Oliver Wigmore is a Ph.D. candidate and presidential fellow at Ohio State University’s Department of Geography, studying the changes in high-altitude Andean glaciers in South America. To help measure how rapidly they are receding, he has had to build his own unmanned aircraft, as most hobbyist systems are unable to cope with the altitudes where he works.
What makes your homemade drones able to operate at high altitudes versus most hobbyist vehicles?
The drones that I build are custom designed to operate well in high mountain elevations over 4,000 meters above sea level where the air density is much lower. We have successfully launched at over 5,000 meters and will push this to around 5,600 meters this year. In theory, they should continue to work well at over 6,000 meters, but we haven’t needed to go that high yet to fulfill our research objectives. Our platforms are built mostly from carbon fiber, which keeps things very light, strong and stiff. They are also much bigger, around one meter in diameter, than typical hobbyist vehicles, which means they can spin bigger props and are more stable in gusting high winds, typical in the mountains.
We could go bigger still; however, these frames also have to be small and light enough that we can carry them on our backs while ascending a glacier with ropes, crampons, axes, etc. Fully loaded, they weigh about three to four kilograms depending on sensor payload. Our drones are fully autonomous,vrunning on 3DR Pixhawk, and often fly out of sight — one to two kilometers from launch site — so we use extended radios for control and telemetry. We give careful consideration to a combination of prop sizes and pitches, motor size and speed, battery voltage, platform weight, and intended survey target while designing our UAS to maximize the capacity to successfully complete our missions.
What advantages do drones give you in your work?
Drones allow us to map the land surface at centimeter resolution, providing a spatially contiguous snapshot of the survey area. This allows us to investigate land surface processes at previously impossible scales, bridging the gap between field instruments/measurements and data derived from satellites. This allows us to better understand spatial heterogeneity within our study areas. Additionally, the advent of this on-demand hyper-spatial imagery allows us to address entirely new research questions.
Can you explain some of the differences between the drones you have made and some of the more hobbyist-oriented drones?
In addition to the differences described earlier, our UAS are fitted with a number of different sensing payloads, as opposed to simply taking photos/video. We collect downward-facing, nadir imagery across five spectral bands. Visible (red, green and blue), near infrared (for plant health) and thermal infrared (land surface temperature). This allows us directly measure and derive secondary indices that correspond to a number of different environmental variables. Through a combination with survey-grade GNSS [Global Navigation Satellite System] ground control, we can create seamless multispectral orthomosaics and generate detailed digital elevation models. In the future, we plan to extend this sensor payload to measure atmospheric variables — temperature, relative humidity, gas concentrations, etc.
How often do accidents happen at such a high elevation? What parts do you find yourself swapping our more frequently?
Accidents happened fairly frequently in our first year, but less so now, thanks to extensive testing and refinement. Though last year I did have one catastrophic crash from about 280 meters above ground when a carbon-fiber propeller sheared in half. Needless to say the result was pretty dramatic. I am now in the process of integrating a parachute recovery system for this type of emergency.
Typically, when we crash it snaps the carbon fiber arm tubes, as these are thinner than they should be for this size of the platform, a necessary requirement to reduce weight and enable high-elevation flight. Luckily, these parts are easily replaced, and I simply take extra tubes down with me each year. The electronic components have generally held up pretty well to the abuse, though I have fried an ESC [electronic speed control] and a motor on two separate occasions. Another issue we face, which has been responsible for at least one crash, is limited sky view that impacts the GPS and navigational solutions. This has resulted in one serious flyaway with a resultant crash into the steep cliff walls.
What kind of field support do you have for your work?
Our study sites are anywhere from one to three days’ hikes from the nearest trail head. If we are going deep into the mountains, we will usually have a local support team — cooks, mule drivers, etc. — as well as a number of different researchers working on different projects. In this case, we will head up the valley with around 15–20 donkeys loaded up with supplies and equipment for a week or so of work from a base camp set up. From the base camp, it’s usually a matter of strapping equipment on our backs and climbing onto the glaciers where we collect ground-control GNSS data and conduct our flights. My adviser, Bryan Mark, has been working in the Cordillera Blanca for a long time, so we have a large network of local and international collaborators, as well as maintaining good relationships with local government entities, such as the Huascarán National Park Office and the Glaciology Office of the National Water Authority.
What have your findings been so far, relating to glacier movement and water supplies?
The high spatial resolution of UAS data/imagery is facilitating a far better understanding of the extreme heterogeneity and complexity of tropical Andean glacier dynamics and proglacial systems that regulate downstream water supply in the dry season. Return surveys over Llaca glacier have measured dramatic and highly variable down-wasting (vertical loss) of glacier mass, while only recording limited horizontal retreat. This is important, as often aerial studies of glacier loss only account for changes in glacier area, which in the case of Llaca doesn’t change much despite significant loss of ice volume. UAS allow us to collect detailed information on these changes at low cost and regular intervals.
Another question that we are interested in is the role of groundwater systems in buffering dry season stream flow. The high-resolution UAS imagery allows us to corroborate earlier findings by other members of our team using traditional hydrology techniques — hydrochemical tracing, seismic studies, etc. Additionally, it allows us to track soil moisture variability and identify both surface and subsurface pathways that regulate water supply to diverse downstream water users.
Do you have future plans to use UAS and, if so, for what work?
UAS will continue to be a significant component of my research. We are only just scraping the surface of how these systems can be utilized in geography/Earth science research. Specifically, I am working on using multispectral data for quantification and analysis of real environmental variables, as opposed to simple image analysis. We will continue to use UAS to understand the glacier hydrologic system of the tropical Andes and, hopefully, expand our use to other mountain regions facing similar issues.