Drones a new method to measure river streamflows

The U.S. Geological Survey used a drone last week to measure streamflows in the Roaring Fork and Colorado rivers.

The crow-sized quadcopter buzzed 100 to 200 feet above the rivers, filming the water’s surface. After it landed, USGS personnel analyzed the footage using a new algorithm that calculates the amount of water flowing between the banks.

A four-person team ran measurements over four locations on the Roaring Fork River (below Maroon Creek and at Emma) and the Colorado River (at and below Glenwood Springs) last Tuesday and Wednesday. After testing the new method here in western Colorado, USGS personnel say it likely will be used to enhance flood-warning and water-management systems across the country.

With a network of 362 streamflow-monitoring sites across the state, the USGS is by far the largest source for streamflow data in Colorado. There are 17 gauges in the Roaring Fork watershed. Measurements from these gauges, updated in real time every 15 minutes, are essential for stream-flow forecasts, flood warnings, dam operations and irrigation diversions in the area.

But the gauges have one issue: They only measure water depth, not velocity. Typically, USGS personnel can estimate streamflow from depth, but they still need to verify those assumptions by periodically updating their water velocity measurements as well.

To do that, USGS’s hydrologic technicians often use an “acoustic Doppler current profiler,” or ADCP. The ADCP, which looks like a tiny orange outrigger with a robotic cookie on its back, uses sound waves to measure the speed and amount of the water passing under it.

To use the ADCP, USGS personnel must either pull it along the water from a bridge, set up a cable pulley or get into the river themselves and drag the floating device across, from bank to bank, about six times. The entire process, including unpacking and repacking, can take 40 minutes, and it’s dangerous when water levels are high.

That’s why USGS is testing its new system, an algorithm that can measure surface velocity from well above the surface. With a drone, the whole process only requires about 10 minutes. Then the technicians can analyze and publish the data from the road.

Brandon Forbes, associate director of the Hydrologic Data Program at USGS’s Colorado Water Science Center, oversaw the tests last week. He said he has been pushing to implement this kind of technique for a decade now, and it is finally happening.

Before working in Colorado, Forbes studied flash floods in Arizona for 15 years. Even during a massive flood, he said, teams of USGS personnel would have to carry out the sometimes 40-minute ADCP operation to calibrate the streamflow gauges, sometimes forcing them to go into the river itself where a bridge wasn’t available.

This method, he said, can be “super dangerous and super difficult.” The new remote measurement technique will be “much safer.”

Forbes said the new method also will allow the USGS’s 65-person water science team, spread across Colorado, to quickly measure rivers during potentially dangerous floods.

“We’ve had such a mild spring and our snowmelt has been slow, but if this all came off quickly, this technology would have been very useful,” Forbes said. Data is used for public safety, potential bridge closures and reservoir operations.

“It’s critical stuff,” he said.

Alan Kirk, a hydrological technician at the Grand Junction office who is personally responsible for managing two of the gauges in Glenwood Springs, said the drone method will add significant benefits for his daily work.

“[The new method] is a huge time saver. It will make the job a lot safer and faster. … It’s a game changer.”

Kirk, who will continue to make measurements at the two sites in Glenwood Springs, among others, said he plans to learn how to fly the drones himself: “It’s a great next step for me.”

John Fulton, a research hydrologist on the team, said in the next few weeks he and Matt Nicotra, a physical scientist, will post the information from the initial measurements. After that, they will officially publish the findings sometime in the fall or winter.

Then, he said, the new method will be ready for implementation by USGS teams across the country. Fulton and Nicotra will write the manuals that inform the implementation.

Austin Corona, Special to the Aspen Daily News