3: The number of fireguard groups experiencing different fire intensities.
Fire severity is likely to be related to vegetation type, and amount of cleared
land. Since fire severity is largely determined in aerial photographs by the
status of the tree canopy (green, scorched or absent), completely cleared
land cannot carry extreme fires. This methodological limitation is, however,
likely to reflect the reality of fire severity since forested areas commonly have
higher fuel loads than unforested areas and are capable of supporting much
more extreme fire conditions.
All neighbourhoods in this study contained some trees and most were
heavily treed. Evidence of ground fires was still evident from aerial
photographs, most of which were taken within days of the fires.
0
5
10
15
20
25
Cleared/
township
Some trees Half bush
/half pasture
Mostly bush Open
woodland
Tall forest
Figure 4:
The number of fireguard groups in different vegetation types
0
2
4
6
8
10
12
14
16
Patchy Low Moderate High Very high Extreme
Community Fireguard in 2009 Victorian fires—DRAFT INTERIM REPORT 12
Figure 5:
An example of extreme fire severity in a predominantly forested area near central Kinglake.
The forest on the left exhibits signs of crown fire (no leaves left), while the trees along the road are
scorched. To the right of the road in cleared areas, the trees are unburnt.
Figure 6:
An example of a township area of central Kinglake which experienced patchy burning. Many
areas are unburnt, while other areas have scorched trees. These areas experienced strong ember
attack. (Source: images.land.vic.gov.au)
Community Fireguard in 2009 Victorian fires—DRAFT INTERIM REPORT 13
Impact of the fires
A total of 51 current Community Fireguard groups were directly impacted by
fire in Region 13. Of these groups seven have been identified as having fire
related fatalities. Further information on fatalities has not been released by
the police and CFA is unable to confirm or identify fatalities within these
groups. Three other groups in region 13 also recorded fatalities. In one case
the fatality was from heart failure and not regarded as a fire fatality. In
another group the deaths were not fireguard members although they
occurred on fireguard properties and did not appear to relate to the activities
or preparation of the fireguard members directly. The third group had only
recently formed as a fireguard group and had not yet met or received any
training. These fatalities are not included in this analysis.
The suppression of fatality data by the police has not allowed any official
confirmation of these figures, or analysis of the circumstance of death. The
police have not used the information obtained by the CFA on community
fireguard membership, level of engagement, household preparedness or
training in their assessment of the circumstances of death. No timeline is
available for the provision of fatality data or the analysis of circumstances of
death.
Group location Township Group size No. of
fatalities
No of households
with fatalities
Bald Spur Road Kinglake 12 13 5?
Ward Street Kinglake 21 4 3?
National Park Road Kinglake 17 3 2
Jacksons Road St Andrews 14 6 3
Wild Dog Creek Rd St Andrews 8 1 1
Old Kinglake Rd Steels Creek 15 4 2
Hargreaves Road Steels Creek 18 1? 1
Table 3:
Possible fatalities in region 13 Community Fireguard groups as currently known to CFA.
The impact of the fires on the groups in region 13 ranged from no loss or
damage to severe loss and damage (see figure 7).
0
2
4
6
8
10
12
14
16
No damage Some damage Few houses lost Some houses
lost
Many houses
lost
Many houses
lost and
fatalities
Impact severity
Number of groups
Figure 7:
Numbers of groups by impact severity for region 13
Community Fireguard in 2009 Victorian fires 14
Nature of the groups
There is considerable variation between groups in relation to how big they
are, how long they have been established and how often they meet. The
average group size is twelve households, ranging from as few as four to as
many as thirty-two (see Figure 8a). The majority of groups are less than four
years old, with a large cluster of groups are three years of age and a second
smaller cluster of much older groups, averaging around eleven years of age
(Figure 8b).
Size of group
35302520151050
Frequency
16
14
12
10
8
6
4
2
0
Mean =12.37
Std. Dev.
=5.586
N =51
Age of group (years)
12.5107.552.50
Frequency
15
10
5
0
Mean =4.61
Std. Dev. =3.476
N =51
Figure 8:
Size and age of Community Fireguard groups
On average groups meet with a CFA facilitator once a year, however some
groups meet more much more often (up to five times a year) particularly
those that are completing the core program. Older groups (over seven year)
meet less frequently on average (0.62 times per year) than younger groups
(1.55 times per year). A number of older groups have not met with a CFA
facilitator since they were first formed, more than ten years ago although it is
possible that they meet more regularly on their own.
Community Fireguard in 2009 Victorian fires 15
Program completion by the groups
The groups also differ in relation to the level of education and program
information they receive. Groups established prior to the core program in
2003 may have received very variable information, depending upon their
willingness to meet and the facilitator presenting. Since the establishment of
the core program, records have been kept of which sessions within the core
program are provided to each group, allowing us to document how much
training each group has done.
Of the 51 groups examined, five had only participated in one session and
could not be regarded as engaged in the program fully. Only sixteen groups
had received all six or seven sessions (fire equipment being regarded as an
additional, rather than core session) and could be considered to have
completed the program (Figure 8).
0
2
4
6
8
10
12
14
1234567
Number of sessions completed
Number of groups
Figure 8:
Frequency distribution of number of sessions completed by groups.
The sessions on fire behaviour, house survival and survival plans were the
ones most likely to have been completed (Table 5).
Session Number of groups
receiving session
Introduction to CFG 25
Fire behaviour 44
Personal survival 33
House survival 38
Street walk 24
Fire equipment 22
Survival plans 36
Table 5:
The frequency with which different session are delivered.
Relationship between group characteristics and losses
The losses experienced by the different Community Fireguard groups were
most strongly associated with the severity of the fire they experienced (r
2
=
0.719, p<0.05, df=49). In other words, seventy percent of the variation in
losses experienced by fireguard groups could be attributed to the severity of
the fire they experienced.
Fire severity was also strongly correlated with vegetation type, with the most
heavily treed areas (as expected) experiencing the most severe fires
(r
2
=0.683, p<0.05, df=49). As a result, vegetation type also correlated with
losses (r
2
=0.410, p<0.05, df=49). It is likely that more accurate
measurements of fire behaviour (including topography and wind) are likely to
have a strong effect on losses experienced.
The were no differences between the losses experienced by the groups in
relation to program completion, meetings attended, size or age.
Community Fireguard in 2009 Victorian fires 16
north east area
Thirteen Community Fireguard groups from North east area (region 12) were
impacted by the Kilmore Murrindindi fires. Three of these groups suffered
fatalities.
Group location Township Group size No. of
fatalities
No of households
with fatalities
Marysville Marysville 12 6 3
Kerami Crescent Marysville 9 1 1
Keppel Court Marysville Unknown Unknown
Table 3:
Possible fatalities in region 12 Community Fireguard groups as currently known to CFA.
Welfare checks were made of all groups by headquarters staff by agreement
with the region. Details of groups on the central database contained
extensive errors which made contact difficult in many cases. Some phone
trees were provided by region which assisted in group contact. Some
groups no longer existed as groups and were excluded from this summary.
One group in Marysville could not be contacted because of the death of the
contact person. All groups were invited to participate in the voluntary survey.
All agreed, and five returned surveys.
Most of the groups impacted in region 12 appear to have been established
some time ago and it is difficult to assess whether or not they have
participated in the same type of training as groups in other regions.
Heathcote Junction/Wandong
There is one fireguard group in this area which reported to be a well-
prepared, well-organised group. They used a scanner to monitor the fire and
activated their phone tree as soon as there was news of a fire at East
Kilmore. Some members left and other stayed. The fire came within 100m
of the group on the other side of the road and the group suffered no losses
or impact.
A second group listed as “Wandong” appears to have the same contact
details as the Heathcote Junction/Wandong group and has been removed
from the database.
Hazeldene/Flowerdale
The Hazeldene/Flowerdale area has four fireguard groups. The groups can
be characterised as being pockets of relatively small subdivisions with close
house proximity in areas surrounded by forest.
The Long Gully Road group started in 1999, but struggled to become self-
reliant and is no longer active. This is a high risk location which has burnt in
the past. There were initial reports of deaths in Long Gully Road from the
February fires, but these have not been confirmed. Houses losses and fire
severity appear to have been greatest at the end of Long Gully Road, rather
than the safer start of the street where fireguard interest was higher.
Three groups cluster around the Silver Creek development in the southern
part of Flowerdale. Households in this area are on very small allotments,
with close proximity to neighbouring houses and to the surrounding forest.
Creekside Drive group had 3-4 meetings several years ago and have a
current phone tree. They lost all houses in their group bar one.
King Parrot Creek Group was established in 2001 and has had six meetings
although some members have joined since them and are only on the phone
tree. This group had a strong expectation that they would be told when the
fire would arrive and to activate their phone tree and felt let down that they
were not warned personally. The same issues had been raised in (CFA
2004) and led to a clarification of CFA policy that fireguard groups should be
Community Fireguard in 2009 Victorian fires 17
advised that they are not likely to be personally warned and that the phone
tree is not activated by the CFA.
Silver Parrot Road fireguard was established in 2004 and is regarded as self-
reliant. They had not met for eighteen months before the February fires.
Most houses were lost with only around twelve reported to have survived.
Fire severity was high in the area with evidence of nearby crown fires.
Figure 10:
Approximate locations of three Flowerdale community fireguard groups,
illustrating small block developments within forest (Source:google maps).
Glenburn area
The Glenburn area includes five fireguard groups, although the status of
these groups has varied over time as large groups have split and reformed.
Most of the area is open farmland with large property sizes and widely
dispersed homes.
Most groups experienced patchy burning, with some property damage but no
houses lost in the fireguard groups. One group noted that they did not see a
fire truck or receive a warning about the fire. Another group experienced
extreme ember attack, but no fire front. They felt that training did not match
reality. Three non CFG houses were reported lost on the street, although
these were not apparent from aerial photography.
Marysville
Three fireguard groups were established in Marysville in 2006. They each
completed two-three sessions before their facilitator was called away for
other duties and the core program was not completed. All three groups in
Marysville suffered high house loss and fatalities, including the contact
person of one group. The two remaining groups returned their surveys and
one contact person who has since moved away from Marysville wishes to
start a new fireguard group in her area. This group also reported difficulty
getting people involved and committed to fireguard and stressed the need for
maintenance sessions at least every eighteen months.
Community Fireguard in 2009 Victorian fires 18
Buxton-Taggerty
There is also a very small dispersed group drawing on a few individuals from
the townships of Buxton and Taggerty. They completed the core program
but had great difficulty gaining members. Further assistance was required to
promote the program.
Narbethong
The only group in Narbethong operates along Tarnpirr Road between
cleared land and forest. Only five houses survived out of 14-15 houses.
There were no fatalities and most of the houses that were defended
survived. The adjoining forest shows evidence of crown fire.
Conclusion
Groups that suffered the greatest losses in region 12 appeared to be those in
areas of relatively high intensity housing with close proximity to forest.
Despite the extreme severity of the fires experienced by some groups, the
general response was favourable towards the community fireguard program.
All three facilitators who delivered programs in this region were praised and
thanked for their advice.
One group reported disappointment that they had not been personally
contacted by the CFA and that no-one had come to tell them what to do.
Another group felt that training was not relevant to the reality they
experienced on the day.
Many groups reported that it was difficult to recruit members. Several
groups asked for assistance promoting the program and a greater emphasis
on recruitment and advertising. There was a sense that with appropriate
support many of the contact people would be willing or were keen to
continue and expand their commitment to community fireguard.
In general there was strong support for the program tempered by
disappointment that it wasn’t better resourced and promoted.
Community Fireguard in 2009 Victorian fires 19
gippsland
Gippsland area had groups affected by two fires. In region 9, three fireguard
groups were impacted by the Bunyip fire, but none returned surveys. In
region 10, seven fireguard groups were impacted by the Churchill fire. There
were no fatalities in these groups.
All fireguard groups were contacted by their facilitators after the fires and in
many cases facilitators contacted all members of the groups individually.
Many of the impacted fireguard groups in Gippsland are surrounded by or on
the edge of large expanses of forest. Properties tend to be large, but some
groups are located in smaller isolated subdivisions within forested tracts.
Churchill
One fireguard groups in the area surrounding the Bunyip fire reported
difficulties with lack of services, particularly electricity and phones, despite
the fact that they were not burnt themselves. Restrictions to road traffic
caused considerable problems for residents, both those who stayed and
those who left. Difficulties were also experienced with incorrect advice
provided by police and lack of fuel reduction on road sides and nearby parks.
Two groups on the edge of the fire area reported property damage including
loss of sheds and fences. One group noted difficulties for fire agencies due
to inconsistent road numbering. This group also stressed the value of the
fireguard program to educating and assisting them to understand and
address their bushfire risk, but noted that those households most in need of
assistance did not participate.
One small group in the fire impact area in Callignee lost two houses, neither
of which were defended, with considerable property damage. They noted
that the group did not meet regularly and new residents were unaware of the
group. The group abuts extensive young blue gum plantations (up to eight
years old) which show evidence of patchy canopy fire.
A second larger group in Callignee lost nine of its fourteen houses. Despite
much cleared land, areas of adjacent forest carried crown fires. This group
noted the importance of keeping the phone tree up to date and the difficulty
of preventing complacency. They also mentioned the burden of
responsibility shouldered by contact people in the group when plans go
wrong, even though preparation is an individual responsibility.
Community Fireguard in 2009 Victorian fires 20
outer metro north west
There were only two fireguard groups in the outer metro north west area
affected by the Kilmore fire, both of those on the eastern border of region 14
in Strathewen and Kinglake West. Both were contacted by headquarters
staff by agreement with the region.
Group location Township Group size No. of
fatalities
No of households
with fatalities
Chads Creek Road Strathewen 11 2 1
Whittlesea-Yea Road Kinglake West 20 2 1
The group in Strathewen was a relatively informal group and not particularly
active at the time. It was reported to be difficult to get people involved and
some difficulties were experienced with early facilitators. The group suffered
high losses in line with the surrounding area. Wind damage, prior to the
southerly change, was high in the area.
The Whittlesea-Yea Road group reported as a highly prepared and
organised group with a high level of support from local brigades. They
received four meetings which were well-attended. although were initially
resistant to participating in the program. This group lost all but two houses
and had two fatalities at one address. This group backs onto forest which
shows evidence of extensive and severe crown fire. Anecdotally, wind
damage was reported as high. One house was lost even with fire trucks in
attendance.
Community Fireguard in 2009 Victorian fires 21
household analysis
Data on loss, defence and CFG attendance was recorded for each individual
household in each group by the group co-ordinator. This self-reported data
provides precise information on the number of households in each group and
their level of involvement in the group, as well as a broad measure of
behaviour during the fire (defended or defended house) and outcomes (lost
or saved).
Five groups which did not return surveys were reconstructed from
membership lists on file and aerial photography. These groups did not
include information on house defence or membership regularity.
Area Region Valid
Surveys
Total number
of groups
Gippsland 9
0 3
Gippsland 10 4 7
North east 12 2 13
Yarra 13 31 57
Outer metro north west 14 2 2
Total 41 82
Table 6:
Survey returns by region
Forty one surveys were completed in total, providing a sample of 50% of all
fire affected groups. The returned surveys represented a cross-section of
groups across fire severity and region, including all but two of the highest
impact groups (see table 6).
House defence
The data on these groups reveals a strong relationship between house
defence and house survival. Houses that were reported to have been
defended were much less likely to be lost than those that were undefended
(X
2
= 45.25, df=2, p<0.05). Eighty percent of defended houses were saved,
whilst fewer than half (46%) of undefended houses were saved.
It is possible that this effect may be strengthened by a bias in reporting.
People may be more likely to report a house being defended that was saved,
and people may be likely to leave a house that has caught fire, thus leading
to it being reported as undefended. However the comments on the survey
forms indicate that many houses that were lost were also actively defended
and well-prepared.
Figure 11:
Relationship between house defence
and survival.
0%
20%
40%
60%
80%
100%
Defended (N=203) Undefended
(N=176)
Saved Lost
Community Fireguard in 2009 Victorian fires 22
Fireguard membership
Many groups reported that they had no regular members. Some reported
only one regular member (the contact person themselves), with most
members only being interested in the phone tree. A few groups, however,
were very active, with all members being regular attendants.
There is a strong relationship between the level of involvement with the
community fireguard group and house survival. Members listed as regular
attendants in the Community Fireguard group were more likely to have
saved their homes than those listed as being occasional members, phone-
tree or newsletter only members (Chi
2
=12.57, df=2, p<0.05). Eighty percent
of regular CFG members saved their homes while only 55% of irregular
members saved their homes.
Non fireguard properties adjoining fireguard
properties (controls) suffered a similar level of
property loss (64%) to the irregular members
of the group. General community house loss
in the fire area is estimated to be between 30-
40% suggesting that irregular membership of a
fireguard group does not confer any benefits.
Figure 12:
House loss rates across regular and
irregular fireguard members compared to non
fireguard groups. (Chi
2
= 15.92, df=2, p<0.05)i
Regular members of Community Fireguard groups were much more likely to
defend their homes during the fire than irregular members (Chi
2
=32, df=1,
p<0.05) which probably explains the bulk of the improved house survival.
However, the undefended homes of
Community Fireguard members were also
more likely to survive (60%) than the
undefended homes of irregular members
(44%) (see Figure 13). This trend suggests
that the homes of regular Community
Fireguard members are perhaps better
prepared, and better able to survive fires, than
irregular Community Fireguard members.
Figure 13:
House survival of
undefended
homes of
regular and irregular Community Fireguard
members (Chi
2
=3.84, df=1, p=0.05)
These results suggest that active membership of Community Fireguard is
associated with a better ability to protect homes during a fire. Increased
house survival is likely to be associated with increased personal survival.
Most of the fatalities associated with Black Saturday fires appear to have
occurred either outside or in a home that burnt down.
Increasing house survival from 65% to 80% in the Black Saturday fires would
have saved in the order of 870 houses, reducing losses from an estimated
2029 residential properties to 1159.
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Regular
(N=188)
Irregular
(N=194)
Controls
(N=467)
Saved Lost
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Regular (N=55) Irregular (N=101)
Survived Lost
Community Fireguard in 2009 Victorian fires 23
Household actions
Many groups provided additional information on their fire experiences in their
survey returns which provide a valuable insight into how Community
Fireguard operates before, during and after a fire.
Many groups not only provided information on whether or not a house was
defended, but also provided details of the circumstances and timing of house
defence or departure. Although the purpose of the question was to control
for the effect of house defence on house loss, the responses to the question
can also provide information on the actions of residents during the fire.
Residents of “defended” homes can be assumed to have stayed (except
where the surveys explicitly state that the home was defended by neighbours
or firefighters). Residents of “undefended” homes either left or where
absent.
Figure 11:
Reported actions of fireguard households
Of the 310 households whose actions were known 69% stayed at their
properties, although two reported doing so even though they had planned to
leave. A small number of households reported that some members left
while others stayed. This mixed strategy may have been more common and
be included in the “stayed” reports. Thirty percent of households left or were
away on the day. Of the 58 households for which information on time of
departure was reported, 75% left late and 25% left early.
213
households
stayed
16 households we
re
away
119 household actions are
unknown
76 households left
429 household surveys
310 household
actions are known
5 reported mixed
strategies (some
members left and some
stayed)
2 planned to leave
but couldn’t
15 reported leaving
early
43 reported leaving late
Community Fireguard in 2009 Victorian fires 24
references
Boura, J. (1998). Community Fireguard: Creating partnerships with the
community. CFA Occasional Paper No 2. Melbourne, Country Fire Authority,
16pp.
Country Fire Authority (2004) A review of community fireguard, Community
Safety, Melbourne,CFA, 30pp.
Country Fire Authority (2009) Community Safety Summer Report 2008-2009,
CFA, Melbourne.
Gibbs, L. et al. (2009) CFA Post-Fire Qualitative Research: A descriptive
analysis of Community Fireguard Group members’ experiences of the 2009
Victorian bushfires, The McCaughey Centre, University of Melbourne.
McWaters, V. (1998). Review of the Community Fireguard program.
Melbourne, Integra Pty Ltd: 13pp.
Petris, S. and P. Potter (1995). A review of all state and federal reports on
major conflagrations in Australia during the period 1939-1994 and a national
bushfire preparedness strategy. Melbourne, Australian Fire Authorities
Council and Emergency Management Australia: 44.
Rohrmann, B. (1998). Assessing the impact of fire risk communication. InFire
Conference - Fire information for the 21st century. Melbourne.
ResearchGate has not been able to resolve any citations for this publication.
A review of community fireguard
Country Fire Authority (2004) A review of community fireguard, Community Safety, Melbourne,CFA, 30pp.
CFA Post-Fire Qualitative Research: A descriptive analysis of Community Fireguard Group members
  • L Gibbs
Gibbs, L. et al. (2009) CFA Post-Fire Qualitative Research: A descriptive analysis of Community Fireguard Group members' experiences of the 2009
The McCaughey Centre
  • Victorian
Victorian bushfires, The McCaughey Centre, University of Melbourne.
Review of the Community Fireguard program
  • V Mcwaters
McWaters, V. (1998). Review of the Community Fireguard program. Melbourne, Integra Pty Ltd: 13pp.
A review of all state and federal reports on major conflagrations in Australia during the period 1939-1994 and a national bushfire preparedness strategy. Melbourne, Australian Fire Authorities Council and Emergency Management Australia
  • S Petris
  • P Potter
Petris, S. and P. Potter (1995). A review of all state and federal reports on major conflagrations in Australia during the period 1939-1994 and a national bushfire preparedness strategy. Melbourne, Australian Fire Authorities Council and Emergency Management Australia: 44.
Assessing the impact of fire risk communication. InFire Conference -Fire information for the 21st century
  • B Rohrmann
Rohrmann, B. (1998). Assessing the impact of fire risk communication. InFire Conference -Fire information for the 21st century. Melbourne.