PERFORMANCE MONITORING OF A NEW INSTALLATION AT APIA OBSERVATORY A.Jackson, A.Beraducci, T.Tafua

Abstract Apia observatory is one of the longest running magnetic monitoring stations in the world and is strategically important because of its remote location in the Pacific. In 2011 it was upgraded to modern instrumentation, having a Gemsys GSM-90F1 Overhauser magnetometer and DMI FGM-FGE triaxial suspended fluxgate variometer installed. Since this upgrade we have monitored the so- called ``closing errors'', the difference between the total field measured by the variometer and that measured by the Overhauser. Two effects are seen: one is a diurnal signal amounting to 1nT peak-to-peak per day, and the second is a long-term drift over several months amounting to some 3nT/month. The distance between the instruments is approximately 4 metres. We recently began some tests with an additional GSM-19 total field magnetometer to try to isolate the origin of these signals. We have additionally tested the effect of raising the temperature in our fibreglass variometer enclosure. In our report we will compare these signals to those seen on Ascension Island when magnetometers are separated by some tens of metres.

Presentation Outline Overview of Apia Observatory 2011 upgrade Problems affecting current operations Comparison with other observatories Conclusion

Overview of Apia Observatory The Apia Geomagnetic Observatory in Samoa is one of the oldest observatories in the South Pacific. The first geomagnetic measurements to be made there were in 1902 and thus making the Magnetic Observatory in Apia one of the longest periods of recordings in the world (>100 yrs). It provides the only geomagnetic data for this region of the earth therefore it is very significant.

Location of Apia Observatory (API)

2011 Upgrade It was upgraded to modern instrumentation, having a Gemsys GSM-90F1 Overhauser magnetometer and DMI FGM-FGE triaxial suspended fluxgate variometer installed. This upgrade was made possible through the endless support from Andy Jackson of ETH Zurich,GNS New Zealand, USGS and Alan Berarducci of Compass Rose Surveying.

Problems affecting current operations Problems due to Meteorological conditions Problems due to artificial interference Unexplained Total Field (F) drifts Changes to Total Field Site difference

Significant Events in Apia 2014 Operations February first visit by Prof A. Jackson to Samoa to check on observatory operations and organise remedial works. March Daniel Whittaker GNS visited to repair faulty fluxgate thermal control. May 2014 visit by Alan Berarducci to : - organise and oversee repair of Absolute Hut - Absolute observations training for local staff - re-measure Total Field site difference between the continuous proton and the D/I observing position - check the continuous proton stability - install drying agent in the DMI fluxgate sensor - check the DMI fluxgate sensor temperature coefficient

Problems due to Meteorological Conditions Following a tropical cyclone in 12 th December 2012 several large palm trees fell on the Absolute Hut damaging the roof. Following this rain shorted out all the hut electronics preventing any absolute obs being made. To prevent a repeat of these problems all trees in the vicinity of the Absolute Hut have now been cut down to avoid future damage.

Problems due to artificial interference During 2013 comparisons between the Total Field (F) measured by the proton magnetometer and F calculated from the fluxgate H and Z components were made. D fg H fg Z fg F fg F ptn Fptn – Ffg These comparisons indicated a step of 1nT occurring between 10:00 and 16:00 every day. November 2013

Problems due to artificial interference This “step” problem continued up to 31 December 2013 and from this date to the present it has not recurred. Locally investigations were carried out to find possible reasons for this interference, the local radio station transmissions, power transformers adjacent to the observatory site, were all investigated but no definite cause was ever traced. January 2014

Unexplained Total Field (F) drifts While carrying out the investigations into the daily “step” interference it was noticed that there was a regular change in the F proton - F fluxgate (Closing Error) difference. Changes in the Closing Error of approximately 2nT/month and the sudden change in the direction of the difference on 7 January 2014 are very difficult to explain. It was thought the drift was possibly due to a faulty proton or instability in the fluxgate analog to digital converter – both these possible sources of error were checked and both instruments were found to be operating correctly

Proton Stability Tests To check the stability of the observatory proton a second proton was set up and both instruments were operated continuously for several days. A plot of the difference between the external proton and the observatory proton shows a constant difference of approximately 40 nT over a 24 hour period.

Fluxgate Stability Tests A comparison between F calculated from the fluxgate data and the two protons was also carried out. In both cases the difference between F measured by the two protons and F measured by the fluxgate changed by 0.2 nT over the 24 hour period. This result agrees with the long term comparisons and indicates that the drift problem is associated with the fluxgate not the proton. Note: The magnitude of the difference in the lower plot is greater than in the upper plot, this is due to ground noise as the external proton was deployed approx. 30 cm above the ground

Unexplained Total Field (F) drifts In May 2014 silica gel was placed inside the DMI sensor dome. As can be seen from the graph below the addition of this drying agent seems to have had a significant effect on the drifts between fluxgate F and proton F.

Total Field Site differences between the recording proton and the D/I observing position The following slide shows the differences and the changes in the Total Field correction between the recording proton and the D/I observing position. For comparison purposes the differences and changes are shown for : Apia Gan Ascension Island The large annual change of 4.7 nT measured at Apia was derived from preliminary measurements and this value is under review.

Site Differences between Recording Proton and D/I position ApiaGanAscension nT Change /year

Conclusion Everything at Apia seems to be working well: All trees potentially causing damage to buildings - removed No daily 1 nT steps No long term drifts

Future of Apia Observatory -Improve capacity of local staff to analyze and apply magnetic data to other geological and geophysical activities such as climate change modelling. - Exposure to international observatories, trainings and work attachments to further the capacity on this particular area of science.