Hi, List,
As will have been noticed by anybody crazy enough to
follow all stuff, I have been using the on-line Impact Calculator
at LPL based on the most complete model of Jay Meosh to
try and approximate what size, speed, and. kind of object
would match the description of the Great Norway Rock
http://www.lpl.arizona.edu/impacteffects/
By now, as reports change and more details come in,
I've run that calculator many hundreds of times trying to create
a match for the descriptions. After that many times, I began
to get suspicious when I began to get some inconsistent results,
such as big changes in its results with only small changes in input
parameters, so I started "testing" the model.
I very quickly detected some very odd behavior on its
part. Taking a specific run of the model which had produced
a good blast (airburst) with seismic effects and plenty of noise,
I altered one parameter by the smallest unit amount: I changed
the angle of entry from 45 degrees to 44 degrees. Obviously,
such a small change should only produce a very slight change
IN THE REAL WORLD, but in the model -- all the blast,
seismicity, and sound vanished from the results!
I was able to get this bad behavior to repeat in other examples,
and with other parameters. Some runs hold up better than others,
but basically what happens when you get more bad than good
answers, you lose all trust in ANY result from the model. You
look at every change you induce and say is this right or too much?
Too little? How accurate? And then, you give up on the model.
So, consider this a caution for anyone else thinking of
relying what seemed to be a very convenient and useful tool
for impact scenarios. Treat its results as a guide only, at best.
While I was fiddling with these incoming fireball simulations,
Chris Petersen sent me some emails off-line, trying to gently
convince me that a spectacular fireball did not have to be a
spectacularly big object and the reported Norwegian fall did
not need to be anything like as big as I supposed. He cited
some very convincing real-world examples:
"Well, there are models and there are models (and there is
reality). That model may have many useful elements, but
consider one actual example. I have a very well characterized
event over southwest Colorado from 2002. The meteoroid
fragmented at a height of 36 km and dissipated 1e10 joules, or
about 2.4 tons TNT. The estimated entry mass was 95 kg.
This event was recorded barometrically at three stations,
the farthest being 720 km away. Witnesses reported sounds
up to 64 km from the terminal explosion. The fireball lit up
entire valleys bright enough to stimulate full color vision (the
absolute visual magnitude was -17). The explosion produced
a signal on a seismometer 325 km away. This event may
have produced up to 2 kg of meteorites, but nothing has
been recovered."
One item that bears repeating is: "There are models and
there are models (and there is reality)." Sadly true. I should
have known better. It's clear from the example that the
Norwegian observations could be produced by an object
much less grandiose than what I suggested. The Melosh
model's 34 ton intruder could end up to be a 34 pound
rock! Well, maybe a 340 pound rock...
In all fairness, I should mention that this on-line calculator gives
users one disclaimer about its performance, "These results come with
ABSOLUTELY NO WARRANTY," so it's not like we weren't warned...
a little. I can second that. Use with Caution.
Sterling K. Webb
-----------------------------------------------------------------
______________________________________________
Meteorite-list mailing list
Meteorite-list@meteoritecentral.com
http://six.pairlist.net/mailman/listinfo/meteorite-list
