2.2.3 Continuous waveform data

In SEISAN one can plot or extract continous data from either a standard SEISAN database (internally many different formats) or from a BUD or a SeiscomP archive (internal format MiniSeed. The archive option is the most common way to use continous data in SEISAN. It is easy to convert most format to MiniSeed with WAVETOOL and the use any group of MiniSeed files to create the archive with program DATASELECT (included in SUP), see section 'getting data into the data base' See also page [*].

Continous data in a BUD or SeiscomP archive

SEISAN reading BUD and SeisComp archives

SEISAN is able to read miniseed from a BUD and SeisComp (SDS) archives data structure. The reading routines using Chad Trabant software have been implemented by Ruben Luis. Reading continuous data:
This works just like reading SEISAN continuous data, except there are no S-files, only the archive files. All the same functions are available:
Plotting, zooming and extracting segments and registering events.
Read archive data as an event from eev:
A reference to a segment is made in the s-file and it is treated as if it was a file. When a keyword for archive (ARC) is found, the reading is directed to the archive instead of to a file. The archive reference is e.g.

   ARC ROSA  BHZ PM    2010 1011 0100 00 14400                                   6
where ARC indicate archive, STAT is station code, COM is component, LO is location code YYYY MMDD HHMM SS is start time and DUR is duration in secs.
Thus the segment in archive with given start time and duration is considered a file. If later plots require less data than the segment referenced, the whole segment is still read, like reading the whole trace in a file in archive with given start time and duration. A mixture of archive references and file names can be used.
The archive reference can also use wildcards to reference many channels, like e.g. just writing ARC and the rest of the line blank, all channels in the archive will be selected.
Station is blank or *, component, network and location are blank: All channels defined in SEISAN.DEF will be plotted with start time and duration given in ARC line (if not blank, see next option). If a component is given, only that component will be selected. If a network is given, only that network will be chosen. If a location code is given, only channels with that location code is selected.

Start time and or durations blank: Start time will be origin time - a time given in SEISAN.DEF (ARC_START), duration will be a time given in SEISAN.DEF (ARC_DURATION).

Station is P: For phase reading - All channels for all stations listed in the S-file found in the archive will be plotted. There is no requirement for the station to have any other information than the station name, component code is not used. So a new station can easily be added. Example:
   ARC P               2010 1011 0100 00 14400                                   6

Station begins with _ indicates virtual network as defined in SEISAN.DEF. This can also be inserted with arc command in EEV. Example of link to virtual network named _CTBTO in sfile, below format line:

   ARC STAT  COM NT LO YYYY MMDD HHMM SS   DUR                                    
   ARC _CTBTO          2010 1011 0100 00 14400                                   6

Plotting all stations without an archive reference line: If parameter ARC_BY_DEFAULT in SEISAN.DEF is set to 1, all channels in the archive will be selected.

NOTE: An ARC line can be inserted/edited in the S-file from EEV by command arc.

The archive is defined in SEISAN.DEF where each channel is defined by ether an ARC_CHAN2 line (from version 12) or an ARC_CHAN line (still possible, this was the only option prior to version 12). ARC_CHAN2 makes it easier to define virtual networks, which with ARC_CHAN have to be defined through ARC_VIRTUAL_NETWORK lines. Another improvement is that with ARC_CHAN2 it is possible to specify the archive with variables that are defined through ARC_ARCHIVE2 lines.

Here is an example of using the ARC_CHAN2 definition and definition of the nnsn_eida archive variable:

ARC_ARCHIVE2                            nnsn_eida /home/seismo/WAV/SLARCHIVE
--------- explanation arc_chan2         S    C  N L  Def Start    Stop     Archive   Vnet1 Vnet2 Vnet3 Vnet4 Vnet5 Vnet6
ARC_CHAN2                               ASK  HHZNS00 1.0 20181106          nnsn_eida _SW   _SN   _NNSN
ARC_CHAN2                               ASK  HHNNS00 1.0 20181106          nnsn_eida _SW   _SN   _NNSN
ARC_CHAN2                               ASK  HHENS00 1.0 20181106          nnsn_eida _SW   _SN   _NNSN
ARC_CHAN2                               ASK  EHZNS00 1.0          20181106 nnsn_eida _SW   _SN   _NNSN
ARC_CHAN2                               ASK  EHNNS00 1.0          20181106 nnsn_eida _SW   _SN   _NNSN
ARC_CHAN2                               ASK  EHENS00 1.0          20181106 nnsn_eida _SW   _SN   _NNSN

The columns in the ARC_CHAN2 line contain:

41-52: station, component, network and location
54-56: start date as yyyymmdd
58-65: stop date as yyyymmdd
67-74: archive name that is pointing to archive path
76-84: name of virtual network that station is part of
86-90:name of virtual network that station is part of
92-96:name of virtual network that station is part of
98-102: name of virtual network that station is part of
104-108: name of virtual network that station is part of
110-114: name of virtual network that station is part of
116-120: name of virtual network that station is part of

The ARC_ARCHIVE2 line gives the archive name (41-49) and the respective archive path (51-). One can add several ARC_ARCHIVE2 lines.

When defining a channel using a ARC_CHAN text string: SSSSSCCCNNLL, where SSSSS is the station, CCC is the component, NN the network and LL the location. These lines can be generated with program sample_read_wav, which will make a list of all channels in a given set of files (it thus requires to have one of several single files with the channels needed).. An alterentive is to use program GET_ARC_CHANNELS which will scan an archvie to list the channels in SEISAN.DEF format. See section 'Waveform file management tools'.

The directory of the archive is given by ARC_ARCHIVE, see e.g.:

ARC_CHAN                                PMOZ BHZPM
ARC_CHAN                                PMOZ BHNPM
ARC_CHAN                                PMOZ BHEPM
ARC_CHAN                                SFJD LHZIU10
ARC_CHAN                                SFJD LHNIU10
ARC_CHAN                                SFJD LHEIU10
ARC_ARCHIVE                             /uibsan/home/s2000/BUDARC
ARC_DURATION                            10000.0
ARC_START_TIME                          5000.0
ARC_TYPE                                1.0                        
ARC_BY_DEFAULT                          0.0
where each channel is defined as well as the location of the archive. The specification is the same for both BUD and SeisComp archives. ONLY one archive type can be used at the same time and the archive type, ARC_TYPE is given in SEISAN.DEF.
Groups of channels can be defined as virtual networks. The virtual network names will be shown in the station selection when using 'arc' option in MULPLT. The definition is done in SEISAN.DEF:
ARC_VIRTUAL_NETWORK                     _NO       BER  HHZNS00
ARC_VIRTUAL_NETWORK                     _NO       BER  HHNNS00
ARC_VIRTUAL_NETWORK                     _NO       BER  HHENS00
The virtual network name has to start with a '_' to make them visible as such in the station selection.
Archive reading works on both Linux and Windows.

Continous data in a SEISAN database

In SEISAN continuous data has no special format. Continuous data is simply ordinary waveform files that follow each other in time. In order to treat the data as continuous, the data can be put into a SEISAN continuous data base. Such a data base is made as follows:

- For each waveform file from a station or network, an S-file is created. The S-files only contain reference to the waveform file(s). Program AUTOREG can be used to create the S-files.
- The waveform data is optionally put into the corresponding waveform station directories, however they can also be in WAV or working directory. For large data sets it is strongly recommended to use the WAV database structure.
- The continuous databases are defined in SEISAN.DEF in DAT.

If e.g. data is to be stored from 3 different stations (three componet files), create 3 databases under WAV and REA with the name of the stations (program MAKEREA). If the continuous data consist of 20-minute files, this would mean about 2200 files pr month, which is a reasonable number. It is now possible for some programs (MULPLT, WAVETOOL) to get access to any or all of the traces in the continuous data base and plot and extract data. If the continuous data is archived from a real-time system it is best to have one database per station as it will at times be necessary to backfill gaps as data may not have arrived in real-time.

Note that waveform files in a SEISAN continuous structure must contain the year and month which is used to located the corresponding year-month structure. The default start of the file names accepted are:

ccmm ccyy-mm ccyymm

where cc is century, yy is year and mm is month. If year and month are placed diffenrently in the file name, their location must tbe specified in SEISAN.DEF, parmeter CONT_YEAR_MONTH_POSTION_FILE.

It is also possible to store the data without having a database for each station:

The waveform files in a continuous data base can have different formats for different stations and one S-file can refer to more than one waveform file, provided they start at about the same time and have the same duration.

A simpler way to use smaller quantities of continuous data is to make a list of these files with DIRF and an application program can then use that list to work with the data. Currently two programs have special options for this kind of continuous data. The MULPLT program will plot data from several files as if it was one file in one continues trace the RESAMP program will resample the data from several files and put it into one output file.