New research, led by Postdoctoral Fellow Dr Lea Dasallas at Te Whare Wānanga o Waitaha | University of Canterbury (UC), shows that even shallow floodwater can be powerful enough to knock people off their feet or sweep vehicles away if it is moving fast enough.

However, most public flood maps still focus almost entirely on how deep water gets, not how quickly it flows.

“Floodwater doesn’t just pool – it flows, and when it flows quickly, even relatively shallow water can become extremely dangerous,” Dr Dasallas says.

As climate change drives more intense rainfall, the researchers say cities need to rethink how they plan for floods, shifting from static flood maps to dynamic models that show how water moves through transport networks in real time.

The research was undertaken as part of a Horizon Europe funded project* called the Minority Report that is dedicated to enhancing the resilience of vulnerable urban populations and their built environments against disruptive climate events. 

Using central Wellington as a case study, the team modelled an extreme rainfall event under future climate conditions. When water velocity was added to the models, previously ‘safe’ roads and intersections emerged as high-risk zones, especially in areas where streets effectively act as channels for fast-moving water.

“These are places people still try to drive through or walk across,” Dr Dasallas says, “but once you account for velocity, it becomes clear that those routes are much more dangerous than they appear.

“When flood velocity is included in the assessment, the areas classified as high risk for people walking increased by more than 80 per cent. Medium-risk pedestrian areas, including for children and older people, more than tripled,” Dr Dasallas says.

The study, published in the Journal of Flood Risk Management, didn’t stop at identifying hazardous streets. The team overlaid flood risk maps onto the transport network to test whether people could still reach essential services during the peak of a major flood. 

The team looked at access to hospitals, public transport hubs and key bottlenecks in the central city. Under depth-only flood modelling, most of the population appeared to retain access. However, when velocity was included, some regions in the CBD that still had access in the depth-only assessment are now shown to be cut off, especially for pedestrians.

In some scenarios, nearly all walking routes to key services were deemed unsafe during the flood’s peak. Vehicle access was also significantly reduced, particularly where steep terrain and narrow routes created choke points.

The findings highlight how quickly urban mobility can be significantly disrupted during extreme weather events even when floodwater does not appear to be particularly deep. 

Creating a framework for safer decisions

Rather than simply identifying risk, the researchers have developed a framework that can support real-world decision-making during floods.

By combining flood modelling with transport network analysis, the approach can identify which streets should be avoided and calculate safer alternative routes, effectively creating a flood-aware routing system for emergency planning.

Dr Dasallas says the goal is to improve urban resilience. “We want to help councils, emergency managers and the public make more informed decisions before and during flood events. 

“That could mean more targeted road closures, clearer public warnings, and better planning for access to hospitals and emergency services that would be based on how water actually behaves, not just how deep it gets.”

The UC researchers warn that as storms intensify, relying on outdated flood assessment methods could increase the risk of injury or loss of life, particularly in cities with steep catchments and dense transport networks.

“Understanding flood velocity is essential to keeping people safe, challenging the common perception that shallow floodwater is safe to cross,” Dr Dasallas says.

* This project has received funding from the European Union’s Horizon Europe Framework Programme for Innovation under grant agreement no 101147385.