A predictor of the September minimum?: "Cole's Correlation"

[Vergent]

The arctic sea ice snow cover is about to be torched.



Here is the current snow cover.



Here is the forecast snow cover for the 12th.



Here was the snow cover on June 12, 2012.



Here is the snow cover on June 12, 2014



Here is June 12, 2013

My conclusion is that there will be high arctic melt pond formation even earlier than 2012. I think that this may be a really bad summer for the deniers and luke warmists.

Wow, nice illustration!!! What I find remarkable is how the snow cover as early in the season as June 12th seems to foretell the ice in September, not only extent, but also the shape of the pack.

In 2013 and 2014, it seems the "green" area's periphery for June 12 snow roughly matches the ice pack three months later, excluding the really low-concentration areas. However, in 2012 by that standard it looks like the whole ice pack is set to vanish - and we of course know that it didn't do so.

One possible explanation for this discrepancy is the double snowstorm that hit the central Arctic on June 16 and 20 or thereabouts. I wonder if we would see the snow cover maps for (say) June 23 or June 25 showing an inclination for that ice to likely survive the season.

Based on the extremely large amount of humid air projected to surround the Arctic by June 11th/12th according to CCI/GFS, it only takes one system to pull a plume of that over the arctic to result in an even bigger spike in surface melt than we already had - and though still too far out to really trust, the long range forecasts seem to call for the 12th and 13th to feature a low in Barents coupled with a high in the Pacific portion of the CAB. If this pans out, we will get a big tongue of "REALLY STEAMY" air (relatively speaking of course) injecting itself right into the heart of the central Arctic...

Note I mistakenly posted June 19, 2012 snow cover in my original post. I have corrected it there and in this quote.

Nightvid Cole,

Brilliant observation!
2012:




Here are links so you can flip between tabs:

http://i.imgur.com/QoaMAaj.png
http://i.imgur.com/dqyYLMu.png
http://i.imgur.com/uHf1x3C.png

An explanation for the Laptev Bight?

2013:




http://i.imgur.com/OjaqsPf.png
http://i.imgur.com/vAbxpVv.png
http://i.imgur.com/cfrUXtE.png

2014:




http://i.imgur.com/ueAoPzH.png
http://i.imgur.com/YlVWyOr.png
http://i.imgur.com/GZSLOQY.png

2015:



http://i.imgur.com/Dq3q388.png
http://i.imgur.com/0TntH9N.png

For 2013, this over predicts the ice, but is accurate for the regions of melt. 2014 is a stunning correlation. 2015? This too is not good.

Credit for this goes to Nightvid Cole.

Verg

[wehappyfew]

I like the analysis... hate the result:

North pole = open water...



...

[Nightvid Cole]

Thanks Verg.   

I would expect that by now Santa would be saving for a boat (or, I suppose a submarine...)

[Vergent]

In the next model run, it does not seem to get better.

This indicator makes sense, though. Evidently it takes 60 days of high insolation, after melt pond formation, to melt out the ice.

Verg

[Nightvid Cole]

In the next model run, it does not seem to get better.

This indicator makes sense, though. Evidently it takes 60 days of high insolation, after melt pond formation, to melt out the ice.

Verg

It's been 24 hours since that forecast - where's the page to check for the newest?

[Vergent]

Well, okay.....It takes some navigation.

Go  here.... http://marine.copernicus.eu/web/69-myocean-interactive-catalogue.php

Select Arctic ocean.

Select "more info" for the top result.

Select "view product".

Select the "+" under the hand.

Select "blue marble north polar"

Select "-" to make the menu go away.

Click on the variable menu field. It will say "sea_floor_depth_below_sea_level"

Select "surface_snow_thickness"

Then select the date.

Verg

[Vergent]

Do you ever get the feeling that you are a rat in a maze?

[Vergent]

If you actually get there, then you can play with the scale on the right... If you click on it you can adjust the color scheme and intervals... You can also adjust the top and bottom of the scales by adjusting the appropriate boxes.

Enjoy, and employ. I do not protect my sources like other people do.

Verg

[Vergent]

Thanks for asking...

Verg

[Nightvid Cole]

Thanks for answering. 

It seems the correlation for June 13th is quite high if the lower bound is set to 0.03 (3 cm) and the upper bound to 0.06 (6 cm), so I have done a crude projection for the fate of 2015.

Each image contains - ON THE LEFT: Snow cover from TOPAZ for June 13th of given year; ON THE RIGHT: Ice cover from Cryosphere Today on September 8th of the given year.
**************************************************************
2012:
http://tinypic.com/view.php?pic=65bng0&s=8#.VXLto9JVhHw

2013:
http://tinypic.com/view.php?pic=2whgiab&s=8#.VXLwLdJVhHw

2014:
http://tinypic.com/view.php?pic=2qjy7o3&s=8#.VXLvcdJVhHx

2015:
http://tinypic.com/view.php?pic=2lucobs&s=8#.VXLwttJVhHw

[Vergent]

Alarming! It is disconcerting that two of the three historical tests of this correlation over predicts the September ice. This is particularly true for the Kara. This was intentional on my part when I started the thread. On both polls, I am the bottom man on the totem pole. That is because I think that albedo feedback will take over. Did you notice the albedo variable?




I wonder how many extra gigawatts we will get this year?

Verg

[Nightvid Cole]

Alarming! It is disconcerting that two of the three historical tests of this correlation over predicts the September ice. This is particularly true for the Kara. This was intentional on my part when I started the thread. On both polls, I am the bottom man on the totem pole. That is because I think that albedo feedback will take over. Did you notice the albedo variable?




I wonder how many extra gigawatts we will get this year?

Verg

Interesting - I didn't notice the albedo variable, though it tells much the same story as the snow cover - that this will be an absolute blowout melting year. I don't know how many gigawatts of gain we'll have - though I can say "too many"!

[Vergent]

I think the grey scale may be more intuitive for albedo.

The wattage absorbed from the sun will be almost double last years.

Verg

Edit: spelling

[Vergent]

For some reason TOPAZ4 has not updated. It is supposed to have a 10 day forecast, but the 13th is the last day. Mabe the storm is making it confused.

http://www.weather-forecast.com/maps/Arctic?over=pressure_arrows&symbols=none&type=wind

Here is a nice GFS visualization that goes out 10 days........Just hit "play".

Verg

[Laurent]

Less snow, less albedo and much more rain forecasted...
I really like this tool, climate reanalizer is slow for me, but this one is fast...really cool. Rain does matter a lot I think.

[Vergent]

Its not just the ice albedo. In a week, the only place in the arctic with snow will be Greenland. 10c (50f) or warmer everywhere on land except Greenland.

The second law of thermodynamics is pissed off big time. The ice is going to get torched no matter which way the wind blows.

Verg

[Neven]

My experience in the past with TOPAZ was that it wasn't all that reliable (even though the research director was a very friendly scientist, consenting to an interview on the ASIB). However, I must say those comparisons with observations are pretty similar. Maybe they get changed afterwards when the data is in?

[Vergent]

Neven,

TOPAZ4 is a very distant relation from the former TOPAZ. The original TOPAZ was mostly model, and a simple one at the. TOPAZ4 is data driven, only using modeling for data gaps. It is pulling in data from ensembles of satellites for each parameter. The hind-cast is subjected to reanalysis.

Verg

[Neven]

Thanks, Verg. Very interesting.

[Vergent]

Less snow, less albedo and much more rain forecasted...
I really like this tool, climate reanalizer is slow for me, but this one is fast...really cool. Rain does matter a lot I think.

Is this the reanalizer you are talking about?

http://cci-reanalyzer.org/Forecasts/

Verg

[Laurent]

yes, this one. I tryed to change the speed with the + button, the orange square move along but image does not update much. It used to work fine but now it doesn't, don't know what is going on.

[Nightvid Cole]

I found another correlation - this time between the region of ice concentration >60% on August 05 , XXXX (also from MyOcean source) correlated with the NSIDC September monthly coverage in XXXX.

2012:



2013:



2014:



I intend to use this for my August submission to SIPN, because the y-intercept for the linear regression is just 0.6452 M km^2 (The ideal value is zero). I can thus argue that the MyOcean-derived red region is my Ansatz for the shape of the ice pack in September, though I will still use the full linear regression against pixel count for my extent guess. Of course this will have to wait for either a forecast or data for August 5, 2015 to be available. In the meantime, my July SIPN prediction is a repeat of June.

[jdallen]

That's rather good, NightVid.

[Neven]

There's nothing for it but to wait until August 5th. 

But looking good as a predictor.

[Pmt111500]

I was reminded of this site http://polarportal.dk/en/havisen-i-arktis/nbsp/sea-ice-temperature/
and had some fun moving the temperature scale 1 degree up and moving bits of loose ice to the Arctic basin pack. Not that this is a valid method for prediction, though at some thickness the ice appears to warm up to melting point all through the ice before melting away completely. I wish all maps would have a scale in them so the area could be approximated from the pixels. No idea what this amounts to.

[Nightvid Cole]

I was reminded of this site http://polarportal.dk/en/havisen-i-arktis/nbsp/sea-ice-temperature/
and had some fun moving the temperature scale 1 degree up and moving bits of loose ice to the Arctic basin pack. Not that this is a valid method for prediction, though at some thickness the ice appears to warm up to melting point all through the ice before melting away completely. I wish all maps would have a scale in them so the area could be approximated from the pixels. No idea what this amounts to.

Interesting...

[Nightvid Cole]

Unfortunately I am currently unable to access the computer my previous work was on, so the image size is different from before - however I still was able to project NSIDC extent for September as (3.90 +/- 0.17) M km^2. Image differs slightly from this figure due to non-ideal slope/intercept but should still represent a good guess for what we will see in September.