19-30 November: Preparation for ascent of Erebus.

We installed our new digital seismograph, consisting of a Compaq 286 Deskpro computer with a 40 Mbyte hard disk, and 1.66 Mbyte of RAM, in the Science Lab beside the existing Sony and San-ei recorders. It is fitted with a Data Translation DT 2821 A/D board with 16 inputs sampled 100 times/s, and was programmed by Dr W.H.K. Lee of USGS, Menlo Park, CA. Although it displays the seismograms continuously on the computer screen, it only records when it is triggered by a large enough earthquake, or by serious electrical interference (such as the D-region Radar). These events are stored and later edited to remove false triggerings, and transferred to floppy disks. This data can be displayed on another computer one channel at a time, filtered, and the onset times picked for focal location using program Hypo71.PC. A full range of seismic programs for IBM PC computers has been provided by Dr W.H.K. Lee.

Static electricity was a serious hazard to the computer, and was possibly the reason that only Howard Nicholson's printer would work with our computer. Our printer was OK with his computer. To prevent further trouble, we earthed the computer table, and zapped ourselves on it before touching the computer. Misinterpretation of some teething problems with the program resulted in the hard disk becoming non-accessible, and we were grateful to Mr Bill Schmidt of the McMurdo Communications Message Centre for help in retrieving it.

30 November to 10 December: Work on Erebus.

Skilton, MacKay, and Dibble flew to Fang acclimatization camp on 30 Nov., but cloud prevented the helicopter returning with Grizzly G2, and it was left at Cape Royds until 2 December. Gentle 17 brought it up, but had trouble releasing the hook, and could not lift us to the hut. Attempts to drive up with 2 people on the toboggan (as laid down by the Operations Manager) were unsuccessful, but when S-081 arrived, we obtained permission for Bill McIntosh to drive up alone.

Continuing cloud deprived us of US Skidoos, and G2 was used to move both parties to the hut on 4 December. The TV camera, infrasonic microphones, and long period horizontal seismograph all needed servicing this season. The plastic coated heated window fitted to the camera box last season had become opaque, and was replaced by an acrylic window.

The camera had an intermittent fault, and had to be sent to Philips in Auckland for repair. Phil Kyle also had a TV system on the volcano to monitor crater activity at the Erebus Hut, but when he discovered his transmitter was unusable, we agreed to connect his NTSC colour camera to our transmitter, and it was recorded at Scott Base. We are sharing this data. Our camera was returned page 9 just in time for Dr Neilya Dunbar of S-081 to reinstall it on 30 December, before the final descent from Erebus.

New infrasonic pressure sensors of type SCX05 DN were installed at E1 and CON in place of the low sensitivity LX0503A types installed last season, and the long period horizontal seismometer was readjusted. Ice buildup in the rock recess around it appears to delevel it. 10-19 December: Re-activating the Windless Bight Infrasonic Array.

The array had been idle since August 1988 because the new tapes had been mislaid, and the off-line analysis computer had faults in the monitor and printer. The Array (Fig. 1) consists of 7 condenser type microphones arranged in two quadrilateral arrays - one large (RTG,ERE,TER,ROS) and one small (RTG.5,6,7), with one microphone common to each. Each microphone has an acoustic filter consisting of a 91 m long pipe with holes spaced along it, for the purpose of reducing wind noise. The pipes, RTG hut, and telemetry antenna are being progressively buried in snow, increasing the low pass filter effect, and threatening the RF signal level. Preamplifiers, VCO modulators and RF transmitters are powered by a radio-thermoelectric generator (RTG) in the hut at the common microphone, ensuring continuous operation. Initially of 40 W capacity, the RTG has decayed exponentially to c11 W, but the loading resistor intended to soak up the surplus power had not been removed.

At Scott Base, the analog signals are filtered to give bandwidths of 1 to 10 s for the small array, and 7 to 70 s for the large array, and digitised, 4 times per second for the small array and once a second for the large array, and after on-line processing in blocks of 128 s, are recorded on 800 BPI tapes lasting 5 days each. I was able to get the off-line computer to scan these tapes by exchanging components between two faulty unused consoles, but there was no serviceable serial input printer. This was my first success in reading the tapes since my 1984 visit to University of Alaska (attempts always stopped at the question "do you have the source code"), and so I requested and received permission to shift the off-line computer to Wellington.

The RTG hut was visited on 14 December by R. Dibble, K. MacKay, H. Nicholson, T. Exley, T Eason, M Van der Sluy, and Bill Schmit. The loading resistor was removed, restoring the 11.5 V supply to 12.0 V, and the RF outputs were measured, being within 65 to 250 mW on all 7 channels.

Currents to all microphones were 19-20 mA, and all were transmitting carrier tone. Apart from being untidy, and buried in snow, the hut and equipment were in excellent order. McMurdo had already checked the RTG itself.

The recorded tapes are being processed in two ways, using the programs provided by University of Alaska. SCAN and AZSCAN search the on-line analysis results recorded on the tapes, and print out times, amplitudes, spectral peak frequency, and arrival direction and velocity when the cross correlation coefficient exceeds a chosen level. These programs run very fast, but it is only the best correlated signal in any 128 second block which is returned. Brief eruption signals from Erebus are ignored unless very large. The other option is to reprocess the original digital waveforms, using BEMSCN, DATPLT, SPCTRM, and others. Of these, BEMSCN was specially modified for me by Mr B. Mckibben of the University of Alaska, to find the cross correlation coefficient of signals with Erebus azimuth and infrasonic velocity, and also the ratio of peak to RMS amplitude in each two minute block. Computer time was 9 s per block, Erebus is a frequent source of wind eddie signals, but only explosions give a high peak to RMS ratio.

Also in this period, a major effort was made to reduce the number of false triggers of the digital seismograph by the D- region radar. The frequent short duration high power radar pulses have a wide bandwidth, and the large horizontal loop antenna practically above the Science Lab induces strong signals in everything. Steps to reduce their effect included: (i) running the entire IMEEMS installation off one power point (because neighbouring points had separate long cables from the fuse box); (ii) running the computer from an isolating transformer and power filter; (iii) having a single common earth point at the power point for all chassis, racks, and table frames; (iv) coiling page 10 and binding all excess wire lengths; (v) shunting inputs and outputs with the lowest value resistors the circuits could tolerate. The result was satisfactory, but precarious. I object to such deliberate sources of interference, and to the stock reply that only badly engineered equipment is affected, especially as the problem is under my control for only 2 months a year.

19 December to 17 January 1989: Recording and analysis of data at Scott Base.

Prof Kaminuma and Kevin MacKay made chart records of all earthquakes well recorded on the Sony data-tape recorded, and printouts of the digitally recorded earthquakes. The latter were also archived onto floppy disk, at a normal rate of 5 per disk. This will allow 1500 earthquakes to be recorded on our 300 disks during the year. They also made continuous video tapes of Erebus TV, and recycled the tapes with no useful information, and made both PAL and NTSC copies of all the eruptions recorded for distribution to collaborators.