ESR Radar data for 2002/01/17

Spectra from ESR RTG

Beware that in these plots, and all plots from the ESR rtg, the frequency axis is inverted, i.e. upshifted frequencies are to the left.

	Header	Header	Header	Header
06:45:55 UT:	spectra 1	spectra 2
06:46:00 UT:	xxx	spectra 2
06:46:05 UT:	xxx	spectra 2
06:46:10 UT:	xxx	spectra 2
06:46:15 UT:	xxx	spectra 2
06:46:20 UT:	xxx	spectra 2
06:46:25 UT:	xxx	spectra 2
06:46:30 UT:	xxx	spectra 2
06:46:35 UT:	xxx	spectra 2

Some interferometry results

From the raw timeseries, we compute powerspectra from each of the two antennas, and the cross-spectrum, which, when normalised, gives the coherence, and also the phase. This is explained in more detail on a separate page.

When plotting these quantities for all available ranges, we obtain a colourplot like the figure to the right. Left to right, top to bottom, the panels are:

• power spectrum, 42 m antenna (transmit and receive)
• power spectrum, 32 m antenna (receive only)
• coherence (normalised cross-spectrum)
• cross-spectrum phase

Powerspectra has been computed in (mostly) the same way as is done in the ordinary ESR receiver system, cross-spectra in the same way, but extended to cover negative lags.

Time development

To see how the event develops in time, we have made plots of power and coherence vs. time. For power, we have separated the covered range into three intervals, and extracted the maximum power per unit frequency and unit range from each interval, separately for up- and down-shifted frequencies.

The figure on the right shows the development of power. We can see how both shoulders increase by two orders of magnitude within one second, and how the downshifted shoulder stays higher as the upshifted shoulder weakens, and that the downshifted shoulder remains higher throughout most of the event.

Since we have not performed background subtraction or range² correction, power from different ranges (or range intervals) should not be compared directly.

The figure on the right shows development of coherence vs. time. It is like the figure above.

There are a couple of caveats with this figure also: No attempt has been made at distinguishing spurious (false) coherences from those that are likely to show real coherent scattering structures. From a single coherence plot (like the third panel of the four-panel plots above) it is relatively simple to assess the noise level of the coherence. This is not so easy from this summary plot. Furthermore, coherence is only extracted with no phase rotation (see below). Some of the coherent structures have been seen to be more prominent when phase rotation is applied.

Combined PAI images and radar spectra

Due to the separation of the radar and the camera, the radar beam, when drawn in the cameras view, will move depending on the altitude we map it to, this is shown in figure Beams

We can compare the time development in the radar measurements, with the time development of the averaged intensity/pixel measured through the circle defined by the radarbeam at a certain altitude, to study the correlation between the two instruments. There is a very poor correlaTion in the two timeseries, and a probable cause of that is the separation of the two instruments. By studying the PAI movie sequences, we see that the situation outlined in the rightmost drawing in figure FoV (where the auroral arc ends up within the radarbeam), is occuring between 06:46:20 -- 06:46:22. This can also be identified from the lower panel of figure Comparison, where the broken line (averaged intensity/pixel in a ring surrounding the radarbeam) is much lower than the solid line (averaged intensity /pixel within the radarbeam). This short time period is interesting because we expects some correlation between the radar and optical measurements here. Obviosly, the correlation is still bad. However, it seems like the optical measurements is dragging behind the observed radar echoes (with less than a second). Asgeir reminded me about the green lines having about a 0.7 - 0.8 s. delay time. During this campaign, the PAI was not using any filter, and it is this state very sensitive to the green. Next thing to do is to test the correlation when a delay time of 0.7 - 0.8 s. is subtracted from the optical time serie.