Skip to content

Releases: swiss-seismological-service/scrtdd

v1.2.0

05 Nov 13:28
Compare
Choose a tag to compare

This is a minor version upgrade that contains only minor changes but also few backward incompatibilities.

ENHANCEMENTS:

  • better handling of airquakes
  • slightly improved cross-correlation performance
  • job scheduler now processes one origin a time and then checks for updates in the messaging system. The previous behaviour, which didn't check for messaging updates, could lead to the processing of stale information
  • Improved relocation statistics: both logs and output files contain now advanced relocation information that can be used for a statistical analysis on the module results

CONFIGURATION CHANGES:

  • renamed parameter keepWorkingFiles to saveProcessingFiles
  • removed parameter cron.wakeupInterval (it is now hardcoded as 1 second)
  • cron.delayTimes default value is now 10 seconds
  • added default groups subscription in default configuration

BACKWARD INCOMPATIBILITIES:

  • relocated-event.csv and relocated-phase.csv file format has changed. This affects ony the command line options: --merge-catalogs and --dump-catalog-xml that are not able to read files generated by former releases

v1.1.1

17 Aug 11:43
Compare
Choose a tag to compare

Minor bug fixes

v1.1.0

25 Jul 15:28
Compare
Choose a tag to compare

This is a minor version upgrade that contains only documentation updates, minor changes and bug fixes

ENHANCEMENTS:

  • Documentation
  • eval-xcorr command line option output has improved (see documentation)
  • source origin id is now stored in the relocated origin as comment (scrtddSourceOrigin)
  • Signal to Noise ration option has improved

BUG FIXES:

  • -O and --ep options do not work together but they are supposed to
  • theoreticalPhaseManualOrigin option doesn't create automatic picks for all manual origins in a catalog

CONFIGURATION CHANGES:

  • removed scrtdd.publicIDpatter: scrtdd uses global.publicIDpattern instead
  • many default parameter values have been updated with more sensible values

v1.0.0

24 Jun 19:21
Compare
Choose a tag to compare

This is pretty much the same as "v1.0.0 Release Candidate 4" but the documentation of some configuration options have been updated and a bug in --merge-catalog command line option has been fixed

v1.0.0 Release Candidate 4

12 Jun 11:25
Compare
Choose a tag to compare
Pre-release

BUG FIX: fixed computation of partial derivatives in the solver that were slightly incorrect

ENHANCEMENTS:

  • improvement to --evel-xcorr option
  • processing logs are now saved to in the working directory

v1.0.0 Release Candidate 3

05 Jun 08:55
Compare
Choose a tag to compare
Pre-release

BUG FIX: the relocated phase file (relocated-phase.csv) generated by the multi-event relocation cannot be properly read back into scrtdd

v1.0.0 Release Candidate 2

29 May 12:37
Compare
Choose a tag to compare
Pre-release

NOTABLE CHANGES:

  • multi-event relocation is now performed per cluster: instead of relocating all events in one single double-difference system each cluster is relocated independently. This results in improved solutions since the parameters of the solver are cluster dependent (e.g. downweighting by residuals). Also each cluster covers a smaller area, which helps in avoiding breaking the solver assumption of a flat earth surface (the lat/lon/depth coordinates are converted internally to an euclidean geometry system by the solver)

ENHANCEMENTS:

  • changed default configuration values: damping factor enabled by default, meanShiftconstraintWeight disabled. Although those parameters influence each other, the main advantage of damping factor is that it is able to constrain mean shift of the events even when relocating multiple clusters, while meanShiftconstraintWeight works better only when relocating a single cluster
  • solver now logs residuals by inter-event distance: this is useful to understand how the relocation is going since the residuals should increase with inter-event distance if the events are correctly relocated (see Waldhauser/Ellsworth)

v1.0.0 Release Candidate

03 May 19:43
Compare
Choose a tag to compare
Pre-release

NOTABLE CHANGES:

  • A new double-difference system solver has been added to the module and the external HypoDD software is no more required. Users who want to keep using the HypoDD software as a solver need to stick to the former v0.99.x release series but keep in mind that is not going to be maintained anymore

WARNING:

  • The module configuration has changes. A new configuration has to be created for this new version

ENHANCEMENTS:

  • Added --eval-xcorr command line option to evalutate the configuration of the cross-correlation settings. See documentation for more details
  • The clustering algorithm is much faster
  • The clustering algorithm selects now only well-connected events (minimum number of observations and neighbours must match all involved peers in a cluster)
  • better waveforms caching mechanism

BUGS:

  • Neighbouring selection by ellipsoid algorithm selects wrong candidates sometimes

v0.99.4

27 Mar 13:04
Compare
Choose a tag to compare

WARNING:

  • The catalog format has changes. Catalog files generated with previous versions are no longer valid

BUGS:

  • since HypoDD supports station ids with a maximum length of 7 chars, sometimes the realocation fails due to the station ids generated by scrtdd being too long. Now the station ids are converted before being used to create HypoDD input files

ENHANCEMENTS:

  • scrtdd is not able to deal with multiple location codes for the same stations. In previous releases errors were produced when a station had more than one location codes.

v0.99.3

25 Mar 15:58
Compare
Choose a tag to compare

BUGS:

  • filtering phase by station distance doesn't work properly during clustering

PERFORMANCE:

  • fixed a regression introduced in release v0.99.1: the waveforms are downloaded multiple times from the configured RecordStream, which result in a increased relocation time
  • -O and -ev options: when relocating an origin from command line do not try to preload the entire catalog into memory, it is useless and the user has to wait for that to finish

CHANGES and ENHANCEMENTS:

  • Automatict phase pick times coming from the catalog are never updated with the results of cross-correlation (this still happens in real-time relocation to the origin phases of the event is going to be relocated)
  • when trying to match the channels of two phases (e.g. HH-HH or HG-HG) during cross-correlation always try to match the real-time origin phase channels with the catalog ones