Electric field components plotted with coning angle (green curve)
and UxB electric field (blue curve): meridional wind for
westward component and zonal wind for north-up component.
Upleg:
E west
E north-up
Electric field components plotted with coning angle (green curve)
and UxB electric field (blue curve): meridional wind for
westward component and zonal wind for north-up component.
Upleg:
E west
E north-up
Downleg:
E west
E north-up
Below are plots of the perp-B electric field assuming
E*B = 0 (black curves) and E*B = 4 mV/m (red curves).
I chose 4 mV/m to show that the oscillation change
seems to be small even for large fields.
E west
E north-up
Next are plots of the electric field with (only) the UxB component of the electric field in the westward and north-up direction plotted in red. The wind speeds are from the Larsen TMA rocket, B = 0.32 G.
E west
E north-up
Note: I think that E*B ~= 0 for this data, since the component
of the parallel field measured by the crossed booms is E_par*sin(phi),
where phi is the angle between the payload axis and the magnetic field.
The phi is about 20 degrees on the downleg and about 110 degrees on the
upleg, so sin(phi) is small for the downleg data, and big for the upleg.
This would mean that the magnitude of the field oscillation on the upleg
should be different than the downleg, which doesn't seem to be the case
(see the figs. in my thesis).
FFT of wave region (black line) and comparable FFT taken below
wave region (red line):
fft_v12_v34
FFT of each downleg channel with k^-n power law; dashed line
is n=2.5, solid line is n=3.5.
fft_v12
fft_v34
Plot of FTP RMS current (big red blob), V1S data (green line),
and gaussian fit to Arecibo radar data (black line) on:
upleg
downleg
X axis is altitude, Y axis units are arbitrary.
Plot of drift wave length scale (grad(n)/n) and wave activity
on channels:
V34 upleg
V12 downleg
V34 downleg
X axis is altitude, Y axis arbitrary units. Maximum growth should be
in region where grad(n)/n is negative.
V12
V34
V12 & V34 (wavelet power only)
Similar plots on upleg, altitude range: 93.4 - 92.7 km:
V12
V34
V12 & V34 (wavelet power only)
V34 5 Hz component
V34 10 Hz component
V34 15 Hz component
V34 20 Hz component
V12 5 Hz component
V12 10 Hz component
V12 15 Hz component
V12 20 Hz component
Wavelets vs. FFTs....
Fourier basis functions are localized in frequency but not
in time. Small frequency changes in the Fourier transform will produce changes
everywhere in the time domain. Wavelets are local in both frequency and time.
(from a good
wavelet page, with the IDL code I used to generate the following plots.)
"...One possibility would be to do a running Fourier transform (FFT), using a certain window size
and sliding it along in time, computing the FFT at each time using only the data within the
window. This would solve the second problem (frequency localization), but would still be
dependent on the window size used. In addition, a windowed Fourier transform introduces
problems with aliasing of high frequencies onto low freqencies, and relies on the assumption that
the signal can be decomposed into periodic components.
Wavelet analysis attempts to solve these two problems by decomposing a timeseries into
time/frequency space simultaneously. One gets information on both the amplitude of any
"periodic" signals within the series, and how this amplitude varies with time. "
Comments on Data:
Plotted below are V34H power spectra on downleg; horizontal axis is altitude, vertical axis is frequency.
V34H downleg, freq. vs altitude (short altitude segment).
V34H downleg, freq. vs altitude (long altitude segment).
Next pass at line spectra: I've been more careful windowing data so as to only include sporadic E activity in the "peak power" plots. Background spectra are taken a few spin cycles above the wave activity. It appears that the peak power is at ~10Hz, with harmonics (or some other modulation) at ~20Hz, ~30Hz and ~40Hz. The wave activity occurs over such a short time scale that the peak frequencies (10Hz, 20Hz, etc.) are very sensitive to the FFT window, and tend to shift a little, depending on where I define the start of the wave activity.
Downleg only:
Average of all four HF channels
X axis is altitude, Y axis is frequency, Z axis is power. This was kind of a neat way of displaying frequency power vs. altitude.
A couple of interesting features: The 100 Hz noise is much larger below the sporadic E layer (at ~93 km). This effect is not as noticeable on the upleg plot because of some ACS-related noise and (probably) noise from payload outgassing. The 100 Hz stuff also reaches a minimum on both the upleg and downleg at ~101 km. There's also some low frequency stuff around ~107 km on both the upleg and downleg.
The upleg spectrum is much noisier, and the amplitude of the 100 Hz signal is higher than the downleg spectrum. Features: the sporadic E wave activity is spread over a larger altitude range than on the downleg; I'm not sure whether it's most likley physical or instrumental. There's also some coning mixed in with the 100 Hz signal on the upleg that's not noticeable on the downleg; the 100 Hz signal is proportional to the total electric field strength, and since the v x B component is large on the upleg, the 100 Hz signal is also.
