Introduction:

IMESS is the brain-child of Dr. Philip Kyle, and was set up in collaboration with Dr. Jurgen Kienle (university of Alaska), Dr. Katsutada Kaminuma (Institute of Polar Research, Tokyo) and the writer. The principal objectives and responsibilities are listed in Appendix 2B.

Results:

Earthquakes in the range 1 to 10 kJ in seismic energy (M = −2.8 to −0.5) were occurring in 1980/1 at roughly half the rate in 1974/5 and 1978/9. Above 10 kJ, where the events on Erebus usually accompany eruptions and exceed the number predicted by the linear regression line for smaller earthquakes, there were also fewer events in 1980/1 than previously. The clean white snow around the vents in the crater showed there were far fewer bombs ejected. Furthermore, the eruptions were usually emissions of gas accompanied by a loud roar, and without an explosive onset. After 27 December, some of the eruptions were explosive, and Kyle was able to collect a few fresh lava bombs from the crater rim. A list of eruptions is given in Table 1.

TABLE 1: Eruptions of Erebus between 20 December 1980 and 9 January 1981. page 23

Figure 5: Comparison of seismograms and microphone recordings for five events.

A distinction between gas emission and explosion seems to be apparent in the microphone recordings. Roaring gas emissions of short (c. 3s) duration, such as at 1443 NZST on 31 December, were recorded as a congressional air pulse of duration about one second (Fig. 5a) followed by a single oscillation of period 1/7 Hz which is probably the organ pipe mode of the crater. The accompanying earthquake was small (160 J) and preceded the air wave by about one second. Fig. 5b shows another roaring gas emission of about 18s duration which occurred at 0949 on 27 December. The air-pulse onset is similar, but the organ pipe oscillation is not obvious, presumably due to the long duration of the discharge. Explosions which eject bombs, such as at 1644 NZST on 31 December, have a strong oscillatory onset of frequency 2 Hz superimposed on the compressional air pulse, followed by the 1/7 Hz coda (Fig. 5d) but the coda is of shorter period (c. 4s). This event was not observed well enough to know if bombs were ejected at the time.

Earthquakes not accompanied by eruptions, such as the 2 kJ earthquake seen recording at 2102 NZST on 31 December 1980, were not recorded on the microphone (Fig. 5e), whereas in 1978/9 the larger earthquakes were recorded (VUWAE 23, 1978/79 Immediate Report). Possible reasons are obscuration by the high wind noise level in 1980/1, and differences in the microphone. In 1978/9 it was a dynamic microphone in a Helmholst resonator, giving a sharply peaked response, whereas now its response to waves of constant pressure amplitude rises proportional to frequency (6 dB/octave) up to the resonance frequency of about 80 Hz - then flattens off and falls at 6 dB per octave rise up to the frequency for which Q = 1. Above this the response falls at 12 dB per octave rise in frequency. Both microphones also act as ground accelerometers to some degree.

The results expected from the figure of eight induction loop were, (i) the detection of magnetic signals from eruptions of conducting magma in the static magnetic field of the earth, and (ii) the detection of spectral effects in the magnetic micropulsations (dominantly of solar origin) recorded separately from the two halves of the loop. These expectations are not realised in the recordings presently available, due in part to the high static discharge noise prior to 1 January, and a break in the Crater loop between 1 January (when someone ploughed through it) and 4 January when the break was located and repaired.

No signals were detected on the Camp loop which correlated with any of the eight eruptions which occurred when correlation was possible. This result was expected. Of the three eruptions which occurred when correlation with signals on the Crater loop was possible, there was only a doubtful correlation with the eruption at 1749 NZST on 1 January. There was also one possible correlation with an earthquake at 19.0 hours on 5 January.

A comparison between the micropulsation spectra on each loop was made from the recordings between 7 and 9 hours NZST on 5 January 1981, when the signal/noise ratio was excellent. The two spectra (Fig. 7), which were determined with a variable 1/3 octave bandpass filter, have almost the same shape. If anything, the level on the Crater loop falls relative to that on page 26

Figure 6: Frequency of occurrence of Erebus earthquakes of all types versus seismic energy. The frequency in 1980/81 was about half that in 1974/75 and 1978/79.

page 27 the Camp loop by about 3 dB per decade rise in frequency between 0.1 and 1 Hz, but this could be due to minor instrumental differences between the two channels. It appears that either the conducting magma has no significant effect on the spectra at 1 Hz and below, or it affects both loops equally. Scale model experiments suggest that the former is the case unless the magma column has radius much greater than 200m and conductivity less than 0.5 mho/m.

Further results must await the processing of the telemetry tape recordings in Japan, following purchase of the playback equipment, and the receipt of tape recordings made over the winter. At the time of writing (12 May 1981) the telemetry signals are still being recorded satisfactorily at Scott Base.

Figure 7: Spectra of magnetic micropulsations recorded on the Crater and Camp loops between 0700 and 0800 hours NZST on 5 January 1981. Difference in level between the loops at each frequency are given in relative units at the bottom of the figure.