Introduction

In 1973 we began development of a radio echo-sounder for use in northern Canada. The need to operate on small glaciers dictated the sounder’s properties. We chose 840 MHz in the ultra-high-frequency (UHF) band as the centre frequency for several reasons. Although dielectric losses and random scattering increase with frequency, a high-gain antenna can offset much of the loss: the small size of a UHF antenna simplifies airborne operation. Most Canadian polar ice masses are expected to be within the maximum range of a UHF sounder.

In August 1976 we tested the sounder on three glaciers in the Steele Creek basin, St Elias Mountains, Yukon Territory, Canada. Extensive glaciological work had previously been done on two of these glaciers, Rusty and Trapridge Glaciers, so the thermal regime and ice depth were known at a number of sites; a network of survey markers provided ground control points. To get similar information about Hazard Glacier we set up a field camp and carried out a levelling survey, ground-based experiments with the radar, hot-water drilling, and ice temperature measurements. Later in August, using a Bell 206 helicopter, we did the aerial surveys of these glaciers. Our results from Rusty, Trapridge, and Hazard Glaciers, which we present here, confirm that aerial UHF radio echo-sounders can function well on smaller cold ice masses, and that the technique can provide complete ice-thickness data with relative ease.

Background

Development of UHF (300 MHz-3 GHz) radio echo-sounders has lagged behind that of sounders operating at the lower VH frequencies (30-300 MHz). Greater dielectric losses and higher susceptibility to scattering have limited the usefulness of UHF sounders on large polar ice sheets, the first regions of extensive radio echo-sounding, Reference WaiteWaite (1966), using an SCR-718440 MHz aircraft altimeter in north-west Greenland, reported loss of signal at depths exceeding 400 m. VHF sounders have no difficulty sounding ice many times this thickness. On small and medium-sized valley glaciers, valley-wall reflections and aircraft manoeuvrability, rather than ice penetration depth, are the main considerations; thus the small size and directivity of UHF antennae become attractive.

In 1970 on Reference DavisRoslin Gletscher, Davis and others (1973) used a UHF sounder to measure ice thicknesses in a cold valley glacier. Man-hauling a modified SCR 718 radio altimeter they were able to measure ice thicknesses up to 400 m; all previous soundings using the SCR 718 had been made on polar ice caps. They reported that scattered returns at many locations often made identification of a distinct bottom echo difficult (Reference Smith and EvansSmith and Evans, 1972).

The first UHF radar designed strictly as an ice-depth sounder was a 620 MHz instrument built for the Department of the Environment of Canada (DOEC) by Reference GoodmanGoodman (1975). Results of man-hauled surveys of Rusty and Trapridge Glaciers demonstrated that the DOEC system was capable of finer resolution than VHF systems, and that valley-wall echoes and rapidly varying bedrock topography did not seriously affect the sounder’s performance, As with the SCR 718, internal scatterers could have made the identification of a distinct bottom echo ambiguous.

The University of British Columbia (UBC) radio echo-sounder, which is similar in concept to the DOEC system, has been specifically designed to sound the smaller polar glaciers and ice caps found in northern Canada (Narod, unpublished). Its centre frequency is 840 MHz, bandwidth 40 MHz, and system performance is 124 dB. The 35% increase in frequency over that of the DOEC system brings increased susceptibility to internal scattering and higher dielectric losses, but. these drawbacks are offset by the high gain and small size of the antenna and the ease with which the entire system is made airborne. Surface echograms demonstrate that scattering returns vary greatly with minute changes in antenna position (Reference BaileyBailey and others, 1964). By moving the antenna while continuously recording, the relatively constant bottom echo can be identified among the highly variable scattering returns. When airborne, the scattering returns vary too rapidly to be visible on an oscilloscope phosphor.

Instrumentation

The UBC echo-sounder comprises a transmitter, receiver, circulator, antenna, and oscilloscope display. Table I lists relevant system parameters; complete system details will appear in a forthcoming doctoral thesis. The system performance of the radar transceiver (transmitter, receiver, and circulator) is 124 dB. Adding the two-way antenna gain of 31 dB, the effective system performance is 155 dB. Both the transmitter and receiver are internally protected against abnormal power surges caused by cable failures. A 120 MHz signal from a crystal oscillator is multiplied to 720 MHz and 840 MHz to provide the local oscillator and carrier frequencies respectively. The carrier is then amplified, gated, and isolated at the 4 kW level. The circulator is a three-port ferrite device which allows the system to operate with a single antenna. The receiver input is protected by a solid-state limiter. The signal is converted to the 120 MHz intermediate frequency and filtered. The I.F. amplifier-detector has a logarithmic response characteristic with a useful dynamic range of 80 dB.

The antenna is a 90˚ corner reflector with two driven dipole elements. Its measured forward gain is 15.5±1.0 dB. The estimated E-plane and H-plane half-power beam-widths are 18° and 44° respectively. The antenna is compact enough to be attached to the cargo hook of any helicopter with high skid gear. On our August 1976 flights we outrigged the antenna from the helicopter in an experimental installation.

The video output from the receiver modulates the phosphor intensity of a Tektronix model 475 oscilloscope. A slow ramp on the vertical input scans the trace through the full screen height in one minute. The operator manually controls the vertical scan and a Polaroid-backed oscilloscope camera records the data. This results in approximately 20% of lost coverage by dead time. (In future, we plan to replace the photographic recording system by a magnetic recorder and graphic recorder.)

Survey procedure

As well as the arrival time of the bed echo, useful sounding data require knowledge of aircraft position and ground clearance. The UBG sounder, when used as its own radio altimeter, receives a strong echo from the ice surface. It is necessary, however, to fly at least 40 m above the ice surface in order to separate the transmitted pulse from the first received pulse. The small dynamic range of the phosphor and film cause this limitation. An improved recording system could reduce the minimum required elevation to less than 25 m above the ice surface. Alternatively, flying as close to the surface as possible, the pilot can use visual references (survey markers, etc.) to control his elevation and maintain a constant clearance above the ice surface. The advantage of the latter procedure is a greater signal-to-noise ratio (Reference SwithinbankSwithinbank, 1968). Its disadvantages are several: reduced air speed, reduced accuracy (the pilot cannot maintain a perfectly constant height above the surface; we have estimated our low-level height at 6±2 m) and poor flight-line photography–not to mention discomfort due to air sickness and increased risk of crashing. Data for Rusty Glacier were collected from high elevation (Figs 1 and 2) ; data from Trapridge Glacier were collected from both high and low elevations; data from Hazard Glacier were collected only at low elevation.

Fig. 1.

The flight path of the profile reproduced in Figure 2. The region shown is at lat. 61° 15' N., long. 140° 20' W.

Fig. 2.

Cross-section of Rusty Glacier showing photographically recorded data. The small circles at the ice surface represent the centre-line survey stakes. The larger circles at the bedrock boundary indicate the ice depths measured with the DOEC system (Reference Clarke and GoodmanClarke and Goodman, 1975). Vertical exaggeration is 4: 1. The data records show clear separation of the transmitted pulse, ice-surface reflection, and ice-bottom reflection, even over ice as thin as 30 m. The dark vertical bars are synchronizing marks with the 35 mm camera or with verbal cues recorded on the flight recorder. The vertical scale is 500 ns division⁻¹. Data for the cross-section were compiled from two passes of identical plan and overlapping proflle (one not shown).

For flight-line records we used a combination of dead reckoning and aerial photography. A tape recording of the aircraft intercom was verbally synchronized with the data collection. Navigation information included compass headings, airspeeds, and visual references. Aerial photography with a 210°-coverage fish-eye lens was also synchronized with the data collection. Our intention was to take photographs every 10 s, which, when compared to large-scale photographs, would provide locations and headings (Fig. 3), Unfortunately photographic flight control was incomplete due to erratic behaviour of the motor-driven Nikon camera caused by aircraft vibrations (the camera has now been replaced by a 70 mm Vinten aerial reconnaissance camera). We have estimated our plan accuracy at better than 50 m for accepted data. Complete survey details will appear in a forthcoming doctoral thesis.

Fig. 3.

An aerial fish-eye photograph taken over Hazard Glacier. These photographs provide flight-line control. The surface drainage feature seen in this photograph is visible on a government aerial photograph taken at 15000 m elevation. The location of this photograph is marked by an asterisk on Figure 5.

Fig. 4.

Trapridge Glacier ice-thickness map. The contours are computer generated by interpolating from the airborne data onto an arbitrary square grid The interpolation uses an inverse-distance-squared weighted average of the nearest datum in each of eight equal sectors. In regions of low density coverage and particularly near the south-west headwall, the map reflects the arbitrary nature of the interpolation scheme.

Results

The profile of Rusty Glacier presented in Figures 1 and 2, except for crossing the terminus of Backe Glacier at the lower end, follows the centre fine of survey markers down the glacier. The profile is in excellent agreement with an earlier survey made by man-hauling the DOEC system (Reference Clarke and GoodmanClarke and Goodman, 1975). For ready comparison with this work we have taken 176 ms⁻¹ as the velocity of electromagnetic waves in ice. At the ten locations where direct comparison is possible, seven of the measured ice depths agree within the resolution of the system. The remaining three discrepancies, all less than 10 m, have been attributed to a deviation of the flight path near the headwall, or to difficulties of interpretation of ground survey data. It is also possible that the airborne system could yield data implying shallower ice, since it illuminates a larger area of the bedrock.

The Rusty Glacier profile demonstrates several features of the echo-sounder. Resolution (about 4 m or one-tenth of a graticule division) is currently limited by the photographic recording scheme. The averaging effect of the photographic film suppresses scattering returns since each grain is exposed by the accumulated output of over 1000 transmitter pulses.

Data for the ice-thickness map of Trapridge Glacier (Fig. 4) were compiled from three of the four flights made during the two days, resulting in more extensive coverage than on Rusty Glacier. Trapridge Glacier also has an array of survey markers useful during the flights as visual navigation aids. Again, where it is possible to compare the airborne ice-thickness measurements with those made from the surface with the DOEG system (Reference GoodmanGoodman and others, 1975), agreement is very good, within the resolution limit of the present airborne system.

Fig. 5.

Hazard Glacier ice-thickness map. The contour internal is 20 m. The major medial moraine is shown as a shaded area. The asterisk marks the location of Figure 3.

There are a number of features in the new ice-thickness map that were not indicated by the ground-survey data. A shallow rib extends from the northern edge, about half-way up the glacier, across the glacier towards an ice fall. This zone of shallow ice may split the glacier into two separate dynamic zones. The major rock ridge which splits the terminus extends under the ice about one-third of the way up the glacier, A basin of ice has been found between the two rock spurs near the centre of the hcadwall. The new features are consistent with the ground-survey data and are results of the more extensive coverage of our airborne programme.

The most serious limitation of the sounder is its inability to penetrate great thicknesses of temperate or debris-rich ice. We believe the ice-thickness map of Hazard Glacier (Fig. 5) owes much of its character to an internal moraine layer formed by the merging of the glacier’s two main tributaries. The map shows, along the northern edge of the glacier, a shallow ledge of relatively constant (60-80 m) ice depth. The boundary of the ledge is defined by a tortuous 80 m ice-depth contour. We do not believe that in this region we have actually measured total ice depth. More likely the radiation has been scattered by an internal moraine layer.

The evidence for an internal reflector is two-fold. Each transverse flight line yielded data showing an extremely steep rise from the “deeper” southern region of the glacier to the northern ledge (Fig. 6). These very strong boundary points do not define a smooth valley floor. Instead they cause the tortuous 80 m contour in Figure 5. We believe that the rises are too steep for a reasonable valley configuration and that the irregular contours result from the occasional penetration of the radiation through the internal layer. Secondly, the edge of the shallow ice depths coincides with the major surface medial moraine. The structure which has caused the formation of the medial moraine could also have caused the formation of an internal moraine layer which terminates in the vicinity of the medial moraine. Possibly ice from the northern arm has overridden ice which originated at the southern headwall.

Fig. 6.

A transverse section of Hazard Glacier. The cross-section is shown looking up-glacier, with the northern boundary to the right. The flight line shown on Figure 5 went from the northern boundary to the southern boundary, The steep rise coincides with the edge of the major medial moraine. The shallow ledge is probably not bedrock but a debris Layer, perhaps brought with overlain ice from the northern arm of the glacier.

Concluding remarks

Our field tests demonstrate that airborne UHF echo-sounders can be used to measure the ice thickness of small and medium-sized cold valley glaciers. The compact size of the antenna and instruments makes them easily deployable from even the smallest aircraft. Using a small helicopter, complete ice-thickness data can be collected quickly and efficiently. The deepest ice encountered was 200 m on Hazard Glacier, though this is not the depth limit of our sounder. During an airborne survey of the glacier lying in the north-west col of Mount Logan, Yukon Territory, we sounded 430 m of ice under conditions that were far from ideal (Clarke and Narod, unpublished). The major drawback of UHF echo-sounders is their inability to penetrate temperate or dirty ice.

Several improvements to the sounder are in progress. These include the organization of the radar into a single rack-mounted package for easier airborne deployment, increased pre-amplication for improved system performance, addition of a graphic display for monitoring and field replay, a magnetic recorder for digital or analogue data, and direct-recording navigation aids.