Article A5 ended with a striking finding from the Sendai Framework monitoring data: the targets that are "on track" are the institutional ones — the number of countries with DRR strategies, the coverage of multi-hazard early warning systems. The targets that are "at risk" or showing "mixed" progress are the outcome ones — actual reductions in mortality, in the number of people affected, in economic losses. The implication is precise and uncomfortable: the international community has proven reasonably good at building the institutional architecture of DRR. It has proven far less good at translating that architecture into measurably safer communities.
This gap matters enormously for geographic policy. It suggests that simply building more warning systems, writing more national DRR strategies, and holding more international frameworks will not, by itself, reduce disaster deaths at the rate needed to meet global targets — let alone to keep pace with the rising hazard intensity driven by climate change. Something else is required: an understanding of what makes the difference between DRR that genuinely reduces vulnerability and DRR that produces impressive-looking institutional documents while leaving the most exposed communities largely unchanged.
Ilan Kelman's argument — that disasters are not natural events but accumulated human choices — is the thesis statement of Package A restated as a policy challenge. If every element of vulnerability is the product of identifiable decisions (where to settle, how to build, whether to invest in warning systems, how to plan land use), then every element of vulnerability is, in principle, reducible by different decisions. The question for A6 is: what do we know about which decisions, made by whom, at what scale, produce genuine reductions in disaster mortality — and what can we learn from the cases where well-resourced, well-intentioned DRR programs failed to produce that result?
Notice the geographic dimensions of this question. DRR operates at multiple scales simultaneously: at the global level (international frameworks, development finance, technology transfer), at the national level (building codes, emergency management systems, land use planning), at the community level (local warning networks, social capital, evacuation drills), and at the individual level (household preparedness, insurance, building quality). Effective DRR must operate coherently across all of these scales, because vulnerability is produced at all of them.
The DRR spectrum: from hard engineering to community resilience
Disaster risk reduction is not a single intervention — it is a spectrum of approaches ranging from large-scale physical infrastructure to community-based social programs, with a range of non-structural policy measures in between. Understanding this spectrum is essential for examination responses that ask students to evaluate DRR strategies, because different approaches have different cost profiles, different time horizons, different beneficiary populations, and different evidence bases. A strong geographic response does not simply list DRR strategies — it evaluates them against the specific vulnerability context.
The "last-mile" problem: the gap between warning and safety
One of the most geographically important concepts in DRR practice is last-mile connectivity. The term comes from telecommunications (the "last mile" of cable from a network trunk line to an individual house) and captures the specific challenge of bridging the gap between a warning generated in a national meteorological centre and the action taken by an elderly woman in a remote coastal community who does not own a smartphone and whose nearest radio is unreliable.
Resilience: what the concept actually means
The word resilience has become the dominant framing in contemporary DRR policy — but it is used with enough imprecision to require definition. In geography and DRR, resilience has two distinct interpretations that carry different policy implications.
The first is engineering resilience: the ability of a system to return to its pre-hazard state after a disturbance. A resilient city in this sense is one whose infrastructure, economy, and social structure recover quickly to their previous condition after a flood or earthquake. This is the sense used when governments promise to "build back better" after disasters — implying a return to a baseline, ideally improved.
The second is ecological or adaptive resilience: the capacity of a system to absorb disturbance and reorganise while undergoing change, so as to retain essentially the same function, structure, and identity. This is a more demanding concept — it accepts that the pre-disaster state may not be recoverable (especially under climate change) and asks instead what new stable state the community can achieve that maintains core functions and values. In the context of communities facing sea-level rise or intensifying fire risk, adaptive resilience may require relocating entirely — acknowledging that returning to the previous state is not possible.
The distinction matters because "building back better" as an engineering resilience goal may sometimes be the wrong geographic objective. In some cases, the "better" the community is rebuilt in the same location, the more valuable the assets that will eventually be lost to the same hazard, more intense, in a future event. Adaptive resilience sometimes demands the harder conversation about whether the pre-disaster location is the right one at all.
The DRR effectiveness evidence: what the research shows
Across the academic literature on disaster risk reduction, a consistent set of findings has emerged about which approaches produce the strongest evidence of reduced mortality and reduced affected persons. These findings are not always what early DRR policy frameworks assumed — and the gap between the evidence and the dominant policy priorities is itself a geographically important finding.
What DRR success actually looks like: Bangladesh revisited
The Bangladesh cyclone shelter program (introduced in A3 as a case study of hurricane mortality reduction) is the most rigorously documented example of integrated, multi-approach DRR producing consistent mortality reduction at national scale over decades. It is worth examining not just as a success story but as a model of what the integration of all four DRR approaches looks like in practice.
Non-structural (policy): National Disaster Management Act creates legal framework for mandatory evacuation orders. Building standards for coastal construction. Land use zoning that restricts development in highest-surge zones. Integration of DRR into local government planning processes.
Early warning (detection & communication): Bangladesh Meteorological Department provides 72-hour cyclone track forecasts. Warning is transmitted through seven tiers: national → district → upazila → union → ward → neighbourhood → household. At the community level, 55,000+ trained volunteers deliver oral warnings door-to-door and lead evacuation to shelters.
Community-based (social resilience): The volunteer network is the critical last-mile mechanism. Volunteers know which households have elderly, disabled, or young children members — and prioritise those households. Evacuation drills are held annually. Communities have participated in designing shelter locations.
• 1970 Bhola Cyclone (before any system): ~300,000 deaths
• 1991 Bangladesh Cyclone (early system): ~138,000 deaths
• 2007 Cyclone Sidr (mature system): ~3,406 deaths
• 2009 Cyclone Aila (mature system): ~190 deaths
• 2020 Cyclone Amphan (full system): ~128 deaths in Bangladesh (vs. ~80+ in comparable Indian coastline with less developed system)
The physical geography of the Bangladesh coast has not changed. The severity of cyclone events is comparable or greater than in previous decades. The reduction in mortality — by a factor of more than 1,000 since 1970 — reflects the investment in all four DRR approaches operating coherently over five decades.
Critically: Bangladesh remains a lower-middle-income country throughout this period. The investment was not a consequence of becoming wealthy — it was a consequence of political commitment to systematic DRR as a national priority.
The failure modes: when DRR doesn't work
Understanding why DRR fails is at least as important as understanding why it succeeds. The research literature identifies a consistent set of failure modes — patterns by which well-designed, well-funded DRR programs produce institutions without protection. These are geographically important because they reveal where the pressure points are in converting DRR investment into DRR outcomes.
Ilan Kelman and "Disaster by Choice"
"Build Back Better" — the principle and its failure in practice
One of the most widely cited principles in post-disaster recovery policy is Build Back Better (BBB). The concept holds that the window immediately following a major disaster — when communities must rebuild anyway, when political attention and international aid are concentrated, and when the vulnerability that caused the disaster is painfully visible — represents an opportunity to reconstruct in ways that reduce future risk.
Managed retreat: the hardest conversation in DRR
It is, without question, the most politically and socially difficult DRR strategy. People have deep attachments to place — to the landscapes, communities, and identities formed where they live. Relocation severs these connections in ways that produce genuine and documented psychological, social, and economic harm. For Indigenous communities, the attachment may be spiritual and cultural in dimensions that make "compensation" an inadequate concept.
Yet in some geographic contexts, the alternative is to repeatedly reconstruct communities in harm's way, spending ever-greater sums on resilience measures that will eventually be overwhelmed by a sufficiently extreme event. New South Wales (2022–present) is actively grappling with this question: following catastrophic flooding of towns like Lismore, the state government is funding voluntary buyback schemes for the most flood-exposed properties — the first significant Australian managed retreat program. The scale is small relative to the exposure; the political controversy is large relative to the scale. Geographers are central to providing the spatial analysis that identifies which properties are at sufficient risk to justify buyback, and to the planning of where relocated communities should go and what support they need to rebuild their social fabric in a new place.
You now have the DRR spectrum (four approaches and their trade-offs), the last-mile concept (the gap between warning generation and community action), the Bangladesh success model (all four approaches integrated over decades), the effectiveness evidence (warning systems and land use planning have the strongest benefit-cost ratios when implemented well), the six failure modes (from false alarm fatigue to the political economy of invisible prevention), and the Build Back Better evidence (showing how reconstruction frequently recreates vulnerability rather than reducing it). The geographic argument you construct must explain the institutional-outcome gap from A5, use the Bangladesh and failure mode evidence to identify what distinguishes effective from ineffective DRR, and evaluate what is required for DRR to actually reduce disaster mortality in the face of intensifying climate change.
Australia's DRR architecture: sophisticated but stretched
Australia has one of the most technically sophisticated disaster risk management systems in the Asia-Pacific region. The National Emergency Management Agency (NEMA, established 2022, replacing the former Emergency Management Australia) coordinates national DRR policy. Each state and territory has its own emergency management organisation — the Queensland Fire and Emergency Services, NSW Rural Fire Service, Country Fire Authority in Victoria, and equivalents elsewhere — with large professional and volunteer workforces. Australia's Bureau of Meteorology (BoM) provides some of the world's most advanced cyclone, flood, and fire weather forecasting. Building codes in cyclone-prone regions of northern Australia incorporate specific wind-loading requirements developed after Cyclone Tracy (Article A4). Risk mapping of flood plains, coastal surge zones, and fire-prone areas is conducted systematically at federal and state levels.
And yet: the 2019–20 Black Summer burned 18.6 million hectares, killed 33 people directly and an estimated 417 from smoke, destroyed more than 3,000 homes, and produced ecological impacts that are still being assessed. The 2022 South-East Queensland and Northern NSW floods caused three successive flooding events across communities still recovering from previous disasters. The Northern NSW town of Lismore — which has been flooded repeatedly through its history — experienced its worst-ever flood event in February 2022, killing seven people and rendering thousands of homes uninhabitable.
The gap between Australia's DRR architecture and these outcomes is not evidence that the system is failing. It is evidence that the physical hazard environment is intensifying — and that a DRR system calibrated against historical hazard parameters will experience progressively more frequent design-basis exceedance as climate change shifts those parameters.
The within-Australia vulnerability geography
Australia's DRR system performs very differently across its geographic distribution of communities. For major urban centres on the eastern seaboard, the system is sophisticated and well-resourced: early warning is reliable, emergency services are proximate, hospitals are accessible, evacuation infrastructure exists, and the economic capacity of households provides buffers. For remote and regional communities — and particularly for remote Indigenous communities — each of these conditions may be absent or significantly degraded.
Remote Indigenous communities often face compound vulnerabilities: extreme heat, flooding, and cyclone exposure is combined with limited road access (making evacuation difficult or impossible in wet season conditions), inadequate housing that provides poor protection from heat and wind, limited community infrastructure for shelter, and health conditions that elevate sensitivity to both heat stress and respiratory impacts from smoke. The Australian DRR system's last-mile challenge is at its most acute in these communities — and the A5 finding about within-country vulnerability distribution applies here as clearly as it did to New Orleans under Katrina.
The Lismore question: managed retreat in practice
The town of Lismore, New South Wales, has flooded 88 times since European settlement. It sits in the bowl of the Wilsons River floodplain, bounded by ranges that concentrate water into the town's drainage catchment with extreme efficiency. After the catastrophic February 2022 flood — in which the levee protecting the town was overtopped by more than a metre — the NSW Government initiated a voluntary property buyback program, offering to purchase the most flood-exposed properties at pre-flood market value.
The Lismore program is the most significant managed retreat initiative in Australian history. By 2024, several hundred properties had been purchased, their owners relocating to higher ground. The program is accompanied by planning controls restricting reconstruction on the acquired land. But the numbers who have accepted buyback represent a small fraction of the flood-exposed population. Many residents — for reasons of attachment to place, dissatisfaction with the offered price, or the complexity of their housing situation — have rebuilt in the same locations. The geographic tension between the DRR logic of relocation and the human geography of place attachment is playing out in Lismore in real time.
The Lismore case is important for A7: the question of whether certain Australian communities should be defended in place or supported to relocate is one of the most significant geographic policy questions of the coming decades — and the 2019–20 Black Summer raised a version of it for fire-exposed communities that is still unresolved.
Connecting to Article A7: the Black Summer as DRR test
Article A7 — the case study that closes Package A — applies the complete framework built across A1–A6 to the 2019–20 Black Summer. Every concept you have built is implicated: the physical fire behaviour from A4, the ENSO and climate change context from A3, the vulnerability dimensions from A5 (particularly for rural communities in fire-prone areas), and the DRR question from A6 (what was Australia's DRR system capable of, and where did it fall short?). The Black Summer was not a failure of Australia's DRR system in the same way that the 2004 Indian Ocean tsunami was a failure of early warning infrastructure. It was something different: a design-basis exceedance event that revealed the limits of preparedness calibrated against historical fire parameters that climate change had already made obsolete.