Join us
Orange welcome sign that reads Royal Geographical Society with IBG.

Become a member and discover where geography can take you.

Join us


Disaster events in recent years have emphasised the complexity of combined hazards. In 2021, Haiti was struck by an earthquake and subsequent tropical storm during the ongoing Covid-19 pandemic. Similarly, public debate in the United States around flooding and tornado hazards continues about events occurring at the same time – even where citizens receive conflicting alerts advising them to enter basements to avoid tornadoes at the same time as alerts to evacuate to upper floors of properties because of impending flooding.

The importance of a multi-hazards approach was seen in the 2011 Tōhoku quake and Fukushima nuclear reactor meltdown – underlining the need for compound hazard warning systems, and this was addressed in the UN’s 2015 Sendai Disaster Risk Reduction Framework.

An expert panel, convened by the Society’s Disaster Risk Management Professional Practice Group and chaired by Faith Taylor (King’s College London) explored compound risks, dynamic exposure and vulnerability in the Fireside chat series.


Watch the webinar


What makes risks and vulnerabilities compound/dynamic?

Professor Bruce Malamud (Professor of Natural and Environmental Hazards, King’s College London) highlighted how some risks are dynamic. Our exposure to these hazards can change over time, or in response to other disasters.

  • These dynamic processes can be multi-stage - for example, an earthquake can push survivors to an area at increased risk of a flood, changing both the vulnerabilities and exposure to hazards.

  • Our vulnerability to risks can also change – for example, a disaster can damage buildings or overwhelm emergency responses, making people more vulnerable to other risks.

  • Understanding and planning for these dynamic relationships can help improve risk mitigation. This may require, for example, better land use planning or co-operation between different organisations.

Iain Willis (JBA Risk Management) pointed to two ways disasters may compound. Sequential events – like the 2011 Tohoku earthquake, which caused a tsunami, which led to a nuclear reactor meltdown – have been modelled before, notably in the cases of fire after earthquake and landslides after rainfall.

However, risks can also compound by coincidence, with Covid-19 being a key example. Iain’s work assessed how flooding impacts could be made worse by covid-19, with scenarios such as covid-19 spreading in evacuation centres during a flood.

To mitigate compound events, it is helpful to use counterfactual approaches and imagine scenarios that have not yet happened. Even with Covid-19, there could be more severe and wide-reaching scenarios in future, just as similar events in the past like SARS did not reach the level of severity and geographical spread of the Covid-19 pandemic.


Dynamic risks in context

Liz Riley (Executive Director, CDEMA) noted that, in the Caribbean context, small island developing states have “unique vulnerability to external shocks…superimposed on existing vulnerabilities”.

Hazard interactions in the Caribbean are complex and compound. Hydrometeorological hazards and vector-borne disease hazards are exacerbated by climate change. Severe droughts have been followed by intense weather, which can worsen landslide hazards.

Such hazards occur across a broad geographical space with multiple countries affected; across different timescales; and even within individual hazards. For example, in Hurricane Dorian in 2019, wind hazards were combined with storm surge and rainfall hazards across the timescale of the storm. As an additional challenge, risk perception varies widely and influences how people are exposed to and vulnerable to hazards.


Modelling compound risks and addressing data gaps

Antoine Bavandi (Disaster Risk Finance team, World Bank) highlighted demand for comprehensive and holistic risk assessments of compound risks – including those that can compound year-on-year such as drought.

New data sources like social media and collaboration between stakeholders can help model such risks across a sector, such as agriculture. For example, modelling the income of smallholder farmers can help understand how multiple risks can affect their income.

Audience members raised data gaps in compound events. Iain noted that modelling typically uses historic data, but compounding risks can change the probabilities around events - such as climate change making a once-in-200-years flood more frequent.

Liz agreed that past data was no longer a guide to future risks, and mentioned the additional challenge of downscaling climate models to better represent regional risks. A key challenge facing practitioners is to apply information to changing hazard dynamics and then risk assessments. Iain reiterated the need for effective downscaling, noting that global climate models were not fine-grained enough to understand extreme events at more local scales.

Bruce added that there was a need to understand changing scenarios around climate change and migration in the next 10-15 years rather than 50 years. Antoine noted that people’s responses to risk also changed vulnerabilities – for example, farmers adopting different seeds in response to drought – and that data on changing exposure was also needed.


Adapting scenario planning for unlikely risks

Liz said that in the Caribbean, dynamic exposure helped inform disaster planning scenarios – which were changing “rapidly and radically” based on the ever-changing hazards and availability of assets to respond. She also emphasised the importance of planning at the regional Caribbean system level to share resources, not just within individual countries.

Iain explained the use of ‘realistic disaster scenarios’ in the insurance industry, which focus on the probable maximum loss – the worst-case scenario. However, 1-in-10000 year events are difficult to communicate well as they are extremely unlikely (a 1-in-10000 year event has a probability of occurring of 0.01% in a given year). Instead, scenarios looking at 1 in 100-250 year return periods, are more realistic and probable (a 1% and 0.4% chance of occurring in a given year). They also use historical examples and counterfactual approaches to examine what the impacts of near-misses or past disasters would be today, which can help frame impacts realistically.

Liz, however, noted that people tend towards a better understanding of scenarios they have seen before, and that hazards can exceed historical extremes. For example, in 2021 St. Vincent and the Grenadines had been impacted by a volcanic eruption, a Covid-19 wave, a dengue fever outbreak, and severe weather after the eruption which caused lahars – demonstrating the need to account for unlikely scenarios. Unexpected hurricane behaviour and rapid intensification have also been observed, challenging expectations from records.


Better public communications

On public communication, Liz reflected on a case in the Bahamas where some people in coastal communities did not evacuate despite accurate forecasts of 20ft storm surges. Mapping can offer one way of helping publics visualise impacts. Antoine agreed that telling a clear story about a sequence of events was key for communicating risks with the public and for working with stakeholders.

Finally, Bruce emphasised the need to use local experiences to support risk planning. He also pointed to access to information, especially in the global South, as a key challenge – but that open source data and sharing research can help address this.



Share this resource

This is an open access article under the terms of the Creative Commons Attribution License (CC BY NC 4.0), which permits use, adaptation, distribution and reproduction in any medium or format, provided the original work is cited and it is for non-commercial purposes. Please contact us for other uses.

How to cite

Royal Geographical Society (with IBG) (2022) Compound Risk: Dynamic Exposure and Vulnerability. Available at Last accessed on: <date>


Featured image: @Marcus_Kaufmann/Unsplash