Support the Arctic Sea Ice Forum and Blog

Author Topic: HYCOM  (Read 4850 times)

interstitial

  • Frazil ice
  • Posts: 472
    • View Profile
  • Liked: 162
  • Likes Given: 70
HYCOM
« on: June 27, 2020, 03:50:42 AM »
I think this model deserves its own thread. I was using another thread but I want to focus on this model. Some experienced users don't think it is very good but I want to follow it for a while and see. Because there is conflicting opinions on its value I decided to keep it off the main melting thread and it doesn't really fit elsewhere. To get started if you compare 2019 on 6/25 and 9/25 final area roughly conforms to anything black or thicker on 6/25. I would like to see if 2020 follows the same pattern. In a few days I will compare other years.

binntho

  • Nilas ice
  • Posts: 1388
    • View Profile
  • Liked: 453
  • Likes Given: 106
Re: HYCOM
« Reply #1 on: June 27, 2020, 05:53:04 AM »
I am surprised at how well the final extent in the top image (25/9/2019) fits the black thickness boundary three months earlier. So many times have I seen people saying that "everything at such and such color will melt out" but this is the first time I've actually seen anybody doing this sort of comparison after the event.
because a thing is eloquently expressed it should not be taken to be as necessarily true
St. Augustine, Confessions V, 6

interstitial

  • Frazil ice
  • Posts: 472
    • View Profile
  • Liked: 162
  • Likes Given: 70
Re: HYCOM
« Reply #2 on: June 28, 2020, 06:30:06 AM »
I have provided a 4/15/2020 map of Piomas vs Hycom as requested. I Also included worldview.

North of Svalbard shows a fair bit of 4 meter ice on Piomas where Hycom shows 2.5 meter ice. In the Beaufort along the Alaska coastline shows a few small pockets of 3 meter ice on Hycom and it does not show on Piomas. The CAA side shows a fair bit of 4 meter ice on Piomas where Hycom shows most of that ice as 3 meter ice with a tiny amount of thicker ice at 4 and even 5 meters. The main body of ice in the Beaufort is between 2 and 3 meters thick according to Piomas. On Hycom that ice looks to be 1.4 to 2.3 meters thick. That is a large volume difference. Near Wrengel island Piomas shows a thick arm starting at 4 meters and thinning to about 2.5 meters. Hycom shows 1.7 meter ice near Wrengel island with the thicker arm shifted south between 1.8 and 2.5 meters thick. Piomas shows a hump in the Greenland sea of thin ice that I could not find on worldview and does not show on Hycom. Overall Piomas shows most of the central Arctic basin being from 2 to 3.4 meters thick. In the same area Hycom shows between 1.4 and 2.3 meters thick.

Most of the differences in shapes can be explained by the difference in resolution between the models. Many of the features are similar though their are some clear differences as noted above. Hycom uses more colors in their scale making a more precise visualization of thickness. Overall Piomas tends to show thicker ice than Hycom. As with any model it needs to validated. Maybe I will track down some Polarstern data and see how it compares.
« Last Edit: June 28, 2020, 09:35:18 AM by oren »

interstitial

  • Frazil ice
  • Posts: 472
    • View Profile
  • Liked: 162
  • Likes Given: 70
Re: HYCOM
« Reply #3 on: June 28, 2020, 07:00:03 AM »
The June to end of year comparisons are interesting but are not as predictive as 2019.

interstitial

  • Frazil ice
  • Posts: 472
    • View Profile
  • Liked: 162
  • Likes Given: 70
Re: HYCOM
« Reply #4 on: June 28, 2020, 07:01:41 AM »
a

oren

  • Moderator
  • First-year ice
  • Posts: 5896
    • View Profile
  • Liked: 1995
  • Likes Given: 1769
Re: HYCOM
« Reply #5 on: June 28, 2020, 09:41:27 AM »
Thank you interstitial for the wealth of data.
I edited the comparative post above and added a Cryosat-SMOS map for mid-April. I am having a hard time comparing them directly but maybe you can do the same visual comparison you did between PIOMAS and Hycom.
I am adding here two maps of April monthly anomalies compared to 2011-2018, one from PIOMAS and one from Cryosat-SMOS, provided by the Polar Science Center as part of their PIOMAS April update. These can help in finding the differences. I wish such a map in the same format  would be available from Hycom, but that would be too much to ask.

oren

  • Moderator
  • First-year ice
  • Posts: 5896
    • View Profile
  • Liked: 1995
  • Likes Given: 1769
Re: HYCOM
« Reply #6 on: June 28, 2020, 09:46:48 AM »
If I may add, the ice distribution in Hycom is very weird, with all the thick ice up to 5m bunched very near to Greenland and the CAA, and the rest of the CAB at a measly 2-2.5m. I find it very hard to believe this represent a true gradient, and Cryosat-SMOS does not support this either.

It now occurs to me that an ASCAT animation, showing where the old ice is and covering the period leading up to mid-April, could be very useful here as well. There should some animations available that uniquorn has posted through the winter and spring, will look for a suitable one.

Andre Koelewijn

  • New ice
  • Posts: 42
    • View Profile
  • Liked: 20
  • Likes Given: 6
Re: HYCOM
« Reply #7 on: June 28, 2020, 10:21:46 AM »
Note that the HYCOM-map serves a specific purpose: to show nuclear submarines operators the maximum probable thickness, to enable safe decisions where to surface.
It is meant as an upper bound throughout the arctic, not an alternative to estimate the actual total ice volume.

OffTheGrid

  • New ice
  • Posts: 46
    • View Profile
  • Liked: 28
  • Likes Given: 31
Re: HYCOM
« Reply #8 on: June 28, 2020, 08:46:10 PM »
If I may add, the ice distribution in Hycom is very weird, with all the thick ice up to 5m bunched very near to Greenland and the CAA, and the rest of the CAB at a measly 2-2.5m. I find it very hard to believe this represent a true gradient, and Cryosat-SMOS does not support this either.

It now occurs to me that an ASCAT animation, showing where the old ice is and covering the period leading up to mid-April, could be very useful here as well. There should some animations available that uniquorn has posted through the winter and spring, will look for a suitable one.
Perhaps we should remember that smos has a 40km grid resolution, and cryosat cannot resolve leads and ridges below a scale of 200-300m. It is unlikely that smos is contributing in the area you reference anyway as it is good at sub 1 m thickness, whole cryosat is better when ice is thicker than that.
So cryosat may assume ridge fields tops as continuous freeboard when things are far more complex at scales below Its resolution. A pity both are confounded during melt season, smos by wet ice, and cryosat by atmospheric moisture, especially cloud and fog, snow, particularly with melted and refrozen layers can cause both to oberestimate freeboard also.

Overall I am fairly impressed by the match between Hycom and crysmos in the april comparison. I am not sure why you find it so weird that the thickest ice is where it gets crushed and stacked by drift pressure against these coastlines, where also it experiences the coldest, with Katabatic winds dropping off tall ice sheets and mountains in midwinter?
At the end i guess, probably the truth may be somewhere in the middle. With unknown it is good I think to keep minds open to All informations, and individuals must balance how many "grains of salt" should each source be taken with.  :P

interstitial

  • Frazil ice
  • Posts: 472
    • View Profile
  • Liked: 162
  • Likes Given: 70
Re: HYCOM
« Reply #9 on: June 29, 2020, 12:06:41 AM »
A number of great comments it may take some time to explore them in detail.


Note that the HYCOM-map serves a specific purpose: to show nuclear submarines operators the maximum probable thickness, to enable safe decisions where to surface.
It is meant as an upper bound throughout the arctic, not an alternative to estimate the actual total ice volume.
I have heard that before on this site and I am not sure where it comes from. I looked for a statement about submarines or maximum probable thickness and found none. I have no doubt that the U.S. Navy is using this model for operations. I would note as of about a year ago the Navy is interested in countering Russian moves in the arctic and is increasing surface operations in the region.
 
In general I wonder how useful it would be for a submarine. I remember talk on this site of using underwater submarines equipped with LIDAR to measure sea ice thickness to validate models. So that statement seems backwards. I have no direct knowledge of U.S. Navel capabilities but I would bet just about anything the submarines have Lidar now and have had it for decades.
 
The objective stated on their website is: he GODAE objectives of three-dimensional depiction of the ocean state at fine resolution in real time, provision of boundary conditions for coastal and regional models, and provision of oceanic boundary conditions for a global coupled ocean-atmosphere prediction model, are being addressed by a partnership of institutions that represent a broad spectrum of the oceanographic community.
 
The partnership does include a number of institutions but I am not naive. Whoever pays most of the bills controls the mission and academic institutions are always looking for funding.
Reading through I also found this bit (Bold is mine): The systems will run efficiently on a variety of massively parallel computers and will include sophisticated data assimilation techniques for assimilation of satellite altimeter sea surface height and sea surface temperature as well as in situ temperature, salinity, and float displacement. They use floats among other things for validation. I am guessing the lidar on submarines is also used for validation though that would be something the navy did internally and would not mention because submarine locations are classified.
 
Finally they finish with this (bold is mine):
This collaborative partnership provides the means to leverage and accelerate the efforts of existing and planned projects, as well as new opportunities for collaboration and cooperation among the partners ranging from research to the operational level. The anticipated results of this effort include integrated analyses/reanalyses for research programs, a foundation for model enhancement based on testing of hypotheses, synthesis of remotely sensed and in situ data, and inputs to coastal, regional, atmospheric, and ice prediction systems. Improved open-ocean nowcasts and forecasts will be applied to search and rescue operations, shipping routes, tracking of icebergs and major pollutants, commercial fisheries, etc.
 

interstitial

  • Frazil ice
  • Posts: 472
    • View Profile
  • Liked: 162
  • Likes Given: 70
Re: HYCOM
« Reply #10 on: June 29, 2020, 12:08:12 AM »
If they are using it for maximum probable thickness and piomas shows thicker ice what does that say?

interstitial

  • Frazil ice
  • Posts: 472
    • View Profile
  • Liked: 162
  • Likes Given: 70
Re: HYCOM
« Reply #11 on: June 29, 2020, 12:37:15 AM »
Thank you interstitial for the wealth of data.
I edited the comparative post above and added a Cryosat-SMOS map for mid-April. I am having a hard time comparing them directly but maybe you can do the same visual comparison you did between PIOMAS and Hycom.
I am adding here two maps of April monthly anomalies compared to 2011-2018, one from PIOMAS and one from Cryosat-SMOS, provided by the Polar Science Center as part of their PIOMAS April update. These can help in finding the differences. I wish such a map in the same format  would be available from Hycom, but that would be too much to ask.

Anomaly maps are a wonderful tool for comparing a point in time with the average of that model. If you compare anomaly maps from different models you are introducing another factor namely the time averaged thickness map for each of the models. If the time averaged thickness varies between models the results get more confused.

It would be nice if Hycom provided anomaly tools.

It may be a good time to mention that the Hycom map varies slightly from Piomas. The differences are not very large.
« Last Edit: July 22, 2020, 10:28:01 AM by oren »

interstitial

  • Frazil ice
  • Posts: 472
    • View Profile
  • Liked: 162
  • Likes Given: 70
Re: HYCOM
« Reply #12 on: June 29, 2020, 12:50:48 AM »
If I may add, the ice distribution in Hycom is very weird, with all the thick ice up to 5m bunched very near to Greenland and the CAA, and the rest of the CAB at a measly 2-2.5m. I find it very hard to believe this represent a true gradient, and Cryosat-SMOS does not support this either.[/size]

This I think represents the biggest difference between the two models and IMO which gets it right is the better model. I seem to recall that ice grows too about 2 meters and then beyond that it grows slowly. I need actual thickness data to validate the model not comparison with another model.
« Last Edit: July 22, 2020, 10:28:39 AM by oren »

interstitial

  • Frazil ice
  • Posts: 472
    • View Profile
  • Liked: 162
  • Likes Given: 70
Re: HYCOM
« Reply #13 on: June 29, 2020, 04:18:07 AM »
I used mosaic news post #741 4/19/20 with T62. The difference between thickness on t62 between 4/15 and 4/19 is negligible so I used this date. I found on the 19th T62 was around 84.2N and 15.5E.I used the grid on Hycomm to identify the location and found it shows 1.6 m thickness. Since Piomas does not have a grid and I already made an overlay of the two and checked the same point on Piomas. Piomas says 2.5 m in that location. There is nothing close to 1.4 m in that area everthing else is thicker if anything. This is only one point on one day and not conclusive but I just picked it based on ease of getting the data. This took a long time since I found a point close to that using measurements and angles.

interstitial

  • Frazil ice
  • Posts: 472
    • View Profile
  • Liked: 162
  • Likes Given: 70
Re: HYCOM
« Reply #14 on: June 29, 2020, 05:11:14 AM »
post #1943 by Frivolous in the melting thread includes this comparison of piomas a model to cryosat. Notice piomas has a high bias.

igs

  • Frazil ice
  • Posts: 231
    • View Profile
  • Liked: 96
  • Likes Given: 70
Re: HYCOM
« Reply #15 on: June 29, 2020, 09:08:11 PM »
post #1943 by Frivolous in the melting thread includes this comparison of piomas a model to cryosat. Notice piomas has a high bias.

If Hycom is a tool for the military ( marine / submarines ) I would rather trust military grade data than civil models. They can simply not afford to err, considering what those datasets are used for.

That proves nothing of course, just a thought in case we never get any proof.

I would bet that once the satellites and pictures taken from vessels and airplanes show zero ice, that Piomas will show ice still, just a gut-feeling based on many occurances of that kind, even though they were regional and minor.
Knowledge that does not increase every day will decrease every day !

interstitial

  • Frazil ice
  • Posts: 472
    • View Profile
  • Liked: 162
  • Likes Given: 70
Re: HYCOM
« Reply #16 on: June 30, 2020, 02:18:48 AM »
Piomass is a fairly simple model it uses wind speed, wind direction, surface air temperature, cloud cover and with the latest update sea surface temperatures. Cloud cover and time of year are used to calculate solar and longwave radiation. Using air temperature and energy input modified by currents, wind and SST’s along with ice thicknesses, concentration and extent are used to calculate ice growth. Currents (from unnamed ocean model) and wind can move ice when the ice is mobile and SST energy. From what I read that is the basics
 
Piomas concerns:
*There is currently no funding to maintain or upgrade Piomas
 *Most of the validation data is older 6 years or older   
  http://psc.apl.uw.edu/sea_ice_cdr/data tables.html
     Exceptions:
BGEP, which is only 4 moored buoys in Beaufort sea.
icebridge quicklook, that prerelease data has a stated accuracy of +/- 1 meter of thickness. The final versions is much more accurate but not the quicklook data.
Cryostat volume is 23% lower than Piomass while this is not thickness it the extent is the similar.
*Piomass shows thin ice in Greenland Sea where none exists.
 
Piomas more info:
*Piomas uses NSIDC ice concentration and snow extent NISE near real time snow and ice extent 25 km
*Reynalds Sea surface temperatures for NCEP/NCAR on ice free areas
*Atmospheric wind, surface air and cloud cover to compute solar and longwave radiation
*Unnamed forced ocean circulation model with input at its open boundaries located at 45 north
*Stated piomas uncertainty +/- 1.35x10^3 km^3 areas in sea of Okhotsk and Gulf of St Lawrence are
   excluded
*spatial coverage 45 N to 90 N grid size 360 x 120 generalized curvilinear coordinate system (Grid shape irregular
Focus only on ice growth
 
Hycom uses hybrid coordinates layered in the open stratified ocean and smoothly transition to terrain following in shallow areas with specific coordinates for mixed layers or unstratified seas. It’s a global ocean model coupled with a global atmospheric model. 
It includes such physical processes as background internal wave breaking, shear instability mixing, double diffusion salt fingering and diffusive instability are parameterized. In the surface boundary layer wind driven mixing, surface buoyancy fluxes and convective instabilities are parameterized. Other factors include nonlocal mixing of temperature and salinity which permits counter gradient fluxes. The Kraus-Turner slab model the dynamical instability model of Price and the MellorYamada level 2.5 turbulence closure of the Prineton ocen model.
 
Hycom algorithms
   advdiff.pdf (horizontal advection/diffusion)
   boundary.pdf (boundary conditions)
   diapycnal.pdf (three interior diapycnal mixing algorithms)
   float.pdf (synthetic floats/drifters/moorings)
   hybrid.pdf (hybrid vertical coordinate adjustment)
   ice.pdf (energy loan ice model)
   KPP.pdf (K-Profile Parameterization vertical mixing)
   KT.pdf (three Kraus-Turner mixed layer models)
   mesh.pdf (horizontal mesh)
   momentum.pdf (momentum equation, including pressure gradient force)
   MY.pdf (Mellor-Yamada level 2.5 turbulence closure vertical mixing)
   PWP.pdf (Price-Weller-Pinkel dynamical instability vertical mixing)
   state.pdf (equation of state, including cabbeling and thermobaricity)
   surface.pdf (surface fluxes, including penetrating shortwave radiation)
   vdiff.pdf (solution of vertical diffusion equation)
   Other algorithms used from MICOM (HYCOM’s precursor) include continuity equation barotropic   
   momentum equation advection algorithm and vertical mode splitting
   https://www.hycom.org/attachments/067_overview.pdf
Hycom does an 8 day run every day.
0.08 degree resolution from 40 N to 40 S and 0.04 degree resolution poleward.
Focus on entire ocean atmosphere with parameterization of many physical interactions.
 

interstitial

  • Frazil ice
  • Posts: 472
    • View Profile
  • Liked: 162
  • Likes Given: 70
Re: HYCOM
« Reply #17 on: June 30, 2020, 02:28:38 AM »
All the buoy information I could find for April 15 showed
ice thickness < 2 m (mostly between 1.5 m and 1.8 m)


Piomas data for the same area > 2.5 m


Hycom varies between 1.6 m and 2 m for the same area

oren

  • Moderator
  • First-year ice
  • Posts: 5896
    • View Profile
  • Liked: 1995
  • Likes Given: 1769
Re: HYCOM
« Reply #18 on: June 30, 2020, 03:08:07 AM »
One thing that PIOMAS does is to model a distribution of ice thicknesses in each grid cell. This enables taking stock of pressure ridges, which contain a lot of volume though only cover a small area. The PIOMAS map shows the average for each cell, but the output files contain the distribution, which Wipneus graphs from time to time.
Looking at the Mosaic buoys, one of them was placed on a 7m thick floe, a real life example of uneven thickness of floes in the same vicinity.
I am sure Hycom models hydrology much much better, but not so sure how well it performs with ice thickness and transport.
Since PIOMAS assimilates NSIDC concentration data, it is rare (but happens) for it to show ice where none exists. I think this mostly stems from coarse resolution and will tend to disappear, so errors don't accumulate. But my knowledge of PIOMAS inner workings is very limited, I just use some of the output provided by Wipneus.

interstitial

  • Frazil ice
  • Posts: 472
    • View Profile
  • Liked: 162
  • Likes Given: 70
Re: HYCOM
« Reply #19 on: June 30, 2020, 04:07:53 AM »
If Hycom was just an intense ocean model it probably wouldn't be as good as Piomass. The Hycom model is coupled with in sophisticated atmospheric model which generates eight day forcasts.
I missed the 7 m data.
The submarine data is interesting with ice spikes pointing down in the water. 


navgem_0.176deg_computagrid_mask_2048x1024.D[/b][/size] is the land-sea mask needed for NAVGEM 2.0.[/font][/size]NAVGEM 2.03-hourly operational output: 20200401 and continuing until it is decommissioned[/color][/size]The 3-hourly files are generally updated once per week (Thursday or Friday) on hycom.org.[/color][/size]navgem2.0_0.176-sea_2020_03hr_mslprs.D[/b]mean sea level pressure (hPa/100 - 1000)navgem2.0_0.176-sea_2020_03hr_soiltm.D[/b]surface (land and ocean) temperature (Kelvin)navgem2.0_0.176-sea_2020_03hr_spchum.D[/b]2 m specific humidity (kg/kg)navgem2.0_0.176-sea_2020_03hr_temp2m.D[/b]2 m air temperature (Kelvin)navgem2.0_0.176-std_2020_03hr_ttlpcp.D[/b]total (large-scale and convective) precipitationnavgem2.0_0.176-sea_2020_03hr_ulwsfc.D[/b]upward surface longwave (W/m2)navgem2.0_0.176-sea_2020_03hr_uswsfc.D[/b]upward surface shortwave (W/m2)navgem2.0_0.176-sec_2020_03hr_dlwsfc.D[/b]downward surface longwave (W/m2)navgem2.0_0.176-sec_2020_03hr_dswsfc.D[/b]downward surface shortwave (W/m2); alias correctednavgem2.0_0.176c-sea_2020_03hr_dswsfc_noaliascorrect.D[/b]downward surface shortwave (W/m2); no alias correctednavgem2.0_0.176-sec_2020_03hr_longwv.D[/b]net surface longwave (W/m2)navgem2.0_0.176-sec_2020_03hr_solrad.D[/b]net surface shortwave (W/m2)navgem2.0_0.176-sec_2020_03hr_uv-10m.D[/b]10 m zonal and meridional wind velocity (m/s)navgem2.0_0.176-sec_2020_03hr_wndspd.D[/b]10 m wind speed (m/s)navgem2.0_0.176c-sec_2020_03hr_Taqa_null.D[/b]Note: no Qr/Qp arrays in this file.[/font][/color]

interstitial

  • Frazil ice
  • Posts: 472
    • View Profile
  • Liked: 162
  • Likes Given: 70
Re: HYCOM
« Reply #20 on: July 03, 2020, 04:10:52 AM »
latest

johnm33

  • Nilas ice
  • Posts: 1475
    • View Profile
  • Liked: 207
  • Likes Given: 70
Re: HYCOM
« Reply #21 on: July 03, 2020, 10:13:29 AM »
The two images show what appears to me to be a massive 'spouting out' incident occuring in the depths of Beaufort, I wouldn't fancy being in a sub caught anywhere in that turmoil.


OffTheGrid

  • New ice
  • Posts: 46
    • View Profile
  • Liked: 28
  • Likes Given: 31
Re: HYCOM
« Reply #22 on: July 04, 2020, 07:28:46 PM »
Latest 30 day model run to 11 July.
Signs of thicker ice limbs in coastal Beaufort, ESS, Franz Joseph regions head towards orphaned.
Ice over 2.5m by CAA wilt out. Or export to sure melt zones. Click to animate.

interstitial

  • Frazil ice
  • Posts: 472
    • View Profile
  • Liked: 162
  • Likes Given: 70
Re: HYCOM
« Reply #23 on: July 04, 2020, 10:20:47 PM »
Johnm33 thanks for the information. Any idea what would cause those? Another question I wonder how much mixing and heat those events caused? Any idea if this is normal?

ajouis

  • Frazil ice
  • Posts: 143
    • View Profile
  • Liked: 49
  • Likes Given: 31
Re: HYCOM
« Reply #24 on: July 05, 2020, 01:12:59 AM »
Interstitial, that last hycom run gives you the perfect opportunity to test it against reality, one, in a general manner, given that widespread white “almost gone” areas, that should mean a peak of losses july 15-22, second, some of those very thin areas should be visible from satellite photos, especially for the north of Svalbard, compared to now (provided it is not too cloudy), similarly we could compare with piomas on ice distribution at the mid july release. The quasi arm in Beaufort has shown to be very similar in piomas and hycom, so we can see if they are really parallel or divergent on some important features.
After a thousand steps on the ice, it cracked.
The Man looked down at the infinite blue of the sea.
On the horizon, standing still, the polar bear had just scented his next meal.

 Less than 3000 cubic kilometers this Piomas minimum.

interstitial

  • Frazil ice
  • Posts: 472
    • View Profile
  • Liked: 162
  • Likes Given: 70
Re: HYCOM
« Reply #25 on: July 05, 2020, 01:41:00 AM »
Interstitial, that last hycom run gives you the perfect opportunity to test it against reality, one, in a general manner, given that widespread white “almost gone” areas, that should mean a peak of losses july 15-22, second, some of those very thin areas should be visible from satellite photos, especially for the north of Svalbard, compared to now (provided it is not too cloudy), similarly we could compare with piomas on ice distribution at the mid july release. The quasi arm in Beaufort has shown to be very similar in piomas and hycom, so we can see if they are really parallel or divergent on some important features.
I agree and the more I learn about it the more informative  it is. I will post more on that in a few days. Another big thing it is saying is the bueafort ice is not as stable as it appears. I expect large late season losses there.

johnm33

  • Nilas ice
  • Posts: 1475
    • View Profile
  • Liked: 207
  • Likes Given: 70
Re: HYCOM
« Reply #26 on: July 05, 2020, 11:36:08 AM »
"Any idea what would cause those?" I think massive movements of of water caused by mslp highs/lows passing from land to sea, or vice versa,  or passing over Lomonosov and some by tidal surges in concert with those movements. The movements set up internal waves in deeper layers, these either cancel out or reinforce each other if the second then events like this occur and they're not uncommon in the Canada basin, looking at amsr2 on polarview you can see signs of these events and similar signs elsewhere but without Hycoms model.
https://www7320.nrlssc.navy.mil/GLBhycomcice1-12/beaufort.html
Amsr2 data browser
 The ice strength gif of Beaufort shows the ice will not hold a crack, suggesting to me it's very weak.

interstitial

  • Frazil ice
  • Posts: 472
    • View Profile
  • Liked: 162
  • Likes Given: 70
Re: HYCOM
« Reply #27 on: July 07, 2020, 04:31:37 AM »
I need to modify my prediction. Hycom does show thicker ice still blocking the northwest passage however it looking at worldview it is thicker floes frozen in place by thinner ice. That thinner ice will allow the channel to clear when the rest of the channel clears.

interstitial

  • Frazil ice
  • Posts: 472
    • View Profile
  • Liked: 162
  • Likes Given: 70
Re: HYCOM
« Reply #28 on: July 08, 2020, 12:47:27 AM »
New Ice thickness. Still on track for melt of large section of Greenland sea and Baffin bay on 7/10-7/17. The CAA can get tricky to predict. I still like this model better.

ajouis

  • Frazil ice
  • Posts: 143
    • View Profile
  • Liked: 49
  • Likes Given: 31
Re: HYCOM
« Reply #29 on: July 08, 2020, 08:29:06 AM »
Today’s bremen comforts the accuracy of the forecast, at least somewhat, the baffin and hudson abrupt losses were even anticipating a bit, but it is to be expected in a predictive model, not everything aligns the model’s way
After a thousand steps on the ice, it cracked.
The Man looked down at the infinite blue of the sea.
On the horizon, standing still, the polar bear had just scented his next meal.

 Less than 3000 cubic kilometers this Piomas minimum.

interstitial

  • Frazil ice
  • Posts: 472
    • View Profile
  • Liked: 162
  • Likes Given: 70
Re: HYCOM
« Reply #30 on: July 08, 2020, 07:19:52 PM »
Agreed. I am thinking a 10 cm especially the last 10 cm is not that much and thinking that it could predict that precisely was to optimistic. That is not what I wanted the model for anyway. Mostly I just like to see how thick the ice is and how much it is actually melting. Otherwise I look at it and wonder why it hasn't melted already or why it all just disappeared. At first I was trying to convince others of its accuracy. After comparing several floats in the CAB I am convinced that it is more accurate than Piomas. Where there are discrepancies I trust Hycom more than Piomas. People have complained it is reporting a maximum thickness but if that is true it still shows thinner ice in more places than Piomas. It is not perfect. In the western side of northwest passage thicker ice froze in place. Hycom shows the thicker ice. Piomas shows it as thinner. Worldview shows reality thicker ice frozen in place.

interstitial

  • Frazil ice
  • Posts: 472
    • View Profile
  • Liked: 162
  • Likes Given: 70
Re: HYCOM
« Reply #31 on: July 08, 2020, 11:16:23 PM »
here

uniquorn

  • Nilas ice
  • Posts: 2486
    • View Profile
  • Liked: 1346
  • Likes Given: 221
Re: HYCOM
« Reply #32 on: July 09, 2020, 12:34:00 AM »
Another big thing it is saying is the bueafort ice is not as stable as it appears. I expect large late season losses there.
Thanks for the study on Hycom. The model is useful for comparison. Despite repeated assertions by some on the melting thread based on extent, amsr2-uhh doesn't show Beaufort looking stable imho.
jun4-jul8

johnm33

  • Nilas ice
  • Posts: 1475
    • View Profile
  • Liked: 207
  • Likes Given: 70
Re: HYCOM
« Reply #33 on: July 11, 2020, 11:49:40 AM »
The sea ice concentration gif shows imho that the whole pack has softened and will no longer hold a crack, an alternative is that the whole Arctic is becalmed? [18:06-17:07]

meddoc

  • Frazil ice
  • Posts: 262
    • View Profile
  • Liked: 11
  • Likes Given: 0
Re: HYCOM
« Reply #34 on: July 11, 2020, 12:02:09 PM »
HYCOM- as I stated couple of times before- switched from ArCc to Glb (whatever it should mean) I think back in 2017. They simply ripped off huge Portions of 2m + thick Ice.

I'd have the previous Models' from 2016, but had to make some space available on hard drive.

Jim Hunt

  • Young ice
  • Posts: 4650
  • Stay Home, Save Lives
    • View Profile
    • The Arctic sea ice Great White Con
  • Liked: 515
  • Likes Given: 42
Re: HYCOM
« Reply #35 on: July 11, 2020, 12:53:25 PM »
HYCOM- as I stated couple of times before- switched from ArCc to Glb (whatever it should mean)

It's short for "Global". As in "Global Ocean Forecasting System".

Quote
They simply ripped off huge Portions of 2m + thick Ice.

Not exactly! They started assimilating concentration, amongst other things. Chapter and verse is available at:

https://forum.arctic-sea-ice.net/index.php/topic,1602.msg84108.html#msg84108
Reality is merely an illusion, albeit a very persistent one - Albert Einstein

Jim Hunt

  • Young ice
  • Posts: 4650
  • Stay Home, Save Lives
    • View Profile
    • The Arctic sea ice Great White Con
  • Liked: 515
  • Likes Given: 42
Re: HYCOM
« Reply #36 on: July 11, 2020, 12:54:32 PM »
I would appreciate a posting of Hycom Apr 15th and Cryosat Apr 15th side by side somewhere, maybe in the new Hycom thread. I would look for differences where the anomalies were supposed to be - near Svalbard, in the Beaufort, and in the ESS.

Here you go:
Reality is merely an illusion, albeit a very persistent one - Albert Einstein

interstitial

  • Frazil ice
  • Posts: 472
    • View Profile
  • Liked: 162
  • Likes Given: 70
Re: HYCOM
« Reply #37 on: July 11, 2020, 01:02:53 PM »
I rearranged the scale bar to emphasize how inconsistent the fading from one color to the next is. Look at each meter and see how many colors it has. I think they did this to emphasis certain threshold values. That is just a guess. 

interstitial

  • Frazil ice
  • Posts: 472
    • View Profile
  • Liked: 162
  • Likes Given: 70
Re: HYCOM
« Reply #38 on: July 11, 2020, 04:33:40 PM »
HYCOM- as I stated couple of times before- switched from ArCc to Glb (whatever it should mean) I think back in 2017. They simply ripped off huge Portions of 2m + thick Ice.
I'd have the previous Models' from 2016, but had to make some space available on hard drive.
I don’t have many of the old ones but I was able to find some. From what I can see in those areas where thickness was reduced in the new model it was not reduced by a set amount everywhere. The change mostly seemed to affect ice thicker than 2.8 meters.  Some of the thicker ice was left unchanged.
It is my understanding that ice can grow to around 2 meters thick at most in one year. After that growth slows way down. If none of that ice melts how much ice would be added next season? A meter? Maybe but I doubt it. That wouldn’t be ice growth slowing way down. I don’t know the answer. In previous years the oldest ice known as perennial ice could get to 9 years old or older. It took about that long to get to 4 meters thick. https://earthsky.org/earth/decline-of-arctics-thickest-sea-ice The amount of ice that is 4 years old or older is almost non existent as shown by graphic below. https://svs.gsfc.nasa.gov/4750
They updated the model for a reason and the data they got after the update was more accurate than the data they got before. Otherwise they would have stuck with the older model.
Looking at the area from 80N in the quadrant facing Europe and not near an island Hycom shows thickness up to about 2 meters on April 15. The same area on Piomas shows not less than 3 meters. The buoy data I have seen showed in that area at that time ice up to about 2.1 to 2.2 meters thick. Frankly Piomas shows about 2/3rds of the arctic thicker than 2.75 meters in April. It also shows a quarter of the central arctic thicker than 3.5 meters thick. There is almost no ice 4 years old or older and I doubt much if any ice younger than 4 years gets that thick. I think Piomas become an inaccurate model about the time the older ice all melted.
 
Primarily Piomas uses an Ice growth equation based on temperature and time and location when the ice moves. Solar flux ice extent and concentration are also factored in. Recently they added surface ocean temperatures on ice free areas to the mix. Salinity is fixed. A fixed current is used. Sea water under the ice is assumed to be a uniform layer in each grid element. No atmospheric or ocean modeling is done. Near as I can figure it is an academic pursuit probably put together by a professor and grad student around 2003.  It does get updated from time to time but there is no dedicated staff. 
The Hycom model is built from first principles. The atmosphere and ocean are modeled globally.  Currents change as conditions change. It also has more than one full time staff member and a budget. I have written more about in this thread and you can dig into it if you really want to know. The Hycom model is frequently updated so comparisons should be made with the same version. They also do not produce a volume total. Which I find disappointing because we like our spreadsheets on ASIF.     

interstitial

  • Frazil ice
  • Posts: 472
    • View Profile
  • Liked: 162
  • Likes Given: 70
Re: HYCOM
« Reply #39 on: July 11, 2020, 04:35:43 PM »
and the rest
note white bar is the sea ice over 4 years old.

Comradez

  • New ice
  • Posts: 71
    • View Profile
  • Liked: 29
  • Likes Given: 0
Re: HYCOM
« Reply #40 on: July 11, 2020, 04:56:39 PM »
The two look surprisingly similar!  Shouldn't this increase our confidence in HYCOM? 

If so, I think I like HYCOM a little more just because it has such higher resolution. 

oren

  • Moderator
  • First-year ice
  • Posts: 5896
    • View Profile
  • Liked: 1995
  • Likes Given: 1769
Re: HYCOM
« Reply #41 on: July 11, 2020, 04:59:53 PM »
I remind again that PIOMAS models pressures ridges when ice is transported, which is how it can get above the thermodynamic growth limit for ice thickness.
I know nothing of the details but can see the results, e.g. near Svalbard. Is that thick ice really there? I honestly don't know.

Laurent

  • Young ice
  • Posts: 2537
    • View Profile
  • Liked: 8
  • Likes Given: 34
Re: HYCOM
« Reply #42 on: July 11, 2020, 05:09:01 PM »
Source: https://www7320.nrlssc.navy.mil/GLBhycomcice1-12/arctic.html
(hycom ARC is an older model that allows US ships to choose an ice-free route - it is slightly less accurate than the new model to ensure that ships will not encounter ice).
https://www7320.nrlssc.navy.mil/hycomARC/arctic.html
The data not being available for 2019, I have recovered the image of the proposed 1-year animation.
The comparison data with the ARC model ( the one for 2012) are to be taken with big tweezers, although it is the same color scale! 2012 because the volume record took place in 2012.

interstitial

  • Frazil ice
  • Posts: 472
    • View Profile
  • Liked: 162
  • Likes Given: 70
Re: HYCOM
« Reply #43 on: July 11, 2020, 06:14:19 PM »
I remind again that PIOMAS models pressures ridges when ice is transported, which is how it can get above the thermodynamic growth limit for ice thickness.
I know nothing of the details but can see the results, e.g. near Svalbard. Is that thick ice really there? I honestly don't know.
I think pressure ridges explain thickness near land masses and some elsewhere but I doubt 50% of the central arctic adds a meter by that mechanism. Hycom shows some thick ice in the same areas around land masses but my main concern with Piomas is it shows much thicker ice in the bulk of the Arctic.

I did not start out intending to doubt Piomas I just thought Hycom had value too. It is so much work to write up what I find that I often don't do it. I don't know how fast other are but for me my last post on this thread took about seven hours. AFAIK    few have attempted to countered my claims and they claimed in the area described in my last post that one buoy was tethered on a 7 meter thick chunk of ice. It is certainly possible but I am not sure if is is even true. I think he may have said that because i didn't post the bouy data. Mostly the main anti Hycom comments are someone else said piomas was a better model and I don't think the ice is that thin. Back in 2003 it probably was a great model and that was probably when the opinion was formed. Things change I am constantly trying to explain that what was true then may not be true now. I once had an argument with a guy who insisted that heat pumps did not work below 0 C. Maybe that was true a long time ago but things change. In the Arctic almost all the 4 year old  or older Ice melted. That implies all the thicker ice melted and younger ice normally doesn't get that thick.

interstitial

  • Frazil ice
  • Posts: 472
    • View Profile
  • Liked: 162
  • Likes Given: 70
Re: HYCOM
« Reply #44 on: July 11, 2020, 06:20:11 PM »
Thanks Laurent. I didn't think that data was still available.

oren

  • Moderator
  • First-year ice
  • Posts: 5896
    • View Profile
  • Liked: 1995
  • Likes Given: 1769
Re: HYCOM
« Reply #45 on: July 11, 2020, 10:05:41 PM »
Quote
few have attempted to countered my claims and they claimed in the area described in my last post that one buoy was tethered on a 7 meter thick chunk of ice. It is certainly possible but I am not sure if is is even true. I think he may have said that because i didn't post the bouy data.
A. That was me...
B. That 7 m thick floe is true. The Mosaic people put the buoy there and they measured it. The information was posted by uniquorn in another thread.
C. I really don't know how reliable PIOMAS and/or Hycom is.
D. I will pay more attention to Hycom this melting season.

interstitial

  • Frazil ice
  • Posts: 472
    • View Profile
  • Liked: 162
  • Likes Given: 70
Re: HYCOM
« Reply #46 on: July 11, 2020, 11:55:15 PM »
Oren: At the time I read the 7m commment it struck me as odd. Sometimes I over think these things. Your comments have been helpful. Thank you. My search for buoys was not exhaustive but that is what I found. I wonder if my commentary has been helpful and people do not have a response or if people are just ignoring it. I can get a bit ......       Sigh......    Moving on. End post.

uniquorn

  • Nilas ice
  • Posts: 2486
    • View Profile
  • Liked: 1346
  • Likes Given: 221
Re: HYCOM
« Reply #47 on: July 11, 2020, 11:59:10 PM »
https://data.meereisportal.de/download/buoys/photos/2020T61_deployment.pdf

There are a lot of other deployment data in the photos directory. Mosaic photos on fomo or twitter show much more ridging than I was expecting but despite being thicker, ridged ice doesn't have the same structural integrity as thermo.

Keep up the good work interstitial :)
« Last Edit: July 12, 2020, 12:05:04 AM by uniquorn »

interstitial

  • Frazil ice
  • Posts: 472
    • View Profile
  • Liked: 162
  • Likes Given: 70
Re: HYCOM
« Reply #48 on: July 12, 2020, 12:02:50 AM »
Thanks uniquorn.

oren

  • Moderator
  • First-year ice
  • Posts: 5896
    • View Profile
  • Liked: 1995
  • Likes Given: 1769
Re: HYCOM
« Reply #49 on: July 12, 2020, 12:14:45 AM »
Your commentary has been very helpful. I am guessing people have nothing useful to add to it at present.