If you examine the barycentric spike multiplets, you'll see that they're obvious RFI - lots of spikes at about the same frequency, spread across different sky positions.

If you zoom out to a longer time range, you'll see that the RFI drifts in frequency; for example, look at this waterfall plot.

The RFI probably drifts because it's coming from a moving object like a satellite or airplane. In any case, our "zone RFI" algorithm doesn't detect it because of the drift. We're currently thinking about how to recognize it. We'll probably use approaches previously used in SERENDIP.

Graph in freq/time coordinates. What about dimensional coordinates (direction)? In post you say signals come from different points on the sky. So, this time/freq graph includes signals from all sky (visible from Arecibo at least) ? If not, how one tell from the picture what part of sky used?
In short, there is obvious order in spikes plotted freq vs time. But how one could say if it's RFI or not just from this picture? How to add direction info to this plot?

---

SETI apps news
We're not gonna fight them. We're gonna transcend them.

Presumably, the only way to demonstrate that this is really RFI, is to check if the same drifting signal was detected simultaneously in some of the other 6 beams?

---

Not necessarily - if the signal is coming from a satellite it could be detected in only 1 beam.

The telescope moves fairly slowly, so signals in a small time range are also in a small position range.

The right-hand graph (sky position) shows how much the position changed during the time period. You can see, for example, if it was tracking a point in the sky. If it was, a column of signals at the same frequency is not necessarily RFI.