[Prev][Next][Index][Thread]
Re: (meteorobs) Sporadics and wavelets (another awkward question)
Lew Gramer wrote:
> I recently reread this email, and thought about the correspondence that I've
> seen since then in meteorobs and WGN about sporadics. So my question is: is Dr.
> Meisel's way of viewing sporadic variations the most valid one procedurally? If
> not, what other approaches are researchers advocating nowadays?
There are a number of approaches to detecting the nature of a cyclic
phenomena. I think Fourrier analysis was the first and has remained
popular because it can handle both continuous and discrete phenomena.
For those unfamiliar with the terms, the old fashioned analogue pen
recorder gathers data continuously, where as recording data values at
various moments in time is discrete sampling. They both have rules
about the degree of fidelity and the implications of experiment design.
The nice thing about discrete sampling is that it is appropriate for
computer analysis.
Note that discrete analysis can be done with missing samples, abritrary
numbers of smaples per cycle, even varaible times between samples. You
can impose more stringent requirements to speed up analysis via
mathematical short cuts.
One requirement is that you need to know the duration of a single
cycle. I would guess that a year is one cycle worth analyzing,
especially for sporadics, but you can always test the data after you
have gathered it to make sure.
I believe much of the software that you need is available on the S&T
site if you wish to persue this further. Please let me know if you do,
because if you don't have acess to a basic to quickbasic translater and
a quickbasic compiler, I do. Compiled code runs much faster. My
compiler is for dos, but you can also buy one for Unix.
> More importantly, if Fourier/wavelet techniques ARE a valid approach, then what
> data gathering and data gleaning techniques make for the best input to such
> analysis? Is there anything in the current IMO gathering and publishing regimes
> which would prevent IMO data from being useful for sporadic analysis like this?
The more precisely the time is measured, the better the opportuniy to
spot minor showers. For analyzing sporatics, I suspect knowing the time
+/- 24 hrs would be adequate to show general trends. I would certainly
welcome professional advice on the matter to ensure that the data is
useful. In view of the variety of meteor detection techniques,
correlation among them could be a big problem. I would suggest a limit
of 1 or 2 techniques. Radio detection with computer recording of data
could generate enough data to analyze on a site by site basis, and then
correlate data among sites just in case any site suffered from say a
loss of sensitivity etc.
If you are trying to detect minor showers, it might be wise to correct
for the daily variation.
Regards,
John
--
John Ohrt, Regina, SK, Canada
johrt@ibmdot net
References: