Last month, the United Nations University released a free tool that generates high resolution maps of floods worldwide since 1985. The new resource comes after a year of historic water-related disasters, including severe floods in Western Europe and the northeastern United States. Experts hope the online tool will aid in disaster readiness and future planning, especially for vulnerable countries with limited access to reliable flood maps.
The tool allows scientists, organizations and curious members of the public to adjust variables to see where floods have occurred in the past. Users can select a location and timeframe, and the tool, which draws on decades of data from remote-sensing satellites, produces a flood map at a 30-meter—about 100-foot—resolution. Viewers can see water inundation images down to street level. Similar mapping tools have been developed to assess floods by type and region, “but you don't see anything happening at a global scale,” says Hamid Mehmood, a remote-sensing specialist at UN University’s Institute for Water, Environment and Health (UN-INWEH) in Hamilton, Canada, and the tool’s lead developer.
A 2019 report from UNU-INWEH noted that around 90 percent of natural disasters are water-related, including cyclones, floods and droughts. According to the report, which Mehmood co-authored, over 5,300 water-related disasters have been reported across the world since 2000, leading to more than 325,000 fatalities and economic losses exceeding $1.7 trillion. Experts warn the financial and personal losses are likely to escalate with climate change, as warmer air can hold more moisture, affecting the intensity and frequency of precipitation. Having easily accessible flood data could help governments, insurance companies and urban planners determine the safest locations for investing in housing and industry. Knowing where past flooding has destroyed agriculture, for example, would help local stakeholders identify safer places to plant crops.
To create the World Flood Mapping Tool, Mehmood and his colleagues relied on satellite images publicly available through Google Earth Engine. The catalogue of raw data comes from nearly four decades of observations by a series of NASA satellites called Landsat, which have been snapping photos of Earth since the 1970s. To generate water inundation maps, the team created “data cubes”—layers of pixels captured by the satellite over a given period of time to identify temporary water bodies. “Based on the frequency of water in that column of pixels, we can classify it as a flood or a permanent water body,” explains Mehmood.
To test their model, developers compared the generated maps to documented flooding events in countries like Australia, Bangladesh, Canada and India, and reported 82 percent accuracy. One of the powers of the project is that it enables individuals to look at water inundation events over years or decades, which can show flood patterns that might otherwise be missed. Mehmood says the tool will be especially helpful for counties in Africa and Southeast Asia, which are rapidly urbanizing and often have limited flood maps. Authorities could use records available through the tool to pinpoint areas most susceptible to floods or plan evacuation routes.
The map also offers overlays of population, buildings and land use, which the tool’s developers hope can be applied to community planning efforts and insurance assessments. If low-lying coastal areas are hit with repeat flooding, for example, that could help officials guide builders to avoid constructing homes and businesses in those locations. But there are limits to the satellite images collected by Landsat, says Paul Bates, who studies hydrology at the University of Bristol in England. Because Landsat can’t see through dense clouds or vegetation, the model could miss some major floods, Bates says. Still, he sees the tool’s benefit, especially for countries in the global South—lower-income regions of Latin America, Asia, Africa and the Caribbean. “People in those countries may not have access to other tools, so providing them with access to an easy to use, free, public tool that allows them to map inundation in their countries to a reasonable level of detail—it's a great thing to do.”
The mapping tool seems helpful across a range of science disciplines, says Andra Garner, a climate scientist at Rowan University. While geographers can use the technology to identify those safer places to expand housing, hydrologists can use it to better understand the increasing risk of coastal flooding for island nations. But the real power of the tool may be in how it can make the escalating risks of climate change more concrete for members of the public. “This kind of tool can potentially be useful to help people really visualize what the kinds of flooding might look like, what the scale is, and the scope, how extensive it can be,'' says Garner. Seeing a map of an inundated area—especially one we have a personal connection to—lets individuals consider how their actions impact the climate. “It could be useful for trying to think about how to develop our society in a climate that is, at this point, changing pretty rapidly.”
There is no single factor exacerbating flooding across the globe, Mehmood explains, but rather a combination of local factors. For example, in some cases, the impacts of flooding are made worse by the geology of the area or way the land is being used. Mehmood says this is only the first version of the World Flood Mapping Tool—his team is already working on a more detailed version for commercial use that will offer resolution at a 10-meter—roughly 33-foot—resolution, which they hope to unveil next year. The updated model will rely on artificial intelligence to generate risk maps under different climate scenarios. “You look at parts of Southeast Asia, areas which never had floods, that get extreme flooding now,” he says. “The intensity is increasing, and we just cannot ignore water-related disasters—specifically floods—anymore.”