Established and emerging uses of predictive services in Victoria

The recent 2019-20 Black Summer bushfire season brought both sector and public attention to the important role that predictive services can play in mitigating and preventing the loss of life and assets from bushfires.

This research study focused on two key questions:

1. What are the key features of effective relationships between Fire Behaviour Analysts (FBANs) and their users within emergency management?
2. How might the predictive services role and predictive services outputs be developed in the future?

We explored these questions by conducting twenty-five (25) semi-structured interviews with key practitioners who either perform the FBAN role or use predictive services in Victoria.

Our findings are categorised into 7 topics:

1. Interviewee background: The interviewees included people who represent the following operational roles: FBAN, State Regional Commander, State Agency Commander, Level 3 Incident Controller, Level 3 Public Information Officer, Level 3 Planning Officer, Level 3 Operations Officer, and Level 3 Situation Officer.

2. FBAN outputs: We collected data from FBANs and other operations staff who occupy roles which interact with and/or use the outputs produced by FBANs from within the broad emergency and incident control arrangements provided for within the Australasian Inter-service Incident Management System.

We found that the FBAN advice and products are used to:

• inform operational decisions, including pre-event planning, decisions during first attack and extended attack;
• inform the development and release of public information and warnings; and,
• support the conduct of planned burning.

Most interviewees agreed that the design and framing of outputs should be standardised for public dissemination, however, there was less consensus that such standardisation is required for operational purposes.

3. Trust in FBAN outputs: We found that in general:

• there is a high level of trust in FBAN outputs amongst participants who are users of predictive services; and,
• there are variances amongst users in terms of the level of trust they place in particular FBANs outputs and advice.

Trust is an important factor in the high-pressure contexts of incident and land management, including both interpersonal trust and trust in particular tools and systems. We found that trust seems to depend, in part, upon users' perceptions of a given FBAN’s skills, level of experience, and knowledge of the area or context in which they are offering predictive intelligence. For FBANs, both having fireground experience and “local knowledge” and being able to convince users that one has fireground experience and “local knowledge” seem to be important factors in eliciting the trust of some users.

4. Timeliness of FBAN outputs: We explored whether FBANs produced advice and outputs in a timely manner. We found that:

• FBAN advice and outputs were produced in a sufficiently timely manner for users’ purposes.

For example, a “quick and dirty” map of potential fire spread was seen to be acceptable during first attack. However, the pressure related to timeliness decreases during an extended attack and is substantially decreased during pre-event planning and planned burning, or in any instance where there is potentially more time for FBANs to converse with end-users, collect and validate intelligence and tailor the outputs to the end-users’ needs.

5. Benefits of releasing FBAN outputs publicly: We explored whether participants felt predictive outputs produced by FBANs (e.g., fire spread predictions) should be released to the community (i.e., affected communities and the general public) to facilitate their decision-making and planning during periods of significantly elevated fire danger (e.g., days of “Extreme” fire danger rating). We found that:

• with appropriate guidance and instructions, these predictive outputs should be used to facilitate community decision-making and planning in relation to bushfire hazards and incidents.
• there was general support amongst both users and FBANs to release FBAN’s predictive outputs publicly.
• the benefits of releasing FBAN outputs publicly outweighed the risks.

Participants noted that there is a moral imperative to release information to assist members of the public in their decision-making in relation to fire hazards. Further, there was a sense amongst many participants that, by releasing FBAN products publicly, emergency management organisations could support trust-building between agencies and communities and better demonstrate their transparency and accountability when working with the community to plan for and respond to bushfires.

6. Risks of releasing FBAN products publicly: We explored with the participants whether they felt that it was a risk to release FBAN outputs to the community to facilitate their decision-making and planning on days of significantly elevated fire danger. In general, participants reported that there were low levels of risk associated with releasing FBAN outputs to the community to facilitate community decision-making and planning.

We found that, where there was unease amongst participants in relation to releasing FBAN outputs publicly, it was because of:

• potential political and legal risks to emergency management organisations and staff. Our analysis suggests that such legal risks are likely poorly founded.
• potential for members of the public to misinterpret these outputs, leading to risks of maladaptive behaviours. These risks, we suggest, are irreducible but can be substantially mitigated through education or explanation via social media and other channels. 

7. The future of the FBAN role: We explored whether participants felt that the FBAN role will continue to be important within the broad emergency and incident control arrangements. We found that:

• the FBAN role had a very important role to play in the future of planning for and responding to bushfire hazards and incidents effectively in the future; and,
• there was agreement amongst participants that predictive services generally and the FBAN role specifically are a vital part of future emergency management arrangements in Victoria and Australia.

With this in mind, it is important to understand that the development of predictive services must ensure that it is further integrated into emergency management arrangements outside the State Control Centre. Specifically, viewed in relation to other project findings, it is clear that for such development and integration to be effective, FBANs will be required to spend more time interacting with users. Such interactions may sometimes occur outside fire seasons, though it is also apparent that they also need to occur throughout the fire season, on firegrounds and in Incident Management Centres and Regional Control Centres, in order to grow FBANs’ skillsets, maintain trust with users, educate users about their outputs and adapt those outputs to user needs. Various users had suggestions about how predictive services might better serve their needs, the vast majority of which related to the refinement of output types and modes of presentation to incidents and situations as they emerge.

Following this feedback also means exploring the different ways that FBAN skillsets and competencies can be used to assist key decision-makers such as Incident Controllers and community information and warning teams which continue to be so important in terms of planning and responding to bushfires. These skillsets and competencies - at both state and regional levels - also have relevant application to conduct of planned burning, and both users and FBANs suggest that more use can be made of FBAN skillsets and competencies to assist the states land and fire management agencies in achieving fuel management goals.

Based on these findings, we make 4 recommendations, detailed as follows:

Recommendation 1: that emergency management agencies explore which FBAN outputs could be released to the community to support their planning and decision-making during bushfire season.

Recommendation 2: that emergency management agencies explore how FBANs and predictive service outputs can be utilised to better support other relevant functions such as the conduct of planned burning and public information.

Recommendation 3: that emergency management agencies explore the ongoing training and development needs of FBANs.

Recommendation 4: that emergency management agencies work with FBANs to develop continuous learning processes which can be used to improve predictive services after bushfire and planned burning seasons.

In terms of utilisation, our recommendations will require an implementation strategy which needs to be mindful of the requirements to continue to plan for and respond to ongoing incidents that require predictive services. Accordingly, we propose a staged approach to utilisation that is channelled through the Bushfire and Natural Hazards CRC Black Summer research group.

Finally, we have confidence that these findings and recommendations are relevant to many other jurisdictions in Australia, including Western Australia, Tasmania, South Australia, Australian Capital Territory, Queensland, and New South Wales, given their similarity in bushfire hazards, incident management systems, and the role and scope of predictive services.