What the Buoys are telling

[FishOutofWater]

That's really interesting. The surface fresh water layer has thinned this summer compared to last and heat from the Pacific water layer is moving upwards under the ice. Although the weather has been pretty cool in the Beaufort sea this summer the stormy conditions have apparently caused generalized Ekman driven upwelling as the surface water layer has been thinned.

That's one possible explanation. Perhaps there are others I haven't considered.

The buoy data shows that the melting season may hold on for a while after the weather cools because there is so much stored heat coming up from the Pacific summer water layer. We could have some surprises in September.

[Bruce Steele]

Uniquorn, Could you make a static day profile like post #2178 above ?  For 443910 ?
I watch the profiles t/s contour but it is a broad brush.
It seems ITP #120 is warming differently than ITP #121

Thanks in advance.

[Michael Hauber]

Drift path will make a difference but the point is made. I wonder if we will see the same surface warming with less open water. At 75N maybe not that much difference.

Am I interpreting correctly - the first time the green line is hit near the surface (~10 m) is about day 580, whereas in 2015 is was about day 560 - so 20 days behind?

[Bruce Steele]

Michael, Yes the ~ -1.2 water in the surface mixed layer is showing up later this year than it did (last year) ( my mistake )I think the buoy locations are fairly close for comparing the two years. There is still probably enough heat to melt out ITP 121 but I am guessing,

ITP 85 was 2015 and ITP 121 is this year. So comparison is between those years, not between this year and last .

[uniquorn]

Uniquorn, Could you make a static day profile like post #2178 above ?  For 443910 ?
I watch the profiles t/s contour but it is a broad brush.
It seems ITP #120 is warming differently than ITP #121

Thanks in advance.

I'll let it develop a bit further before running the numbers. I see what you mean though, itp120 is warming from the bottom up and itp121 more top down.

That's really interesting. The surface fresh water layer has thinned this summer compared to last and heat from the Pacific water layer is moving upwards under the ice. Although the weather has been pretty cool in the Beaufort sea this summer the stormy conditions have apparently caused generalized Ekman driven upwelling as the surface water layer has been thinned.

That's one possible explanation. Perhaps there are others I haven't considered.

Possibly warmer saline meltwater falling through the freshwater layer but sitting on the denser pacific layer.

[uniquorn]

Drift path will make a difference but the point is made. I wonder if we will see the same surface warming with less open water. At 75N maybe not that much difference.

Am I interpreting correctly - the first time the green line is hit near the surface (~10 m) is about day 580, whereas in 2015 is was about day 560 - so 20 days behind?

ASI was a very different shape in 2015. a lot more open water in the Beaufort

[Jim Hunt]

Something to experiment with when I have an hour or two to spare.

It took more than a couple of hours, but I eventually got basemap working under Windows using the pre-built binaries kindly provided by Christoph Gohlke at:

https://www.lfd.uci.edu/~gohlke/pythonlibs/#basemap

The result didn't look exactly like yours though. See below.  It was marginally easier to get things working on a Raspberry Pi, but the results were not dissimilar. When I have another hour or two to spare it looks as though I need to experiment with font sizes?

[uniquorn]

Keep at it Jim

Here is a closer look at 300234066897240 recently dropped in the Lincoln Sea.
Tides, inertial oscillations or both?

[SimonF92]

Something to experiment with when I have an hour or two to spare.

It took more than a couple of hours, but I eventually got basemap working under Windows using the pre-built binaries kindly provided by Christoph Gohlke at:

https://www.lfd.uci.edu/~gohlke/pythonlibs/#basemap

The result didn't look exactly like yours though. See below.  It was marginally easier to get things working on a Raspberry Pi, but the results were not dissimilar. When I have another hour or two to spare it looks as though I need to experiment with font sizes?

Jim, I really appreciate you testing this for me, its very useful.

My first point is, the size issue is a problem with the line I have highlighted- try tweaking these parameters- its a little strange to me as to why this happened but i suspect its because those parameters control the "size in absolute dimensions" of the figure. Thus, you may be using a much larger or smaller display screen than me?

Irrespective I need to take it into consideration.

My second point is that your trials and tribulations with getting Basemap installed cannot be ignored. The whole idea of these tutorials are their accessibility. I suspect not many people would have the patience to work so hard to install a library. I just finished a program for my job so now I have more free time i will work on a Cartopy port .



Lastly, hopefully with that line fix, now you should have a working program, id like to point out that you can colour the buoys on any parameter within the data by changing the line:


colours=simb_df['air_temp'].values.tolist()



to any numeric variable in the dataset, from this list:

Index(['wdt_counter', 'program_version', 'time_stamp', 'latitude', 'longitude',
       'air_temp', 'air_pressure', 'bottom_distance', 'water_temp',
       'surface_distance',
       ...
       'dtc_values_186', 'dtc_values_187', 'dtc_values_188', 'dtc_values_189',
       'dtc_values_190', 'dtc_values_191', 'buoy_id', 'snow_depth',
       'ice_thickness', 'date'],
      dtype='object', length=213)


If you wanted to colour on ice thickness for example, the line would be;


colours=simb_df['ice_thickness'].values.tolist()


Dont forget to change the label of the colour legend to reflect

You might not get the results you expect with ice thickness, this is because some days the ice thickness is erroneously calculated and a result bugs the range of the colourmap, ill attach an example image. Considering applying rolling ball smoothing or outlier detection but I always try to avoid data manipulation.

Anyway, hope this helps and thanks for trying to run it

[SimonF92]

While I have it loaded, 441910 and 443910 on the same plot.

The latter was installed on a thinner floe but seems to have suffered less recently, presumably due to its relative position


The final plot is also inclusive of 859790, which is no longer reporting. Jim, you need to add a line of code to the program to use 859790 as it has bugged coordinates. Line shown in images.

If anyone has any advice about how to make these plots less 'messy' then i would welcome it . In these plots it becomes clear why choppy and erroneous thickness calculations can skew the colourmap away from useful ranges

[Jim Hunt]

Curious. My previous attempt at a reply seems to have broken the forum.

Let's see if this one works.

[Jim Hunt]

Let's see if this one works.

It didn't, until oren removed my previous reply. To reiterate:

Further experimentation revealed that the map saved to disk looks very different to the one displayed on my laptop screen. I won't show the image yet, since that might have been the cause of the "issue".

There also seems to be a problem with conversion from the Excel internal date format? Here's my quick hack to get around the Excel leap year bug:

Code: [Select]

simb_df['date']=pd.TimedeltaIndex(simb_df['time_stamp'],unit='d')+pd.to_datetime('1899-12-30')

[SimonF92]

Thanks for the addition Jim, I'll do my DD on your fix and if it makes sense ill add it to the Cartopy port i started yesterday- seems to me currently that it does.

Fyi I dont think the data is ever handled in excel- those are unix timestamps, so the excel bug is possibly not a factor here, but happy to be told im wrong on that

[Jim Hunt]

Uploading a full size image failed again, so this is a version resized to 25% and converted to a .JPG

[Jim Hunt]

Fyi I dont think the data is ever handled in excel- those are unix timestamps

When I loaded the CSV file into Excel the dates displayed properly! Surely UNIX timestamps are long (milli)seconds, not float days?

[SimonF92]

Fyi I dont think the data is ever handled in excel- those are unix timestamps

When I loaded the CSV file into Excel the dates displayed properly! Surely UNIX timestamps are long (milli)seconds, not float days?

Good to know- I will fix that, thanks.

First line here rounds the unix timestamps to just days- so that the plots only cast one reading per day as opposed to several. This is for memory reasons, though im sure uniquorn would disapprove of me not showing the full data  .

simb_df['time_stamp']=simb_df['time_stamp'].round(0)
simb_df['date']=pd.TimedeltaIndex(simb_df['time_stamp'],unit='d')+pd.to_datetime('1900-01-01')

[uniquorn]

Depends what time of day it is. With thickness perhaps it doesn't matter but with air temps it would. If memory is a problem maybe take an average.

443910 suddenly losing thickness at surface. Perhaps another draining event. Bottom melt continues erratically. Trend is between linear and polynomial.

[Michael Hauber]

There is a bit of warm air being pulled into the Beaufort at the moment and EOSDIS channels 3-6-7 look like the eastern half is one of the few places in the Arctic currently experiencing significant surface melt.

[uniquorn]

443910 roughly in the middle  https://go.nasa.gov/3m6a3VR

looked a bit different yesterday on polarview  (or iwsviewer) also centre

[uniquorn]

Chukchi, 052460 bottom distance becoming unstable. Shorter timescale, aug1-17
Jim's map above shows the rough location

[uniquorn]

9 simb3 buoys awaiting deployment

[uniquorn]

052460 had a 3rd event at ice bottom. Here is a colour map overview of the melting season so far. Some kind of temperature inversion at surface today or perhaps a bit sunnier. Enough to warm the vertical buoy but not the ice.

[uniquorn]

443910 hanging in there at 0.63m thickness, a touch faster than the linear loss trend.
Large chart to show small changes in 3m water temp.
Note that bottom melt started on jun15, about 3days after top melt

[uniquorn]

A quick look at Neumayer T86 down south. Cooling from surface down but there are some anomalous large numbers in the recent data.
Haven't sorted out date labels yet, mar17-aug19

[uniquorn]

052460 update. The colour chart runs from aug6-23 attempting to match the drop in surface distance, probably due to snow, to temperatures. The same has been done to the thickness chart. Without the ~15cm of snow, ice thickness is only ~60cm and below the linear trend. Hopefully the buoy is sitting in the middle of an ex melt pond full of snow and surrounded by much thicker ice.

The white horizontal lines are initial deployment thickness on mar9

[uniquorn]

052460 update. More snow, another event at ice bottom. Nearly all the ice is at the same temperature as the water.

[uniquorn]

4 new itp buoys on whoi ftp, they haven't been posted on the website yet.

[uniquorn]

052460 update. Perhaps the ice bottom measurements become more erratic as the ice becomes more 'see through' to the sounder. Seems unlikely that bottom distance is changing so much.

[Tor Bejnar]

I happened on this data source: https://www.ndbc.noaa.gov/
Here (from here) are some current Arctic ship reports (Latitude of -90 would be from the South Pole).
There are Ice Accumulation and Sea Ice Observed options, but not reported by these ships. 
(Ships are just big buoys with people on them, right?)

[Jim Hunt]

Here are some current Arctic ship reports.

I find the "box search" handier. This one for the Northwest Passage for example:

https://www.ndbc.noaa.gov/box_search.php?lat1=70&lat2=80&lon1=-180&lon2=-60&uom=E&ot=S&time=8

I suspect that the ID column explains the apparent demise of SailWX. The individual ship IDs are no longer reported "in the interest of security of the vessels". However you can cut'n'paste the lat/long info into WorldView and then have fun guessing.

[Bruce Steele]

WHOI ITP 120 and it’s associated Simb3 buoy 443910 moved from the Canada Basin unto the Beaufort shelf a couple weeks ago and has now reversed course and headed back out over the Canada basin. The ITP t/s contours clearly show the mixed waters ,surface down to ~ 35 meters ,were colder over the shelf . There is now -1.2 water about ten meters below the surface and because it is in the mixed layer that heat can be mixed towards the surface.
 The heat from the Pacific Summer water is affecting the surface mixed layer and although the salinity difference keeps Pacific water below 35 meters there still seems to be heat moving up into the mixed surface waters from below.
 The 443910 buoy’s position logger has more detail than the ITP 120 buoy and clearly shows the track up onto the shelf and then back again.
 I suspect the ITP 120 and 443910 will melt out .

https://www2.whoi.edu/site/itp/data/active-systems/itp120/

[uniquorn]

Argo float 6904087 surfaced north of Svalbard with a 5m temp of 1.39C on station 15
gif of stations 6-12
https://fleetmonitoring.euro-argo.eu/float/6904087

[uniquorn]

itp120 temp and salinity as it passes over the shelf break and back, day226-250
4.4MB

beige dot is rough location

co-located with 443910, now 34cm thick

[Bruce Steele]

We have some new buoys today. ITP 126 went in with Dartmouth 2021 #3( also new )The ITP has a Sami and PAR sensors. ITP 124 isn’t running profiles yet, stuck at depth. ITP 122 is working but no mass buoy co- located . Also Dartmouth 2021 #2  is new and sending data.

[uniquorn]

new buoy locations. itp128 still to come...

[uniquorn]

After looking at the two new SIMB3's for a while, squashing the buoy (dtc) temps from -2.4C to -0.2C works well with these colours. Basic rule appears to be that cooling from top down is good, cooling from bottom up is bad.

The first two buoys were deployed in floes that melted out though 052460 survives to tell the tale.

551610 and 569620 are new.

[uniquorn]

Argo float 6904087 surfaced north of Svalbard with a 5m temp of 1.39C on station 15
gif of stations 6-12
https://fleetmonitoring.euro-argo.eu/float/6904087

Quite rare for a buoy to make it up the WSC.  1.6C at 3m on sep7

[uniquorn]

I like this float
Argo float 6904087 surfaced a little further north east of Svalbard with a 5m temp of 1.865C on station 17 yesterday

rough location in open water at brown dot

[Brigantine]

I like this float
Argo float 6904087 surfaced a little further north east of Svalbard

Do we know what depth it drifts at in between profiles? Compared with the other floats that drift off to the NW instead

[uniquorn]

https://fleetmonitoring.euro-argo.eu/float/6904087 in tech plots
400-500m in cycles 10-17. I think the default is 1000m

[uniquorn]

443910 west of Banks either slipped in its hole or lost a lot of snow yesterday.
Snow melt seems unlikely

[A-Team]

West Spitsbergen Current core is said to be 300 m. The parking depth of most Argo floats is 1000m but that would be scraping along NW of Svalbard. This float did not round the Yermak branch. The reported pressure in dbar is conveniently ~ depth in meters.

Note there seem to be 4 other Argo floats currently reporting as open source in the Chukchi and Laptev:
https://www.bsh.de/EN/TOPICS/Monitoring_systems/Argo_floats/argo_floats_node.html

Argo floats are autonomous measuring instruments with a length of 1.6 to 2.0 m and weight between 20 and 30 kg. Most of these floats are drifting for most of their time at a depth of 1.000 m, the so-called parking depth. Nearly all floats have a measuring cycle of 10 days which starts with a further descend from their parking depth to 2.000 m. From there they rise slowly to the surface measuring temperature and conductivity of sea water and the pressure of the water column (depth).

Floats descend or ascend by changing their volume, i.e. their density. Therefore they pump mineral oil from their pressure case into an external synthetic bubble under the float or from the bubble into the pressure case. Currently, about 4.000 floats are drifting in the world ocean providing temperature and salinity profiles from the upper 2.000 m of the ocean for marine and climate research. The map shows the positions of all floats which transmitted data within the last 30 days

Published ice and current studies on Greenland's east side go back nearly 900 years to Viking days:

The East Greenland Current north of the Denmark Strait II.
Aagaard, K., and L.K. Coachman
Arctic, 21, 267-290, 1968
http://pubs.aina.ucalgary.ca/arctic/Arctic21-3-181.pdf

A southward-setting current carrying large amounts of ice runs along the east coast of Greenland. In Viking times, the vast extent of ice along the southeast Greenland coast was well known. Thus one reads in Konungs Skuggsjg (written about 1250) that upon proceeding west from Iceland: “AS soon as one has passed over the deepest part of the ocean, he will encounter such masses of ice in the sea thatI know no equal of it anywhere else in all the earth. They extend so farout from the land that it may mean a journey of four days or more to travel across them. There is more ice to the northeast and north of the land than to the south, southwest, and west.These ice floes have peculiar habits. Sometimes they lie as quiet as can be, though cut apart by creeks or large fjords; at other times they travel with a speed so swift and violent that a ship with a fair wind behind is not more speedy; and when once in motion, they travel as often against the wind as with it.” (Larson 1917, p. 138.)

https://oceancurrents.rsmas.miami.edu/atlantic/spitsbergen.html

"The WSC is the northernmost extension of the Norwegian Atlantic Current. It flows poleward through eastern Fram Strait along the western coast of Spitsbergen. A mainly barotropic current, the WSC appears to be predominantly steered by the bathymetry. It is about 100 km wide and is confined over the continental slope, where it reaches its maximum current speed of 24 to 35 cm s-1 at the surface. Because it transports relatively warm (6 to 8°C) and salty (35.1 to 35.3) Atlantic Water, the WSC keeps this area free of ice.

At around 79°N the WSC splits in two. The Svalbard branch stays close to the continental shelf of Spitsbergen, flowing north and east and eventually sinking and spreading at intermediate depths. It travels around the polar basin and returns, after much modification, to the Atlantic with the EGC. The Yermak branch first follows the western side of the Yermak Plateau. Then, north of 80°N, it detaches from the plateau and flows more directly westward. It usually loses its Atlantic Water signal quickly due to enhanced tidal mixing in the area.

The Yermak Branch eventually recirculates southward as the Return Atlantic Current. The warm, saline Return Atlantic Current flows southward along the eastern edge of the EGC to a depth of about 300 m. The sharp temperature and salinity gradients between the two form the East Greenland Polar Front. In addition to the two major branches of the WSC, at least three others have been identified in the literature, but they have not been found consistently in every study....

[uniquorn]

working on combining simb3  569620 with whoi top4 data. Depth scaling needs some work.

[uniquorn]

WHOI ITP 120 and it’s associated Simb3 buoy 443910 moved from the Canada Basin unto the Beaufort shelf a couple weeks ago and has now reversed course and headed back out over the Canada basin. The ITP t/s contours clearly show the mixed waters ,surface down to ~ 35 meters ,were colder over the shelf . There is now -1.2 water about ten meters below the surface and because it is in the mixed layer that heat can be mixed towards the surface.
 The heat from the Pacific Summer water is affecting the surface mixed layer and although the salinity difference keeps Pacific water below 35 meters there still seems to be heat moving up into the mixed surface waters from below.
 The 443910 buoy’s position logger has more detail than the ITP 120 buoy and clearly shows the track up onto the shelf and then back again.
 I suspect the ITP 120 and 443910 will melt out .

https://www2.whoi.edu/site/itp/data/active-systems/itp120/

443910 is likely to go first as 29cm ice may struggle to support a 4m long buoy. Another erratic change in data today so melt out is probably near. Winds are light though

[uniquorn]

https://www2.whoi.edu/site/beaufortgyre/expeditions/2021-expedition/2021-dispatches/dispatch-12-chasing-ice/

For ref, my emphasis

Isabela Le Bras
September 2, 2021

Along the northern stretch of our cruise track, our plan was to deploy four ice-based observatories. These included 4 Ice Tethered Profilers (ITPs), 3 Tethered Ocean Profilers (TOPs), 2 Seasonal Ice Mass Balance Buoys (SIMBs), and 1 Arctic Ocean Flux Buoy (AOFB). ITPs measure ocean salinity and temperature along an 800m wire hanging below the sea ice. Two of the ITPs aboard also included oxygen sensors and stationary Submersible Autonomous Moored Instruments for partial pressure of CO2 (SAMI-CO2). TOP is a new profiler designed to bump into the sea ice and measure the near-ice water properties, which the ITP does not measure. SIMBs measure ice and snow thickness, and AOFBs measure under-ice heat and momentum fluxes as well as atmospheric properties. Co-locating multiple platforms allows us to cross-calibrate them and to better understand the full air-ocean-ice system.

The challenge before us was to find four suitable locations for these platforms. For each station, we needed to find a floe large enough to park the Louis in and walk on safely, with roughly 1m thickness, and as few ridge features as possible. We wanted to space the stations about 100km apart to get good coverage of the area. Since setting up these systems takes many hours, ideally, we would find a spot every morning before breakfast.

Things didn’t look too promising on September 1st in the northwest corner of our clockwise cruise track. The sea ice was sparse and every floe we encountered crumbled beneath the Louis. By lunchtime, we still had not found anything and decided to do an open water ITP deployment to ensure appropriate spacing. Of course, we could not co-locate any other platforms using this method. Our hope was that this ITP would freeze in soon as they have successfully in the past.

September 2nd got off to a more auspicious start. First mate Nick Houle pulled into a potential floe at about 6:30 in the morning and Jeff O’Brien and Cory Beatty completed a sea ice survey before breakfast. They were lowered over the side onto the sea ice and Jeff tested the ice with a pick before stepping onto it. Cory followed in his footsteps and together they drilled through the ice and measured its thickness, reporting back over radio. The first message was “Seven-Five centimeters”. Not quite one meter, but thick enough to proceed with the survey. Luckily, the sea ice at the next locations was a little thicker, 80, then 95, then 85cm. Because we were aiming to recover a nearby ITP that same day, and the sea ice was likely to be thicker to the east, it was decided that this ice-based observatory would include three out of the possible four components: an ITP, a SIMB, and a TOP.

After breakfast the deck crew lowered the gangway and craned over instruments and tools directly onto the sea ice. Five hours, several snow flurries, and a lot of hauling, shoveling, drilling, winching, tying, and final computer-checks later, the station was complete. The day was not over yet, though! We still had an ITP that had ceased functioning to pick up.

Last year, ITP 119 had been too far west to recover, but had drifted back over the Northwind Ridge into the JOIS 2021 area. The Louis was receiving hourly ITP 119 position updates and was able to track it down and dislodge it out of the small floe it was found in by first ramming the floe with the bow and then sending the starboard bubblers at it. Once dislodged, the ITP shot along the Louis and seemed to have a mind of its own, evading Roger Carew, who managed to hook onto it from the hanging (hu)man basket. Another hour later, the muddy ITP119 was aboard and the Louis headed east in search of the next ice floe, filled with tired but happy passengers.

[SimonF92]

working on combining simb3  569620 with whoi top4 data. Depth scaling needs some work.

I can help with that if you like

[Brigantine]

https://fleetmonitoring.euro-argo.eu/float/6904087 in tech plots
400-500m in cycles 10-17. I think the default is 1000m

I see... Some interesting stuff in there!

[Bruce Steele]

Uniquorn, Thanks for a great melt season of watching the buoys. The simb3 443910 seems like it is melted out. There seem to be lots of simb3 and ITP and TOP WHOI buoys working right now and hopefully there will be lots of survivors next spring. 
 The simb3 give us such a beautiful detailed look at the ice . It will be interesting to have the new TOP buoys to close the current gap between the top of the ITP runs at 9 meters and the bottom of the simb3 buoys at 3 meters.
 I keep waiting to see heat coming up from below all the way to the ice.  I think it happens I just haven’t seen it in the buoy data but part of the problem has been that 6 meter gap in the surface water.
 Heat from above via insolation seems obvious in the simb3 plots . So I will be looking forward to next year !

[uniquorn]

Bruce, the TOP buoys measure temp at 2m depth assuming the ice is not too thick. Gives us a chance to compare sensors at that depth if they are located near to each other.

@SimonF92, will be in touch when I get stuck. A bit busy for coding at the moment

Here is a quick look at SIMB3 551610. Surface cooling vs bottom cooling. Thickness is steady at 0.77m

[A-Team]

 It's a tough environment for buoys right now above Banks Island where the ice has been strongly affected by winds out of the south. Note the ice east of Patrick Island forced to pivot from the differential advection. Click to animate.