Subject: Morrison replies to critics
From: David Morrison
Submitted: Wed, 26 Feb 2003 17:25:48 -0800
Message number: 85
(previous: 84,
next: 86
up: Index)
LSST SWG Members:
I had not expected such a stimulating set of questions and comments
in response to my "sound bites" on the NEO hazard. But that is part
of the value of sound bites: they are more likely to be read to
elicit a response. I am happy to enter into this dialogue, and I hope
the other members of the LSST SWG will forgive me for prolonging this
discussion by attempting to respond to these comments.
David Morrison
---------------------------------
AL HARRIS:
>Morrison wrote: A large country (like Russia, China, USA) is more
>likely to be hit, but its citizens are no more at risk than those of
>a small country.
I would clarify "its citizens individually are..."
Reply: OK, I accept the clarification
>Morrison wrote: We have a choice: Deal with the small impacts that
>hit more frequently but pose very low risk, or deal with the large
>ones that are infrequent but are more likely to kill you. (The
>current Spaceguard Survey deals with the large ones that pose the
>greatest risk)
With the technology at hand (optical surveys), we actually don't have
a free choice: by the very nature of the survey, we deal with the
largest ones first. Even a highly capable survey like LSST finds the
biggest ones fastest, whether you want to or not. It is just a lucky
(?) coincidence that it is the largest, and thus easiest to discover,
objects that pose the greatest hazard. Even if that were not true,
we would still find the biggest ones first. It is really quite
unrelated to priorities. That's the point to capture in a "sound
bite"
Reply: Al is correct, and I apologize for this one. There is no way
to find the small ones without first (or simultaneously) finding the
big ones. The only choice is whether, having found most of the big
ones, we should now go for the smaller ones.
>Morrison wrote: Impacts are the only known natural hazard that could
>destroy civilization.
I would question this. We see evidence of volcanism of far vaster
proportions than anything we have experienced in historical times.
Thus I think it is entirely plausible that a really massive volcanic
event could bring down a civilization. If you include disease as a
natural disaster, one could imagine an epidemic of something as
deadly as AIDS and as contagious as the flu could bring down
civilization, or even cause an extinction.
Reply: While the issue is debatable, I stick to my sound bite. As far
as we know volcanoes (especially the very large flood basalt
eruptions) operate on much longer time scales than impacts, and I
suspect that civilization would survive and adapt to even a long
series of large eruptions. As to a new disease, please note the term
"known" natural hazard.
----------------------------------
DAVID WITTMAN
Morrison wrote: There is no last-minute warning system: Either we discover an
asteroid and predict its impact decades or centuries in advance, or
else it strikes without warning -- the first you will know is when
the sky lights up and the ground starts to shake.
I would understand if you said we couldn't divert it at the last
minute. But we can't detect it either? (Keep in mind I don't know much
about asteroids and may be representative of the people you are trying
to educate.)
Reply: I have tried to envision a system that would detect objects
on final approach (say in the last couple of weeks) before impact. It
would need to cover the whole sky (including the directions toward
the Sun) and reach to the magnitude of the NEA (whatever size is
targeted) at a distance of order 10 million kilometers. As far as I
can tell, this is more difficult than implementing the Spaceguard or
the proposed LSST survey. Even if you wanted to find NEAs on final
approach, the chances are that before you ever found one on final
approach you would have found and cataloged everything. Put another
way, since any asteroid that is going to hit will repeatedly come
close to the Earth, we are likely to pick it up on an earlier pass
rather than on final approach. And as you note, it isn't useful to
find them at the last minute anyway, so what is the point of
contemplating such a survey?
More importantly, what is the implication for the LSST of statements
like "You have more to fear from the residual 10% of big ones that
Spaceguard will not have found by 2008 than from the sum of all the
smaller asteroids."
Reply: This is a new result that had been guessed earlier but now
seems supported by the recent the work of Harris and Chesley. The
risk from the smaller NEAs is so small that the residual 10% of big
ones still dominate.
How long will it take for LSST to catch these last 10%? How long would
Spaceguard take? After that, how often do new threats appear?
Reply: Good questions -- I have not seen either case documented in
any detail. The challenge is that as you get down to the last few
percent you increasingly find yourself looking for NEAs in odd orbits
(otherwise they would have been picked up already). As for new
threats, since we are in equilibrium, new threats appear on the same
time scale as the dynamical removal of existing NEAs -- or about once
per million years for NEAs >1 km.
-----------------------------------
NICK KAISER
Morrison wrote: In the real world, impacts are deterministic: There
either is or is not an asteroid that will collide with Earth during
this century.
But our knowledge is probabilistic. Based on our current
knowledge there is some probability per unit time that we will be
struck by an object of some energy, composition etc. The correct way
to quantify the utility of future NEO searches is to compute how that
probability will be reduced if, as is most likely, no objects are
found.
That this is hardly a "sound-byte". Moreover, a common reaction, even
from technical folks, on hearing this is to exclaim that "just making an
observation cannot change the risk". The problem is that people often
forget that a probability is a statement of our knowledge about the
world, rather than some objective reality. I understand then why you
don't like to use the P-word, but it is really the right way to look
at this, so we have a communication/education problem.
Reply: These are subtle issues, and you probably understand how to
properly use the term "probability" better than I. The impacts are
deterministic but our knowledge can be described in probabilistic
terms. All I know for sure is that if we ever found all of the NEAs
at a given size (and we think we have done this already for those 5
km or larger), and if none of them is on a threatening orbit, then we
have in fact reduced the hazard to zero. I make this argument because
it is very common for folks to write, as Mike Strauss did in his
recent telecon minutes, as our objective is to improve our knowledge
of the statistical risk, rather than to actually track down these
guys one at a time. We don't need to know more about the
probabilities, but rather to know (as certainly as possible) which
(if any) will hit the Earth in the next century, and exactly when and
where it will hit.
Morrison wrote: Many science objectives can be met with statistical
samples, but planetary defense requires a complete survey.
But no survey will be complete. That's why we need to explain the
probabilities, even though this is evidently a challenge.
Reply: In my example above of the 5 km NEAs, the survey is complete.
That is the ultimate objective.
Morrison wrote: The public does not care if the annual risk of dying
from an impact is one in a million or one in two million -- they want
to know if it will happen to them. AND When you cross a busy
intersection, you don't ask what the statistics are on pedestrian
accidents -- you look to see if a car or truck is headed toward you.
That seems a bit condescending. A good analogy is screening for
cancer. Surely people understand that this is probabilistic, and
can weigh up the costs and benefits (or at least their policy making
representatives can). I don't see that NEO searches are any different.
Reply: I don't mean to be condescending; my experience is that this
is a good sound bite for making my point. The impact risk is not
easily compared with other risks (earthquakes for example), because
the laws of planetary motion are strict -- no one throws the dice to
see if an asteroid will go careering out of orbit this year and head
for the Earth. If anything is going to hit in the 21st century, it is
already on a collision course. That is as deterministic as you can
get!
Morrison wrote: There is no last-minute warning system: Either we
discover an asteroid and predict its impact decades or centuries in
advance, or else it strikes without warning -- the first you will
know is when the sky lights up and the ground starts to shake.
Well, one could operate an LSST in such a mode as to detect "death-plunge"
objects. It would dictate a very different observing strategy and would
tilt the balance towards distributed apertures, since one would want to
go very broad. We looked into this for Pan-STARRS (and our review panel
wants us to reexamine this) but we concluded that the utility was small.
As it happens tough, about 10% of the incoming collision risk comes from
the trailing "sweet-spot", so there is some chance of detecting incoming
stuff. One would get about a month warning of a 30m impactor.
Reply: See answer above to Wittman. I can't see why anyone would want
to forgo a survey that will provide decades of warning in favor of an
approach that will provide only days of warning, and as far as I can
tell the decades-warning-survey is actually easier to implement the
"death-plunge" survey. Also, I have never understood the interest in
finding the "sweet spots" from which final approach to the Earth is
likely to come. Our objective is to find the darned things in
advance; if any are ever in the sweet spot then we have failed.
--
+++++++++++++++++++++++++++++++++++++++++++
David Morrison, NASA Ames Research Center 240-1
Tel 650 604 5094; Fax 650 604 4251
david.morrison@arc.nasa.gov or dmorrison@arc.nasa.gov
website: http://astrobiology.arc.nasa.gov
website: http://nai.arc.nasa.gov
website: http://impact.arc.nasa.gov
LSST LSST LSST LSST LSST Mailing List Server LSST LSST LSST LSST LSST LSST
LSST
LSST This is message 85 in the lsst-general archive, URL
LSST http://www.astro.princeton.edu/~dss/LSST/lsst-general/msg.85.html
LSST http://www.astro.princeton.edu/cgi-bin/LSSTmailinglists.pl/show_subscription?list=lsst-general
LSST The index is at http://www.astro.princeton.edu/~dss/LSST/lsst-general/INDEX.html
LSST To join/leave the list, send mail to lsst-request@astro.princeton.edu
LSST To post a message, mail it to lsst-general@astro.princeton.edu
LSST
LSST LSST LSST LSST LSST LSST LSST LSST LSST LSST LSST LSST LSST LSST LSST LSST