MOSAiC news

[blumenkraft]

We are moving south-west ATM.

Edit: Ah, and apparently we had northern lights.

[blumenkraft]

Geo-colour & I5 band

More south-west movement and new cracks around.

[blumenkraft]

I'm supposed to study PHP. So i did what people do in this situation: Procrastinating by playing around with the RAMMB-SLIDER.

I found what i think to be the best setting to track the Polarstern if you don't like or know how to interpret the infrared bands.

This is M13 (4.05 µm "Fire Detection") & Geo-Color. It is obvious what the clouds are while also seeing the cracks and of course the bright dot that is the Polarstern.

Link >> https://rammb-slider.cira.colostate.edu/?sat=jpss&z=5&im=6&ts=1&st=0&et=0&speed=30&motion=loop&map=1&lat=0&opacity%5B0%5D=1&opacity%5B1%5D=0.6&hidden%5B0%5D=0&hidden%5B1%5D=0&pause=20191225021728&slider=-1&hide_controls=0&mouse_draw=0&follow_feature=0&follow_hide=0&s=rammb-slider&sec=northern_hemisphere&p%5B0%5D=cira_geocolor&p%5B1%5D=band_m13&x=15295&y=15906.7919921875

[blumenkraft]

Yesterday we moved north-east.

[blumenkraft]

Nice!

The latest radar images of the ship’s surroundings

Link >> https://data.meereisportal.de/maps/animations/Iceradar/

[blumenkraft]

The position of the PS via Day&Night band, 60 frames, 51-minute increments.

I chose this band because it's the most complete set of frames.

We are moving very slowly in the wrong direction ATM (North is right). 

[A-Team]

Two significant new offerings at Meereis Portal: the interactive buoy GIS display discussed earlier in #437 and release of sea ice motion radar cam scenes looking out from the Polarstern bridge  (link in #454).

Note the comparable radar on the antique RV Lance digitized an image a minute with a 360º view and range of 9 km during N-ICE2015 whereas with Mosaic the 280º field of view excludes the Ice Cam, reports every six hours (ie at 1/360 the frame rate) with only 5.2 km of range.

Note the 06:00 daily nadir S1AB coincides with the 06:00 oblique ship radar, meaning the two cannot be color-composited though the Sentinel at 200% can fill in the missing ice camp (2nd image).

We looked at marginal ice zone action and a radar cam article back in #83: https://doi.org/10.1002/2016JC012387

Presumably the PS has more sophisticated radar than they're letting out because you can buy better recreational boating radar for a few $k on ebay. The blackout on the Ice Camp could plausibly be attributed to rfi with all the other equipment out there. Dedicated instruments and strain panels monitor conditions there but that data has not been shared except anecdotally (and via S1AB analysis here). 

The full dataset is sent daily over the Polarstern's fast internet (still being denied on Dec 27th) to Bremerhaven where it is bundled into rolling windows of two weeks extent (14 days x 4x = 56 frames served as slow avi movies). The archive goes back to Nov 1st but increments daily. No explanation has been offered for the delayed upload on Dec 15th and first mention on "Follow Mosaic" on Dec 28th.

It's not clear if the Oct 4-31 floe motion is being held back or just delayed. The ice would have been in total chaos back then going by the chaos of the last two months in strengthened thickened ice, below. The minute-by-minute research grade terabyte dataset is not available which is appropriate.

The archive began duplicating the avi with identical and fully interchangeable m4v formatted videos on Dec 15th, reminiscent of the head-scratching duplication of 'mosaic_multisensor' with all-purple sea ice concentration overlays.

The file sizes are kept to a few MB using avi even though they are really just gif slide shows. ImageJ is very unusual in having an avi reader. This opens them as image stacks, with helpful options for grayscale and 180º rotation (to better align with S1AB and other 'greenland down' satellite imagery).

ImageJ is also very good at sharpening the periphery, reducing the overly bright returns from ice near the bow, changing palette, and concatenating bimonthly bites into a single image. Surprisingly converting the avi to forum-friendly mp4 at https://cloudconvert.com give as very helpful further reduction in file size.

The smaller gif at the bottom features the dramatic lead that opened a few km south of the Polarstern on 15 Dec 2019. Moderate benefit can come from contrast enhancement, sharpening and an indexed palette.

The bridge radar is confusing on first view, so set it on loop and (after download) click repeatedly on the double arrow to speed it up. The white scale bar is intrusive. Its central tip is the center of the fixed polar coordinate frame in which the bridge radar is stationary. It isn't clear if the bow-stern axis is also fixed nor if it points north-south. (The integrated navigation system records this information but it is not provided.)

The timestamp would be better with seconds and minutes rounded away, repositioned lower into vacant black space. The stern may wag about causing the view to shift when little is actually happening; the ship is also drifts and rotates quite a bit over the 60-day time frame of the video.

Overall, it is astonishing that scientists have been able to keep equipment deployed on the ice running at all. Actually, we have no idea how Leg 1 data collection went as little can be gleaned from short reports.

Expeditions like this were already being squeezed out of the picture by advanced calibrated satellites, massive buoy arrays and long range autonomous gliders. The ice seems too far gone for the floe platform concept to work.

Just because the ship will be drifting aimlessly until early January (when winds sweeping up from Siberia will finally send it Fram-ward) doesn't mean the ice will be quiescent. Worse, record lows have arrived with -34.7ºC recorded on 19-12-28 at 23:00 utc, making damaged equipment that much harder to repair.

And it seems that we are far short of the 8-10 extreme weather events expected in winter at the Polarstern's location (below or see #368). It's been quite stormy in the North Atlantic but so far extreme weather has not moved up past Svalbard to any extent.

[uniquorn]

drift update dec16-31. Overall drift is quite small over the last 2 weeks.

[gerontocrat]

NSIDC's MOSAIC website gives you a weekly log of where the PolarStern is/was

https://mosaic.colorado.edu/education/mosaic-mondays/mosaic-monday-december-30-2019

Qu: Goosey goosey gander, wherewil you wander?

A: Since October, not very far.

[blumenkraft]

Here for a year.

Bury your hurry.

[A-Team]

NSIDC's MOSAIC website

That has nothing to do with NSIDC. It is educational material for teachers of kindergarten-6th grade put together by CIRES from 'follow mosaic' blurbs. Note it makes no sense to give lat lon out to two positions but not the timestamp when longitude can jump by 4º in a single day.

Since October, not drifted very far. Drift quite small over the last 2 weeks.

Correct. The Polarstern is way behind their anticipated location for this date. This matters, as drift across the Eurasian Basin was one of the three core pillars of the expedition. So far, even including the 45-km wide swath, very little of the Arctic Ocean has been sampled.

The second pillar, evolution of a fixed floe and its surrounding ice over a full annual cycle, has effectively been abandoned as the floe is already in shambles. You can see that from S1AB daily turmoil on the Mosaic floe, from the nrt ship forward-looking radar, and from plaintive postings about downed deployed equipment.

The third, continuous ice, water and atmospheric observational data not obtainable from satellite, buoys or autonomous gliders is in wait-and-see mode as it's not clear what fraction of the time deployed instrumentation has been operational. Since each year of ice weather is one-off, especially these days, the biggest benefit may come from calibration of satellites and better understanding of what autonomous tools are reporting.

The Polarstern will finally be getting a couple days of favorable tailwind (below) but winds are not strong enough to move the multi million sq km block of MYI (Modis/Ascat up-forum) sitting between them and the pole.

I foresee them reaching 87.1ºN (still 335 km short) and perhaps 112ºE which is actually more Fram-favorable as circumpolar drift has been far more frequent than transpolar over the last decade.

On nullschool, checking the 'grid' overlay helps visualize the angle of wind relative to circles of constant latitude. The ice generally moves 'to the right' of streamlines, per Nansen. So it is only in the latter frames that the wind is providing a poleward stress component.

Here the position of the ship needs to be adjusted frame by frame to account for how much drift has taken place as the forecast progresses. Otherwise the wind bearing and speed won't be appropriate to the actual position. Those can be estimated quite well by finding an applicable period of the wind speed and bearing on the 1940-line ship hourly wx report. An iteration might improve the drift forecast,

https://www.awi.de/fileadmin/user_upload/MET/PolarsternCoursePlot/psobsedat.html

The second attachment below provides 168 high precision locations of the PS from mooring until today along with built urls to S1AB and GFS and rotation angles to put north up for the given location (so as to match orientation of the dramatic new observation deck radar).

That archive is being actively kept up though the 1024x1024 size of the rolling avi video needs adjustments to display here. Dec 29th is missing entirely. Hopefully the missing month of October will eventually be supplied. It would be of interest to annotate the jigsaw puzzle action to see if fissures repeat over time and how big events correlate with GFS wind stress.

https://data.meereisportal.de/maps/animations/Iceradar/

Technical note: This unusual radar imagery is sub-optimal as archived, mostly because it is too bright near the bow, too dim overall and too fuzzy on the periphery. These problems are intrinsic to how the radar works.

However because ImageJ has an avi frame reader, complex video editing software is not needed and some yet-to-be determined combination of our usual tools (frame deletion, cropping, resizing, moving of timestamp, adaptive contrast, unsharp mask, gamma-corrected contrast, indexed palettes etc) can improve the product.

[A-Team]

Below, another version of the Polarstern's drift so far based on once-a-day precise S1AB positions from the csv attached above. We are seeing a great many wheel reinventions but few improvements on what Google Earth freeware has long provided in web browser form from a quick template paste and click. The kml file has to be attached as txt for the forum.

The ascat mp4 shows the Polarstern's position within plastically deforming ice. Actually the ice is quite strong and brittle at sustained -30ºC temperatures but appears almost liquid on ascat because over-rafting, ridging and lead opening take up any slack.

The Polarstern's drift has been contained so far within the small rectangular overlay; the last frame shows daily drift positions. While the resolution is much less than S1AB or bow radar, ice motion still appears quite dramatic at the ship's location even at ascat's scale.

Cryo2Smos ice thickness has progressed up to what Mosaic thought it would be back on Oct 1st, three full months ago. The regional buoy array could provide more accurate pointwise data on growth; thickness depends on where and how the buoy was initially installed. Snow depth, if any, slows transmission of cold from the air into the ice.

Boundary conditions, in this instance Greenland, can very much constrain ice response to wind stress.  Right now, with 15 straight hours of 8-10 m/s wind from a ship wx station bearing of 120º, the Polarstern should be (and is) moving poleward and to the west.

However there is a lot of thick ice ahead of it and that ice is pressing up against solid land. This means less actual displacement than if the wind were still at 240º. The wind field is not homogenous however and has been moving ice rapidly along north Greenland towards the Fram Strait (above) which may ease the push-back..

Folding in forecast uncertainty, it isn't feasible to predict ice motion more than 3-4 days out. Broad-brush seasonal weather forecasting may be on the horizon but halfway reliable predictions of trans- or circumpolar drift are lacking, even as hindcasts. Buoy trajectories are largely inapplicable to Mosaic drift because only 0-1 are launched per year in the central Arctic Ocean and those mainly in early August.

Sea ice motion will be a major driver of the 'end game' for Arctic sea ice; our lack of understanding effectively derails all attempts at predicting this from trend lines.

[blumenkraft]

Two decades in one GIF.

60 frames, 51-minute increments.

[uniquorn]

update on the simb ice mass balance buoys. Two of them were deployed onto thinner ice and appear to be having some trouble. Perhaps simb4 has tilted or fallen.
https://www.cryosphereinnovation.com/386850

Also of interest are two meereis Rbuoy charts from buoys with thermistor chains. These show surface temperature, the temperature gradient through the ice and water temperature below the ice. There are 208 sensors spaced at 2cm intervals. It's not easy to tell ice thickness from the charts but by counting the number of thermistors from today's yesterday's surface temperature at ~-32.2C to water at -1.8C they indicate ice+snow thickness of 1.1m to 1.2m
https://data.meereisportal.de/download/buoys/2019R8_300025060111910_TEMP_proc.csv
https://data.meereisportal.de/download/buoys/2019R9_300025060015720_TEMP_proc.csv

MoreSome old info from Mario Hoppman on temperature chains. cffr

[A-Team]

That poster concept above is from M Hoppmann, 05 Apr 2017. It is outdated in the sense the thermistors today do not need the ship LAN and can report cheaply to other satellites than Iridium. I do wonder about "powering by regular lead batteries" which inevitably get dumped as the floe melts.

Chinese researchers have developed better, cheaper thermistor electronics and have deployed over 50 in the Arctic Ocean over multiple years.

https://epic.awi.de/id/eprint/44433/1/Hoppmann_MOSAiC2017_DTCNetwork_poster_small.pdf
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6308795/

Almost looks like they are pushing a small wooden dinghy along rather than a nansen sled. Hopefully the 'remote sensing equipment' being relocated away from a new pressure ridge or open lead (?)does not need the jumble of cables and twisted wires (bottom center) to function.

Temperatures plus wind chill have made it unsafe to take gloves and face masks off to repair deployed equipment, though conditions are not more extreme than those seen (somewhere!) in the continental US this time of year

The Nov-Dec close-up of sea ice motion around the Polarstern is a stretch for Ascat resolution but conveys the instability of the ice pack on a daily level.

Just testing below whether the Meereis Portal m4v format loads to forum display. No. Even though Neven has it enabled. It is just an iTunes DRM version of mp4. They are duplicating their avi files with it. Changing extension to mp4 is supposed to work but does not. Lots of threats posted at meereisportal under German copyright law, not allowed in the US in view of NSF (taxpayer) funding. Avi is best for us because ImageJ has a frame reader only for it.

[uniquorn]

Radar animations appear to have stopped at dec31. Hopefully they will resume.

drift update jan2-3, some data is missing over the new year. Some white space filled with latest polarview, S1B_EW_GRDM_1SDH_20200103T063526_141B_N_1.final-mos. KapD exit trail still visible.

Some may have use for animation data attached

[A-Team]

There's been a good series of Leg 1 blogs up at UA Fairbanks by R Rembler that supplement those of M Shupe and BBC. These represent all that we really know -- or will ever know -- about how sea ice motion is affecting operations on the Mosaic floe.

AWI has censored and silenced all other sources in favor of relentless, uninformative . The Carl Zeiss culture of quality is long gone; Germany today is VW diesel, AWI information management, and 21 of 103 buoys dead within days including 8 set out by AWI and 2 of 6 placed en route by the Kapitan Dranitsyn. An additional 24 floe buoys have been silenced and do not report to the international buoy tracking service.

These buoys aren't AWI's property as they seem to think; the buoys, the ship, all the equipment and the data itself are owned by taxpayers who funded Mosaic. AWI itself does not have one pfennig of its own.

As an aside, note the KD's exit swath is still quite visible today on S1AB. Scientists from Leg 1 just disembarked in Tromsø on Jan 1st; no information is available on what they observed in transit (ice thickness is always measured).

Big plans for daily bow radar release have seemingly foundered: still no coverage of Oct 4th to Oct 31, all coverage stopping on Dec 31, no response at all to polite inquiries from the designated on-shore scientific contacts, web page promises have been shelved, and annoyance expressed by the Meereis Portal communication leader that the (non-forum) public dares question map quality and delays.

Apparently quite a few outside scientists are unhappy with the kindergarten-level portal; they may not be aware Germany invented kindergarten back in 1837.

In other words, download a complete bow radar set now before it disappears forever (or gets locked down in a threatened proprietary format). It takes five to make a complete non-overlapping set: 20191101_20191114.avi, 20191115_20191128.avi, 20191201_20191214).avi, 20191215_20191228).avi and the largely redundant 20191218_20191231).avi. These are small files found at https://data.meereisportal.de/maps/animations/Iceradar/

Open each as grayscale with the free Imagej frame reader. Do not check 'virtual stack'. Under Images --> Stack --> Tools --> Concatenate, join the avi in temporal order, save as avi or gif. In ImageJ, Command-Shift-D pauses to let you to duplicate any subset of frames, for example ones flanking a certain date or all those showing major ice action.

18 Nov 2019 - A big storm By Rob Rember
https://uaf-iarc.org/author/rob-rember/

… On Friday 15 Nov 19, we were notified that the weather was going to get substantially worse over the weekend and that by Sunday winds would exceed gale force (14 m/s). On Saturday the 16th, we quickly went out to sample the first- and second-year ice sites in case our Monday coring was cancelled.

The storm came on Saturday night and continued all day Sunday the 17th as predicted. Winds consistently exceeded gale force for most of the day. The ice opened and closed in front and to the side of the ship several times. At one point from the bridge there was 50 m of open water with waves beginning to form, while on the [starboard] side the ship continued to be moored to the ice floe.

There were large cracks that opened all over our local area (ocean city, remote sensing, ROV site etc.) that stayed open for many hours. We had substantial damage to several power lines as the flow separated and came back together.

This morning the wind is still up around 15 m/s and the damage assessment has begun. We surveyed our coring sites, they are completely intact and can be reached with minimal detours. UAF science has luckily fared well. Other sites will require some major rebuilding with 700 kg power nodes tipped over as ice drifted/sheared and pulled on cables that were fixed to the power distribution hubs.

In the next 30 minutes we will have a meeting to discuss the new steps. The ice is still moving and we are not in a stable situation to begin rebuilding. Even so, the UAF team will head out this afternoon to continue the sea ice coring times series we started approximately a month ago.

I looked at GFS winds, sailwx, ship weather, Rammb, bow radar and seven enveloping S1AB images for confirmation, first noting a 2º longitude lurch in Polarstern position overnight. The big ice radar action came later, on the 19th-20th under continuing strong and strongly shifted winds.

The explanation: ship weather is measured on high and provided as hourly averages so ground level gale force gusts are depracated; GFS 'surface' winds are non-observational so the 1000hPa may be better. Both are in good agreement with Rembler's blog.

Bow radar captures 360º but 80º has been blacked out, notably the entire Mosaic floe and near-ship instrument area observed by Rember whose times, dates and wind speeds may be off because the Polarstern does not use UTC and he does not use the metric system.

I've observed a puzzling non-grasp of causality at Mosaic. Arctic winds and ice motion stress are basin-wide, not local; nothing can be predicted from ship wx point forecasts alone. One glance at the GFS forecast -- a strong, sharply curved cyclone passing repeatedly over the Polarstern at different bearings should have warned them (indeed we called it out here well in advance). It is not homogenous high winds that cause destructive ice motion but high curvature and rapid temporal change in streamlines and their bearings.

Available S1AB: note imagery from the key date of Nov 16th is missing
S1B   2019 11 22 0546   85.7341   120.8084
S1B   2019 11 22 0408   85.7353   120.8119
S1B   2019 11 22 0230   85.7352   120.8192
S1B   2019 11 21 0643   85.7429   120.9425
S1B   2019 11 21 0327   85.7446   120.9824
S1B   2019 11 20 0602   85.7656   121.6001
S1B   2019 11 20 0424   85.7643   121.5311
S1B   2019 11 19 0700   85.8011   120.5766
S1B   2019 11 19 0521   85.8020   120.4943
S1B   2019 11 19 0343   85.8009   120.4512
S1B   2019 11 18 0619   85.8510   120.7805
S1A   2019 11 18 0530   85.8556   120.8340
S1B   2019 11 18 0441   85.8609   120.9092
S1B   2019 11 17 0538   86.0462   122.4998
S1A   2019 11 17 0449   86.0532   122.4685
S1B   2019 11 17 0400   86.0595   122.4405
-+-   2019 11 16 0600   86.1243   120.4086
S1B   2019 11 15 0554   86.1891   118.3766
S1B   2019 11 15 0416   86.1897   118.3522
S1A   2019 11 15 0327   86.1897   118.3386

Hourly ship weather: 60 straight hours of >10 m/s (last column)
86.1   122.4   17 11 19   0000   15
86.1   122.4   16 11 19   2300   16
86.1   122.4   16 11 19   2200   13
86.1   122.3   16 11 19   2100   14
86.1   122.2   16 11 19   2000   16
86.1   122.1   16 11 19   1900   16
86.1   122.0   16 11 19   1800   19
86.1   121.8   16 11 19   1700   19
86.2   121.6   16 11 19   1600   20
86.2   121.5   16 11 19   1500   21
86.2   121.3   16 11 19   1400   20
86.2   120.9   16 11 19   1200   19
86.2   120.8   16 11 19   1100   18
86.2   120.6   16 11 19   1000   17
86.2   120.4   16 11 19   0900   17
86.2   120.3   16 11 19   0800   17
86.2   120.1   16 11 19   0700   16
86.2   120.0   16 11 19   0600   15
86.2   119.9   16 11 19   0500   14
86.2   119.7   16 11 19   0400   12
86.2   119.6   16 11 19   0300   12
86.2   119.5   16 11 19   0200   12
86.2   119.4   16 11 19   0100   13
86.2   119.3   16 11 19   0000   13

GFS nullschool cyclonic winds:
245° @ 11.2  m/s  2019/11/16/0000Z
250° @ 12.9  m/s  2019/11/16/0300Z 
250° @ 14.3  m/s  2019/11/16/0600Z
250° @ 13.5  m/s  2019/11/16/0900Z
255° @ 15.8  m/s  2019/11/16/1200Z
260° @ 14.1  m/s  2019/11/16/1500Z
275° @ 15.0  m/s  2019/11/16/1800Z
290° @ 11.1  m/s  2019/11/16/2100Z 
320° @ 13.4  m/s  2019/11/17/0000Z 
310° @ 10.7  m/s  2019/11/17/0300Z
315° @ 11.5  m/s  2019/11/17/0600Z 
325° @ 11.6  m/s  2019/11/17/0900Z 
355° @ 10.1  m/s  2019/11/17/1200Z 
350° @ 13.3  m/s  2019/11/17/1800Z 
360° @ 14.7  m/s  2019/11/17/2100Z 
360° @ 14.9  m/s  2019/11/18/0000Z
005° @ 16.4  m/s  2019/11/18/0300Z 
005° @ 13.9  m/s  2019/11/18/0600Z 
010° @ 14.4  m/s  2019/11/18/1500Z 
010° @ 10.2  m/s  2019/11/18/0900Z

[blumenkraft]

60 frames, 51-minute increments, D&N band

More of this please!

[A-Team]

if we all clap our hands, Tinker Bell can fly again (?)

Nothing to cheer about here, the Polarstern is drifting into a a long-stagnant zone (per Ascat), getting ever farther behind the eight-ball in terms of hitting their target next October.

Random drift north is not to be conflated with systemic transpolar drift -- the weather pattern for the latter is specific, very different, and yet to kick in.

RAMMB is cool but what is it good for? Not aurora borealis (via reflection on clouds); nullschool has a great tool for that already under its ‘Space’ setting.

Real time Polarstern position? The ship already posts its lat lon online once an hour in real time, that's 2000 times since mooring. It's easy to heuristically predict the ship's position 3-4 days ahead using GFS-ns winds as on Dec 31st in #460.

https://www.awi.de/fileadmin/user_upload/MET/PolarsternCoursePlot/psobsedat.html

Some of the buoys post much more frequently than the 51 minute JPSS satellite and some have 5 dp of positional data. However the public has been denied access to the really high accuracy/high frequency ship and floe buoy data.

We also have four dp precision lat lon posted 2-3x a day at 41 m S1AB resolution attached here back to Oct 4th, 175 scenes to date. RAMMB can’t be used to patch gaps in S1AB coverage (eg Nov 16th) because its archive only goes back two weeks.

2020 01 04 0538  8  86.9848 115.4547
2020 01 04 0400  8  86.9780 115.4551

Determining the maximal resolution setting of RAMMB requires a scale screenshot that includes both the 85º and 87.5º parallels. Those are separated by 698 pixels along the 115º meridian which has length 278.0 km between those two latitudes. Thus the maximal resolution of RAMMB is 398 m/pxl.

However clouds and atmospheric turbulence lead to a 5x5 pixel blur/wobble in the Polarstern's dot. Picking the central pixel gives a 3x3 uncertainty so the actual ability to determine ship latitude is 1200 m, so about 30x worse than the resolution of S1AB which in turn is 2x worse than bow radar.

On 20-01-04 08:00, the PS was at 87.0,115.4 which means 37 pixels between 115 and 117.5, the next graticule line over on RAMMB maximal resolution. That distance is 14.55 km. This means that that the longitudinal resolution is about 0.1º, perhaps a slight improvement on the censored location at AWIwx.

Ship radar images of the last 14 days. Actual radar-image sequence of RV Polarstern from the last 14 days. More videos can be found here. From January, these videos will be available here by a drop down menu.

Live stream from on board: The latest radar images of the ship’s surroundings. Every day, here at meereisportal.de we show from now on a video sequence of ship radar images. The RV Polarstern’s marine radar systems will continue to operate throughout the drift phase and images produced by the system are transmitted to Bremerhaven several times a day.

The imaging system, which shows Polarstern in the centre of the screen, offers vital information on floe movements, deformation and formation of cracks in the ship’s immediate vicinity.

Well, that is a decent plan but higher-ups may have killed that initiative already. It's only a matter of time (days?) before the above text and two month archive are deleted from the Meereis Portal server.

My guess: some antique boomer at AWI is obsessed with putting all Mosaic reports into a thick special-issue journal in 2023. However there are hardly any print journals left now and there won't be any by then.

Like open source buoy data, bow radar can be processed as it comes in, published in a timely way, and contribute to our understanding of icepack deterioration in the Arctic Ocean. In 2023, no one will have the slightest interest in 2019-20 sea ice motion.

Alternatively, all is well and the delay in posting is due to the lead scientist on bow radar only returning to port Wednesday on the KD and just back at work on Monday. I looked at their bow 2017 bowradar paper from the N-ICE2015 five month drift in the very different marginal ice zone (https://doi.org/10.1002/2016JC012387), earlier work cited there, and the 19 forward citations.

They could readily clone this paper for leg 1-2 and then again for legs 3-4 and 5-6. However it's not clear whether ice motion metrics really describe what is going on (see below), nor whether seasonal or year-to-year comparisons can be made, mostly because no previous data exists for the Polarstern's location.

SHEBA, an earlier year frozen into the ice, took place 22 years ago; did the Des Groseilliers even have bow radar then, was it strictly analog, was it filmed, is it archived somewhere online? SHEBA and N-ICE2015 both experienced severe ice motion but more? less? about the same? as the Polarstern is seeing now.

We have a great opportunity to add value to Polarstern bow radar because  it is optimized for the bridge when the ship is underway, not for later perusal. Right now it is over-exposed close to the ship where the return signal is strongest (radial contrast mask?), blurry on the periphery (radial unsharp mask?), higher resolution close in (better detail in 2-3x scale enlargement), offered in grayscale (more interpretable in indexed palette?) and so on.

Still, making a 2x-invert pair below for the tight cyclone of 16 Nov 2019. I don't see a path forward to really describing the complexity of sea ice motion, especially the repeated openings and closings of the same floe line, causing multiple collisions in the same pressure ridge. The extent of ridge venation is just astonishing at the resolution and incident angle of bow radar.

Polarstern helicopters could be imaging the bow radar surface elevations every week with lidar. This is imperative for determining ridge elevation and change. The ROV could be measuring keels from below (though it has scarcely been operable). Lidar data is apparently taken but not disclosed, other than leaked, unscaled, undated, ungraticuled images such as in Rembler's October 23 blog. The blue may just be over-tint on grayscale DEM but with the noxious overlays and scale again not provided(!), it isn't too useful.

The Jan 2020 lidar (airborne lidar scanner ALS) came back with some new cracks and ice dynamics in the heavily instrumented L-site. We have no idea where the site is or where new/old cracks map; no lidar has been released. The L-site is off-limits to bow radar. Its past and current locations have not been released.

Were buoys deployed within the bow radar viewing fame, can we overlay tracks on the Nov-Dec footage, what became of buoys during the storm chaos, is the turmoil seen representative of a larger area?

[blumenkraft]

... Not aurora borealis...

But you can!

[blumenkraft]

Thus the maximal resolution of RAMMB is 398 m/pxl.

They claim:

Select the map and domain to fit the native resolution of the data. The images were plotted using IDL's Lambert Azimuthal projection, which I think has the least amount of distortion and does the best job at maintaining the native resolution of the instrument. The size of the domain was chosen so that each pixel in the resulting image represents the ~375 m resolution of the high resolution imagery bands, while trying to keep the resultant file size down. (These results are animated GIFs over 20 MB in size as it is.)

Note the ~.

Link >> http://rammb.cira.colostate.edu/projects/npp/calval/

[A-Team]

This is the Mosaic forum. If you can find an application for RAMMB, great. Otherwise perhaps start a separate forum for it?

The aurora borealis is off-topic and done vastly better on nullschool; on the Polarstern, people report spending two months at sea without seeing even a single star because the ship has to be so bright. Your image lacks corroborative evidence yet that should have been easily available because the AB view (from OVATION / SWPC / NCEP / NWS / NOAA) updates every 10 minutes or 5x the frequency of RAMMB.

Lightning flashes? Rare in the central Arctic but done infinitely better by a Finnish company detecting weak radio waves. They recently compiled a surprisingly number of AO events out of their vast global archive. RAMMB, can't go back in time. That's a big problem for PolarView too.

Snow vs rain? Yes, with interpretation but that is somewhat a false dichotomy. The Polarstern has lots of snow buoys set out and much fancier ice camp equipment measuring density, internal melt, flooding history, salinity, impurities and so forth. Not going to get at those with RAMMB.

Better time resolution of lead opening and closing than S1AB? Possibly, though RAMMB resolution is not a good match to intrinsic dimensions of ice features. Six-hour bow radar shows how much S1AB itself misses, despite its 30x better resolution than RAMMB.

Even recreational bow radar is digitized once a minute so meereis could be providing 3600x the time resolution or 51x better than RAMMB resolution. Once an hour is probably the sweet spot in terms of file size and correlating with hourly wx data and 3-hourly GFS. However for abruptness of kinematics, a bit of once a minute or continuous analog would be interesting.

Storms? That seems to have been the motivation underpinning RAMMB. For the Polarstern, what we most need are surface wind forecasts and surface roughness. The top priority would be getting ECMWF again; windy dropped the Arctic because there weren't enough clicks.

Transit of relief icebreakers? Yes if only we knew why their progress varied: were they encountering thick ice, stopping for polar bear photos, or putting out buoys? So far none of the ships have made their logbooks or ice thickness data available. Here 51 minutes is better than S1AB burst mode but still not great for velocity estimation. Then there are is the issue of no one digitizing location off RAMMB frames and estimating error with that.

As explained here many times, Lambert azimuthal is used is for its equal area pixels, not because it represents the AO with "less distortion" than conformal polar stereographic (angle preserving). PS, Greenland down, is standard in almost all direct satellite products we use. With netCDF we can interconvert in Panoply.

Lambert thus seems a bad idea because each round of interconversion degrades already blurry data, ie initial PS to RAMMB to PS for overlay say with WorldView night bands or S1AB leads loses detail we cannot afford to lose, the roundtrip is appalling.

The bow radar products above exceeded 150 MB as gifs; 20 MB gifs are a nothing-burger. Maybe RAMMB could offer mp4 and ditch the horrific download code too. A 10 terabyte external drive is very affordable (skip starbucks, make your own coffee and avocado toast).

You might also consider basic contrast adjustments before posting -- two or three clicks can vastly improve clarity of RAMMB animations. For leads and ridges, wavelength matters. If day'n'night is looking at escaping heat in infrared and active radar is looking at roughness backscatter, the two won't correlate that well.

The RAMMB project deserves our applause for putting a lot of opaque satellite channels within practical nrt reach. I am open to it having an application to Mosaic but just don't see what that is. It's hard to find an angle where it is 'best in category'.

[blumenkraft]

This is the Mosaic forum. If you can find an application for RAMMB, great. Otherwise perhaps start a separate forum for it?

Are you serious? A GIF showing the Polarstern is off-topic in the MOSAIC thread? Is it that what you are saying?

The aurora borealis is off-topic

You said it's not visible and i showed a pic where it's visible. That's all. Nothing to be concerned about.

I appreciate your contribution A-Team, really. But with all due respect, i reject your parenting.

[castaway]

Hey guys, thanks for all the hard-working and posting reports here, and the tips to other reports..
It seems data will be unevenly distributed amongst researchers and not all of it may be public.
There should be enough data to many scientists about the globe to produce results, some results would overlap which should be great as to confirm or put in check one's another methodology and rationale. Even though Polarstern data may be broken into pieces and communal science may be harder than necessary, we shall have an interesting time trying to understand what is going on in the Arctic and more about the recent history of planet Earth.

[uniquorn]

Polarstern estimated drift calculated from 10/17 distance between p201 and p207. Full path since oct11 at 16frames/sec. (p207 deployed on oct10) ctr
Some may have use for the data attached as csv

[Tor Bejnar]

Thanks, Uniquorn!

[uniquorn]

The animation above stretches the limits of the forum and my computer and while it provides a 'feel' for drift some detail is lost, so here is a larger static image with a selection of points in red calculated from S1 images as a rough check. cffr
that glitch at 86.6 looks wrong

[uniquorn]

Posting this in the hope that the thread gets more hits
and

[be cause]

  ' boys a boys ' as we often say here in Ireland .. greetings , blessings and  a whole lot of lovings to a few of my favourite poster buoys .. non p.c. b.c.

[uniquorn]

drift update

[uniquorn]

drift update with latlon and distance, a small retrace - see below

[uniquorn]

retrace of the retrace. No big change in the distances. It would be nice to see how the bow radar looked over the last few days.

[uniquorn]

The meereis Rbuoys mentioned in post 463 are called radiation buoys. Here are the latest temperature charts from the 11 working thermistor Tbuoys.
Snow ice thickness can be roughly estimated from the charts, thermistors are spaced 2cm apart.

https://data.meereisportal.de/gallery/index_new.php?active-tab1=method&buoytype=TB&region=all&buoystate=active&expedition=MOSAiC&submit3=display&lang=en_US&active-tab2=buoy

edit: Taking a wider look to view the location of the Tbuoys we can see that the retrace is milder at the periphery, missing t68 completely. T56 looking like it could be quite close to Polarstern.
p201 and p204 left in to help get our bearings.

[blumenkraft]

Polarstern on an express train right now.

[J Cartmill]

'Frightening.' Scientists contemplate the melting Arctic
Some thoughts from some of the MOSAIC scientists:
https://www.eenews.net/special_reports/ice_bound/stories/1061411229

[binntho]

I guess I'm a bit slow, but I just realised that Polarstern is German for the North Star!

[blumenkraft]

Polar star to be precise/literal.

[gerontocrat]

Polarstern on an express train right now.

I think the express train will keep running for a few days more, winds consistently between 6 & 7 m/sec. If anything different, direction will change from towards FJL to towards NE Greenland

[A-Team]

Fascinating to see those buoys doubling back twice in #481. That took place over 34.5 hours between 09:00 on 07 Jan 2019 and 19:30 on 08 Jan 2019. In a subsequent post, uniq's animation will be cropped down to its relevant half-hourly time frame sand compared to the temporal resolution of hourly ship weather, 3-hourly GFS, 2-3x clustered daily S1AB, and six-hourly bow radar.

What caused the double retrace in buoy motion, did it result in severe ice deformation around the ship, was deployed equipment damaged, can these events be predicted in advance, are they anomalously frequent in this year's ice, what are the consequences for melt season?

The forward-looking ice radar is about the only scientific data released during the first four months of the Polarstern cruise, other than ship weather that comes with severely dumbed down location coordinates.

It came as a burst in late December, got updated to Jan 9th, but went into limbo again (despite MeereisPortal reporting downloads to Bremerhaven "several times a day" over the ship's fast Kepler internet (which AWI continues to deny even having).

They did not repair the Nov-Dec record on the occasion of the Jan 9th update, nor provide the missing Oct 04-31 (because frightening levels of thin ice motion are off-message?). The timestamp is to the second, though the radar takes longer than this to complete its 360º sweep (of which 80º are blocked).

There are 4 forward radars visible in moored images, three on the P-deck over the bridge and one on the very top of the mast (crow's nest) that is being called bow ice radar here. This is mostly used for route-finding when an icebreaker is underway, though the Norwegian N-ICE2015 also used it to capture ice motion during the five winter/spring months it was moored to floes.

You won't ever see N-ICE2015, Chinese, or Russian expeditions (like Mr. Federov's year on a floe) mentioned in Mosaic posts or interviews (eg #484), only Nansen and Sheba. N-ICE2015 covered a lot of the same ground as Mosaic and produced two dozen high quality journal papers  addressing every aspect of sea ice a couple degrees north of Svalbard that significantly undercut the originality of Polarstern science (to appear in 2023).

Ice radar (and the secret ship bearing and location) have by far the best time resolution, on the order of minutes. The 2019 radar images have very irregular timestamps generally 6 hours apart but include numerous anomalies such as 35 gaps out of 295 expected, 7 redundancies, 3 one-off times, days missing altogether, and instances that prove the radar capture has excellent time resolution, instances of 3, 5 and 24 minutes are provided below.

09  Nov 2019   12:00
09  Nov 2019   12:03

10  Dec 2019   06:03
10  Dec 2019   06:27

11  Dec 2019   05:57
11  Dec 2019   06:02

The full bow radar time series to date is attached as a cvs for everyone's convenience. I made this by reading the portal's avi into ImageJ (which skips over frame duplications), cropping out the timestamps, tiling them up, and applying free online OCR to get the times to avoid data entry. OCR works great these days. Another application here is to transcribe the daily 'follow Mosaic' blurbs which they've blocked from text-copy.

It is one thing to observe and another to understand the astonishing level of leads opening, leads slamming back together forming pressure ridges, large blocks of ice shearing past each other and all three types of motion occurring together and repetitively at the same and new zones of weakness.

The idea is to look at what all was going on during especially drastic episodes of ice deformation. First though, are fractures almost instantaneous or do they play out over hours or days? If the former, none of our open source tools have sufficient time resolution. Only non-averaged ship weather (eg wind gusts over 100 km/hr) and ship ice radar have the necessary minute-scale resolution.

However it turns out that most of the events play out over 3-4 days. That gives ~14 ice radar frames, 24 GFS-nullschool Arctic Ocean weather displays, maybe 6 relevant S1AB 41 m images, 42 sailwx condition posts, and 72x3 tracking points of the three close-in deformation proxy buoys.

In terms of spatial resolution mis-match, the bow radar is 10 m/pxl so similar to buoy location precision with S1AB at 400% enlargement barely compatible. At maximal scale, nullschool GFS has a resolution of 2094 m/pxl, meaning the green location circle only moves noticeably after major Polarstern displacements like the one underway this coming week, with wind streamlines taking the PS towards the Barents via the FJL/Svalbard gap.

Thus the release of 1-minute bow radar would be excessive but six-hourly is sub-optimal versus  1-hourly (matches weather readings) or 3-hourly (matches GFS releases). The avi media files are not being prepared properly, meaning three times the time resolution need not result in a larger download archive.

The scientific value of the archive would greatly benefit from regular time sampling, disclusure of the radar wavelength, removal of bulk ice pack rotation (ie use of bow-stern coordinates, not ship vote), clarification of UTC vs changing ship time vs fixed CET, radial contrast improvement (too bright off the bow, too dim elsewhere), zonal unsharp mask (to improve focus of periphery, placement of rounded-off timestamps in a dark lower corner, subdivision of the distance scale, cropping to reasonably lit and focused regions, and coordinated ice thickness and surface elevation overflights.

If the white lines indeed represent pressure ridges (or growing surface roughness lineations), there has been prodigious damage to the FYI surface. This could greatly affect the coming melt season if elevation barriers fragment, delay or prevent melt pond drainage, in turn keep albedo low and insolation high. (Alternatively, melt ponds could drain more rapidly if new fractures from ice motion provide drainages.)

The last two-frame animation shows the dramatic complexification of the ice surface around the Polarstern between 01 Nov 2019 and 09 Jan 2020. The four-attachment limit means the whole avi movie has to go in another post.

[binntho]

Polar star to be precise/literal.

Well, yes. But the star Polaris aka Alpha Ursae Minoris is usually called the North Star in English, nordstjernen in Danish but polstjernen in Norwegian and pólstjarnan in Icelandic, étoile polaire in French and polyarnaya zvezda in Russian and réalta thuaidh ("star of the north") in Gaelic. Interesting how some nations are more precise than others (The North Star always points to the north, i.e. 100% accuracy, but the Pole Star points only towards one of the two poles, a disappointing and totally useless  50% "accuracy").

[A-Team]

Below, all 252 bow ice radar frames are concatenated into a single image running from Nov 1st to Jan 9th at 36 fps (50% faster than movie film). The frames are about 6 hours apart. Thus the 70 days (1680 hrs or 100,800 min or 6048000 seconds of actual elapsed time have been collapsed to 7 seconds of mp4 video, which is a speed-up of 864,000-fold relative to a person standing on the bow watching the ice move. Glacial motion of Jakobshavn and Petermann have been depicted many times at similar rates on their respective forums.

The timestamp has been lowered to wasted space next to the bow-stern axis scale. Blurred areas on the periphery have been cropped away for a better fit of file dimensions to forum constraints. Dim areas have been preferentially brightened with linear, gamma and adaptive local contrast correction and focus sharpened with slight unsharp mask.  The initial size of 90MB reduces after conversion of ImageJ to 18 MB avi to a final very reasonable 4.3MB mp4, thanks to free online cloudconvert.com by Munich’s Lunaweb GmbH).

Ice movement can sometimes be seen more easily in a fast inverted image; that is provided at 48 fps with more periphery removed. Be sure to set the controller to loop and move the mouse off the mp4.

One surprising aspect throughout this imagery is that leads close up before they can freeze over, even at very low air temperatures (-30ºC). Open leads historically have been viewed as important direct conduits between ocean heat and its loss in winter to the atmosphere, with refreezing reducing but not totally blocking the effect until an insulating snow cover can build up.

Here none of the leads have stayed open more than a day or two before slamming shut, with refreezing and heat loss having minimal time frames in which to operate. However we don't know how representative the bow radar area is relative to the overall ice pack. Notably, giant lead fields in the Beaufort can stay open for weeks before freezing over.

[psymmo7]

As promised earlier here is an English transcript of the Audio logbook that can be found on the MOSAIC website and is otherwise only available in German. It was posted on the MOSAIC website on Wednesday 8 January, but the recording itself was made on December 14 last year just before the changeover between the first and second leg teams. So it doesn't contain anything particularly newsworthy.

For the record here it is.

MOSAICAUDIOLOGBOOK CHAPTER 9: TEAM CHANGE


Commentator: Artic Drift the Audio logbook.

Markus Rex:  We have now coms as far as eighty-six degrees twenty-six minutes North on the direct route to the North Pole [Ed.: The Polar Stern was at this latitude on either late on 10th or early 11th December]

Commentator: In the Arctic wintery temperatures are gradually being seen. The researchers of the MOSAIC expedition are currently measuring temperatures down as far as minus 30 Celsius. Markus Rex, leader of the expedition explains what further falls in temperature are expected  in the coming days and months.   

Markus Rex: It’s certainly becoming colder as MOSAIC transitions into winter and just in the last few days we reached the lowest temperature record of the season with minus 31 Celsius. Before that the temperatures were mostly above minus 30 or sometimes just touching it. That’s to be expected. The lowest temperatures will be recorded sometime in the period between January and March. The lowest temperatures usually occur in February.

Commentator: In comparison to the expedition of  Fridtjof Nansen, who carried out a similar drift experiment in 1893, these temperatures can be considered warm. The big differences to that time can be traced back to the ice cover and climate change

Markus Rex: Back then Fridtjof Nansen measured temperatures of minus 51 and 52 degrees Celsius at this time of year. Those are temperatures that we won’t observe on this expedition. From all the temperature measurements that we have obtained from automatic drift systems in the arctic we know that temperatures under minus forty-five degrees Celsius practically don’t occur anymore. That’s clearly an indication of change the Arctic is not the same one that Fridtjof Nansen saw then. Our ice is much thinner, it is only half as thick and is therefore very dynamic and full of cracks. The temperatures are higher as well. Climate warming isn’t the first thing that springs to mind, but the change is omnipresent  in the ice and in the temperatures, even if it is  taking place at very low temperatures. It’s not so easy to explain how temperatures below minus 50 could occur in Fridtjof Nansen’s time. After all, even in Fridtjof Nansen’s time the temperature of the ocean under us was only minus one point five to minus  one point seven Celsius, then the water freezes. In other words, we have an enormous underfloor heating system under us an enormous heat reservoir only a thin layer of ice separates from this heat reservoir. The low temperatures are also affected by the ice being thinner and because the ice was thicker in Fridtjof Nansen’s time. He had a thicker insulating layer between the warm ocean and the low temperature air. Our ice layer is thinner, so more heat comes out of the ocean and therefore such low temperatures can’t be reached anymore.

Commentator: The measurements in different scientific experiments are also making progress. Although there is no perceptible difference between “daytime” and “night-time” during the polar night most of the team keep regular working hours. Despite this there are also measurements that are made at night or round the clock.

Markus Rex:
For the most part, we keep usual working hours, simply because the participants also need to take breaks, so we decided to synchronise. But naturally we also want to carry on research round the clock and, in several instances, we are active 24 hours a day. There are scientific reasons for this because there are certain species of organisms in the ecosystem  that have daily rhythms in lower latitudes, migrating back and forth between deeper and shallower layers of water in the course of a day. This raises the question of what these organisms do during the polar night when there is no variation in conditions, when there is no daylight? To investigate that we naturally have to keep making measurements throughout the night.

Our remote-controlled underwater robot is looking at underwater life in the ocean throughout every twenty-four-hour period and I am very curious to know what the results will be. There are really quite a lot of activities that take place during the night. There are automatic instruments that measure continuously anyway and we also have our tethered balloon that takes samples from the 200 metre layer of  the atmosphere above the earth’s surface throughout the night and there are also ocean instruments that operate through the night.

Commentator: The thickness of the ice cover increases continuously during the arctic winter. As a result, larger and heavier equipment can now be brought onto the ice from the icebreaker. This is necessary, in order to prepare for future stages of the expedition, especially when supplies and the exchange of personnel and materials can only take place by air,  which requires a landing strip.

Markus Rex:  As the winter progresses the ice has become thicker. It is still dynamic but, as expected, it is becoming thicker. As a result we have now been able to put our largest and heaviest piece of equipment onto the ice. Our “Pistenbully” [Ed.: also known in English as a snow-groomer] that we use to move other large pieces equipment  and in particular that will be used to prepare the landing strip, that will be used to make changeover between the third and fourth legs of the expedition in March or April. In April, I myself will be here again in April. I have appointed two very experienced colleagues as expedition leaders who will hold the reins in the meantime. In April when I will be back again and  will need the landing strip to get here and the Pistenbully is essential for making it. The ice is now so thick that we were able to put this 16 ton machine onto the ice without the ice creaking and cracking noises and it’s now operating securely.

Commentator: The first leg of the Expedition is now at an end. A new research team is moving into the Polarstern and will carry out the measurements during the next Phase of the expedition. The infrastructural prerequisites for this have been established.

Markus Rex: We have completely finished the construction of the ice camp for our research city on the ice. Every instrument is now in deployment on the ice ..-and working.  We experienced setbacks in the meantime, because instruments fell over or broke down as a result of ice drift causing repeated opening of leads and the formation of pack ice ridges. We have now overcome these setbacks. Everything is functioning and our 30 metre meteorological mast is standing again. Actually, it is only 23 metres high because the rest of it was destroyed, but that won’t affect the measurements, so we will be able to hand over a very well-functioning research city to the participants in the next leg of the expedition.

Commentator: In the course of changing over the teams there was a delay. The Russian supply icebreaker that was carry the new team into the arctic was unable to leave Tromsø harbour because of a storm.

Markus Rex: Just after it set off the Dranitsyn ran into a very, very severe storm. She had to wait in the Tromsø fjord because she couldn’t risk being exposed to the raging waters. There were wind gusts of Storm force twelve. No one would choose to expose themselves to that. So, the  Dranitsyn waited in the fjord for  five days.

Commentator: The Mosaic expedition didn’t miss out on Advent or Christmas either. Despite a shortage of supplies the team enjoyed a sweet surprise.

Markus Rex: On the first and second Sunday in Advent we got chocolate. On St. Nicolaus day [Ed.: 6 December] we all got a little bag with a surprise gift. All of this was extremely pleasing, because the ship had run out of chocolate earlier and we were all extremely happy to get a bag with Christmas cookies, chocolate and all the other nice things. Pretty Christmas decorations were also put up in the messrooms and a Christmassy atmosphere developed.

Commentator: The expedition leader Markus Rex is now returning home will only come on board again in one of the later legs. He is handing over on site responsibility to his colleague Christian Haas.

Markus Rex: Now that the first phase of the expedition has been very successfully concluded, I can return home for a couple of weeks with a good conscience. I will fly back here to take over the leadership again at the beginning of April, possibly even the end of March. I will be with the ship till the end of June. Then I will have another four weeks at home and from the end of July until the end of October I will take over the reins here on board again. In the various periods in between I have appointed very experienced colleagues as expedition leaders and I have great confidence that they will direct the operations here extremely well and competently, so that the expedition will be able to carry on making measurements continuously throughout the year in a uniform manner without any interruptions.

Commentator:
In the next instalment of “Arctic Drift” the new on-site expedition leader Christian Haas will tell us about his impressions and current events in the arctic.

Arctic Drift - the Audio logbook.

[psymmo7]

Some insider impressions here:
https://www.eenews.net/climatewire/stories/1062039461

[binntho]

Scientific American with an article, not all that informative.

[A-Team]

Mosaic released a helpful observational ice thickness/ice surface elevation map in the context of describing a temporary landing strip on the port side. The ice proved thick enough to support the tracked Piston Bully needed to plow it. A longer strip that the Polar-5,6 planes can actually land on will be made in mid-February.

The ice thickness is measured by a towed electromagnetic induction sensor (of salinity); the surface elevation by lidar (called ALS for airborne laser scanner). The thickness was about 1.25 m, what they had expected at the time of the Oct 4th mooring. This is also very close to inferences in posts above based on cryo2smos weekly ice thickness reports (from calibrated satellite freeboard and dielectric).

Note that pressure ridge jumbles are considerably thicker (whiter) but they lack mechanical strength, contain considerable air, and melt very differently from true multi-year ice, a perennial misunderstanding on the melt season and piomas forums.

Once again, the bluish DEM depicting ice heights does not come with a scale. This is unfortunate because it would help with interpretation of oblique bow ice radar. Thin white lines indicate lat lon but they were not labelled! However these are probably 87.0º and 114.º given the date (which itself is uncertain but probably January 6th or somewhat earlier).

The overlay shows bow ice radar of the same date out to 2.5 km (dotted circle). Because of possible rfi (radio frequency interference) and separate dedicated instrumental coverage of the selected floe in the deployment area, the bow radar omits about 80 degrees of its normal coverage. A similar radar on N-ICE2015 provided full 360º coverage.

Mosaic released a second interesting image obtained by TLS (terrestrial laser scanning) which is lidar from elevated ground stations (a 3m tripod mounted on pressure ridges). An area of approximately 1 sq km can be mapped repeatedly with highly accurate ranges, heights and geo-referencing.

Areas behind objects are concealed from the light beam, casting shadows. Once the model is built from combined scans, it can be rotated interactively in open source software within a web page. Only a still is provided here. It's not clear how much of the area covered by bow radar has been imaged with TLS.

The moon has some unfamiliar attributes at very high latitude: it is bright enough under clear skies to read a newspaper and see colors, reflecting well off snow and ice and not setting for two weeks out of each month. The bright moon-lit scene provided on Jan 10th supplements the TLS and contradicts the long-playing sob story of researchers working for months in total darkness.

The ship's directional searchlights can illuminate a broad scene continuously but how far depends on their angle and objects in the way. Regardless, no searchlight or moonlight photos of the bow radar region have been released. With some control points out on the ice, the changing heights of the web of pressure ridges could have been estimated. However this is better done quantitatively from helicopter lidar and em bird as in the first image below.

The growth of the web of pressure ridges between Nov 1st and Jan 9th is shown in #488 above and and 2x enlargement below. How much of daily and special ice motion events was supplementally documented by ALS, TLS or ordinary photography is not known.

Mid-winter on the ice creates special safety issues (ie medevac for serious accidents and health problems) and logistical problems concerning fuel, food, and personnel exchange. These are at their worst in late winter when the ice is expected to be too thick for icebreaker access

leg 2/3 mid-February 2020, commercial Russian icebreaker Admiral Makarov arrives
leg 3/4 mid-April 2020, airplane personnel exchange, no fuel or food delivered: ice too thick
leg 4/5 mid-June 2020, research icebreaker Oden from Sweden arrives
leg 5/6 mid-August 2020, research icebreaker Xuelong II from China arrives

The expedition started on September 20th 2019 in Tromsø; it will finish on October 14th 2020 in Bremerhaven. Allowing two weeks for transits, this results in a full year of experimental coverage (though not exactly of the 'same' floe as it is ever-changing due to months already of almost daily ice rearrangements.)

[A-Team]

Polarstern Update:

-+- The easiest way to follow the incredible effects of a prolonged sub-gale force tailwind on Polarstern drift — westward at almost a tenth of a degree of longitude per hour and a tenth of latitude per day — is to simply scan the hourly online MET columns.

Wind in the 8-12 m/s range with 80º-100º bearing will push the ship towards the FJL-Svalbard gap. But this weather is moderating and perhaps reversing towards the end of the forecast period. The Polarstern has not seen storms coming up from the North Atlantic unlike N-ICE2015 which was beset by them from January on.

Then, by using MET's latest lat lon figures and UTC hour, initialize GFS nullschool, and follow its forecast out a few days with ‘shift-k’. GFS has been surprisingly accurate this fall and winter even though it does not appear to have ingested nor assimilated in situ buoy and ship weather data.

GFS gives an immediate picture — and explanation — of the latest Polarstern drift and a prediction of Polarstern position five days out. 'Passing lows' have limited explanatory powers. There haven't been any of note --  the lowest hPa since September 20th was an unremarkable 992.6 on Nov 16th. (The GAC cyclone bottomed out at 962 hPa with peak sustained winds of 130 km/h.) Drift is determined by near-surface winds, ice sails and edges, and resistance from keels. The current run has to do with strong southern winds off Siberia, not cyclonic lows.

https://www.awi.de/fileadmin/user_upload/MET/PolarsternCoursePlot/psobsedat.html
https://tinyurl.com/uwb9q9p (loads nullschool at 87.3 106.7 on 20-01-13 at 15:00)

-+- Hourly MET reporting from the ship to ECMWF has not improved as eight gaps occurred in the last week, most recently 20-01-12 01:00.  Sometimes gaps can be repaired from sailwx. The problem is not plausibly in collection (sensors down) but with failed transmission (ie no redial).

-+- Bow radar coverage resumed out to midnight on Jan 12th but with 10 gaps in the 6-hourly coverage and only 4 new images and no infill of earlier gaps. Significant new ice shearing can be seen on the central-starboard side. Wind speeds have been fairly high but steady with no noticeable gradient or curvature.

Perhaps certain ship operations are incapable with ice motionimage capture and the associated radar operations have higher priority. It is certainly not storage — a single consumer-grade 256 GB SSD card could capture an entire year of images at one minute intervals. Transmission ashore cannot be an issue either given the pipeline. The Polarstern is not actually capturing that much data — all of leg1 fits on a single 20 TB external drive. Samsung sells a 768 TB rack of SSDs, enough for a half dozen Mosaic expeditions.

The first compound mp4 shows ice motion that preceded the remarkable double retrace that Uniq documented with three close-in buoys in post #481 above. The dramatic buoy drift reversals surprisingly did not have much effect on ice deformation in front of the bow. Instead a major ice movement cycle occurred a few days earlier as the buoys and Polarstern were drifted smoothly along. This suggests that the causes of ice motion are complex, not entirely local and possibly regional (or even basin-wide) meaning knowledge of the near-ship stress field will prove insufficient.

This early January sequence of events makes for an interesting case study of shear sliding and sticking. The lower section C is largely stationary; the upper section A is quite active. The B-B’ unit initially moves back and forth between A and C but later B’ gets stuck on A. As A-B’ move rapidly to the left, B’ slams into its former partner B and deforms.

-+- S1AB radar coverage resumed on 13 Jan 2020 at 03:35 after missing Jan 8, 9 10, and 12 (making those dates very unfavorable for joint coverage of ice motion). This is the 189th image in the 84 days since mooring. The Polarstern is currently not far enough north at 87.3º to be in the pole hole of Sentinel satellites at 87.8º which is ~60 km closer to the pole.

The exit track of the Kapitan Dranitsyn is still highly visible. It arrived on Dec 12th and left the 17th at 14:30 ship time  which is 12:30 CET which is 11:30 UTC [not making this up]. The KD possibly returned in part via the same lead it created coming in.

At prevailing temperatures, the entrance lead would have refrozen over the five day stay at the Polarstern but the return trip might have damaged the newly formed ice. The KD's beam is 27 m wide. Pack ice compression episodes should have closed the lead by now leaving a pressure ridge composed of re-jumbled new ice.

This unusual sequence of events has left a white scar on Sentinel-1AB radar that represents strong radar reflection from rough-textured ice (Mie rather than Rayleigh elastic scattering). The line can serve as a detector of shear and rotation at a regional scale far from the Polarstern. This can be seen already in the Jan 1st to Jan 13th comparison below: an left-center offset is visible.

[uniquorn]

drift update jan6-14, less meandering. S1B from jan13.
drift speed was incorrect on #481 edit: checked again, is good

[A-Team]

drift update jan6-14, less meandering. S1B from jan13.

Looking at sailwx, the GFS wind forecast driving Polarstern drift, the pole hole beyond 87.54º associated with Sentinel's orbit and swath, and the competition for its operating mode, it is safe to say we can say goodbye to open source satellite coverage for the foreseeable future.

The collection of 191 image links and precision Polarstern locations for the expedition to date are attached as a csv. The records also direct to the GFS of that date at the appropriate lat lon.

The ship's radar has also fizzled out again. While restricted to an area of ~20 sq km of a 7.1 million sq km  Arctic Ocean, it provides by far the best picture of local ice motion. The quality seems worse than similar radar on the RV Lance and some of the processing steps here are ill-advised.

It's not clear what is causing the erratic stop-start coverage. It's at the point of ruining the scientific value. This was not some impulsive impromptu hack of navigation radar ... capture is explicitly discussed in the pre-trip planning document:

In addition to the installations in laboratories in the ship, different kinds of instrumentation are installed at suitable positions outside on deck and along the railing of Polarstern. Most
prominent are opportunities to use the P-deck (observational deck) and the crow’s nest for
installations that require a view of the sky or access to higher altitudes:
• Radar for sea ice movement and deformation...

The scenes intrinsically blur and darken in proportion to their distance from the rotating mast radar beam. The former can't really be fixed by unsharp mask and the like. However contrast in the over-exposed bow area and under-exposed periphery can be evened out. Usually a radial transparency mask would be used. However it's easier and works about as well to apply a bimodal contrast LUT followed by mild adaptive contrast even though that doesn't act radially. The initial underlying image is unfortunately zoned, the scale bar off-center, blacks not black and so on. The goal here is to retain image quality after a 2x enlargement and improve clarity farther out from the bow.

[A-Team]

Meereisportal posted another drift segment colored by velocity using software not up to the task: there is no indication of time or dates and dithered colors are uninterpretable (ie do not correspond to anything in the palette).

The color jumble may be due to using one-minute GPS time resolution which is much more than can be displayed. They may have fitted a rolling 4-point cubic curve or spline and then colored continuously, with the jumble resulting from downscaling compression.

Why not just use GoogEarth where color and segments can be controlled without code and the data is shareable as kmz? Or Uniq's animation which is a lot livelier and carries the essential data within its frames?

The main use of one-minute GPS time would be determination of fine-scale acceleration (slope of the velocity curve). The icepack's inclination is always to decelerate because of frictional drag of ice keels with seawater underneath. This resistance goes up like the cube so whatever the wind, the icepack will top out at a very low terminal velocity, perhaps ~1.5 km/hr.

Since Mosaic measures the near-surface wind at co-temporal intervals, that could have displayed on the same graph, not as separately but integrally as a height. We cannot fix the graphic because Mosaic doesn't share either parameter at the resolution at which they are measured, ie the sailwx wind is an hourly average.

Ship drift velocity cannot respond instantly to an applied force but only with a lag as the icepack gets up to speed. Wind direction is an additional complexity factor. It's been steady since 01:00 on 20-01-09 which presents a great opportunity to isolate variables. Is the drift independent of wind direction or do issues like immovable land boundaries along the CAA give rise to anisotropy?

In terms of scientific co-visualization, ship and wind bearing need to be displayed cyclically. That is, 360º and 0º are the same, not opposite extremes as on the MET graph. That suggest a color wheel ROYGBV where red and violet are perceptually equated, the wheel being displayed as a linear background gradient along the time axis. The more common wind rose display has lost track of its time coordinate. However looking down the time axis, the bearing could be displayed as angle, wind as length and velocity as color.

Despite the coming unavailability of high resolution lat lon from Sentinel, the three close-in P-buoys provide the Polarstern's location as shown in the base frame below. Assuming the geometric relations between the ship and buoys are constant, it's still tricky to extract future lat lon numbers for the PS. Graphically, circles of fixed radius from P201, P204 and P207 intersect at the PS.

The mp4 shows that the mad dash west and north will morph into a strong move back east and south on the outer highly curved edge of a moderate cyclone.

The overall ice drift on the Eurasian side has been indecisive this year; Fram export has been good but is originating on the eastern Canadian side of the icepack. The Kara Sea is injecting its ice north of the FJL-Svalbard line again but not as much as a year ago.

[uniquorn]

Drift of 1.4km/h seems high at ~87.3N. The highest I could find from nearest buoys reporting drift speed (in m/s) are below. edit: 0.212*3.6 is only 0.76km/h

P201   2020-01-13T10:30:25,87.2753,107.3853,0.211
P204   2020-01-13T10:30:25,87.2957,106.7855,0.211
P207   2020-01-13T09:30:27,87.3551,106.6847,0.212

buoy data here

Highest P204   2019-11-17T23:30:26,85.8660,121.2518,0.424m/s    1.53km/h

Something for spreadsheet analysts perhaps.