Satellite data helped the flood observatory compile this 29 August map of flooding (red) caused by Hurricane Harvey.

In Colorado, a global flood observatory keeps a close watch on Harvey’s torrents

BOULDER, COLORADO—While the media is gripped with pictures of waters swallowing freeways and pouring into homes and offices in southeastern Texas, geoscientist Robert Brakenridge is waiting for images of his own.

Brakenridge directs the Dartmouth Flood Observatory at the University of Colorado in Boulder, which creates real-time maps of flooding events around the world. The lab, which he founded in 1993, uses satellite imagery to monitor changing water levels, and is experimenting with other methods—such as using satellite-based microwave sensors—to track river levels. Such efforts have not only helped identify areas that are prone to floods, but also can provide real-time help during emergencies, guiding responders away from impassable areas and toward people that need help.

During past disasters, the lab has provided public maps and geospatial data that have been put to humanitarian use, and worked with the World Food Program to guide food delivery convoys. During Hurricane Harvey, the lab has been part of daily conference calls in which government agencies, researchers, and other groups share data on the storm and work to develop new and more useful maps.

That is the promise, but the lab’s early work on Harvey is also highlighting the obstacles that still confront researchers trying to fully harness available technology. For instance, despite the extensive attention that Harvey is getting from federal agencies, getting useful satellite images has been slow going for Brakenridge. It took until a few days after landfall for the lab to get workable images to build out its map of the extent of Harvey’s damage; by Tuesday evening the lab had incorporated several other data sources but was still working hard to get detailed data in fast enough to be useful to responders.

“Right now for disaster response, people need to know where the water is,” Brakenridge says. “One satellite image, one timely way of showing where the water is in some detail, that’s better and easier for them to use.”

One big problem is the availability of satellites. NASA’s Moderate Resolution Imaging Spectroradiometer can capture images twice a day, but its sensors can’t see through cloud cover, a problem in the days after Harvey. A European Space Agency satellite equipped with synthetic aperture radar does pierce clouds, but makes larger orbits around the planet and took several days to send an image of the Gulf Coast. Some private and international satellites will send images to the government for free during an emergency, but sometimes strips useful geolocation data, making it burdensome for researchers like Brakenridge to incorporate them into digital mapping systems or link them to topographic maps that can help identify the depth of floodwaters. As the storm clears, however, Brakenridge expects to receive better optical images from government and private satellites.

Ultimately, the Harvey data will be made available on the lab’s website, joining historical records of floods around the world since 1989. The lab’s maps are similar to the flood risk maps produced by the Federal Emergency Management Agency (FEMA), which uses historical data and modeling based on topography to predict future floods. (The maps also often inform land-use policies and home insurance rates.) But where FEMA is projecting hundred-year floods—defined as an event with a 1% chance of happening in a given year—Brakenridge is more concerned with highlighting where a flood has happened recently.

That difference—between FEMA’s approach of building a database based on modeled floods, versus maps based on observed floods—highlights a “really deep question” about how best to think about and visualize flood risks, Brakenridge says. And it’s one that regulators are increasingly reckoning with, as governments around the world and agencies affiliated with the United Nations start to integrate more satellite data into efforts to reduce flood risks and improve responses.

Brakenridge believes that, when it comes to floods, the recent past can tell us a lot about risk in quickly changing landscapes, such as rapidly sprawling cities such as Houston, Texas. “If the past is the key to the future, isn’t the recent past an even stronger key to the future?” he asks. “If your car was robbed 10 years ago in one spot and it hasn’t been hit there since, you’d feel pretty secure. If it’s robbed there last year, you’re not feeling as safe.”

Now, as recovery from Harvey begins, just how safe southeastern Texas will be in the next flood is sure to be a question on the minds of many.