Mar 25 2011 - Locating earthquakes - being faster

Locating an earthquake involves a lot of data and while many of the processes are automated, the final review and location still involves a person having to review the data and location before it becomes available via the website and on social media. The fastest we can typically locate an earthquake is around ten minutes but it's usually more like fifteen to twenty minutes. During the night it can be a real challenge – a pager beep wakes the Duty Officer from a dead sleep, they get up and boot a computer, login, bring up the seismograms, review, locate and finally post the event. Any small problem with the internet or difficulty locating the event can slow the process further. The effort required in the process also places a limitation on how many earthquakes the Duty Officer can locate – there's no way we can get to every single felt earthquake in a timely fashion.

It's been a long held goal in the project to make earthquake locations available faster. Now that the network roll out is nearing completion we're starting to have enough data to try some improved approaches to earthquake location that will be faster and more accurate.

Earthquake location is a fairly specialized field of software development and existing solutions tend to be heavily customised to any one country's own network operations. Since the Indian Ocean tsunami and a greater global emphasis on earthquake detection and tsunami early warning, this situation has been changing with an evolving community of software that makes adopting new techniques easier.

We are investigating replacing our current earthquake location system with SeisComP3 which was developed within the GITEWS ( German Indonesian Tsunami Early Warning System) project by GeoForschungsZentrum Potsdam and is now maintained by Gempa. We're testing SeisComP3 against our current system to see that it does at least as good a job of locating earthquakes and will then move on to see how it and the downstream processes can be tuned to make an earthquake location available faster. As part of this we are also investigating the possibility of making automatic, un-reviewed, earthquake locations publicly available with review being done later by the Duty Officer. This is a big shift in thinking – New Zealand is a long thin country with complicated geology and a lot of seismic noise from the ocean and wind which greatly increases the chances of automatic earthquake locations being incorrect. However, through our testing we're seeing that the better network coverage and greatly improved software techniques are showing a significant improvement in automatic location accuracy. This raises the possibility of a preliminary earthquake location being available in not much more than the length of time it takes us to receive sufficient data.

Once the testing phase is complete and successful, we will move to a long term test implementation of SeisComP3 which the Duty Officers can use to gain experience with the new system. We can then look at re-engineering how we deliver information to the website to take full advantage of the faster information flow (no more tweets beating web site updates) and add a further range of OGC compliant web services.

It's a lot of work and there are some daunting challenges to keep performing our core tasks while completely changing the way we do it but we will be as happy as anyone to stop shouting "hurry up GeoNet!".