By Chiara Ryals
Despite the increasing probability of flood events, people continue to settle along coastlines and high-risk areas, a trend evident by the roughly 1.5 billion people worldwide facing flood risk. Floodbase is building resilience against the changing climate with the largest collection of maps tracking flooding globally, making flood insurance possible for populations left out by traditional maps. The company’s Global Flood Database, using over 15 years of flood data combined with human settlement data, creates the first satellite-based platform for global flood risk management and mitigation. By using both historical flood data and near real-time monitoring, Floodbase identifies who is and will be most affected by flooding. We spoke with Dr. Colin Doyle, Principal Scientist and Chief Innovation Officer at Floodbase to learn more about how the company is building resilience in vulnerable communities and responding to floods that previously did not exist.
From Ancient Belize to Modern Flood Prediction
Before leading efforts to predict the future of floods, Doyle focused on water management of the past. Starting his academic career in science, his first experience with geography as a discipline wasn’t until grad school. For his Master’s thesis, he conducted a geoarchaeological project in Belize, where he explored ancient agriculture and water management systems using Light Detection and Ranging (LIDAR) technology. In 2016, his research team flew LIDAR on a plane over Belize, uncovering one of the largest ancient Maya wetland fields. His work involved excavations in swamps and extensive research at the soil and water lab in the Department of Geography and the Environment at the University of Texas at Austin. By analyzing carbon isotopes, geochemistry, pollen, and soil composition, Doyle reconstructed the history of these wetlands, revealing how the Maya transformed them into highly productive agricultural systems. His findings featured the timelines of canal construction and their eventual abandonment.
At Floodbase, Doyle merges his scientific expertise with GIS to advance flood mapping and data utilization for emergency and insurance purposes. As principal scientist, he builds core technologies and collaborates with insurance and other partners to develop innovative methodologies that address their problems. Doyle’s experience in applied science is evident as he emphasizes the importance of connecting geospatial data to real-world insights and actions. When asked what he enjoys most about geospatial technology, he notes that both natural and social worlds are inherently spatial, and having GIS in his tool belt allows him to answer a wide variety of questions about the world.
Pioneering Projects
Doyle’s work with Floobase’s technologies shifts away from the traditional flood technologies that use models to predict the future. With an accurate look at the state of flood risk, Floodbase uses its data for emergency and insurance purposes. The company provides parametric flood insurance, which uses a predetermined threshold for the magnitude of flooding, instead of relying on an adjuster to determine payouts from a loss. With this technology, they cover business interruptions and aid governments with a variety of complications associated with flooding. Contrary to traditional flood models, Floodbase’s technology reduces uncertainty due to a lack of information on rapid anthropogenic changes or poor topographic data.
Floodbase’s unique approach is characterized by its early adoption of machine learning and the integration of diverse data sets and disciplines. The team comprises engineers, earth scientists, hydrologists, and remote sensing technicians, all working together to frame new models with the best available data sets. When asked about his favorite projects, Doyle refers to the company’s Magnum Opus, the Global Flood Database, which was done in collaboration with NASA as well as the Flood Observatory, and funded by Google. The project resulted in over 900 publicly available flood maps and a Nature cover story, demonstrating that increased flood vulnerability over the past two decades is primarily due to population growth in floodplains rather than increased flooding itself.
Floodbase’s findings on the global population exposed to floods
Beyond just publishing their research and datasets, Floodbase has elevated the potential uses of their data by making contributions to open-source projects and collaborations with humanitarian organizations. One example Doyle discusses is a project with the World Food Programme in the Republic of Congo, where a recently developed refugee camp was at risk of flooding. With Floodbase’s help, they mapped historical flooding with remote sensing data to assess its risk, leading to the camp’s relocation.
Floodbase has another collaboration currently underway with USAID, a project which will provide parametric flood insurance for farmers in Mozambique and Malawi, where nearly all smallholder farmers are uninsured against agricultural risk. Modeling risk and flood insurance systems have been difficult to implement or prohibitively expensive in agricultural economies. The upcoming project aims to automatically check for floods and trigger a payout in the event of a flood, creating a safety net for farmers. Through projects like this one, Floodbase recognizes that climate adaptation requires innovative solutions that combine leading technologies with local expertise. The company has effectively implemented this approach for several at-risk countries across the world by collaborating with governments and humanitarian organizations.
The Future of Geospatial Technology and Learning
Looking ahead, Doyle sees two major ways that the potential of artificial intelligence and cloud computing will revolutionize the field of geospatial technology. First, AI removes the limit of how much data we can process and classify. Now, the question he asks is, what can you do with the classification? The second major change Doyle looks forward to is being able to combine different new forms of data that wouldn’t otherwise be able to be combined. He explains that with the growing size of datasets, machine learning functions can extract insights from that information. Advantages of machine learning for him, along with using them to create more lively models for prediction, and for climate. He’s also excited by the possibilities of cloud computing, by removing computational limitations, allowing for rapid data processing and new avenues for inquiry. The increasing availability of earth observation data from commercial companies further enhances the scope of geospatial analysis. With the weight of those datasets lifted, Doyle looks forward to our ability to ask new questions.
These questions may relate to environmental challenges, as he explains that understanding the spatial nature of the Earth is fundamental to addressing environmental challenges. To him, geospatial technology is a backbone of sustainability, allowing us to see what is working and what is not, so we can build tools to fix our most persistent problems. He advocates for an approach that combines satellite data with ground data and fieldwork to create resilient systems that not only aid in disaster risk reduction but also build safety nets to help communities recover and rebuild their livelihoods after catastrophic events.
Advice for Aspiring Geospatial Scientists
For those entering the field, Doyle advises developing literacy in AI and gaining hands-on experience. He believes in the importance of practical application and encourages aspiring scientists to pursue questions they are passionate about, even if they don’t yet know how to answer them. According to Doyle, the best learning happens in a collaborative environment where ideas can be shared and developed collectively. He advises young people to find others who are motivated and passionate to work with, a move which will set you on a path towards continuous learning. Doyle’s own work with dedicated teams has allowed him to publish pioneering studies and now apply his expertise at Floodbase, where he’s advancing flood modeling to address one of the greatest environmental risks to the global population.