When K6PZB and I first started up on 2m and 70 cm WSPR, we immediately saw interesting "Doppler" components in the waterfall displays. Sometimes, but not usually, these were such that signals wouldn't decode due to the QRM from the aircraft-scattered signal (ACS). We also saw some strange double-sided signals which we couldn't identify.
Now that a larger group has been on 2m WSPR in California for a couple of months, we have had a chance to both examine ACS better and also to study examples of these double-sided signals. I am becoming of the opinion that these are due to wingtip vortices being generated by large commercial aircraft. These pressure waves are well known and can be a problem for later aircraft. http://en.wikipedia.org/wiki/Wingtip_vortices
Because they are a cause of significant drag to powerful aircraft, a lot of energy can go into producing them and they can persist for many seconds or even several minutes after an aircraft passes by.
It is my suspicion that the large pressure differentials produced by these "curlicue pumps" may be sufficient to cause a sufficiently large and intense change in the dielectric constant of air such that significant 2m energy can be scattered from them. Since they are both (counter) rotating and moving it seems plausible that they may be presenting the two-sided Doppler components visible in the attached SignaLab shot of the strong 2m WSPR signal from K6PZB, located only 8 km to the west. The precise mechanism of the scatter, the sources and other details isn't clear. At this point, I'm only speculating. It does almost look like the aircraft is generating rapidly circulating volumes of air that can even criss-cross at times and go from net-pathlength-increasing to net-pathlength-decreasing vectors. This certainly could happen when oppositely moving WTVs cross the line between the 2m WSPR stations. There are also other possibilities.