Observations 2024-04-06 17:30-19:30 UTC

The same setup as in earlier observations was used. This time we had KAIRA, NenuFAR, LWA1, LWA-SV, LWA-NA, LWA-OVRO (unfortunately no useful data this time), DE602,603,604,605,609, FR606, SE607, PL610, PL612, IE613, LV614. DE601 was out of order, and the other DE stations used an incorrect pointing (position of Jupiter one month earlier), but still show signals.

Some results were shown at conferences:

Talk by Olaf Wucknitz on 2024-06-07 at the LOFAR Family Meeting 2024

Poster at the 16th EVN Symposium in Bonn, 2-6 September 2024

I selected about four seconds (starting 18:20:45 UTC) of data with good Jupiter signals in most stations for the initial analysis. Data were channelised from LOFAR/NenuFAR subbands by a factor of $5^3=125$ and from the 19.6-MHz LWA bands by a factor of $2^8 * 7^2 = 12544$, which produces a final channel width of 1562.5Hz. A simple FFT was used instead of a PFB, even though the latter is also implemented. In this process the data were shifted towards the centre of the Earth as reference by applying an integer sample shift in the time domain and a fractional shift and fringe stopping (post-F) as phase correction after the channelisation. Small frequency shifts ($\le 1$Hz) were applied to the LWA data to make the channels exactly compatible with the European channels. The software for this entire procedure is not well tested, and all results should be interpreted with care.

For the plots below I uses a polarisation combination that should be close to the dominant circular sense. This is $X+{\rm i}Y$ for NenuFAR and the LWA and $X-{\rm i}Y$ for LOFAR and KAIRA. I only plot a limited part of the band.

Historical note: Fringes on EU, US and EU-US baselines were first seen on the screen on 21st May 2024. The plots shown below were produced later.

The positions of the LOFAR stations can be found at https://www.astron.nl/lofartools/lofarmap.html. Unfortunately KAIRA is not included there, but you can easily find it on google maps at https://maps.app.goo.gl/9DmXfwPr4h3JS6Xm6.

We start with the shortest baseline NenuFAR-FR606. FR606 actually sits within the NenuFAR array (click to show in full size):

The panels show the amplitudes for (top) NenuFAR and FR606 on a logarithmic scale, the phases of the correlation between them and finally a measure for the correlation coefficient (bottom). This is calculated as the phasing efficiency (0 to 1) when integrating the complex visibility over each pixel. The numbers in parentheses on the top right show how many bins in time and frequency are in each pixel.

What do we see? The FR606 data are affected by strong RFI, probably leaking in from lower frequencies. This shows that NenuFAR is much more RFI-robust. In the NenuFAR amplitudes on top, we see the structure of Jupiter's emission modulated by interplanetary scintillation (e.g. dark lanes towards the right).

The phases are very stable, as expected, with the exception of RFI-affected parts. There are about two phase turns per MHz, which corresponds to a residual delay of about $2.2\,\mu\rm sec$. The cause of this is not entirely clear.

The phasing efficiency approaches 0.9 at the strongest parts, which means that the signal strength is well above the SEFD of the baseline.

This is a very long baseline. Here is the full range:

Unfortunately the bright patches at KAIRA coincide with dark patches at NenuFAR, so the baseline does not have much sensitivity. Nevertheless can we see consistent phases, even though at a low SNR, when zooming in:

At this zoom level (only two samples per pixel!) the correlation coefficient is not meaningful anymore, but the fact that we still see phases means that the signal must be very strong.

In the full range (left) we can again see how the modulation differs between the stations. Fringes can be seen nicely in the zoom in (right):

The correlation is stronger on this slightly shorter baseline. In the zoom in we can see how the phases jump at the minima of the modulation.

Here is a long baseline towards the north-east:

This baseline shows the power of interferometry. Even though the PL610 data are badly affected by RFI, the correlations still show nice signals. The zoom in shows another interesting effect: The phases of the narrow pulses (diagonal lines) are clearly different from the phases of the more diffuse emission. This means they must come from a different position on the planet. This can be seen best on the leftmost burst between about 2.3 and 2.4 sec.

Here we see similar effects as on the baseline to PL610:

Here are all three LWA-internal baselines (OVRO did not deliver useable data this time). On the LWA_SV baselines, we can see how the modulation pattern shifts a bit in time, which comes with interesting phase effects on the LWA1-LWA_SV baseline.

In the zoom ins, some fine horizontal striping is visible in the phase and coherence plots for the LWA1 baselines. For LWA_NA-LWA-SV it is also there, but appears wider in these plots.

Some investigation of these baselines and LWA-Europe baselines indicate that this is related to how the delay compensation is done in the stations. If my understanding is correct , the shift is done purely in the time domain for the LWA1 station, which means there are no funny frequency dependent effects. For the other stations, it is apparently applied in frequency space after channelising to 0.025MHz for LWA_SV, and a bit less (probably 196/8192=0.02392578125MHz) for LWA_NA. The bandpass of this process can be seen in higher-resolution plots of the amplitudes. In addition, the phases are very noisy, and the coherence is much lower, in a narrower regular pattern. Plots not shown here indicate that the period is exactly 1/4 of the channel width (of the order 6kHz).

Here we show this by using a frequency resolution that is 4 times higher then before (50176 channels over the 19.6MHz band, channel width 390.625Hz.

The plots for all three baselines show exactly the same frequency range:

In the amplitudes we see the filter bands in LWA_NA and LWA_SV, but not in LWA1. The patterns from the former can be seen in the baselines to LWA1. The baseline LWA_NA-LWA_SV shows a combination of both, with a beating pattern (particularly low coherence at the lowest frequencies).

The filter pattern is very plausible, but why we have this additional finer (factor of 4) modulation of the coherence is not clear to me.

We start with baselines to NenuFAR, because it is by far the most sensitive European station. Residual delays are still large, therefore we use again the higher spectral resolution of 390.625MHz.

In the zoom in, fringes can be seen at all frequencies.

In the zoom in, we do again see the regular pattern in frequency of higher and lower coherence and oscillating noise in the phases.

In the zoom in, we see how the high/low noise frequencies are a different from the ones seen in LWA_NA.

With 7970 km this is the longest projected baseline:

KAIRA is very special, because it is not part of the general astronomical LOFAR array, but mostly used for ionospheric work. Thanks to the tremendous help of Derek McKay, it often participates in astronomical (VLBI and otherwise) observations. The projected baseline length to LWA1 is 6775 km.