Table 4 Summary of the decision-making analysis on a complex, practical example: the 2015 Odena Fire, Spain. Real situation assessment (A), and two different fire suppression strategies considered by the incident managers of the fire: the defensive strategy (B), and the creative decision-making strategy that was finally implemented (C)

Values at risk

(1) Complex scenario with pressure to avoid fire burning the WUIs and to avoid flames reaching the national park (Fig. 8).

(2) While wind is blowing, this type of wind-driven fire is really difficult to stop at its head. Windows of opportunity need to be assessed in order to get certainty to manage the wind scenario. The fire was divided into two situations: (2a) west winds would push a fast and intense fire toward the WUI area around Montserrat National Park. (2b) late afternoon, the fire would shift following the north-northwest wind, it will drive away from the WUI and National Park, leaving a new and much wider fire head running into the RUIs and the more sensitive forest ecosystem that could change to shrubland once burned.

Emergency management

(3) Scenario 1: protect WUI and National Park but risking RUI and a fragile ecosystem (left flank polygons, Fig. 8).

(4) Scenario 2: protect RUI and fragile ecosystem and minimize the damage in WUI (protection and prevention measures applied before). The national park would burn: the area has a resilient ecosystem that would easily recover; nevertheless, it would create a visual impact jeopardizing tourism.

Fire ecology management

(5) The area affected by 1986 fires is a young pine forest (Pinus halepensis). Before, part of this land was covered by Pinus nigra Arnold and Quercus faginea Lam., but after those fires, more xerotinic species dominated the landscape.

(6) Considering resilience in this fragile ecosystem, compared to the coastal ecosytem (more resilient), is a priority.

(1) If efforts are concentrated on protecting the national park and the WUI, then fire would hit both as a head fire, and chances of losing the WUI, RUI, and national park are high. There are too many values at risk to defend, and fire can easily grow. Realistically, there are two main connections between polygons where the fire cannot be stopped once it gets there before the wind shift (red connections, Fig. 8a) Focusing on them (Known Unknown) means working on uncertain resourcedemanding scenarios than would likely cause collapse.

(2) Initially, resources will be allocated to protect the WUI at polygons 4.1and 5.2 but without stopping the fire, then WUI at 7, 8, 9 and 18. The protection of the national park would be required if fire hits polygons 7 and 8 (Fig. 8).

(3) While protecting the WUIs and the national park, the fire will shift north and northwest, bringing RUI and sensitive ecosystems into the equation.

(4) The defensive approach of chasing flames based on traditional value-at-risk approach would then drive the fire service to collapse. Once there, basic values of life and property would not be defended at all, anywhere.

(5) Under this scenario, the fire is a major incident that would only be solved due to weather improvement, after losing thousands of hectares. It’s a classic situation that causes frustration for firefighters and distrust among the population of the country.

(1) Look for the safe and certain scenario using the uncertainty matrix (Fig. 3). For instance, the fire service knows that it can succeed in breaking connectivity between 1 and 3.1 and 6 (north flank) at the beginning. If the fire is not controlled then, it will unlikely be stopped after the wind shift. The fire service also knows that they can successfully protect the WUI and the east side of the national park once the wind shifts. Those are the certain tactics.

(2) The creative tactics proposed are to use windows of opportunity and accept the cost of them.

(3) Breaking the connection between polygon 1 and 3.1 (Fig. 8a) leads to major spread of head fire going to 4. Flank connectivity to 5.2 can also be broken. This will leave a narrow and slower wind-driven fire.

(4) Focus on the left flank before the wind shift, as a way to avoid the possible head fire to polygon 6.1 and 6.2.

(5) If the fire is kept narrow, it will spread with much less energy and can develop less convection. Power of fire (Rothermel 1991) stays lower until the wind shift, and allows the fire service to be ready to stop the connection from 4s to 8s, national park and WUI.

(6) Once the wind shifts, connections to polygon 6 will be active (Fig. 8b). The narrower, the better. What happened is that, the connections to polygon 6 were only two narrow strips under control before connection to 6 was established.