SeisComP in the new minor version 7.4 was released on 04 September 2026 focusing on a few important fixes and many key improvements. They include:
Consistent ML-type magnitudes when computed in IASPEI mode
Upgraded LOCSAT locator,
Enhanced synchronization of event parameters between SeisComP instances
connected by QuakeLink,
Generation of unique event IDs consistent across two SeisComP instances,
Improved graphical user interfaces: scmvx and scconfig are largely upgraded
for advanced interactions and scolv now allows the integration of felt reports
for events.
Many more valuable new features are available which we describe along with
instructions for installation on our website:
Enjoy working with this new SeisComP while we are excited to implement for you
further.
Hi, I was wondering specifically about the LOCSAT changes.
I had a look at gempa GmbH - SeisComP released in version 7.4 and Change Log — SeisComP Release documentation and there are statements like “Furthermore the LOCSAT code has been fixed to properly initialize variables and to correctly read travel-time tables. This change can lead to different locations when existing locations are relocated or when processing an identical pickset compared to older versions.”
Are there any guidelines about what kinds of differences in location we’re talking about? Of course it will depend from case to case, and everyone can test different versions of SeisComP to see for themselves. But some kind of sense of the scale of the implications would be useful given the large catalogues some users have.
Unfortunately I cannot tell you what exact differences this bugfix will cause. It heavily depends on the travel-time tables used. With the default tables there shouldn’t be an issue. Furthermore, the uninitialized variable issue could have led to errors such as “maximum iterations reached” or “SVD routine can’t decompose matrix” in cases where it actually should not. This effect was more or less random and there aren’t related studies as far as I know.
