(meteorobs) Fireball in Texas

[Mike Hankey]

Rob,

The self entered observer rating is far from perfect, but I think it
helps, even if just to cut down on the number of reports so you can at
least see whats going on a little. Of course there are people who
think they are experts, but don't know north from south. As you point
out the blatantly wrong reports are easy to spot.

When you look at this event for all reports (and other big ones)
there's so much data its a cluster and you can't make heads or tails
of it. But if you filter on level 3 things start to take shape. In
about 5-10 seconds of looking I can see a path the meteor took and a
general area where it ended.

http://i.imgur.com/9Adai.jpg

Now obviously the trajectory is just an estimate and could be off by
quite a bit, but it only took 5 seconds and its told me a lot.

>From the witness reports we are alerted to the event's occurrence, the
time of the event, we're able to ascertain movement direction, and we
can usually guess the terminus location down to at least the county
level. That's pretty good! Obviously not as good as a radar return
that tells you exactly where to go, but it would be harder to find the
precise radar returns without the general location tip off provided by
witness sightings.

We've made a lot of improvements to the data entry interface, but
folks are folks and still have a hard time reporting accurately. I
think where we can improve is having some filters to weed out the bad
reports. Coming up with a method to automatically identify and flag
the bad reports or qualify the good ones is the tricky part. The gold
stars is a good idea, just need to figure out a series of criteria
that will assign the star while not knowing the true solution. Ideas
are welcome.

We continue to tweak the site and make improvements.

Thanks,

Mike

On Fri, Dec 7, 2012 at 6:17 PM, Matson, Robert D.
<ROBERT.D.MATSON at saic.com> wrote:
> Hi Mike,
>
> Yes -- it shows the merit in averaging the reported times!  I wonder if
> the standard deviation improves at all with the witness's self-reported
> experience level?  If I go to level 3 and higher, there are 28 reports,
> with times varying from 12:37 to 13:45 (!).  (Obviously the 13:45
> witness
> entered the wrong timezone.)  Even discarding the bad 13:45 time,
> another
> level-3 witness reported the time as 13:25 -- off by 42 minutes.
>
> There are only two witnesses out of the 104, both level 1, whose
> red terminus vectors triangulate to the actual location of the fall.
> They deserve gold stars.  :-)  --Rob
>
> -----Original Message-----
> From: meteorobs-bounces at meteorobs.org
> [mailto:meteorobs-bounces at meteorobs.org] On Behalf Of Mike Hankey
> Sent: Friday, December 07, 2012 2:47 PM
> To: Meteor science and meteor observing
> Subject: Re: (meteorobs) Fireball in Texas
>
> Hey Rob,
>
> Great work again with the Galactic Analytics find.
>
> While its unfortunate that human beings can't determine the fall time as
> well as a doppler radar, the average time for the fall calculated from
> the witness reports is: 2012-12-07 12:44 UTC
>
> Off by less than a minute :)
>
> http://www.amsmeteors.org/fireball_event (see avg UTC column for event
> 2085)
>
> Now you bring up an interesting concept with regard to scoring each
> witness account based on their ability to accurately tell and report
> time. I'll have to think about this more.
>
> Thanks,
>
> Mike
>
> _______________________________________________
> meteorobs mailing list
> meteorobs at meteorobs.org
> http://lists.meteorobs.org/mailman/listinfo/meteorobs