Seventy-six young adults stepped into a water flume, faced six combinations of depth and speed and then had to answer two simple questions: how fast is this water moving and how dangerous does it feel? The study found that most people overestimated water speed, especially as the current got faster or deeper. It also found an important difference between participants. People who had lived through more storms judged the moving water more accurately than people with little or no storm experience.
The experiment, published in PLOS ONE, focused on a problem that affects both flood safety and storm-surge warnings. People often have to decide, in seconds, whether a flooded road, stream crossing, or surge-covered street is survivable. If their eyes and bodies read the scene poorly, a dangerous choice can feel reasonable in the moment.
Official warnings already stress how deadly moving water can be. The National Weather Service warns that shallow floodwater can sweep people and vehicles away and NOAA Ocean Service explains that storm surge becomes especially destructive when water depth and forward motion combine. The Florida-led experiment adds a human-perception layer to that physical danger: before people lose balance, they may already be misreading what the water is doing.
Six test conditions let the researchers vary both depth and speed
The research team immersed participants in a controlled moving-water flume at two depths, 0.45 meters and 0.90 meters and at three speeds, 0.4, 0.8 and 1.2 meters per second. Each person experienced all six conditions in randomized order. After about 20 seconds in each run, the participant estimated the water speed and rated the risk of personal injury on a 0 to 10 scale.
Researchers at the University of Florida equipped each volunteer with waders, a raincoat and a safety harness attached to a metal cage. The design let the team expose people to stronger water while keeping the test controlled enough to compare one condition with another. The same setup could not recreate the full chaos of a real flood or storm surge, yet it gave the authors a cleaner way to isolate the two physical variables that matter most for stability: water depth and water speed.
Before the flume trials began, the team also asked about prior experience with rip currents and tropical cyclones. That detail matters because the paper was not only asking whether people fear moving water. It was also asking whether earlier exposure teaches the body and the mind to read a dangerous scene more realistically.
Most people saw the water as faster than it really was
The broad pattern was clear. Average speed estimates rose above the true speed in all six water conditions. At 0.45 meters of depth, participants estimated the 0.4 meters-per-second flow at an average of 1.19 meters per second, the 0.8 flow at 4.04 and the 1.2 flow at 5.26. At 0.90 meters of depth, the average estimates climbed to 1.23, 4.78 and 6.36. The current felt faster than it was and the mismatch generally widened as the test became more intense.
The depth effect is important because many public warnings talk about flood height in plain numbers. Height alone does not describe what a person experiences once water starts pressing against the legs and torso. In this experiment, deeper immersion changed both perception and danger ratings. A waist-high flow can feel far more forceful than a shallower one even when the measured speed is unchanged and the data show that people folded that bodily sensation into their speed estimates.
One detail deserves care. The paper says people became less accurate as actual water speed increased or as water depth increased, but the distortion was not identical in every measure. At the highest tested speed, 1.2 meters per second, the slope of perception was fairly close to the one-to-one line even though the average level of the estimates still sat too high. In plain terms, participants were still overshooting the number, yet their sense of how much faster the water was becoming had improved by the top condition.
Prior storm experience improved calibration, but it did not erase the bias
The clearest moderation result involved the link between actual water speed and perceived water speed. Participants with no storm experience showed a much steeper, less accurate response curve than those who had experienced 10 or more storms. In the paper’s simple-effects model, the slope for the zero-storm group was 7.65, while the slope for the 10-or-more group was 3.90. The experienced group still overestimated, but the overestimation grew less sharply as the water sped up.
A similar pattern appeared for rip currents. People without rip-current experience had a less accurate slope than people who had encountered them before. The study therefore supports a narrow, useful claim: prior exposure can improve calibration. It does not show that experienced people become precise human instruments and it does not show that prior exposure makes anyone safe in strong flow. The advantage was relative, not absolute.
The authors also flagged an important weakness inside the storm-experience variable itself. Four participants reported experiencing 10 or more tropical cyclones, which the paper says was unlikely given their ages and the historical record. When those four people were removed, the moderation of the actual-to-perceived speed relationship weakened from conventional significance to marginal significance. The direction stayed the same, so the signal did not vanish, but the evidence became less firm.
Perceived speed helped explain risk, though the direct danger signal stayed strong
The study did more than compare guesses against the true flow. It tested whether perceived speed partly explained why faster water felt riskier. The answer was yes. Water-speed perception partially mediated the relationship between actual speed and perceived risk, which means some of the danger rating came from what participants thought the water was doing, not only from the physical force acting on them.
Even so, the mediation result needs careful reading. For people with no storm experience, the model attributed about 42 percent of the total effect of actual water speed on risk to the indirect path through perceived speed, with 58 percent left in the direct path. For people who reported 10 or more storms, the split was about 41 percent indirect and 59 percent direct. Those percentages are close. The paper therefore does not support a dramatic claim that experience completely changes how risk is constructed. It supports a smaller claim that experience improved the actual-to-perceived speed link and that this slightly altered the balance between direct and indirect risk pathways.
The discussion section adds another nuance. People with more storm experience appeared to have a somewhat stronger direct relationship between actual water speed and felt risk, while the perception-based mediation was slightly stronger for people with fewer storms. That sounds contradictory only if direct and indirect effects are treated as a winner-take-all contest. In practice, both paths operated together. Experienced participants may have read bodily danger cues more directly, while less experienced participants relied a bit more on their distorted speed estimates.
Why the findings matter for flood and storm-surge warnings
The practical message is straightforward. Public safety campaigns often tell people how high floodwater may rise, how fast a surge may arrive, or which roads may go under. The experiment suggests that a person standing in moving water can still misread the scene badly enough to make a poor choice. That is one reason the National Hurricane Center emphasizes storm-surge risk in impact terms and why weather agencies also describe moving-water hazards through what they do to people rather than through raw measurements alone.
Because the study used a controlled flume, it could not capture the turbulence, debris, uneven footing, panic, darkness, or cold that often define a real flood. The sample was also young, mostly male and drawn from university students. Those limits matter. Older adults, children and people with different body sizes or mobility could respond very differently. The paper itself calls the findings preliminary and warns against stretching them too far.
Still, the value of the work is easy to see. It gives disaster researchers a better starting point for asking how people interpret danger before they fall, before they are swept away and before a rescue becomes impossible. Better warnings may eventually combine simple physical numbers with human consequences, such as how a given flow can knock down a person or move a vehicle. If that kind of message helps people align perception with reality, a laboratory result with 76 participants could support safer decisions long before the next flood arrives.
