Oh, Transpolar Drift has stopped working?!? Front cover story!
As mentioned, TPDrifters never post evidence supporting their views, though nothing could be easier. It seems for recent years, there isn't any. They just keep repeating what they were taught long ago in school, even in the face of conflicting current data.
Many floe tracks do exit the Fram, others the Nares, still others via the Barents islands or CAA garlic press (like the single functioning Obuoy), and in recent days, some left to the south out to the Bering Sea (reverse TPD?) per Modis. Many central traces never leave, they simply mill about or stall for as long as they can be followed.
We followed the journey of Big Block over many months up the AK coast before it stalled, broke up and melted out in the warmer Chukchi. Six of the last eight years have shown the same local pattern. The 'goat's head' too was followed for a long time in an earlier winter.
This year, Fram exiting was practically non-existent during the summer and early fall but picked up suddenly around day 298 (Oct 25th). That can be seen on either on Ascat or the Piomas volume export chart that wipneus posted the day before:
Fram export area this fall, winter and spring will be dominated at the 85% level by export of the Kara Sea intrusion, ice that was formed there in late October (TMI or third month ice). Some FYI from the Laptev will also contribute, as it often has in recent years.
Additionally, small but thick areas of Greenland and CAA-flanking ice will exit, along with some deformed central ice. These later losses play a disproportionate role in pinching out trends to zero in multi-year ice age class proportions.
Not a single floe from the Beaufort-Chukchi-ESAS western sector is at risk of TPD transport out the Fram/Barents/Nares over the July 2017 to May 2018 time frame. Disagree? Show me the floe.
It would be great to have nullschool-like sprites tracking ice movement. These feature sprites would initiate, flow along their variable length path for variable lengths of time, terminate upon breakup, melt or exit.
Buoys can contribute multi-year drift tracks but there aren't nearly enough of them. Worse, there are very significant biases in shipboard release position, season of installation and cessation of function.
It is easy to automatically characterize day-on-day ice movements but much harder to do path integrals of the resulting velocity vector field. That's because there's not necessarily ROI handoffs one day to the next.
However fancier (but still sub-optimal) contrast post-processing (up-forum mp4) reveals a near-saturation density of potential sprites lurking in ordinary Ascats. It doesn't seem very dignified to track these manually but in fact it can be done very rapidly in ImageJ. That's because it's primarily cell biology software and those scientists have forever been wanting to track movement of cells and interior stained features, so it's been developed to the nth degree of convenience (in 5D no less).
TPD is ill-posed, boring. It is the sprites product that is needed. From that trajectory database, the rest follow: convergence, divergence, deformation, volume loss by thickness class, etc. From there, comparative mobility measures.
Below, the image flicker from the two past days (before full storm onset) show significant one-day movement of the pack in the direction expected from strong persistent winds, 'ENE' to 'SSW' as the map is laid out. Indeed there seems to be a central swath of deformation, not just bulk translation, between essentially stationary ice. Sentinel-1AB can differentiate between shear and plastic deformation along the interfaces.
Technical note: 16-bit 3-channel Sentinel-2AB revolutionized cryosphere imagery processing (unlike land scenes, ice is basically white on white). Ascat is offered as 24-bit color png but the three channels are identical, thus it is really 8-bit grayscale. The question is, does changing to 16-bit gray in ImageJ options prior to global and locally aware contrast stretching bring any benefit? That could nuance two intermediate round-off steps but ultimately the monitor will be displaying an 8-bit gif. So the gains may be modest except when zeroing on in special small regions where every bit helps.
