Re: [APML]: Atmospheric Turbulence

[patrickg@nosc.mil (Patrick W. Greenlee)]

>Since my discussion with my meteorologist friend, I've been thinking about
>how data
>such as that could be attained. If I had a perfect mount, and my scope had
>perfect
>optics, then I could just put a star on my CCD chip and record the (x,y)
>locations
>without issuing any guidance corrections using CCDOPS' track log
>functionality. The
>times series of x-data and the y-data could be pushed through a FFT, and
>the spectral
>density functions estimated. This would show the strength of any harmonics
>in the
>data,

More importantly it should show the time varying nature of the seeing.  Any
spikes and harmonics are likely to be artifacts of the equipment as you say
or FFT algorithms and aren't really the object of interest.  There are
various sources of disturbances in seeing with different spectral content.

>but if John and myself are right, the wandering about of a star image due to
>atmospheric turbulence is truly random, so the spectrums should have no
>spikes (ie, no
>significant cycles in the data).

Truly random?  Now there is a topic.

Not having spectral spikes is fine.  What is needed is a quantified
characterization of the variation.  Recall the old adage about, "you can't
control what you can't measure"?  Of course measuring something is only
necessary not sufficient.

Now, about random.  If you were to measure the variations in height above
mean sea level of the surface of the water at a particular location it
might appear random at first.  Then you might notice the tides, then later
that there is a modulation envelope for the tides related to the sun and
moon, then later a seasonal effect to that, then a weather (barometric
effect), then a weekend and holiday boat wake effect, not to mention the
waves from distant storms, and the high freq. low amplitude effect of rain
making small ripples and on and on.

I don't suggest there are lots of similar periodic effects.  I do suggest
that there are likely several sources that sum to effect seeing and that
their contributions may have different spectral characteristics.  Measuring
them is a good step toward understanding or compensating for them.  Random,
well maybe, but all random isn't created equal.  At any given instant,
seeing may be the sum of several different pairs of dice with different
numbers of dots on them which are being thrown at different intervals.

> My mount certainly isn't perfect, nor are my optics,
>so if I would actually do this, I'm sure I'd see spikes, but these would
>not be due to
>seeing but rather to the periodic error of my drive. I was intending to do
>this one
>day anyway, to quantify the periodicity of my drive and really get a
>handle on how
>often I should be issuing guidance corrections.
>
I bet this would be of interest to several techies lurking quietly around
the edges of this forum.  Please give it a whorl Steve and or anyone else.

Thanks for the idea of how to get data.  Unfortunately I don't have the
means at hand of performing the experiment, so for now I will have to shout
encouragement from the sidelines.

Thanks again,

Patrick