[Prev][Next][Index][Thread]

(meteorobs) Limiting magnitudes and more

to: Meteorobs.
subject: Limiting magnitudes and more
cc.: Peter, Rainer


from Marco Langbroek


Hello everybody,


Great to see all those replies and discussions, both via Meteorobs and in
private mails, about ZHR-calculations, correction factors etcetera! This is
actually what I like about Meteorobs: that it is lively and people respond
to each other, and that constructive debate is possible.

To some of you, these discussions might be all abracadabra. Don't bother
about that. This is something for freaks (and freaks are odd, as you know.
They need a psychiatrist). But still, if you try, it is all less esoteric
than it sounds, and it concerns things fundamental to meteor science,
though you can perfectly enjoy your hobby without it, let me add. So don't
bother if you think you don't understand, or if you think that it just
doesn't interrest you.

There are two contributions to the discussion to which I want to answer in
particular: that of Rainer Arlt and of Jim Richardson. Both were very
constructive and I enjoyed them very much.

I fully agree with most of Jim Richardson's statements, also his statements
about the nature of science (this is riding my hobby too). Indeed, in
science all 'truths' are provisional and models are not reality itself.
'Objectivity' does not exist, it is all a matter of conviction, culture and
attitudes (more about that later, in my reply to Rainer actually). So I
should have been more carefull in my statement, though I also think that
Jim takes it a little bit different than it was meant. What I wanted to
make clear is that in my opinion the problems Norman touched upon in his
critique on Joseph were in this particular case not so much due to the
nature of the limiting magnitude correction itself, but to other elements
in the reduction procedure, most notably perception differences and a wrong
applyance of a particular element in the Lm correction (the r-value). This
is something different than to say that the Lm correction equation itself
is beyond discussion, though I admitt that my statement can be read that
way.

There is one particular passage in Jim's response that I don't quite get
and in which he, I think, might misinterpret some things. Jim wrote:

> "This brings up one of the problems which Norman pointed out. If your
limiting magnitude is 6.5 or less, than the correction factor is 1.0 or
greater. If you have extremely dark skies, as he does, then the correction
becomes less than 1 and reduces the total number of meteors seen to match
the standard conditions. Marko [sic!] corrected for this effect by assuming
[sic!] that Norman had less than standard perception, multiplying the total
by 2.5 to bring the number back up again."

Now, I don't see the point about the fact that for Lm larger than 6.5 the
correction REDUCES the number of meteors. Because that is what it actually
SHOULD do indeed! With the radiant in zenith and being the 'average'
observer, you SHOULD see more meteors than the ZHR with Lm above 6.5. The
reason that you don't in most cases, is that you seldom have the radiant in
the zenith. Also, it is certainly not true that I corrected for THIS when I
applied that Cp-factor amounting to a factor 2.5 upgrading to Norman's
data. I should have applied that same 2.5 factor when he had Lm 6.0, so a
Lm lower than 6.5. Also, I didn't 'assume' this factor 2.5. I INFERRED it
(calculated it) from calibrating his sporadic rates to match those of the
'median' observer observing with a similar Lm. So it has a base, in his
data. Rainer already commented that IMO found a similar result for Norman's
data. It shows that Norman, as every other observer, has a SYSTEMATIC
difference in observed numbers compared to the 'median' observer. This
systematic difference should be more or less independant of the Lm (though
there are some pittfalls in this, which I will not discuss now for reasons
of space): it is the same with Lm 6.0 or 7.0. And it is for this that I
corrected, not for the reduction due to Lm larger than 6.5 'in order to
bring the number back up again' as Jim commented. In this aspect, I feel
that Jim might use the term 'perception' different from what I do in his
contribution. In the way I (and Peter Jenniskens) use it, it has nothing to
do with shifting detection probabilities for magnitude classes with
shifting Lm, but with systematic differences between observers. Be carefull
not to confuse things up in this respect.

I fully agree that many things can be improved, also concerning r-value
determinations. But for the time being we have to work with the current
methods. Also, I don't see things so pessimistic as Jim. Peter Jenniskens
recently correlated visual and video magnitude distributions to obtain r-
values (for the alpha Monocerotids and Quadrantids) and they were well in
line. Also, there are other techniques for establishing r-values than
probability functions (and for establishing detection probability functi-
ons).

To turn back to the original subject: I think (and this is what I meant in
my comment to Norman) that there are elements in the reduction procedure
which are more uncertain and are more messing up the results than the Lm-
correction that Norman deemes the cullprit: for example the radiant
altitude correction (another cullprit Norman rightly deemed), but also
systematic under- or overestimates of the Lm by some observers and certain-
ly the whole issue of observer perception. Rainer already made some very
valuable remarks in his contribition regarding radiant altitude correction,
so I won't go into that again. But I do want to add something about
observer perception:

There ARE differences in perception between individual observers (put two
obsevers together at one site and you'll note) which are in principle Lm
independant. They are roughly systematic and I think you therefore can and
should correct for them. One means of doing this is by using the sporadics,
because you can reasonably assume that there is a correlation between your
observed 'fraction' of sporadics and streammembers compared to the 'median'
observer. There are other ways too, of which Rainer advocates the combina-
tion of many observers to counter perception effects. Largely, this is a
matter of conviction and scientific culture and there probably is not one
'sacred' way of doing it. But personaly I have some doubt if the power of
statistics is in sheer numbers, as Rainer argues. Me thinks some idea and
control of the quality of your data (by normalizing for effects known to
influence them as good as possible) is more important than sheer numbers.
Anyway, I don't think the 'right' values pop up by combining many obser-
vers, because of the thing put forward by Peter Jenniskens in his paper on
annual stream activity (A&A 287 (1994), 990-1013): he notes 'a skewness of
the Cp to high values'. Anyway, Rainer already notes that a scarceness of
data can cause problems. Systematic differences between different parts of
a global activity profile due to group-based perception differences become
particularly risky. This is one of the reasons why I (and Peter) have so
little fate in minute, small scale detail in activity curves (subtle sub-
peaks etcetera) and tend to talk more in the line of large scale profile
trends in our analysis.
In addition, I want to make clear that I by no means pretend that the DMS
system of accounting for perception differences is fool-proof or the
ultimate 'all time all weather' solution. They do not perfectly align
results (Jenniskens states that they alligns results to 'roughly within
20%'), but when applying them you can just note that the scatter indeed
becomes considerably less which for me is a sign that things work at least
to some extend, which is better than nothing.

Best wishes to everybody,


Marco Langbroek
(the Netherlands)


PS: To be fair, problems with using the sporadic rates for calibration
might be for example:

- differences in r-value between sporadics and stream meteors. This becomes
a concern when an observer under- or overestimates his Lm. On the other
hand: while not fully correcting it, our Cp system takes of the sharpest
edges of any systematic under or overestimate of Lm by an observer!
- It might be that a faint meteor coming from a radiant has more chance to
be judged as a genuine meteor than a faint sporadic. This is psychology
involved. It leads to different perceptions for sporadic and stream
meteors.

(note to Rainer: I think a similar effect, though the other way round,
caused that 'hick-up' in my sporadic rates in my Nov. 17 Leonid data. When
you positively see many faint meteors coming from a radiant, you might be
more easily convinced that that other faint flash that did not come from
the radiant was a genuine (sporadic) meteor too. I think this might have
triggered my high sporadic detection in that sample of data).


*--------------------------------------------------------------------------*
|     Casper ter Kuile, Akker 145, NL-3732 XD De Bilt, the Netherlands     |
|    Phone: (31)-30-2203170; Fax: (31)-30-2202695; GSM: (31)-6-54723974    |
|                         E-mail: pegasoft@cc.ruudot nl                       |
|         DMS-website: http://www.pidot net/~terkuile/meteors/dms.htm         |
*--------------------------------------------------------------------------*