[wxqc] Pressure Frequency Analysis

[Evan Bookbinder]

Victor,

If you'll check inland locations you'll see a classic single sine wave 
for a given day. The second peak you're seeing in the afternoon hours, 
given your location, most certainly would be correlated to pressure 
rises associated with frequent diurnal convection/outflow boundaries 
that affect the Keys.

Evan


Victor Engel wrote:
> First, thanks, everyone for engaging me in this discussion. I find it
> all very interesting. Googling around for information on the
> phenomenon, I've found sites explaining it as a result of solar
> heating. If that is the case, then why are there two daily pulses? I'd
> expecte just a single pulse, such as we get with our radiation meters.
>
> I've thrown together a quick and dirty page showing the results of my
> analysis. It is located at http://the-light.com/weather/research/ . I
> don't know if attachments are allowed on this list, but I figured it's
> better to post attachments elsewhere in case there are members who
> don't wish to receive them. Now some replies to others' comments.
>
> Mark said:
> Having a little background in this myself, just about everything but weather
> systems is "buried in the noise". What I mean by that is there is so much
> going on in the normal atmosphere that it is quite difficult to extract
> anything with real meaning.
>
> Victor replies:
> I'm not sure what you mean by "meaning" here, but fourier analysis can
> find such patterns rather efficiently. I would also bet that if you
> saved barometric data as an appropriately scaled WAV file, a typical
> person would be able to discern a humming corresponding to these
> patterns. It's not any more difficult than listening to a whisper in a
> crowd. We can do it, despite the overwhelming amplitude of the noise.
>
> Mark continues:
> Your 12 and 24 hour peaks are likely due to
> heating of the sensor during the day causing some error. You need a very
> good sensor set to compensate for this.
>
> Victor replies:
> I doubt it, especially since barometric pressure is recorded by the
> console located INSIDE the house, where insolation effects are
> negligible.
>
> Mark continues:
> While you see certain peaks for the dataset you tested, the next two years
> may bring something entirely different, and depends entirely on the
> long-term weather patterns at the time.
>
> Victor replies:
> I certainly expect that to be the case for longer frequencies, which
> is why I didn't consider anything past the 177 hour maximum. Even that
> one may be problematic, but I could not ignore the 1/4 lunation value.
>
> Mark continues:
> About 10 years ago I tried making a
> prediction system based on the periodic information which already existed
> combined with polynomials, and tried to extrapolate what is coming next. It
> was a lesson in futility. The code worked, but not the atmosphere.
>
> Victor replies:
> I understand there is a chaotic nature to the data, but not all of it
> is chaotic. I'm simply making a first stab at trying to identify some
> of the nonchaotic data.
>
> Mark continues:
> May I ask where you are getting your dataset from? I may try this myself if
> I can get the pressure data extracted from the Davis data files. It is easy
> enough once extracted to put a massive FFT onto it.
>
> Victor replies:
> I extracted it from the .csv files produces by Virtual Weather Station
> software. It updates the database from my weather station (Davis
> Vantage Pro 2) once per minute.
>
> Gerry said:
> I've not looked at this, but if you have a longer dataset, have you
> looked at the frequency domain with a Fourier analysis and seen what the
> frequency component really looks like?
>
> Victor replies:
> Yes, but only to confirm the results I obtained using the method
> described in the link, above.
>
> Steve said:
> There is a language issue which is confusing this discussion. Consider
> that one one definition of tide at dictionary.com is "anything that
> alternately rises and falls, increases and decreases, etc.: the tide
> of the seasons.". In that light, the diurnal variation is certainly
> "tidal". You appear to be using tidal to mean caused by the moon.
> There is an atmospheric tide, but it is not caused by the moon.
>
> Victor replies:
> No. By tidal, I meant caused by the sun and/or moon. I expect the
> tidal influence of the moon to be present but reduced when compared to
> the influence of the sun.
>
> Steve continues:
> To think about the effect of the moon on the atmosphere, consider that
> the average depth of the ocean is thousands of feet of a much higher
> mass material, yet the tide, in the absence of resonance caused by
> coastal geography, is only a foot or two.
>
> Victor replies:
> I'm not sure what your point is. A feather is equally affected by
> gravity as a bowling ball or even the air. The difference is that the
> air is compressible. Gravity acting on water sloshes it around.
> Gravity acting on air, it seems to me, can act on it in a different
> fashion, stretching it and compressing it instead of sloshing it.
>
> Steve continues:
> There is a lunar tide of the
> atmosphere, but it is negligible. Your numbers actually prove that the
> moon is not the cause of what you observe, if it were you would see a
> frequency of 23.3 hours, yet the rise and fall is very close to 24
> hours. The rise would happen at a very different time every week.
> (I'll get to the 12 hour thing in a moment.)
>
> Victor replies:
> But I didn't say the moon was the sole cause. I indicated it was
> tidal. By the way, I make it closer to 23.2 hours. There is also
> considerable variation in the length of a lunation -- much more than
> there is in the length of a day (which for the sake of this argument
> is the time of earth's rotation). It varies with the fumocy (full moon
> cycle) of about 412 days, that is the beat pattern between the
> anomalistic and synodic months. Given a long enough data set, I expect
> to see a fumocy-related spike in an FFT. Probably very small, but not
> absent.
>
> Steve continues:
> The diurnal variation of pressure is caused by heating in a way, but I
> dislike language that links it too closely. If you take a number of
> atoms of gas, it has the same mass at absolute zero and at a million
> degrees, so you can't say the pressure (which is the weight of the
> atmosphere above a point) changes because of the solar heating. For
> quick proof, consider that the pressure falls in the early afternoon,
> just when the atmospheric temperature is reaching its maximum.
> Simplistically, what happens is the hotter gas expands, and there is a
> shift of mass (and heat energy) from the hotter area to the cooler
> area. The atmosphere becomes less dense in the warm areas.
>
> Victor replies:
> I'm skeptical of this explanation. Were this explanation to be valid,
> there would have to be atmospheric motion traveling around the world
> at roughly 1700 kph! Even a shock wave doesn't travel that fast.
>
> Steve continues:
>
> No, the diurnal variation in atmospheric pressure is well documented,
> it truly is a tidal phenomenon. It is not even buried in the noise, it
> is clearly discernible in any plot of pressure in areas with
> significant insolation:
>
> http://www.findu.com/cgi-bin/wxpage.cgi?call=CW0925&last=480
>
> The barometric variation due to weather is obviously of greater
> magnitude, but the effect is clear to the eye even without Fourier
> analysis. The effect is greater in tropical regions than the poles,
> and in higher latitudes greater in summer than winter, because these
> areas receive more solar heating.
>
> Victor replies:
> If the phenomenon was caused by heating, then on a seasonal basis, I'd
> expect it to be strongest where there is the greatest insolation,
> which does not correspond to the tropics.
>
> Victor
> _______________________________________________
> wxqc mailing list
> Post messages to wxqc at lists.gladstonefamily.net
> To unsubcribe or change delivery options, please go to:
> http://server.gladstonefamily.net/mailman/listinfo/wxqc
>
> The contents of this message are the responsibility of the author.
>