Digital Dialogues

Expert ideation to solve policy challenges

May 2019

Digital Dialogue No. 3:

Communicating Changing Risk

The Challenge

Flood risk in many parts of the United States is projected to increase in the coming years due to both continued floodplain development and climate changes. This is especially true in coastal areas where hurricane patterns are shifting and sea level rise is both increasing nuisance flooding and pushing storm surge further inland.  These impacts are already being felt.  For example, tidal flooding in coastal communities has increased over 200% over the last 20 years.

Scientists now have a deep understanding of potential impacts to flood risk over the coming decades, but many residents do not have this information.  While there are multiple sources online that clearly communicate the risk and allow residents to examine future risks to their property and community, many fail to seek out this information, have trouble internalizing what it means for them in terms of financial impact, and are unsure what the information implies for any risk reduction measures they should adopt.

This digital dialogue, undertaken with the First Street Foundation, addresses the following question:


How can changing flood risk be better communicated to households and communities?


Professor of Hydrology, University of Bristol

Understanding risk and chance is difficult: if it were easy, casinos would make much less money.  For relatively infrequent events, such as flooding, the complex geography of which places get hit and which do not, along with the considerable uncertainty in defining these zones makes this especially hard.  Faced with these difficulties, scientists often ask themselves how to better communicate uncertain risk information to the general public, but this misses an important point: there is no such thing as the ‘general public.’  Instead, we have many ‘publics,’ each of whom will have preferred ways of receiving risk information.  Single modes of risk communication are therefore likely to fail for at least some groups in society and instead we need a plurality of different, and tailored, approaches.

Communicating risk when nothing is changing is tricky enough, but when risk changes over time then the challenges become very great indeed.  This is particularly true for rare floods that should, on average, only occur perhaps once in a decade or once in a century in any given location.  Here, individual experience often cannot conclusively discern if floods are becoming more or less frequent, and even for scientists this determination is hampered by short observational records and data errors.

What we do know is that flood frequency can change over time because of man-made climate change, natural climate cycles operating over decades to centuries, and physical changes within catchments such as dam construction, urbanization and changing land management.  However, we also know these changes do not apply everywhere uniformly and this complicates the risk communication problem.  Most places will see increases in flood frequency in the future, but some may actually experience decreases.  Moreover, it is statistically possible for rare floods to occur close together in time at the same place simply by chance.

Clearly, we need to develop property-level risk communication because, to misquote Tip O’Neil, all flooding is local.  This would be a radical change from current FEMA maps which designate broad zones as being equally at risk.  We also need to develop multiple ways of communicating risk information to the various ‘publics’ and put this in more human terms, such as risk of experiencing a flood over the lifetime of a mortgage.  Lastly, as scientists we need to be honest about the uncertainties in our predictions, and what we both know and do not know.


Former Mayor of Coral Gables, Florida

As the former Mayor of a prosperous, well-educated, South Florida community of 52,000 with 47 miles of exposure to the sea, I found that city officials need to take the lead in communicating facts about sea level rise risk to their residents. It is incumbent on those elected to deal with problems, to understand what lies ahead, inform community groups, and prepare city infrastructure for the future.  Realtors will not jeopardize sales by revealing sea level rise risk to potential clients, many of whom are recent immigrants with no awareness of the dangers ahead.  Much of the tax base of my city lies in parcels below 4 feet of elevation. And our city sits atop limestone and a low- lying aquifer. Our ability to maintain the high quality of life for which we are renowned depends on preparing for the climate challenges ahead, and on alerting residents to crucial facts so they can manage the risk of future flooding.

We began by making LIDAR measurements at 4-inch resolution of our city, placing critical infrastructure, including schools, on maps showing elevations and then held community meetings to inform residents.  Turnout was excellent.  We sent videos of our presentations to homeowner groups and schools, and prepared a detailed legal study (free to all) that outlined obligations of homeowners and the city as the waters rise.  All of our studies are freely available on the city website.

The most receptive audiences will be our children and grandchildren, as they will inherit the challenges, and bear much of the tax burden of building up and better. Residents understand flood risk better when it is described as their home’s likelihood to flood over the life of their mortgage (e.g. 30 years), rather than likelihood to flood in a single year.  Realtors should be obliged to disclose such flood risks to their clients. City officials should prepare for possible retreat from certain areas and not abstain from discussing that possibility for the lowest lying areas.  As much information as possible should be provided by the city about historic facts and low, medium, and high probabilities of future flooding.  Individuals will make their own assessments of what to do if given good information, but elected officials should not hide unpalatable facts for fear of affecting the real estate market today.


Assistant Professor of Behavioral Science, Booth School of Business, University of Chicago

Behavioral scientists have demonstrated repeatedly over the last 50 years that people are not good at understanding risk and uncertainty, and this may explain why so few people take action to protect themselves against flood damages. As a behavioral scientist, my colleagues and I ask this question: can we communicate flood risk information in a way that improves understanding and motivates people to take action? To this end, we have conducted several surveys and lab experiments in which we tested whether certain types of flood risk information are more likely to spark action than others. Based on our findings, I have three recommendations for effectively communicating flood risk:

  1. When conveying the probability of a flood, either communicate the risk over a long, but relevant period of time, such as 10 or 30 years, or provide qualitative risk information, such as “high risk” or “medium risk.” In incentivized experiments, we have found that risk framed over longer time periods significantly increases the number of people taking costly protective action in the short-run compared to risk framed over a 1-year period. Moreover, in a survey, people report that the qualitative risk information (e.g., “high risk”) is even more motivating than the extended time horizon.
  2. If choosing between communicating the probability of flood or the potential impact of a flood, go with impact. In a survey, we found that people report being more likely to take action after being given impact information about a potential flood (e.g., potential monetary damage to their house if a flood occurs) than after being given probability information about a potential flood (e.g., 1% chance of flood)—even if that probability is framed over 30-years.
  3. An alternative approach to risk communication is to provide strategies that make protective actions easier (e.g., simple instruction graphics) or ways to reduce the cost of protective actions (e.g., specific information about where to obtain a low-interest loan for elevating one’s home). This is based on a survey we conducted that found the following: Whether a person decides to take on a protective action (e.g., buy sandbags or raise the wiring in one’s home) is largely driven by the perceived cost and perceived ease of the action, not by how effective the action is perceived to be.


Chauncey Starr Senior Fellow, Resources for the Future

Hurricane Agnes hit Florida in 1972. Its inflation adjusted dollar damages ($12.5B) made it the worst Atlantic Hurricane then on record. The next record was set by Hurricane Hugo (1989, $14.1B), and set again by Andrew (1992 , $47.8B), which in turn was dwarfed by our current record Katrina in 2005 ($160B) (damage data here).  People making housing decisions, insurance decisions, business decisions or life decisions should take such facts into consideration. They should also think about something else; the Next Worst Case. When will it come? How bad will it be? How should we prepare? Damage scenarios up to the current worst case are things we’ve seen before. The next worst case will take us beyond what we know.

To start thinking about how to answer these questions, suppose the tallest person you have ever seen was 6’8”. Someday you may meet a taller person; how tall do you think that person will be? Maybe 7’? What is the probability that the first person you meet taller than 6’8” feet will be more than twice as tall at roughly 13.5’? Surely that probability is infinitesimal. The tallest living person, Sultan Kösen of Ankara Turkey, measures a bit less than 8’3”.  People’s height is a “thin tailed” distribution.  Hurricane damage is not; it is “fat tailed.”  You could meet a “13.5 foot” hurricane.

Hurricane loss data has been “de-trended” so that the “normalized loss data”  adjusts for growth in population, wealth and exposure. The normalized data is very fat tailed: if we average the normalized damages of all hurricanes with damage at least $X, the answer is about 2 x $X . If you need to prepare for the next hurricane at least as bad as Katrina, using twice the normalized damages of Katrina will give you an even chance of covering the damages. Growth in population, wealth and exposure should be factored back in.  When will the next worst case arrive? Its arrival time is very hard to predict but arrive it shall. The take away: Don’t prepare for yesterday, prepare for tomorrow.


Senior Fellow and Strategic Adviser, Wharton Risk Center

Climate Change is the consummate ‘boil the frog’ dilemma: we will always find something legitimately more pressing to address—affordable housing, schools, crime—that is, until the biblical rain, arid forests, or encroaching seas, from one day to the next, wipe out our livelihoods. Some in Houston, Texas and Paradise, California have already reached the moment of tumbling, bubbling water.

However, for most, the crucial boiling point won’t arise for this generation of adults, but instead will force difficult decisions on the next generation. For the millions of residents that can see or smell an ocean, bay or inlet from their home, the viability of their home, the reliability of their plumbing, the certainty of access, and the value may all shift and likely decline. With these changes, neighborhoods and communities will have to adapt. Retreat will be the only option for many.

For these reasons, children’s geography lessons today must layer on a factor no earlier generation has contemplated: that coastal boarders will change. Lands will contract across the globe, directly impacting all but those landlocked countries – Zimbabwe, Laos, Austria, and others. Whether in Brooklyn, Venice, or Dakha, topography, subsidence, and currents will combine with changing sea heights to transform the landscape.

Because adults are saddled with the burdens of finding work and building lives, moving or change is naturally exhausting to contemplate, let alone to execute. For children, new landscapes often incite wonder, curiosity, questions, and creativity. Time also allows solutions and alternative visions to emerge. Keeping children in the dark about risk is like holding the lid on the heating pot. If, instead, we incorporate monitoring and longitudinal studies in grade schools, for example, students will understand changing threats over time.

I have explained the impact of king tides on neighborhoods in Florida to my 9 and 7-year-old boys. They immediately asked why houses can’t be built on jacks: “Just like cars! When they have a flat tires!”

All this is to say that as we contemplate the impact of climate change on real people, it is our children who will bear the brunt of our negligent carbon management. We must equip our young frogs with the knowledge of what’s happening to their pot and give them time to jump before it’s too late.


Professor and Director of the Institute of Earth, Ocean & Atmospheric Sciences, Rutgers University

While sea-level rise projections for the end of this century are characterized by deep uncertainty – associated both with uncertainty in the course of future greenhouse gas emissions and in the physics of polar ice sheets – the path of sea-level rise over the next thirty years is becoming increasingly clear. You can use NOAA’s projections; you can use IPCC projections; you can use those from Rutgers, or those from the handful of other research groups around the world that work on localized sea level rise projections and come to the similar conclusions: It is extremely likely that global average sea-level rise over the first half of this century will total will be between about half a foot and 1.5 feet – and larger in regions already experiencing faster-than-global rise.

But magnitude of sea-level rise – global or local – isn’t the most salient metric for most people. Far more salient is to project how often flood waters will overtop key features. For homes sitting just about the tidal flooding threshold, it’s fairly straightforward to estimate the expected frequency of flooding over the lifetime of a typical mortgage – and, more generally, to calculate the frequency of flooding at elevations of salience to a community. That road in front of your shore house that currently floods about once a year? Without adaptive measures, by 2050, it’ll most likely be flooded about 40 days per year; there’s about a 10% chance it’ll be flooded only ten days per year, and about a 10% chance it’ll be permanently flooded.

This information should not be hard to access. It shouldn’t require homeowners to be familiar with tools like Rutgers’ NJ Flood Mapper or Climate Central’s Surging Seas; nor should it require homeowners to fork over cash to an unregulated private sector firm that may be overselling their capabilities. For homeowners, one possibility may be to require disclosure of projections of neighborhood flood frequency over the lifetime of a mortgage, using state-sanctioned sea-level rise projections.


Director and Principal Researcher, Susanne Moser Research & Consulting; Research Faculty, Antioch University New England

Let’s start from here: “tidal flooding in coastal communities has increased over 200% over the last 20 years.” As a statement it’s very impressive and sounds scary, but… no, not really, because “what is tidal flooding”? And “how is this relevant to me in my location?” And “200% – that’s a lot, I think… but is that double or double of double of whatever the flooding was at a time when I didn’t live here? And “what can I do about it anyway?”

In a nutshell, that’s the trouble with trying to figure out effective risk communication. To make risk communication come alive and become useful, it needs at least the following:

  • It needs to be place-specific: specify where exactly in a coastal community the risk has changed from x to y;
  • X and y need to be translated into socially meaningful terms: for example, instead of stating a percentage increase, a more meaningful rendering of the same quantitative information might be how many times a year or month access to a local business district or cultural landmark or a hospital is disrupted. With changing risks, this is easily updated over time: while the flooding-related disruption happened only once a year in 2000, it now happens nearly every month and experts expect it to happen more and more often in coming years;
  • Risk communication is best created with the intended audience: which local references are socially meaningful, how they are best conveyed and by whom is best determined by researchers and risk communicators working with intended audiences. For example, using social media- and web-based dashboards of flood risks may work for younger audiences; irremovable risk information signage using images and words (possibly in multiple relevant languages) installed in prominent places in rental properties, hotel rooms, and public spaces may work for transient audiences; face-to-face meetings may work with permanent residents; engaging employers may work to reach busy people earning a living;
  • Risk communication should never come without solutions communication: the cardinal mistake of risk communicators is to assume that audiences know what to do and know how to implement an effective risk response without being told. Solutions need to be specific, audience-appropriate and meaningful, motivational, doable, and clear. Solutions communication must spell out the pathway of how an enacted solution actually leads to the desired risk reduction. What would make solutions audience-specific, meaningful and doable should also be assessed in research engaging the audience; and finally,
  • Communication of growing risks from climate change will get harder over time: As sea levels rise and extreme events worsen, adapting in place will become increasingly challenging and eventually untenable. No one reacts to existential threats, threats to livelihoods, and loss of property without some resistance and fear. Risk communicators thus must become skilled at face-to-face interaction with intensely emotional audiences. (For more help on that see publications listed at