Confiance / Trust

A Life of Science: A Series by New Scientists

Satellites hold real potential to help societies weather the impacts of flood events.

  

A salty breeze drifts in from the Atlantic, carrying the faint scent of the ocean through the streets of Rabat. Unlike the labyrinth of bustling, narrow alleyways found in many other Moroccan cities, the capital offers a different rhythm—one of oceanfront promenades, green gardens, and a blend of historic markets and modern government buildings. This design gives the city a sense of calm authority. Yet inside the polished government boardroom where I anxiously await the outcome of an important meeting, the air is tense with deliberation and expectation.

After two hours in the boardroom, the air conditioning—cranked high against the Moroccan summer—has turned the sweat beneath my arms icy. I glance around the long conference table, where 15 ministry officials sit, their expressions unreadable. Across from me, my colleagues from the advisory firm in London are waiting—just as I am—for a decision that could change the way Morocco prepares for and protects itself against floods. With several hundred thousand dollars of insurance premiums at stake, my colleagues and I feel a strong incentive to “close the deal” with the Moroccan government and secure our firm a healthy cut of the profits.

Six months earlier, the Moroccan government had sent out a request to a number of advisory firms, including my own, to help the government establish an insurance program to provide its citizens with effective disaster relief in the case of a national flood emergency. In my then-role as a disaster analyst, it was up to me to help design a new approach that could quickly determine who had been affected by floods and how badly. Using this information, the government could distribute cash and other assistance to affected communities to help them get swiftly back on their feet again.

How were we going to achieve this ambitious goal? With satellites—not dissimilar to the GPS satellites we rely on every time we open Google Maps, or the communications satellites which allow us to make phone calls to friends in faraway places. The spaceborne instruments, called environmental sensors, that are used to monitor environmental hazards such as floods also orbit the Earth each day collecting data. However, unlike navigation and communications satellites, environmental sensors are specifically designed to take measurements of the Earth’s surface. If you have ever used Google Earth—or even opened the “satellite” option in Google Maps—you will have seen an image taken by a satellite gazing down from hundreds of miles overhead. The technology we use to track floods works in the same way. Except, instead of showing only what our eyes can see, they use special sensors to detect wavelengths of light beyond human vision, which is helpful for locating features like flood waters.

I take a sip from the cup of sweet mint tea in front of me to wet my dry mouth. As my mind wanders momentarily to imagine the feel of the warm ocean breeze outside, I am abruptly returned to the present by the sound of a senior government official loudly clearing his throat in anticipation of an announcement. The next words to leave his mouth feel like a sucker-punch: “nous ne faisons pas confiance à la technologie”—they do not trust the technology.

While it was customary for multilingual colleagues to translate for the English speakers in the room, I understood well enough what this meant for the outcome of the meeting. The ministry officials had chosen not to pursue our ambitious, satellite-based insurance program. And the reason? The officials believed that the satellite technology would be unable to detect their ephemeral, fast-moving floods. One minute the floods are there, the next they are gone. Morocco’s landscapes could also challenge the satellites. To a satellite, the roads of its dense urban metropolises “look” similar to water. So does the shiny, reflective surface of Morocco’s desert terrain. Reliably monitoring floods for primarily urban and desert areas was too challenging a task for the satellite capabilities that were available.

For weeks I looked back on that day with intense disappointment. However, the longer I dwelled on it, the more I agreed with the officials that the satellite technology wasn’t ready yet. Those who live by a stone-dry riverbed or in a concrete urban center during a rainy season will know about “flash” flooding. The kind of ephemeral flooding that, following an intense rainfall, rapidly overwhelms drainage infrastructure only to recede hours later, leaving a trail of destruction and disruption in its wake. While some environmental satellite-sensors “see” the Earth as often as once every 12 hours, it was entirely feasible that a fast-moving flash flood could have wreaked havoc and then “disappeared” before the satellite even had a chance to detect the floodwaters. Today, I see the decision of the Moroccan government as the spark that ignited my desire to improve what we know about floods and to find ways in which satellite technologies can help us better protect ourselves against the destruction that floods bring on communities—not only in Morocco but in other flood-prone parts of the world.

Five years on from my experience in Morocco, I have swapped my job in the insurance industry for a research position at the University of Arizona investigating the role of satellites in helping us to understand and manage the impacts of floods around the globe. Through my research, I have been fortunate enough to work with flood risk, agricultural, and development practitioners in varied settings, including in the low-lying Haor basin of Bangladesh and in the U.S. Midwest—two places with drastically different socioeconomic contexts but which have both suffered the devastating and enduring consequences of extreme flooding.

Along with my research group, I now spend much of my time testing the latest computer and AI algorithms that translate satellite images into useful information for decision-makers. Using powerful computers and lines of code written in specialized programming languages, we write automated data-processing  “pipelines.” These pipelines of computer code mean that when floods strike, at the click of a button, we can create detailed satellite-based maps indicating where flood waters are most likely currently affecting communities, which can support relief and recovery efforts in the days, weeks, and months following flood impacts.

Outside of the computer laboratory, I work closely with partner organizations—including Bangladesh’s Flood Forecasting and Warning Centre—to integrate these novel types of flood hazard data into applications that can help at-risk communities better prepare for and respond to floods. Many of my favorite experiences have taken place during capacity-building workshops, in which my research group and I demonstrate ways that interested organizations can incorporate satellite technologies into their own workflows. The best part of my work is witnessing first-hand the excitement of a workshop participant after they have created their very first satellite-derived flood map. While I am grateful to have gained a new, more holistic perspective beyond the overwhelmingly profit-driven world of insurance sales, I am still motivated by the need to increase accessibility to insurance to manage the flood risk faced by governments, societies, and individuals. Safe, fair, and inclusive insurance is essential to support post-flood recoveries alongside other risk management and mitigation strategies. Satellites—whose extensive vantage point of unfolding disasters can be used to activate insurance payouts without needing to rely on inefficient claims processes—hold great potential to overcome the problems which plague today’s mainstream flood insurance and facilitate more timely and affordable disaster relief. 

Since I was in Morocco in 2019, researchers have made great strides in applying satellites to monitor floods. This is in part thanks to impressive new satellite instruments capable of imaging the Earth more frequently and with greater precision than ever before. However, as I learned on that day in Morocco, new technologies can do little to help communities without a deep understanding of the local context. For my work, that means engaging with local people and organizations to understand how flood risks present themselves in specific landscapes and for different parts of society. This recognition is also why, in my research today, I feel a strong sense of responsibility to communicate both the opportunities and the limitations of emerging satellite technologies, taking heed of the biophysical, socioeconomic, and political factors that shape the environments in which people live.

If the Moroccan government were to ask me again today whether to invest in a satellite-based insurance program to protect its people against the impacts of floods, I would respond with caution. Despite the immense progress that has been made, detecting fast-moving floods across Morocco’s arid deserts and urban jungles presents a challenge that remains unmet. However, in other parts of the world satellites already hold real potential to help societies weather the impacts of flood events—including through new approaches to insurance that offer timely relief when it’s needed most.

In 2023, I visited rural communities in Sylhet, a region in the far northwest of Bangladesh where expansive lowlands butt up against the rugged, mountainous state of Meghalaya in neighboring India. Among the wettest parts of Bangladesh, Sylheti communities are accustomed to the immense influx of water that arrives each year during the summer monsoon. However, the prior year’s monsoon had brought unprecedented floodwaters that ruined unharvested rice crops, leaving many Bangladeshi communities stranded without income and food supplies.

During my visit, I heard from members of the community about how entire villages had taken shelter in local schools for several weeks before flood waters receded. I also heard about how the post-flood distribution of cash relief and seeds from the government was only accessible to the very largest farms. These firsthand accounts brought home to me the need to find better ways to support communities through the enduring impacts of floods, including more inclusive access to cash and other forms of relief quickly after floods hit.

Returning to the computer laboratory after my time in Bangladesh, I checked the outputs of the algorithms my research group and I had developed. I wanted to see if the disastrous magnitude of the recent Sylhet floods was evident in the satellite data. I was pleased to find that it was! Consistent with the accounts of local people, the satellite data showed that the scale of the recent floods was beyond anything that had previously occurred in the 20 years of satellite observations. What’s more, when disastrous flooding occurred in another part of Bangladesh in 2024, outputs from the satellite-based computer algorithms once again showed loud and clear that these floods were the worst the area had experienced since the satellite record begun.

While the validation of our computer algorithms against real-world accounts provides confidence that satellites can provide trustworthy information about flooded areas, accurate data cannot erase the harsh impacts that continue to be felt by many thousands of Sylheti people following the disastrous floods in the summer of 2022. The next step towards improving post-flood conditions in Bangladesh and other flood-prone regions around the world will require researchers like myself to carefully connect our human-validated satellite data with organizations who can put it to use to help those in their communities who need it most.