Data Quality Control / Data Availability and Statistics
Geophysical Network station equipment is constantly monitored to mitigate break downs or outages and reduce the impact to our clients.
For tsunami warning purposes, where real-time data is critical, redundancies are built into the system. Alternative communication providers and back-up off-site processors, to reduce the impact of outages.
To monitor data availability and quality Geoscience Australia staff can analyse the real-time waveforms to identify abnormal signals. One of these tests is called a power density function plot which characterises the noise on a particular channel at a seismic station. This noise is made up of:
- internal instrument noise
- cultural background noise such as mining, fences etc
- natural background noise such as microseims and ocean waves
- any problems with the instrument such as reboots or calibration pulses.
Other state of health information includes:
- number and length of data gaps
- short-term average/long-term average (STA/LTA) data counts to find any sudden changes in instrument sensitivity
- changes in background noise or small local seismic events
- timing issues when GPS equipment fails
- and a range of other tests.
If a station is identified as having data quality issues, it is removed from the production environment and moved into a testing area. This reduces the likelihood of false seismic events being identified.
These data quality tools help us to identify possible problems with the equipment before they completely fail. This is extremely useful when technicians are already in the vicinity as it can take days and even months to ship replacement equipment to some of our sites.
Until recently, calibration of instruments was performed at site after installation or during configuration changes. Presently, theoretical calibration values are used for each sensor type. The calibration values, response files, site location and name (metadata) for each piece of equipment are bundled together into a data-less seed file. This method ensures accuracy in the metadata from each station and can be easily incorporated into data management centres around the world.
Statistics are provided at quality management meetings, held every quarter, These include number of data downloads from the IRIS Data Management Centre internal network availability statistics, Nuclear Monitoring Station values, any new stations or improvements to the network.