Why flood evacuation plans must account for real human behaviour

23 April 2025

Disaster planning often overlooks how people truly respond during emergencies. New research using insights from schoolchildren shows how social data can make flood evacuation strategies more effective and inclusive

In July 2021, devastating floods struck Germany and Belgium due to extreme rainfall, submerging many towns. In Germany, over 190 lives were lost, particularly in the small town of Schuld, where the River Ahr overflowed. Homes were destroyed, escape routes were blocked, and many residents were trapped with no means to communicate. Survivors described waking to the sound of water rushing through the streets, having received little to no warning about the impending disaster.

This tragedy raised serious questions about how a wealthy nation can fail so dramatically in flood response. Warnings were issued, but they did not lead to effective evacuations. Factors such as poor communication, lack of public preparedness and outdated evacuation strategies turned a potentially manageable situation into a deadly disaster.

Unfortunately, this story is not unique. From Pakistan to the United States, from Sri Lanka to the United Kingdom, people are dying in floods not only because of extreme weather, but also due to failures in disaster risk reduction. More recent floods in Spain and Scotland have also resulted in avoidable deaths, highlighting the fallacy of believing that developed countries are safe from extreme environmental phenomena.

How do we learn from mistakes?

The aftermath of the 2021 European floods revealed a crucial truth: current approaches are inadequate. Evacuation models exist, but they often fail to account for real human behaviour. People do not always evacuate when ordered. Some wait for neighbours to confirm information, while others overestimate their ability to survive at home. Vulnerable populations, such as children, the elderly and disabled people have specific evacuation needs. Yet many models take a one-size-fits-all approach based on limited data.

Our research seeks to address this gap. We focus on enhancing evacuation models by integrating real-world social data – especially insights from schoolchildren. By understanding how different demographics respond during disasters, we can develop models that lead to more effective preparedness and evacuation strategies.

Interdisciplinary research team L-R: Dr Hideyuki Shiroshita (Kansai University), Maciej Pawlick (UCL), Dr Kaori Kitagawa (UCL), Dr Ravindra Jayaratne (UEL), Dr Tomoyuki Takabatake (Kindai University) and Dr Hebba Haddad (UEL)

Our interdisciplinary approach, in collaboration with colleagues at UCL, Kansai and Kindai universities, saw a diverse team co-create a novel approach, bringing together expertise from the fields of coastal and civil engineering, disaster safety sciences, computer modelling, environmental psychology and education.

Japan – a nation ready for disaster

Japan understands the power of water. Having faced the 2011 Tohoku tsunami and frequent typhoons, the country has invested heavily in disaster risk reduction. A key feature of Japan’s strategy is educating children about disaster preparedness. Our study, made possible with the support of Inami Town Council in Wakayama prefecture, combines novel insights such as the inclusion of precise social data related to people’s behaviour. These data were collected by the junior high school children into a computer simulation model that can significantly improve emergency planning.

We visited Inami Junior High School and conducted interactive workshops on climate change, flooding, evacuation and computer simulation modelling, collaborating with Japanese schoolchildren between the ages of 11 and 12 to explore their understanding and reactions to flood risks. As part of their homework, students surveyed their local community about evacuation behaviour.

Student workshops in Japan

Feedback from the workshops was overwhelmingly positive, with children demonstrating improved understanding of their safety during emergencies. The experience was enriching: instead of fear, the students exhibited a practical mindset. Many participated in regular evacuation drills and recognised the importance of evacuating quickly following a flood warning. Discussions about family evacuation plans revealed diverse perspectives.

When we asked about their hypothetical decisions during a flood, their answers varied – some would wait for parents, others would focus on rescuing pets or valuables. These insights, often overlooked in traditional flood models, are essential for understanding real evacuation behaviour.

Data, such as actual evacuation times, were incorporated into our computational model. Each data point enhanced how the model simulated evacuation movements. Future models could include factors such as evacuation speeds based on age, gender and disability, as well as obstacles on evacuation routes to inform direction.

The research team explores evacuation routes

By including social data, our models can more accurately predict how people will act in emergencies. This helps local authorities determine where to prioritise resources before a flood occurs and identify high-risk areas needing additional support. Ultimately, incorporating social data into our models will better equip communities to respond to disasters and save lives.

The UK context

While the UK does not face tsunamis, flooding presents an increasing threat. In 2020, Storm Dennis caused record-breaking floods in England and Wales, costing the governmentan additional £500 million in preventing floods in Wales in the future. Much of the damage could have been avoided with better planning.

In the UK, we conducted similar workshops to those in Japan, but with younger children, who, even at age seven, have a basic understanding of flood risks and personal safety. This suggests that by adjusting resources and tailoring understanding, our modelling techniques can be well received and effectively applied within the UK contextwhere flood risks are rising quickly.

Unlike in the Japanese context, however, the UK school curriculum often lacks direct engagement with disaster preparedness, leaving gaps in understanding. Therefore, as risks rise due to climate impacts, there is a pressing need for curriculum changes to integrate disaster risk reduction information, especially pertaining to evacuation.

Why current approaches are insufficient

One major shortcoming in flood evacuation planning is the reliance on idealised human behaviour. Many models assume people will immediately follow evacuation orders. However, research shows that during real disasters, people often hesitate to evacuate for various reasons, including confusion, underestimating the risks, and social pressures.

The financial costs of inadequate planning are staggering. Every year of flood events can cost the UK economy at least £6.1 billion, yet investment in proactive flood defences and preparedness remains low, less than the £1.5 billion recommended by the National Infrastructure Commission. Research indicates that every £1 spent on flood prevention saves roughly £8 in recovery costs. Despite this, governments globally continue to prioritise post-disaster response over prevention.

The impact of climate change on flood preparedness

As climate change accelerates, flooding will become more severe. Rising sea levels and more intense storms mean that events once thought to occur ‘once-in-a-century’ are now becoming more frequent. Without urgent action, more lives will be lost, and recovery costs will escalate.

By improving flood evacuation models and incorporating real social data, we can create systems that reflect actual human behaviour during disasters. This leads to more targeted warnings, better evacuation planning and ultimately, fewer deaths. The time to act is now, before the next flood disaster strikes.

Interested in learning more about flooding? CIWEM's Flood & Coast 2025 Conference offers a chance to learn and connect with peers, industry leaders and key stakeholders from the world of flood and costal erosion risk management. Take a look at the programme here.

Want to stay up-to-date with the latest CIWEM news? Sign up to 'The Environment' newsletter, our free monthly news round up. Subscribe here.

Maciej Pawlik is a doctoral researcher in disaster science at University College London
Dr Ravindra Jayaratne is reader in coastal engineering at the University of East London