Hot Research Cold Climate

GNSS plays a central research role in surveying one of the planet’s major icefields.
By Kris Carber

Spending the summer on an icefield may not be the ideal vacation for most people, but for researchers from around the world, it’s the opportunity of a lifetime. Each year in July and August, 40 to 50 professors, scientists, and students travel to Alaska for the Juneau Icefield Research Program (JIRP), the longest continuous research study of a glacial system in the Western Hemisphere. Located near Juneau in southeast Alaska, the icefield—the fifth largest in the Western Hemisphere—covers approximately 4,400 square km (1,700 square miles) and feeds 35 large glaciers, including the Mendenhall and Taku glaciers.

JIRP and its parent organization, the Foundation for Glacier and Environmental Research (FGER) based in Seattle, Washington, grew out of research first conducted on the Juneau Icefield in 1946 by Dr. Maynard Miller. The mission then, as now, is to understand the dynamics of glacier systems as they relate to climate change. Central to this understanding is the measurement and monitoring of glacier velocities and surface elevation changes using high-precision GPS/GNSS. Tracking these changes through time allows scientists to gain key insights into climate change.

In 1981, the University of German Armed Forces, the Technical University of Munich, and the Bavarian Office for Surveying and Geoinformation started collaborating with JIRP by sending their own geodesists and equipment to provide support for the many interrelated disciplines represented on the icefield: geology, glaciology, geophysics, botany, hydrology, meteorology, remote sensing, cartography, and geoinformatics. Last summer, both new and veteran researchers participated in the program.  Although JIRP has used GPS since 1992, the 2010 field season saw the first use of GNSS gear on the Juneau Icefield.

While the Juneau Icefield has been studied in detail since 1946, the ways in which it has been measured have constantly changed and improved. For the first 40 years, measuring positions on the ice relied on optical observations using theodolites. The emergence of total stations in the 1980s improved results significantly, but the optical methods were still limited by instrument range, extreme refraction over the glacier surface, and the frequently poor weather conditions. This changed in 1992 when researchers began using GPS receivers. Since then, GPS and GNSS have become the enabling technology for increasingly thorough and accurate scientific research in the quest to understand the complex dynamics of the Juneau Icefield. Such understanding requires fundamental information about the mass balance (the difference between the amount of snow and ice accumulating on the glaciers and the amount lost to melting, evaporation, and sublimation). Other key information includes depth and temperature of the ice, meteorological conditions, surface velocity vectors, strain rates, surface gradients, and changes in the surface elevation.

Ronny Wenzel, a graduate engineer from the University of the Bundeswehr in Germany and a JIRP researcher, says, “GPS/GNSS enables the surveyors to capture data over the entire icefield rather than at just select regions. This provides researchers with a much more comprehensive view of the icefield and its behavior.”

Challenging Environment

While Juneau offers easy access to supplies and air transit, the research takes place in remote alpine sub-polar areas accessible only by helicopter, hiking, or skiing. After arriving in Juneau last summer, the JIRP survey team faced an eight-day weather delay, which prevented delivery of essential GNSS equipment to the main base.

Once on the icefield, travel is by snowmobile, skis, or on foot. Survey teams can cover up to 80 km (50 mi) a day, with temperatures ranging from -10º C to +15º C (14º to 59º F) and conditions as diverse as sunshine, rain, fog, and snow. Wind speeds in excess of 100 km/hr (62 mph) are not unusual. Accommodations are in FGER-installed buildings firmly anchored to bedrock outcrops. These camps typically feature generators for electricity, kitchens, library, sleeping quarters, and classrooms. Occasionally, research teams sleep in tents when their work requires overnight stays on the glacier.

Conducting the Surveys

The use of GPS/GNSS equipment was essential to the survey team’s success. Used extensively were Trimble 5700 GPS and Trimble R7 and R7 GNSS Receivers, with Trimble TSC1 and TSC2 Controllers running Trimble Survey Controller Software for the static and RTK measurements. In camp, the surveyors used Trimble Geomatics Office and Bernese GPS software to process and analyze the GPS/GNSS data. Microsoft Excel, ESRI ArcGIS, and MATLAB software were used for analysis of movement and changes in the glacier.

During July and August of 2010, JIRP participants worked together as a multi-disciplinary team, with surveyors playing a key role in the overall research efforts. Surveyors concentrated on geodetic work as well as supporting the other disciplines by providing GPS positioning information. However, the bulk of the surveyor’s activity was focused on collecting data from longitudinal and transverse profiles on the glaciers to determine the icefield’s annual height changes.

On a typical day, the surveyors would leave camp around 9:30 a.m. to set up the base station over an established benchmark. This base station would then support RTK surveys for two rover teams, each of which would survey a nearby profile by navigating to within 50 cm of the profile’s stakeout points that were uploaded to the rover system the night before. The navigated-to position was then recorded on the rover.

Glacier velocity surveys required measuring the surveyed point a second time, so when the point was measured the first time, a bamboo wand was stuck in the snow at the exact measured point. This allowed the measured point to be re-occupied and measured a second time some 5 to 10 days later. Each rover team could typically measure 40 to 60 points over the course of 6 to 10 hours before returning to camp between 6:00 and 8:00 p.m. They then used the evening hours to recharge batteries using the camp generator (which was also needed for evening lectures), download the day’s data, and upload profile stakeout points for the next day’s work. “Lights out” was generally at 11:00 p.m.

As with all surveying and research projects, accuracy is important. “We tried to achieve the typical RTK accuracies of 2 to 3 cm (0.07 to 0.1 ft),” Wenzel notes. “Even in areas with poor radio communication due to topographic obstacles such as icefalls or mountains, our precision was always better than 5 cm (0.2 ft) horizontally. For the vertical component, the RTK accuracy of several centimeters is normal, especially on the variable icefield surface. The surface is rough with so-called sun cups that can be half a meter deep or more, so when we stake out the profile points, we make sure to not be in a sun cup. So the point might be up to 0.5 m (1.6 ft) from the calculated position. For the second measurement, we use the same spot where the stake was placed during the Epoch 0 survey.”

In July and August of 2010, the surveyors covered 2,620 km (1628 mi) by snowmobile to take measurements at 1,030 separate points. The teams surveyed 380 km (236 mi) of profiles in 26 days, collecting points at intervals of 500 m (1640 ft) on longitudinal profiles. The intervals between points on transverse profiles ranged from 40 to 400 m (130-1312 ft). An additional 4,650 points were collected in continuous topo mode to contour supra-glacial streams (meltwater streams on the surface of the ice). The points surveyed in 2010 were added to the existing database of surveyed points, which comprises GPS/GNSS surveys over a 19-year period from 1992 to 2010. The database now contains approximately 5,500 discrete stakeout points and around 32,000 surveyed coordinates.

Analyzing the Data

The Life and Death of Glaciers

Filmmaker Chip Duncan and his assistant, University of Alaska Southeast webmaster John French (Frenchy), made a surprise visit to the JIRP 2010 summer program. The two have been working on a PBS documentary titled The Life and Death of Glaciers and spent three days with the JIRP researchers, filming the digging of a test pit and the GNSS measurements on that profile; they also interviewed students and scientists, including one professor whose first JIRP expedition was in 1969. The PBS special is scheduled to air in 2011.
Collecting the data was just the first step. It was then necessary to analyze it to understand the dynamics of the icefield. This analysis typically focuses on several areas: glacier velocity and direction, surface height change, surface strain rates, isostatic rebound of the Earth’s crust, and movement due to plate tectonics.

Determining glacier velocity and direction is accomplished by performing two surveys at each of the points of a profile. The observed coordinates of the Epoch 0 and Epoch 1 surveys are then compared with each other, and simple trigonometry determines the distance and direction that the glacier moved.

Scott McGee, a cartographer and one of the principal surveyors with JIRP for the past 23 years, has been tracking glacier velocities on the Juneau Icefield since 1988 and reports that, “velocities on the icefield have remained steady throughout the years. Velocities vary from profile to profile on the icefield, but it’s amazing how remarkably consistent the velocities at each profile have been from year to year, and indeed for the past several decades. For example, the maximum daily surface velocity at Profile 4 in the middle of the Taku Glacier is always in the range of 60 to 62 cm/day.” McGee goes on to state that “today’s GPS/GNSS surveys corroborate surface velocities obtained via theodolite and electronic distance measuring (EDM) surveys done in the 1950s to the 1980s.”

“In order to have an exact annual surface elevation change, the heights of a given profile point have to be interpolated to a reference day, which means we usually have to reduce and add the impacts of slope and daily ablation data from surveys and mass balance work,” says Wenzel. Monitoring this aspect of its behavior provides valuable information relating to both short-term weather patterns and long-term climate change. McGee reports that “our GPS surveys of Profile 4 in the middle of the Taku Glacier over the past 19 years reveal a 3 to 5 year periodicity of alternating surface elevation increase and decrease, while the undeniable trend over the entire 19-year time period reveals that the Taku Glacier at Profile 4 is now about 3.5 meters thinner than it was in 1992. And the cyclic patterns of surface-elevation increase and decrease appear to correlate fairly well with the eastern Pacific Ocean’s El Niño and La Niña episodes.”

While the JIRP survey data typically is focused on the glaciers themselves, other data analyses can reveal changes in the Earth’s crust. As a glacier melts, its mass decreases and the Earth’s crust can rise, resulting in post-glacial crustal uplift. For this analysis, the multi-day static data from the base stations is evaluated. As part of this work, the base station’s coordinates are transformed into the International Terrestrial Reference Frame. By tying the data to permanent stations provided by International GNSS Service, researchers can compile a time series of coordinates and determine both horizontal and vertical displacements of the bedrock upon which the benchmarks are located. Surveys thus far reveal an uplift rate of 0.9 to 1.1 cm per year.

Combining GNSS with Terra-SAR-X

A key part of the summer’s work involved the German high-resolution radar satellite TerraSAR-X (provided courtesy of Astrium GEO-Information Services - Infoterra GmbH), which in combination with its “twin” TanDEM-X is collecting data to create a global DEM of unprecedented accuracy, quality, and coverage. This unique dataset will feature a relative vertical accuracy of 2 m (absolute accuracy 10 m) within a 12 x 12 m horizontal “pixel” and will be available for the entire land surface of the Earth from 2014 onwards.

Initially used in 2009, TerraSAR-X generated a radargrammetry DEM for the research areas in the icefield; the DEM has been validated with the 2009 GPS results determined by JIRP. For further investigation and development of SAR applications in glacial regions, other ground-truth measurements using satellite corner reflectors were conducted in 2010.

The two 28 kg (60 lb each) devices that Wenzel brought to the icefield were designed to reflect the radar beam back to the satellite. The results were used for backscatter analysis in areas with permanent snow and ice coverage, for change detection research, and for validation of ground-control point extraction. Wenzel set the corner reflectors up in different locations and at different times to capture a number of satellite overhead passes. He used GNSS to measure the elevation of the corner reflector’s centerpoint with centimeter precision. As the TerraSAR-X satellite passed overhead it recorded the 3D position of the corner reflector from space; the position was later compared with the ground measurements taken with GNSS receivers.
The use of corner reflectors for glacial research in combination with data acquired by TerraSAR-X was a JIRP first and an important success. The RTK position of the corner reflectors provided a high-accuracy link between the data provided by the radar satellite and ground-based observations. Radar satellite imagery can provide a valuable enhancement to monitoring and detecting changes in non-accessible icefield regions.

The Benefits of GPS/GNSS

“Since JIRP introduced GPS to its program, we have been able to gather additional and better information,” Wenzel says. “With GPS we don’t have to wait for clear visibility. Since GPS/GNSS is not affected by atmospheric refraction over the glacier surface, the data are much more accurate than what could be obtained using theodolites and total stations. By using continuous mode for collecting points, we can essentially scan the surface and create accurate grids from glacial structures. One thing that has not changed, though, is that it’s still all about X, Y, Z, and time, the four parameters that provide the most information about the behavior of the glaciers.”

The efficiency of the GPS/GNSS equipment allowed the teams to overcome their initial eight-day weather delay in 2010 (25% of available measuring time) and expand the surveys to include 50 km (31 mi) of new profiles. The expanded areas included the upper accumulation zones of Taku Glacier (the icefield’s largest glacier) and the Llewellyn Glacier in British Columbia. While the Taku Glacier is one of the few advancing glaciers in Alaska, the Llewellyn Glacier in the northeast sector of the icefield shows an alarming loss of mass; the Llewellyn has contracted by about one third (approximately 10 square km or 3.9 square mi) between 1984 and 2010.

“There is still much research to be done,” Wenzel says. “The glaciers on the drier Canadian side of the icefield show a larger magnitude of thinning than the Taku Glacier.” Researchers want to concentrate their efforts there, combining historic data with the newer GPS/GNSS data, while also incorporating sensors such as a fully automated meteorological station and electronic stream gauges.

Scott McGee, who began his research on the Juneau Icefield in 1988, vividly remembers just how little survey work was actually accomplished back in the days of theodolites and EDMs. “It was normal for us to ski three or four hours in the fog to a location, set up the theodolite over a benchmark, and sit back and wait, sometimes all day, for the fog to lift so that we could do our surveys. A good portion of the time we’d just sit in the fog all day and ski back to camp in the fog, without collecting any data at all.” Now, with the extensive use of GPS/GNSS on the icefield, McGee says, “We can go out every day, in any weather, including total whiteout, and know that we’ll return to camp at the end of the day with tons of new data to add to our ever-growing survey database. And that’s pretty cool.”

For more information on JIRP visit www.juneauicefield.com and www.crevassezone.org.

Sinning Vermessungsgeräte GmbH and Trimble Germany generously supported the 2010 JIRP field season by providing Trimble GNSS equipment.
Kris Carber is a writer living in northern California.

» Back to our March 2011 Issue

Website design and hosting provided by 270net Technologies in Frederick, Maryland.