Using a custom-designed drone, a researcher from The Ohio State University mapped glaciers and wetlands in Peru’s Cordillera Blanca mountain range. Despite a discovered drop in glacier growth, the findings from the unmanned aerial vehicle (UAV) revealed a slightly more optimistic picture of the region’s water supply, which relies in part on the ice formations.
In July of 2014 and of 2015, Oliver Wigmore, a doctoral student in geography at The Ohio State University, used the high-altitude UAV to fly 100 meters above the mountains. Equipped with a standard consumer-grade visible (RGB) camera, the UAV took hundreds of overlapping photos.
The “Structure from Motion” process reconstructed three-dimensional models of land surface elevations. Wigmore compared the models from 2015 and 2014 to determine the glacier’s volume change.
An Eye on the Ice
The custom-designed, multi-rotor hexacopter is fitted with a 3D robotic autopilot and GPS. The UAV includes the visible RGB standard camera; near-infrared and thermal infrared cameras, like the thermal core from DRS Technologies, have also been placed onto the aircraft. The UAVs provide a 10-centimeter image resolution.
“In this case, you have thousands of viewpoints, all looking at different locations on the Earth’s surface,” said Wigmore. “You can see changes in the glaciers at really high resolution — centimeter- scale imaging.”
By evaluating the digital models from the aircraft, Wigmore observed the sections of glaciers melting at the fastest rate. Using the UAV, Ohio State University researchers analyzed the dramatic collapse of the Llaca Glacier’s calving front. Wigmore recorded an average of 0.7 meters of thinning in one year, with a maximum of 18 meters of loss in some locations.
Cordillera Blanca typically only has two to three months of reliable clear-sky days, a small window for above-cloud satellites to take measurements. By going under the clouds and above rough terrain, the UAV allowed Wigmore to more easily examine the glaciers.
Traditional methods of glacier analysis call for a team to install ablation stakes directly into the ice to record how surfaces have dropped during a given period. Over debris-covered glaciers, however, melt rates vary, depending on surface cover characteristics like melt ponds, ice cliffs, and rocks. Results also depend upon the handful of locations where ablation stakes are installed. The UAV provides a more thorough and instantaneous assessment of a whole glacier.
“Using a UAV, we have spatially configured data across an entire surface, and you can get a really good estimate of how much these glaciers are changing and how fast,” said Wigmore, “without having a bias of where you're sampling.”
Glacier change is calculated in two ways: volume and velocity. To determine velocity data, feature points on the glacier surface are mapped across multiple image dates. By measuring the distance between the points over time, the researchers achieve velocity vectors to track horizontal movements of the glacier.
Vertical, or volume, ice change is found by comparing the digital elevation models collected at two different times. The spatially continuous geodetic, or land-surveying, method measures all points on the glacier directly. A value of vertical change is recovered for each 10-centimeter square across the glacier, achieving a total volume of ice change in the surveyed area.
The Technology Onboard
The drones include high-speed motors and long propellers. The low-cost UAV — thanks to advancements in hardware, open-source software, and computer vision — costs about $4,000 to make.
To operate the UAV, Wigmore used GPS and compass capability to plan a series of navigation waypoints. The flight plan was stored and loaded onboard the UAV. Once activated, the UAV then sent back position, speed, and elevation measurements, allowing Wigmore to monitor the drone’s progress. The mission could have been overridden both via laptop and remote control if needed.
The nonprofit geoscience consortium UNAVCO provided the differential GPS system in 2014. For the 2015 research, Wigmore used a differential GPS from The Ohio State University’s Byrd Polar and Climate Research Center.
Wigmore has added onboard sensor systems that map the Earth in visible light, near-infrared, and thermal infrared. In addition, The Ohio State University geography researcher is currently working on a fixed-wing platform, which will allow operation in a longer range.
“The great limitation of the multirotor platform is that you’re fighting gravity the entire time, so you use a lot of battery,” said Wigmore. “If we can get a fixed wing that has a long enough flight time, we can send it further.”
Measuring Soil Moisture
The UAV’s findings corroborated hydrochemistry data determining that Cordillera Blanca’s groundwater system contains both precipitation and glacier melt — a possible buffer to the region’s downstream water supplies.
With the help of visible RGB, nearinfrared, and thermal infrared cameras, Wigmore received the multispectral data necessary to measure soil moisture. Different spectral indicies from Normalized Difference Vegetation Index (NDVI) and Temperature Vegetation Dryness Index (TVDI) bands correlated to measured soil moisture values at calibration points within the scene.
Although Wigmore said there are significant error bars for the method, the approach provides a valuable picture of the highly heterogeneous spatial variation in soil moisture storage within the pro-glacial wetlands.
The spatially distributed map offers a more accurate view of soil moisture compared to traditional satellite images, which are coarser in resolution. The Landsat satellite, for example, has a 30-meter spatial resolution, compared to the 10-centimeter resolution of the UAV image.
“In that Landsat pixel, you may have a stream, a wetland, some dry ground, or bare rock. You’re pretty lucky if there’s any correlation,” said Wigmore. “We used UAV data to bridge that scale gap.”
During the dry season, inhabitants of the Peru mountain region rely in part on glacier melt for their water. As these glaciers recede, Wigmore said there will be less water supply in the dry season and increased variability in the supply. The researcher will explore the role of groundwater, and soil moisture storage, in buffering the dry season supply as well.
“If we have large potential within these wetland systems and within the groundwater aquifer systems to store water at least over an annual cycle, it may be less dire than previously predicted,” said Wigmore.
Wigmore will return to Peru in August to continue the research. The Ohio State University doctoral student will test new UAVs, deploy new atmospheric instruments on the vehicles, and continue mapping the area’s changing glaciers.