seismic data management at the canadian national data centre jim lyons geological survey of canada
TRANSCRIPT
Seismic Data Management at the Canadian National Data Centre
Jim Lyons
Geological Survey of Canada
Outline
Evolution of the National Seismic Network
POLARIS (cooperating network)
CNDC processing
Waveform archive
Data dissemination
Ongoing challenges
Evolution of the CNSN - 1
1st Digital Data: regional seismograph networks:
operated Eastern Canada Telemetred Network (ECTN) and WCTN (in southwest BC), from 1975 1995
only “event files” comprising triggered stations with fixed pre- and post-trigger header/trailer were kept for computer-based analysis & archive
archived on 12" SONY WORM optical disks
data stored as 10-minute Full SEED volumes
Evolution of the CNSN - 2
1st archive of continuous data:
Yellowknife Array Upgrade 1988-9
CNSN Upgrade
Rollout started in 1992, based on new in-house digitizer (GD-1)
CNSN packet structure, data compression scheme and serial transmission protocol
Evolution of the CNSN - 3
CNSN Upgrade we reverted from SEED to a simple file format
“Canadian Archive” (CA) ½ hour network volumes based on a contiguous array of CNSN data packets with short header
• maintains full original data integrity including checksum, SOH and error/status codes
• excellent compression (subsequently adopted by CTBTO for data payload in CD-1.x format)
Evolution of the CNSN - 4
YKA revisited
started converting YKA data packets to CNSN packets in Yellowknife and transmitting as a real time stream to Ottawa for processing alongside CNSN data
allowed consolidation of CNSN & YKA onto common acquisition/archive systems
Evolution of the CNSN - 5
CNSN station map - 2005
POLARIS Network - 1
Portable Observatories for Lithospheric Analysis and Research
Investigating Seismicity
“a Canadian geophysical research consortium focused on investigation of the structure and dynamics of the Earth's lithosphere and the prediction of earthquake ground motion”
POLARIS Network - 2
initially to comprise 90 BB seismographs in 3 sub-arrays, using NMX digitizers and Libra VSAT satellite telemetry
subsequently augmented to ~105 HBB Libra sets with 3 VSAT earthstation downlinks
deployments to last 4-5 years
includes co-located MT systems; GPS to come
POLARIS Network - 3
GSC Role:
acquire, archive and disseminate waveform data
• we reformat real-time data streams on-the fly to pseudo-CNSN packets and process/archive alongside CNSN data (daily volume now exceeds CNSN)
• host POLARIS website ( www.polarisnet.ca )
• house and operate two NMX VSAT hubs
Canadian Seismograph Stations
CNDCProcessing Overview
Waveform Archive - 1
robotic DLT-tape library system installed in 1999
allowed us to put all waveform holdings (near) on-line
reformatted all earlier data formats to CA
developed standard tools to handle and reformat (SEED, CD-1, GSE-2.0, IMS-1.0, Mk2, Integer)
Waveform Archive - 2
moved to 12 TB all-disc system in April 2005
now all-online (no tape access delays)
CD Archive Holdings
1 TB = 10244 bytes
Data Dissemination - 1
Continuous data:
as CTBT NDC, forward IMS Primary SHI data to Vienna IDC in CD-1 format (currently ~194 MB/day)
share real-time data with USGS/NEIC (10), PTWC/ATWC (9), Montana Bureau of Mines (MB - 2) and UWashington (6) via orb2ew (Internet)
10 FDSN stations to IRIS-DMC (daily ftp)
Data Dissemination - 2
Request data:
AutoDRM has run continuously since May 1994
NetDC since 2004
Event archive
developed Web interfaces to ease usage
Data Dissemination - 3
CNDC AutoDRM Data Shipments Number of requests processed Volume in GB (1GB = 10243 B)
Event Selection
Ongoing challenges
serving non-specialist clients
archive and dissemination of non-seismic data (Infrasound, MT and GPS data)*
tracking network metadata over time*
automated quality control and display*
Global Infrasound
Archive(G I N A)
stored in CSS-3.0 wf format
convert to SEED, MSEED, SAC, & Integer on demand
Magneto-telluric Data
Tracking Network Metadata
tracking the complete history of Canadian and foreign stations of interest wrt: agency, affiliation(s), network/deployment(s), location, timing, instrumentation, and response over time has required rethinking our db schema
adopted some features of CSS-3.0 db schema
now reviewing and implementing some Antelope extensions to CSS (e.g., stage table)
CNSN Station Book - 1
CNSN Station Book - 2
CNSN Station Book - 3
Automatic Quality Control
need to track various measures of data quality including availability (completeness), timeliness (latency), stn timing, signal degradation (waveform quality), and metadata quality (e.g., degree to which response is known and trusted)
how to automate and display for potential data users?
RUDY
Bitter cold and wind affecting power (TEG) at YKW4
Station Quality Plot
Conclusion
While the seismological community is reasonably well served, a number of challenges remain:
making data useful to a broader range of people
managing an increasingly heterogeneous data set
automating and displaying data quality
Hopefully we can both contribute to and gain by the GEOSS initiatives.
The End