Thread: magnitudetype in FDSN Event

Started: 2015-06-29 17:31:14
Last activity: 2015-07-08 15:47:13
Philip Crotwell
2015-06-29 17:31:14
Hi

I have had quite a few SOD users report trouble that ends up being related
to how the magnitudetype parameter is used by the FDSN Event web services
at IRIS and elsewhere. The trouble is both that the acceptable values of
this are unspecified in the FDSN documentation and that specific
implementations often differ from common seismic usage. For example,
suppose a user wants events with a moment magnitude (Mw) >=6. There is an
event on 2015-01-07 that satisfies this (I think) in the IRIS web service:

http://service.iris.edu/fdsnws/event/1/query?nodata=404&includeallmagnitudes=true&minmagnitude=6&starttime=2015-01-07T05:00:00&endtime=2015-01-17T06:00:00

The catch is that the preferred magnitude has type MWW and the list of
magnitude types for this event are MWB, MWC, MWW, MWP, MI, MO. So a user
searching for events with type MW and mag >=6 would get get no events
returned. Moreover, there is no way in the FDSN spec to get a list of known
magnitude types from the service.

http://service.iris.edu/fdsnws/event/1/query?nodata=404&includeallmagnitudes=true&minmagnitude=6&starttime=2015-01-07T05:00:00&endtime=2015-01-17T06:00:00&magnitudetype=MW

I understand that magnitude types are really up to the catalog, but it
would be really nice if there was a way for the service to help translate
from common magnitude type nomenclature to catalog specific. We have, for
example, a way to query for known catalogs and contributors, so it would be
might be helpful to query for known magnitude types. In addition, if the
service internally had a list of known/valid magnitude types, then it could
return a "unknown magnitude type" error instead of simple "no events
found", which would help users understand why they were not getting any
events.Even better would be for each instance to document the list of
magnitude types used and what they mean.

The optimum solution, I think, would be for the magnitudetype parameter to
support both exact matches, so users can search for MWW if that is what
they really want, and a way to search for "MW-style magnitudes" that would
be an alias for all of the various magnitudes that are subtypes of the
moment magnitude. Perhaps this could be accomplished if the given type
matched the start of the magnitude type, so MW would match MW and MWW or
perhaps MW* could be used as a type.

It would also be helpful for ease of interacting with various data centers
if the fdsn specification denoted whether the magnitudetype parameter
should be case sensitive or not.

thanks,
Philip

  • Chad Trabant
    2015-07-07 23:49:38

    Hi Philip,

    Ah, the troubles with magnitudes types. I think a pragmatic partial solution is to simply search magnitude types case insensitively, which is exactly what the DMC's service does. We would support such a proposal to FDSN Working Group III.

    Your concept of searching for *-style magnitudes is interesting. The example with MW works, there might be some dark corners with other types but I cannot think of any offhand. I like the addition of wildcards as a way to specify a pattern versus an exact search best. Might also be worth pitching to WG III.

    At the DMC we are getting out of the event information distribution business. With NEIC and ISC catalogs, representing the bulk of event parameters we distribute, both available via fdsnws-event services from their respective producers, we can exit stage left so to speak.

    Chad

    On Jun 29, 2015, at 7:35 AM, Philip Crotwell <crotwell<at>seis.sc.edu> wrote:


    Hi

    I have had quite a few SOD users report trouble that ends up being related to how the magnitudetype parameter is used by the FDSN Event web services at IRIS and elsewhere. The trouble is both that the acceptable values of this are unspecified in the FDSN documentation and that specific implementations often differ from common seismic usage. For example, suppose a user wants events with a moment magnitude (Mw) >=6. There is an event on 2015-01-07 that satisfies this (I think) in the IRIS web service:

    http://service.iris.edu/fdsnws/event/1/query?nodata=404&includeallmagnitudes=true&minmagnitude=6&starttime=2015-01-07T05:00:00&endtime=2015-01-17T06:00:00

    The catch is that the preferred magnitude has type MWW and the list of magnitude types for this event are MWB, MWC, MWW, MWP, MI, MO. So a user searching for events with type MW and mag >=6 would get get no events returned. Moreover, there is no way in the FDSN spec to get a list of known magnitude types from the service.

    http://service.iris.edu/fdsnws/event/1/query?nodata=404&includeallmagnitudes=true&minmagnitude=6&starttime=2015-01-07T05:00:00&endtime=2015-01-17T06:00:00&magnitudetype=MW

    I understand that magnitude types are really up to the catalog, but it would be really nice if there was a way for the service to help translate from common magnitude type nomenclature to catalog specific. We have, for example, a way to query for known catalogs and contributors, so it would be might be helpful to query for known magnitude types. In addition, if the service internally had a list of known/valid magnitude types, then it could return a "unknown magnitude type" error instead of simple "no events found", which would help users understand why they were not getting any events.Even better would be for each instance to document the list of magnitude types used and what they mean.

    The optimum solution, I think, would be for the magnitudetype parameter to support both exact matches, so users can search for MWW if that is what they really want, and a way to search for "MW-style magnitudes" that would be an alias for all of the various magnitudes that are subtypes of the moment magnitude. Perhaps this could be accomplished if the given type matched the start of the magnitude type, so MW would match MW and MWW or perhaps MW* could be used as a type.

    It would also be helpful for ease of interacting with various data centers if the fdsn specification denoted whether the magnitudetype parameter should be case sensitive or not.

    thanks,
    Philip



    ----------------------
    DMC Web Service Developer (http://ds.iris.edu/message-center/topic/dmc-web-service-developer/)

    Sent via IRIS Message Center (http://ds.iris.edu/message-center/)
    Update subscription preferences at http://ds.iris.edu/account/profile/


    • Philip Crotwell
      2015-07-08 15:47:13
      Hi

      Understand the "exit stage left" and agree. I was more thinking of this as
      a general problem with FDSN Event, not just IRIS. Would the fdsn email list
      be a better place to take this up?

      What do you think about proposing a "known magnitudes" call on the service?
      At least then clients have a chance of figuring out what the possibilities
      are.

      thanks
      Philip


      On Tue, Jul 7, 2015 at 7:50 PM, Chad Trabant <chad<at>iris.washington.edu>
      wrote:


      Hi Philip,

      Ah, the troubles with magnitudes types. I think a pragmatic partial
      solution is to simply search magnitude types case insensitively, which is
      exactly what the DMC's service does. We would support such a proposal to
      FDSN Working Group III.

      Your concept of searching for *-style magnitudes is interesting. The
      example with MW works, there might be some dark corners with other types
      but I cannot think of any offhand. I like the addition of wildcards as a
      way to specify a pattern versus an exact search best. Might also be worth
      pitching to WG III.

      At the DMC we are getting out of the event information distribution
      business. With NEIC and ISC catalogs, representing the bulk of event
      parameters we distribute, both available via fdsnws-event services from
      their respective producers, we can exit stage left so to speak.

      Chad

      On Jun 29, 2015, at 7:35 AM, Philip Crotwell <crotwell<at>seis.sc.edu> wrote:


      Hi

      I have had quite a few SOD users report trouble that ends up being related
      to how the magnitudetype parameter is used by the FDSN Event web services
      at IRIS and elsewhere. The trouble is both that the acceptable values of
      this are unspecified in the FDSN documentation and that specific
      implementations often differ from common seismic usage. For example,
      suppose a user wants events with a moment magnitude (Mw) >=6. There is an
      event on 2015-01-07 that satisfies this (I think) in the IRIS web service:


      http://service.iris.edu/fdsnws/event/1/query?nodata=404&includeallmagnitudes=true&minmagnitude=6&starttime=2015-01-07T05:00:00&endtime=2015-01-17T06:00:00

      The catch is that the preferred magnitude has type MWW and the list of
      magnitude types for this event are MWB, MWC, MWW, MWP, MI, MO. So a user
      searching for events with type MW and mag >=6 would get get no events
      returned. Moreover, there is no way in the FDSN spec to get a list of known
      magnitude types from the service.


      http://service.iris.edu/fdsnws/event/1/query?nodata=404&includeallmagnitudes=true&minmagnitude=6&starttime=2015-01-07T05:00:00&endtime=2015-01-17T06:00:00&magnitudetype=MW

      I understand that magnitude types are really up to the catalog, but it
      would be really nice if there was a way for the service to help translate
      from common magnitude type nomenclature to catalog specific. We have, for
      example, a way to query for known catalogs and contributors, so it would be
      might be helpful to query for known magnitude types. In addition, if the
      service internally had a list of known/valid magnitude types, then it could
      return a "unknown magnitude type" error instead of simple "no events
      found", which would help users understand why they were not getting any
      events.Even better would be for each instance to document the list of
      magnitude types used and what they mean.

      The optimum solution, I think, would be for the magnitudetype parameter to
      support both exact matches, so users can search for MWW if that is what
      they really want, and a way to search for "MW-style magnitudes" that would
      be an alias for all of the various magnitudes that are subtypes of the
      moment magnitude. Perhaps this could be accomplished if the given type
      matched the start of the magnitude type, so MW would match MW and MWW or
      perhaps MW* could be used as a type.

      It would also be helpful for ease of interacting with various data centers
      if the fdsn specification denoted whether the magnitudetype parameter
      should be case sensitive or not.

      thanks,
      Philip



      ----------------------
      DMC Web Service Developer (
      http://ds.iris.edu/message-center/topic/dmc-web-service-developer/)

      Sent via IRIS Message Center (http://ds.iris.edu/message-center/)
      Update subscription preferences at http://ds.iris.edu/account/profile/




      ----------------------
      DMC Web Service Developer (
      http://ds.iris.edu/message-center/topic/dmc-web-service-developer/)

      Sent via IRIS Message Center (http://ds.iris.edu/message-center/)
      Update subscription preferences at http://ds.iris.edu/account/profile/



Page built 16:51:08 | v.34dba33a