[dhi-servers] PREF catalog in IRIS_EventDC

[Sue Schoch]

Hi Philip,

I tweaked the algorithm I was using a bit.  It is the same as we will  
use for event associations down the line.  DHI network server will  
return the most authoritative catalog origin available for an event.   
Authoritative catalog is determined the usual way, with each being  
more authoritative - FINGER NEICALRT QED WHDF MHDF ISC.  Events are  
determined by grouping origins based on time and location.  Origins  
within +- 30 seconds and within 0.5 degrees are considered the same  
event.  My tweaking should eliminate the problem you saw yesterday as  
I was only looking at equal times.  Hope this helps you!

Sue

On Sep 5, 2007, at 11:31 AM, Philip Crotwell wrote:

>
> Hi Sue
>
> My understanding from way back when Sandy created the original iris  
> eventdc, was that the PREF catalog was more of a rule of thumb than  
> anything real. PREF meant something like NEICALRT for the first 2  
> weeks, then QED for the next 4 weeks, then WHDF for a month or so,  
> then MHDF, and finally after 2 years have passed ISCCD. This gave  
> an easy way for users to generally get the "best" event location,  
> without getting a lot of duplicates. I did not depend on any event  
> grouping. Is this not how PREF works?
>
> thanks,
> Philip
>
> On Sep 5, 2007, at 10:48 AM, Sue Schoch wrote:
>
>> Morning Philip,
>>
>> In our current event scheme, we do not try to associate events.   
>> That means that an event must have the exact same time for the  
>> order of precedence to come into play.  In the case below, you can  
>> see that the times are different.  The preferred catalog will  
>> actually return both the NEICALRT and MHDF catalogs as the times  
>> are different so it thinks these are different events.
>>
>> We are working on a new event scheme which will do event  
>> associations, assigning one as primary.  This is in the  
>> preliminary stages and will not be available for a while yet.
>>
>> Sue
>>
>> On Sep 5, 2007, at 10:19 AM, Philip Crotwell wrote:
>>
>>>
>>> Probably should give you the event info I get from DHI. Query is  
>>> for just 2006-7-17, mag > 7,
>>>
>>> With catalog=PREF:
>>> time, latitude, longitude, depth, depthUnits, magnitude,  
>>> magnitudeType, magnitudeContributor, catalog, contributor, name,  
>>> flinnEngdahlRegion, flinnEngdahlRegionType
>>> 2006-07-17T08:19:25.000Z, -9.334, 107.2629, 10.0, kiloMETER, 7.7,  
>>> MO, NEIC, NEICALRT, NEIC, 13685414, 282, 1
>>>
>>> With catalog=MHDF
>>> time, latitude, longitude, depth, depthUnits, magnitude,  
>>> magnitudeType, magnitudeContributor, catalog, contributor, name,  
>>> flinnEngdahlRegion, flinnEngdahlRegionType
>>> 2006-07-17T08:19:26.680Z, -9.284, 107.419, 20.0, kiloMETER,  
>>> 6.1:7.2:7.7:7.2, MB:MS:MW:MW, NEIC:NEIC:HRV:NEIC, MHDF, NEIC,  
>>> 13902698, 282, 1
>>>
>>> thanks,
>>> Philip
>>>
>>> On Sep 5, 2007, at 10:02 AM, Philip Crotwell wrote:
>>>
>>>>
>>>> Hi
>>>>
>>>> I just sent in a query for events on 2006-7-17 (over a year ago)  
>>>> using the PREF catalog. The event I got back had a preferred  
>>>> origin in the NEICALRT catalog. I did a similar query (I think)  
>>>> via seismoquery using the PREFERRED catalog and got back an  
>>>> event in the MHDF catalog.
>>>>
>>>> Any idea of why these would be different? I would expect that  
>>>> NEICALRT should only be "preferred" right after the event, to be  
>>>> superseded by the weekly and then the monthly.
>>>>
>>>> thanks,
>>>> Philip
>>>> _______________________________________________
>>>> dhi-servers mailing list
>>>> dhi-servers at iris.washington.edu
>>>> http://www.iris.washington.edu/mailman/listinfo/dhi-servers
>>>
>>> _______________________________________________
>>> dhi-servers mailing list
>>> dhi-servers at iris.washington.edu
>>> http://www.iris.washington.edu/mailman/listinfo/dhi-servers
>>
>> _______________________________________________
>> dhi-servers mailing list
>> dhi-servers at iris.washington.edu
>> http://www.iris.washington.edu/mailman/listinfo/dhi-servers
>
> _______________________________________________
> dhi-servers mailing list
> dhi-servers at iris.washington.edu
> http://www.iris.washington.edu/mailman/listinfo/dhi-servers