How urban conflagration works and why it matters for Australia (Part I)

Author: Iuliia Shustikova, Product Manager - Insurance, Moody's

In the context of bushfires, an urban conflagration (UC) is a catastrophic fire event that can happen when spreading from bushland to and across densely built environments under extreme weather conditions, often overwhelming emergency services and resulting in widespread destruction of structures, infrastructure, and livelihoods. 

Unlike traditional bushfires or wildland-urban interface (WUI) fires, UCs propagate within the built environment—often via embers and radiant heat, rather than through vegetation alone. 

Some well-known recent UC events include:

In response to recent events, scientists, research institutions, and catastrophe modelers have investigated the environmental drivers and conditions that contribute to wildfire-induced urban conflagrations.

Through laboratory and field testing, as well as drawing lessons from similar fire behavior observed following earthquakes, they have developed the first generation of models to better account for this risk in wildfire assessments.

These fires are not limited to the WUI; under specific conditions, they can penetrate deep into densely built urban areas.

Insurers and reinsurers must recognize that UC is a physical phenomenon that can be analyzed and quantified. It should not be regarded as a scaled-up bushfire—a UC is distinct, driven by system interactions and infrastructure fragility. In catastrophe modeling terms, it represents a tail-risk event, where multiple variables produce order-of-magnitude loss escalation.

 

Urban conflagration as a cascading hazard

From a catastrophe modeling perspective, bushfires are potential cascading events because they can trigger subsequent catastrophes, including UCs, a conditional cascading hazard, when multiple environmental, structural, and situational factors must align.

This type of hazard results in a cascade of impacts, not only due to the number of burned structures but also more broadly, on destroyed infrastructure and weakened regional economies.

These events are low-frequency and high-severity. Below are the principal components that can enable UC:

 

1. Extreme fire weather conditions

At the core of any UC event lies the presence of sustained severe fire weather and severe antecedent conditions. These conditions are characterized by high temperatures, low relative humidity, and critically, strong winds, which are also causes for a sudden drop in fuel moisture (e.g., katabatic winds - downslope winds flowing from higher elevations towards lower elevations).

Under these conditions, fuels such as vegetation, building materials, and urban debris become highly combustible, reducing the energy needed to ignite them and allowing fires to spread with explosive speed.

Wind speeds exceeding 35-45 kilometers per hour (21-28 miles per hour) with higher gusts can dramatically alter fire behavior. Wind accelerates flame fronts, lofts embers, and interacts with urban layouts to increase hazard, especially in areas with narrow streets or continuous rows of buildings.

Although these conditions are necessary, they are not sufficient as causes of urban conflagration. There is a random component that depends on the speed at which the fire in the vegetation moves, and on the likelihood that the embers produced by this fire will propagate in the right locations, possibly creating onsets of urban conflagration because of a lack of fire suppression, and leaving emergency response incapable of suppressing the fire in urban areas.

 

2. Windborne ember transport (Spotting)

Windborne ember transport, also known as 'spotting,' is a mechanism for long-distance ignition.

In UCs, embers often ignite structures well ahead of the main bushfire front, bypassing firebreaks or defensible space. Ember ignition does not require direct flame contact, and it is often a first step toward a large-scale structural loss. 

While it's common to think of forests filled with ember-producing trees as the primary threat, history shows that fires spreading through grasslands can also trigger devastating urban conflagrations (e.g., the 2021 Marshall Fire in Colorado, and the 2023 Lahaina Fire in Maui, Hawai’i).

In urban settings, embers may accumulate and ignite rooftops, eaves, decking, or landscaping, or burn the structure from the inside out if there are openings.

Reports have documented embers travelling up several kilometers under strong winds. For instance, during the 2017 Tubbs Fire, embers crossed a six-lane highway, igniting the Coffey Park neighborhood of Santa Rosa, CA, a residential subdivision with no direct forest proximity.

 

3. Urban fuel continuity and structure vulnerability

Another critical enabler of UC is fuel continuity within urban areas. This includes close structure spacing (e.g., less than 10 meters), combustible cladding or roofing materials (e.g., timber, untreated fences, vinyl eaves), flammable landscaping and mulching, and accumulated debris in gutters, under and on decks, or against walls, etc.

Once a structure ignites, nearby buildings may be exposed to radiant heat sufficient to cause secondary ignition and/or a sudden shower of embers of burning debris. Under windy conditions, this can occur in minutes, creating a cascade effect.

The 1991 Oakland, CA, firestorm, for example, resulted in the loss of over 3,800 dwellings within less than a day, primarily due to extremely rapid spread (insured loss of the event estimated to be over US$5.2 billion, trended 2023)

Moreover, even a single point of failure—such as an unsealed eave or timber feature (e.g., a deck) may cause ignition. Structure-to-structure ignition has been observed in areas with minimal vegetation, as illustrated in the pictures from the four events at the start of this blog, highlighting that the built environment can serve as a continuous fuel source.

 

4. Structural density and urban topology

Observations have shown that UCs are more likely to occur under extreme weather conditions in environments with medium to high-density development. Some of the factors that increase the potential for urban fire spread are:

• Row housing or duplexes with shared walls, terraced houses
• Narrow setbacks or side-access only lots
• Topographic funneling (valley suburbs, slopes)
• Cul-de-sacs or one-way ingress/egress that hinder firefighting access and slow evacuations

These factors reduce the effectiveness of fire suppression and promote faster inter-structure ignition. In such a case, density is not just exposure—it is a fire propagation vector.

 

5. Suppression overload and infrastructure failure

Even highly capable fire services can become quickly overwhelmed under UC conditions as their mission switches from saving properties to saving lives. Fires can spread so rapidly that the initial response fails, with local brigades being pushed toward evacuation and perimeter control rather than suppression.

In situations where three or more houses burn in the same street, hydrants become overwhelmed and are unable to function effectively to extinguish fires.

Under high wind speeds, small fires spread within minutes across entire parcels. Alongside this, the communication and utilities powerlines break down, which complicates emergency response.

Afternoon and nighttime ignitions complicate response further—visibility, coordination, and aerial operations become more difficult. While emergency services might adapt to these constraints (e.g., using lighting and thermal imaging tools), capacity remains finite.

 

UC cascade is a systemic chain

All these factors, when combined, create an environment favorable to rapid fire spread. A system that behaves one way under normal stress behaves very differently when those stressors stack.

What makes UC so devastating is the typical sequence of events: 

Figure 1: Sequence of events for an urban conflagration event from ignition to longer-term economic effects

This transition is nonlinear. The probability of UC is low—but once it exceeds a certain threshold (e.g., sustained winds, ember density, sufficient urban density, and extreme fire weather), the containment probability drops sharply, and the loss potential skyrockets.

 

Implications for Australia

Fortunately, no UCs were reported in Australia, although Australian communities share many of the characteristics observed in recent UC events in different states across the U.S.:

• Suburbs with timber fencing and flammable landscaping
• Close-set dwellings and aging building stock
• Increasing peri-urban development into WUI and urban pattern density

Bushfire risk is already well-recognized in bushfire-prone areas and wildland-urban interfaces; however, UC shifts attention to what happens after fire crosses into the built environment.

The hazard is then less about proximity to bushland and more about how structures and urban layout interact under stress. 

This reveals a critical blind spot in risk assessment if UC scenarios are not properly modeled. As urban growth continues and fire seasons become more intense and prolonged, the relevance of UC scenarios increases.

Understanding this cascading hazard can inform land-use planning, building standards, and risk modeling.  Having the means to assess the additional risk associated with UC is one component of a model to assess bushfire risk when making risk assessments at different future time frames.

Read Iuliia's second blog: How urban conflagration works and why it matters for Australia (Part II) here.