Introduction
The present report describes radio echo-sounding equipment for use on temperate glaciers. The equipment has been designed and built at the Raunvísindastofnun Háskólans (the Science Institute of the University of Iceland).
For the last 10–15 years radio echo-sounding with 30–60 MHz signal frequency has been successful on Arctic glaciers in Greenland and Antarctica (Reference BjörnssonBailey and others, 1964; Reference GudmandsenGudmandsen, 1969; Reference Evans and SmithEvans and Smith, 1969). But several attempts to use the same technique on temperate glaciers (Reference Goodman and GoodmanGoodman, 1970; Reference Smith and EvansSmith and Evans, 1972) proved to be unsuccessful until Reference Watts and EnglandWatts and others (1975) reported successful soundings using 5 MHz frequency. A description of the equipment has not been published. Reference Watts, Watts, England, Vickers and MeierWatts and others (1975) explained that the problem encountered in radio echo-sounding of temperate glaciers could be attributed to water-filled voids in the ice. Reference Watts and EnglandWatts and England (1976) analysed the total scattering cross-section of water-filled voids in the ice as a function of frequency. The results suggested that sounding at frequencies below 8 MHz would eliminate or greatly reduce the total return power due to scattering and it might be possible to penetrate thick temperate ice (1000 m) and still get a good ratio of bottom return to scattered return.
Following the success of Reference Watts, Watts, England, Vickers and MeierWatts and others (1975) a joint British–Icelandic expedition carried out radio echo-sounding experiments on Vatnajökull (Reference Björnsson, Björnsson, Ferrari, Miller and OwenBjörnsson and others, 1977). The experimental instrument was built at Cambridge University (Reference Ferrari, Ferrari, Miller and OwenFerrari and others, 1976). After the tests on Vatnajökull in 1976 the equipment was designed and built at the Science Institute. Further development during the last two years has resulted in the design of two devices—Mark I and Mark II. Mark I operates in the frequency band 2 to 5 MHz. The overall range is from 100 m to 1000 m, which is required to encompass the ice thicknesses expected on ice caps in Iceland. This equipment has been used for routine soundings on the ice caps Vatnajökull and Mýrdalsjökull. A preliminary map of the topography under Mýrdalsjökull and a profile across western Vatnajökull have been published (Reference BjörnssonBjörnsson, 1977). Further, the Grímsvötn and the Bárdarbunga areas in Vatnajӧkull have been mapped in details (paper in preparation by Reference Björnsson, Björnsson, Ferrari, Miller and OwenH. Björnsson). Mark II operates at 2 to 10 MHz and has a range from 30 m to 400 m. The device was designed for use on valley glaciers and outlets from the large ice caps, and has been tested on Tungnárjökull and the valley glacier Gljúfurárjӧkull, north Iceland, during the summer of 1978. Further, Mark II has been used for routine soundings of Storglaciären, Isfallsglaciären, and Rabotsglaciären in Swedish Lapland in March–April 1979.
Radio Echo-Sounding Equipment
The radio echo-sounding system consists of a transmitter, a receiver, and two antennae. In the present system, the transmitter and the receiver together with 12 V car accumulators are mounted on two sledges. Each sledge was placed at the centre of an antenna. The antennae were towed on a line behind a skidoo as seen in Figure 1 A bicycle wheel is mounted on the transmitter sledge and a built-in a.c. generator generates pulses which ensure that the repetition rate of the transmitter is proportional to the speed at which the sledge is running. The distance travelled along the glacier surface is obtained by integrating in the receiver the repetition rate of the transmitter. The altitude of the glacier surface can be recorded by a barometric altimeter. A barograph which records continuously at the base camp is used to calibrate the barometric altimeter. The receiver is connected to an oscilloscope with a camera for z-mode (intensity modulation) recording. Fig. 2 shows typical z-mode pictures. The glacier surface is seen in the upper half of the picture and the bottom return in the lower half. The X-axis shows the distance travelled along the glacier surface. The γ-axis shows the elevation of the glacier surface and the height of the underlying bedrock above sea-level. The white vertical lines on the z-mode picture are due to interference from short-wave radio stations.
Radio echo-sounding at Vatnajökull. A skidoo is towing the device. The receiver sledge with a hut for the operator is placed at the centre of the receiver antenna.
Z-mode film records. Continuous records along the line of echoes, (a) A record of Mark I from Grímsvötn, Vatnajökull; ∝ = 0.1 for the antenna. The ringing phenomenon visible in the record is due to the fact that the signal is band-limited. (b) A record of Mark II from Storglaciären, Sweden: x 1.0 for the antenna.
The A-scope (signal versus time) representation is used to monitor the output signal from the receiver. Figure 3 shows a typical A-mode picture.
A-mode film record from the valley glacier Gljújurarjokull, north Iceland. The transmitted pulse, internal reflections, and the pulse reflected from the bedrock can all be identified.
The main characteristics of the equipment are summarized in Table I
Characteristics of the Radio Echo-Sounding Equipment
The Antenna
A broad-band antenna is required to transmit a short pulse without distortion. Reference Wu and KingWu and King (1965) and Reference Shen and KingShen and King (1965) describe such an antenna. The internal impedance of a dipole antenna is given by
where z is the position along the antenna;
The constant ψ is found from
where
If the antenna is placed on the surface of a glacier, the effective permittivity is
The antenna is constructed by using lumped resistors to approximate the continuous impedance function. If one divides the antenna into n sections of equal length h/n, and a lumped resistor is inserted in each section, the value of the pth resistor can be found by integrating Equation (4) over the section
If one re-numbers resistors by writing
R q given by Equation (6) is the qth resistor counted From the end of the antenna towards the centre. R 0 is chosen as
In the present study one selected the constant α = 0.1 for the 2 to 5 MHz antenna. For the 2 to 10 MHz antenna h = 7 m and the constant α = 1. The same type of antenna is used for the receiver and the transmitter.
The feed point impedance of the antenna is given by Reference Wu and KingWu and King (1965) for α = 1 as
The impedance is a resistance R 0 in series with a capacitance C 0
Where
and
Inserting values for the present antennae in Equation (8) one gets
for the antenna of resonance frequency 3.8 MHz and
for the antenna of resonance frequency of 8. 1 MHz.
The feed-point impedance of the antenna with α = 0.1 was estimated on the basis of values for α = 1 and α = 0 (ordinary dipole), and the matching network was adjusted in the field.
In order to make the transmitter load impedance resistive, a compensating network is connected across the antenna terminals. The network consists of an inductor L 0 in series with a resistor
Antenna matching circuit.
The antenna wire and the resistors were put inside 16 mm flexible plastic tubes which are easy to drag across the surface of the glacier.
The Transmitter
The basic transmitter circuit is shown in Figure 5a. A capacitor C is charged to 1200 V by the high-voltage supply HV. When the thyristor SCR is fired, the voltage across the resistor R first rises from zero to 700 V in 0.1 μs and then decays exponentially. The output pulse is shown in Figure 5b. This circuit is capable of giving out pulses with a peak power around 10 kW at a repetition rate greater than 1 kHz.
(a) Basic transmitter circuit.(b) Voltage pulse across the antenna terminals for Mark I, ∝ — 0.1.
A block diagram of the transmitter is shown in Figure 6. The transmitter is powered by a 12 V accumulator. The HV converter converts the 12 V to 1200 V. The capacitor is charged to 1200 V in 500 μs. The thyristor is fired giving out a short pulse. 100 μs after the thyristor is fired, the trigger circuit starts the HV converter and in 500 µs the capacitor is charged to 1200 V. Then the trigger circuit is ready to receive the next pulse from the pulse-rate multiplier PRM. The PRM receives an input from an a.c. generator which is placed inside a bicycle wheel that is attached to the transmitting sledge. The output pulse rate from the a.c. generator is directly proportional to the angular velocity of the wheel, thus making the pulse rate of the transmitter a measure of the speed of the transmitter sledge. The PRM converts the pulse rate from the a.c. generator to 214 pulses per 100 m advance of the sledge. The pulse rate of the transmitter is integrated in the receiver. The integration controls the deflection of the γ-axis in the Z-scope and marks a distance scale along the γ-axis.
Block diagram of transmitter.
The antenna impedance is matched to the transmitter by a broad-band transmission line transformer and an antenna compensation circuit.
The power consumption of the receiver is proportional to the repetition rate, being 12 W at 1 kHz.
The Receiver
The receiver block diagram is shown in Figure 7 The signal from the antenna-matching unit goes through a 2 to 5 MHz band-pass filter or, alternatively, it is connected directly to the video amplifier input. The video signal is amplified 20 dB in each stage of the video amplifier before it goes to a 0.2 μs delay line. Leaving the delay line, the signal is further amplified 15 dB in the A and Z drivers. The overall amplification is, therefore, +75 dB and the band-width is 0.1 to 10 MHz. An attenuator can be inserted between the antenna matching unit and the amplifiers to reduce the overall gain when required. An output from the first stage in the amplifier goes to a triggering circuit which senses the start of the transmitter pulse. An output pulse from the triggering circuit unblanks the Z driver and thus enables the intensity of the oscilloscope beam to be controlled by the video signal. The trigger pulse also starts a sweep generator. The sweep generator gives out a rising voltage at the rate of 0.835 V/μs which controls the X-axis deflection of the oscilloscope. The result is a scale of 100 m ice depth per volt.
Block diagram of receiver.
The starting position of the X-axis beam is controlled by the voltage from a barometric altimeter. The surface elevation of the glacier is, therefore, plotted on the screen and the bottom echo appears at the true height above sea-level. Figure 2 shows typical Z-scope pictures.
Pulses from the triggering circuit are divided by 2N, N ranging from 1 to 7, and counted in a 12 bit binary counter. The 12 bit output number is converted to analogue voltage in a digital-analogue converter. The μ-axis deflection is controlled by the analogue voltage and, dependingon the value of N, the scale of the μ-axis can be varied from 25 m/V to 3200 m/V.
In the A-scope mode, the trigger pulse T goes to the external triggering input on the oscilloscope and the A input to the γ input of the scope. A 35 mm reflex camera is mounted on the scope. In the Z-mode the camera is held open while the beam is scanning the screen. After scanning the whole screen, which takes several minutes, the film is advanced one frame and the γ counter is re-set, enabling the next frame to be scanned.
The receiver is powered from a 12 V accumulator and a d.c./d.c. converter supplies the various circuits with appropriate voltages. The total power consumption is 6 W for the receiver and 72 W for the Tektronix Model 465 oscilloscope.
Experience
The device has shown good performance. Routine soundings are only limited by driving conditions and the visibility on the glacier. Sounding profiles of up to 50 km per day have been obtained on Vatnajökull. Navigation on the ice cap was done by LORAN-C and satellite navigation.
The maximum thickness measured so far is 800 m on Vatnajökull, but one presumes 1000 m can easily be sounded. Crevasses show up on the records but only in exceptional cases has the bedrock reflection been wiped out.
Operating time must be chosen when the sky-wave propagation from short-wave radio stations is at a minimum, that is during daytime in the summer. Ground-wave propagation is usually not a problem because glaciers are mostly situated in remote areas. Interference from medium-wave broadcast stations has not caused problems for soundings at the glacier surface.
Acknowledgements
The authors are indebted to R. L. Ferrari, K. J. Miller, and G. Owen from the Department of Engineering, University of Cambridge, for valuable collaboration during the first stage of radio echo-sounding in Iceland. Further, we are indebted to R. D. Watts, U.S. Geological Survey, Denver, Colorado, for information on his antenna design.
For the last three years the work has been supported by grants from the Science Fund of Iceland and by Eggert V. Briem.
