Between BAL-29 and BAL-40, How Much Bushfire Risk Must a House Bear?

Geoffrey Chen

When Numbers Start Making Decisions · Season Five, “Things That Have Not Happened Yet” · Article 3

1. Moving a house a few metres can change its institutional identity

Consider a residential lot in a bushfire-prone area. An initial plan places the house near woodland and receives a BAL-40 assessment. Move the building several metres towards the road and maintain more separation from vegetation, and the result may fall to BAL-29. The owner is unchanged and no future fire has acquired a date, but materials, approvals and cost can change with that category.

Bushfire Attack Level is not the probability that a house will burn down. It classifies the severity of attack a building may experience in a design bushfire scenario, including embers, burning debris, radiant heat and, at higher levels, possible flame contact. AS 3959 organises this exposure into BAL-LOW, 12.5, 19, 29, 40 and FZ and connects categories to construction requirements.

Western Australian guidance describes BAL-29 as high risk, with increasing ember attack and heat flux between 19 and 29 kW/m². BAL-40 is very high risk, rising to 40 kW/m² with an increased likelihood of flame exposure. BAL-FZ adds direct flame exposure. Western Australian Government: Building in bushfire-prone areas

The visible number resembles a heat measurement, but its institutional consequences extend far beyond heat.

2. BAL measures a design scenario

Assessment considers vegetation class, distance between vegetation and the proposed building, slope under vegetation and relevant fire-danger assumptions. Standard methods use these inputs to estimate the attack at the building location if the design scenario occurs.

BAL is therefore neither a simple historical statistic nor a forecast of the next fire. It is an engineering answer to a conditional question: if a specified bushfire environment develops, what might this position experience? Moving the house, changing maintained vegetation or correcting a survey can change the category.

Many survival factors are outside the number. Maintenance may lapse. Doors and windows may be open. Combustible material may collect in gutters. Fire appliances may not arrive. Residents may evacuate too late. Wind may exceed the design assumptions. Complying with a BAL construction standard improves expected performance; it cannot guarantee survival in every fire. Official guidance expressly states that the unpredictable behaviour of fire and extreme weather prevents such a guarantee.

That limitation is essential. BAL is suitable for assigning design obligations. It must not become a safety promise to occupants.

3. Why 29 and 40 become separate categories

Continuous radiant heat must be converted into buildable requirements. A code cannot prescribe a different set of doors, screens, walls and roofs for every decimal point of heat flux. It selects categories and connects each to tested assemblies and construction practices.

Twenty-nine and forty are not natural divisions of competence or moral worth. They are structural nodes in an engineering standard. They allow designers, builders, regulators and product testing bodies to share a language and prevent cost pressure from redefining risk project by project.

Once a category enters planning, however, the consequence may jump from stronger construction to a different approval pathway. In NSW, some development on BAL-40 or BAL-FZ land cannot use the complying-development route and requires fuller assessment. The boundary can therefore change not only materials but administrative scrutiny.

That has a defensible purpose. At very high exposure, standardised self-certification may not adequately address access, evacuation, water and landscape. But a category alone can mislead: BAL-40 does not always mean “building is forbidden”, and BAL-29 does not mean an ordinary house is safe.

4. Physical differences can be smaller than institutional differences

A calculated exposure of 28.9 kW/m² and one of 29.1 may be physically very close while belonging to different categories. Continuous risk has been forced into classes, so consequences near the boundary can differ more than conditions.

That does not justify abolishing the classes. If every applicant can renegotiate “close enough”, the standard loses force and risk will be interpreted downward. The better response is to retain the boundary and demand stronger evidence near it.

Reports should preserve measured locations, vegetation classification, slope, separation, the edition of the standard and site photographs. For borderline results, an independent review or sensitivity analysis can ask what happens if vegetation is a metre closer, slope is measured differently or a management practice is not maintained.

A single favourable result should not hide a plausible range. If a dwelling remains BAL-29 only when every assumption is optimistic, that fragility belongs in the decision.

5. The model can reshape the site

One benefit of BAL is that it points towards verifiable risk reduction. Moving a building, increasing separation, designing an asset protection zone and improving construction can reduce attack or increase resistance. The number does not merely label a fixed site; it helps designers change relationships on it.

But risk can also be managed merely to obtain the number. Vegetation may be cleared before assessment and later regrow. A management zone may have no durable responsible party. Future owners may never learn the maintenance conditions. The category then completes an approval without preserving the structure that supported it.

Any reduction that depends on management should therefore become a traceable obligation: who maintains it, how often it is inspected and when reassessment is required. Saving only “BAL-29” while discarding the conditions turns a relational assessment into a misleading permanent label.

Nor can all responsibility be placed on one building. Road capacity, refuge, communications, water, public vegetation and fire-service deployment belong to a shared system. More expensive glazing cannot compensate for a community that cannot evacuate. The house can bear responsibility only for what building design can actually control.

6. Error, review and remedy

Disputes can concern different kinds of error. A wrong distance or slope is data accuracy. An unsuitable vegetation class or method is measurement validity. Why BAL-40 changes the approval pathway is threshold justification. Whether development should proceed is planning judgement. Calling every disagreement “a wrong assessment” prevents a focused review.

Applicants should receive the full report and be able to commission a qualified review. Authorities should identify the accepted method and explain why alternatives are rejected. If site conditions change before construction, the assessment must be updated rather than treated as timeless.

Remedy also depends on timing. During design, a house can move and materials can change. Discovering an underestimate after construction makes correction far more expensive. Finding after occupation that a maintenance condition never existed may require both public and private intervention. The less reversible the decision becomes, the stronger the front-end verification should be.

7. Safety is not a property of the house alone

BAL reveals how safety forms across house, vegetation, slope, weather assumptions, materials and maintenance. Treating the category as an intrinsic property of a dwelling forgets that it is a relational result.

Forced structure has real value here. A complex spectrum of bushfire attack is compressed into a few levels that can be built and inspected. The risk lies not in structuring the problem but in forgetting the dependencies after the structure is created. A BAL detached from its report, version and maintenance duties retains authority without understanding.

Responsibility also remains distributed. Designers answer for inputs and method, authorities for planning, builders for construction, owners and agencies for maintenance and shared infrastructure. No party can say after a disaster that numerical compliance absorbed every obligation.

Conclusion: keep the categories and keep the conditions that made them

BAL-29 and BAL-40 should remain. They connect bushfire science to specific construction obligations and are more workable than a vague “high risk” label. Fuller planning review at very high exposure is also proportionate.

But BAL must not be treated as a survival probability or a permanent identity. Borderline sites should disclose inputs and sensitivity. Vegetation management used to lower a category needs durable responsibility. High-level decisions must consider roads, evacuation and water rather than substituting a stronger individual building for community planning.

There is no natural wall between 29 and 40. Institutions draw a line so action can occur before a future fire exists. They must also remember that the name of the category is not what protects the house. Protection comes from the whole set of relationships the category organises and keeps in place.

Primary sources

Discover more from Geoffrey Chen

Subscribe to get the latest posts sent to your email.