The research, conducted by Joshua P. Nicholas, Amy Donovan, and Clive Oppenheimer of the University of Cambridge and published in the International Journal of Disaster Risk Science, combines survey data from 159 adults with 27 in-depth qualitative interviews to answer a deceptively simple question: who prepares, and why? Squamish is an ideal laboratory for such an investigation. Nestled at the head of the Howe Sound fjord, the community of roughly 24,000 people sits within the unceded traditional territory of the Skwxwú7mesh Úxwumixw (Squamish Nation) and faces an unusually diverse hazard portfolio: earthquakes, volcanism, landslides, floods, wildfires, and storms. The region’s hazard history is sobering. The 1915 Jane Camp rock slide near Britannia Beach killed 56 people, the deadliest landslide disaster in British Columbia’s record, and a 1921 landslide-dam outburst flood swept away more than 50 houses and killed 37 others. More recently, a 2020 wildfire triggered evacuations, and the 2021 “heat dome” pushed temperatures in Squamish to nearly 43 degrees Celsius.

What makes the town particularly instructive is its relative affluence. The median after-tax household income was approximately CAD 96,000 in 2020, well above the national median of CAD 73,300. If wealth, education, and modern infrastructure guaranteed preparedness, Squamish should be a model community. Instead, the researchers found preparedness levels that could only be described as modest: respondents had completed, on average, just 4.21 out of 10 standard preparedness actions, ranging from signing up for the community alert system to assembling evacuation kits and practicing home evacuation plans. This mirrors a national pattern—a 2021 survey found that 74 percent of Canadians believe they live in a low or moderate risk area, and only about 11 percent have taken concrete household steps to mitigate weather-related disaster risk.

To explain the gap, the team built an integrated analytical framework drawing on three theoretical traditions that are rarely combined. The first is protection motivation theory, a cognitive model originally formulated by R.W. Rogers in 1975, which holds that protective behavior emerges from two appraisals: a threat appraisal, reflecting how severe and personally likely a hazard seems, and a coping appraisal, reflecting whether a person believes a protective response will actually work, whether they can personally execute it, and what it will cost them in time, money, and effort. The second is the capabilities approach associated with Amartya Sen and Martha Nussbaum, which reframes preparedness as contingent on the practical freedoms people genuinely possess—knowledge, time, money, health, mobility, and social support. Motivation without capability, in this view, produces no action. The third is social contract theory, stretching back to Hobbes and Rousseau, which treats political authority as legitimate only insofar as citizens believe the state will uphold its end of the bargain by providing protection. In a disaster context, perceived institutional credibility shapes whether individual preparation feels rational or futile.

The statistical results are revealing in their details. Using linear regression with Huber-White robust standard errors, the researchers analyzed preparedness as a composite score summing ten binary actions, including learning about local hazards, obtaining insurance, creating home and car kits, attending disaster training, and participating in drills. The single strongest negative predictor of preparedness was lower formal education—a finding with significant equity implications, since it suggests that even in a wealthy community, the people least equipped to absorb disaster losses are also least likely to prepare for them. Meanwhile, two psychological variables emerged as consistent positive predictors: direct experience of hazards, and hazard concern, the felt emotional weight of risk. Age, gender, and occupation showed weaker or less consistent relationships, echoing previous research suggesting that gender differences in preparedness often reflect household roles rather than systematic preparation gaps.

Perhaps the most counterintuitive result concerns perceived hazard likelihood. In the models, simply believing a hazard is likely to occur did not predict preparedness at all—and in some analyses restricted to permanent residents, perceived likelihood was actually negatively associated with participation in drills and training. This finding aligns with a growing body of empirical work indicating that coping appraisal components—self-efficacy, response efficacy, and perceived response costs—predict preparedness behavior far more reliably than threat appraisal. In plain terms: people do not prepare because they believe disaster is coming; they prepare because they believe they can do something about it. Concern motivates, but only where people feel capable of acting. This distinction between intellectual awareness of probability and felt concern has been flagged by risk perception scholars for decades, but the Squamish data demonstrate how sharply the two can diverge in their behavioral consequences.

The qualitative interviews add crucial texture to these statistical patterns, and here the social contract dimension comes into focus. Institutional trust, the researchers argue, does not operate simply: distrust does not uniformly suppress preparedness. Instead, it can produce two opposite outcomes depending on whether individuals believe their own actions remain meaningful without institutional backing. Some respondents responded to distrust with extreme self-reliance, doubling down on personal preparation; others slid into fatalistic disengagement, concluding that preparation was pointless if official systems could not be counted on when disaster struck. The perceived social contract thus governs the broader context in which both motivation and capability operate—shaping whether hazards feel personally urgent and whether individual effort seems worthwhile. Residence status mattered as well: full-time residents, embedded in local knowledge and social networks, tended to prepare more than transient seasonal visitors, consistent with long-standing findings about community attachment and resilience.

The methodology itself deserves attention for its rigor and its candor. Risk perception was measured across ten hazards—later refined to eight after interviews revealed that respondents conflated tsunami with coastal flooding and did not perceive avalanche as a town-level threat—using 7-point Likert scales rather than the conventional 5-point format to improve sensitivity given the modest sample size. Principal component analysis condensed correlated trust items, and thematic analysis of the interviews was conducted inductively in MAXQDA. The authors are forthright about limitations: the sample was recruited through Facebook community groups, which likely under-represents Indigenous residents, seniors, and lower-income households—precisely the groups whose preparedness is most consequential. Indigenous communities in British Columbia face disproportionate disaster and health risks shaped by the legacies of colonialism, a disparity thrown into stark relief by the 1862–1863 smallpox epidemic, which killed an estimated two-thirds of the province’s First Nations population. The authors explicitly caution that their statistical estimates should be interpreted with this representativeness problem in mind.

Why does this matter beyond one scenic fjord-side town? Because it reframes how disaster managers should think about their task. The conventional assumption embedded in international frameworks is that if people understand the risks they face and receive appropriate guidance, they will act. The Squamish results suggest this causal chain frequently breaks at two other links: the conversion of concern into capability, and the conversion of both into action within a trusted governance context. If lower education is the strongest barrier, then preparedness campaigns that simply broadcast risk information may be reaching least effectively the very households that most need help. If hazard concern and coping confidence are the real engines of action, then programs that build practical skills, reduce response costs, and demonstrate institutional reliability may achieve more than any amount of probabilistic messaging. The authors’ earlier work in the same region found that formal hazard expertise did not translate into greater preparedness—an absence of difference between experts and non-experts that directly motivated the present study and underscores that information alone is not the bottleneck.

The study ultimately delivers a synthetic conclusion that is greater than the sum of its statistical parts: preparedness in Squamish is shaped by an interaction of three things—an individual’s hazard concern, what they are materially able to do, and whether they believe institutions will uphold their side of the social contract. None of these alone is sufficient. A concerned resident without resources cannot act; a capable resident without concern will not act; and a concerned, capable resident who believes the state will abandon them may conclude that preparation is either futile or, alternatively, that everything depends on them alone. Understanding the preparedness gap, the researchers argue, requires all three dimensions simultaneously—and designing interventions that address all three may be the only way to close it, in Squamish and in the many hazard-prone communities worldwide that look far more like it than like the resilient ideal imagined in international policy documents.

The low-elevation coastal zone, or LECZ, consists of coastal land lying close to sea level and is among the most densely settled and economically productive environments on Earth. In China, it includes major cities, industrial corridors, river deltas, ports and agricultural areas, placing millions of people and vast amounts of infrastructure within reach of coastal and inland hazards. Its geography creates the possibility of compound disasters: a tropical cyclone can bring destructive winds, intense rainfall and a storm surge at the same time, while high river levels can prevent floodwater from draining into the sea. A hazard, however, does not automatically become a disaster. Risk depends on what lies in harm’s way, how susceptible people and infrastructure are to damage, and whether institutions and communities can reduce losses. The new framework therefore treats risk as a system rather than as a simple map of dangerous weather or geological events.

Wan, Jin and Zhao evaluated four interconnected dimensions: hazard, exposure, vulnerability and adaptive capacity. Hazard represents the potential intensity and distribution of damaging events, including earthquakes, floods, storm surges and tropical cyclones. Exposure measures the people, buildings, economic activity and other assets located in threatened areas. Vulnerability describes how severely those exposed elements may be affected, reflecting factors such as the quality of construction, demographic conditions and socioeconomic disadvantage. Adaptive capacity captures the resources and capabilities available to anticipate, withstand and recover from hazards. Combining these dimensions into a composite risk index allows the researchers to compare places and years even when the underlying threats are different. In technical terms, the approach transforms multiple indicators into standardized values and integrates them into a spatiotemporal assessment, making it possible to examine not only where risk is highest but also how its components have changed over two decades.

The national result is encouraging at first glance. The study reports that the average multi-hazard risk index fell from 0.51 in 2000 to 0.38 in 2020, a reduction of 25.3 percent. That decline does not mean that China’s coast became safe, nor does it imply that every hazard weakened. Instead, the index reflects the balance among changing threats, growing or shifting exposure, declining vulnerability and improved adaptive capacity. Coastal populations and economic assets may continue to increase even while risk falls if construction standards, emergency systems, flood control, public services and institutional preparedness improve more rapidly. Conversely, a low average can conceal dangerous local hotspots. A national score is therefore best understood as a broad signal of changing conditions, not a prediction of the damage from any single future storm, earthquake or flood.

The strongest regional warning comes from the central section of the coastal zone, which recorded an average risk index of 0.86—far above the northern region’s 0.30 and the southern region’s 0.21. The researchers also found that the central region experienced the most substantial improvement over the study period. Its vulnerability declined by 0.54, while adaptive capacity increased by 0.32, changes that appear to have driven much of the overall reduction in risk there. This combination may reflect advances in infrastructure, disaster management, public services and the ability of local systems to absorb shocks, although the assessment itself does not attribute each change to a single project or policy. The central coast’s high remaining score nonetheless underlines the danger of densely populated, economically important lowlands where rivers, cities and the sea interact. A reduction in relative risk can coexist with enormous absolute consequences when a hazard strikes a heavily developed region.

The study’s most distinctive finding is that the dominant driver of risk differs by latitude. In the north, earthquake risk is the leading influence. This means that coastal resilience there must extend beyond sea walls, drainage systems and storm forecasting to include seismic monitoring, land-use planning and structures engineered to withstand ground motion. In the central region, storm surges and floods are the primary drivers. Storm surge occurs when a cyclone’s winds and low atmospheric pressure push seawater toward the coast, temporarily raising water levels above the predicted astronomical tide. When that elevated sea level coincides with heavy rainfall or high river discharge, drainage can fail and flooding can spread inland. In the south, tropical cyclones dominate the risk profile, bringing a combination of extreme winds, rainfall, coastal inundation and potential cascading failures in transport, electricity and communications. These differences make a one-size-fits-all risk policy scientifically inefficient and potentially dangerous.

The framework also changes how adaptation success should be judged. Traditional assessments often focus on a single hazard—such as flood depth, earthquake probability or cyclone wind speed—and then rank locations according to physical exposure. But communities experience disasters through interacting pathways. A storm can damage power networks, disrupt hospitals, close transport routes and delay evacuation, multiplying the original impact. Likewise, an earthquake can trigger fires, industrial accidents or infrastructure failures that are not captured by a narrow seismic map. By incorporating vulnerability and adaptive capacity, the researchers distinguish between places facing similar hazards but possessing very different abilities to cope. A well-resourced city with strong buildings, reliable warnings and effective evacuation networks may face a lower realized risk than a poorer community exposed to a less intense event. This systems-based perspective is particularly important for coastal governance, where responsibilities are divided among environmental, emergency, infrastructure and urban-planning authorities.

The authors used spatiotemporal models to trace how risk evolved from 2000 through 2020 and regression analysis to examine the factors associated with those changes. Spatiotemporal analysis is designed to detect patterns across both geography and time, revealing whether risk is concentrated in persistent hotspots, moving between regions or changing as development and adaptation proceed. Regression methods then help estimate relationships between the composite risk index and possible driving factors. Such statistical results do not by themselves prove that a specific intervention caused a decline, but they can identify the variables most closely linked with changing risk and expose regional differences that national averages hide. The result is a governance-oriented assessment: rather than simply asking which places are most hazardous, it asks why their risk is high, which components are improving and what kind of intervention might address the dominant source of danger.

The findings arrive as coastal development, sea-level rise and extreme-weather concerns intensify worldwide. China’s coastal lowlands are particularly consequential because they combine environmental exposure with major concentrations of people, infrastructure and economic production. The assessment does not provide a forecast of future risk beyond 2020, and its composite index cannot capture every uncertainty surrounding climate change, future urban growth or changes in hazard behavior. Nor can a declining historical index guarantee that losses will continue to fall as coastal conditions evolve. Instead, the study provides a baseline for more precise planning. Its message is that resilience should be organized around regional hazard signatures: earthquake preparedness in the north, integrated flood and storm-surge management in the central coast, and robust tropical-cyclone planning in the south. The broader lesson is global: when several hazards overlap, reducing vulnerability and strengthening adaptive capacity may be as important as estimating the next extreme event.

For policymakers, the practical implication is a shift from generic coastal protection toward differentiated, continuously updated risk management. Central areas may need coordinated river-basin planning, urban drainage upgrades, surge barriers or natural coastal buffers, alongside evacuation systems designed for compound flooding. Southern regions require cyclone-resistant buildings, rapid warnings, shelter capacity and infrastructure capable of recovering from wind and water damage. Northern communities need risk-sensitive construction and emergency planning that treats seismic events as a central threat even in coastal settings. Across all regions, the assessment supports better integration of hazard data, population information, economic indicators and measures of institutional capacity. The decline in China’s average risk index demonstrates that adaptation can change the trajectory of danger, but the regional disparities show that progress is uneven. The coast is not one problem with one solution; it is a living risk system whose most effective defenses must match the hazards, vulnerabilities and capabilities of the people who live there.