Author Topic: Home brew AMSR2 extent & area calculation (Read 1981462 times)

Wipneus · « **Reply #150 on:** July 12, 2013, 08:46:06 AM »

Region of the day is the "torch" Chukchi sea. I posted an image of the Nares break-up in the Nares thread.

(log in to see attached image)

Neven · « **Reply #151 on:** July 12, 2013, 08:56:41 AM »

There'll be more of that in days to come. The northern Chukchi already has melt holes in it.

jdallen · « **Reply #152 on:** July 12, 2013, 09:25:49 AM »

Quote from: Neven on July 12, 2013, 08:56:41 AM

There'll be more of that in days to come. The northern Chukchi already has melt holes in it.

I'm anticipating a "tip over" event, where thin ice within some areas of extent vanishes in a very short time (less than a week) over a large area, leaving isolated floes of MYI.

Portions of the Beaufort and CAA are my favorite candidates for this.

Wipneus · « **Reply #153 on:** July 13, 2013, 08:19:31 AM »

Update 20130712.

Extent: -151k6 (+391k vs 2012)
Area: -72k2 (+384k vs 2012)

I have to be brief, train to catch.

If not for Hudson, Baffin and Greenland resuming the melt, the situation is much the same as yesterday. Declines on the Siberian side, much of it in Chukchi. Slow Beaufort and CAA.

The details in 1000 km²:

Extent: Arctic Basin East Siberian Sea Laptev Sea -12.2 -10.1 -4.3 Kara Sea Barents Sea Greenland Sea -19.1 -1.2 -21.2 Baffin/Newfoundland Bay St. Lawrence Hudson Bay -30.5 0.0 -21.5 Canadian Archipelago Beaufort Sea Chukchi Sea -5.5 -0.2 -24.9 Bering Sea Sea of Okhotsk Total Extent -0.6 -0.2 -151.6 Area: Arctic Basin East Siberian Sea Laptev Sea -58.2 22.6 -6.4 Kara Sea Barents Sea Greenland Sea -16.3 -1.0 -21.1 Baffin/Newfoundland Bay St. Lawrence Hudson Bay -17.1 0.0 -16.3 Canadian Archipelago Beaufort Sea Chukchi Sea 30.3 23.8 -11.6 Bering Sea Sea of Okhotsk Total Area -0.5 -0.3 -72.2

Wipneus · « **Reply #154 on:** July 13, 2013, 08:26:12 AM »

Image of the day, Greenland Sea, shows the large chunck of fast ice breaking of NE greenland.

(log in to see attached image)

Wipneus · « **Reply #155 on:** July 14, 2013, 08:25:32 AM »

Update 20130713:

Extent: -74k5 (+340k vs 2012)
Area: -113k7 (+239k vs 2012)

For extent it is a relatively quit day. ESS takes lead, followed by Chukchi. Changes for area are in comparison bigger in the CAB and Beaufort. CAA shows small increases for both extent and area.

The details (in 1000 km²):

Extent: Arctic Basin East Siberian Sea Laptev Sea -3.0 -18.3 -2.5 Kara Sea Barents Sea Greenland Sea -11.0 -0.7 0.1 Baffin/Newfoundland Bay St. Lawrence Hudson Bay -6.9 0.0 -17.3 Canadian Archipelago Beaufort Sea Chukchi Sea 3.2 -3.1 -13.1 Bering Sea Sea of Okhotsk Total Extent -0.1 -1.9 -74.5 Area: Arctic Basin East Siberian Sea Laptev Sea -58.7 4.9 -11.8 Kara Sea Barents Sea Greenland Sea -6.9 -0.4 17.9 Baffin/Newfoundland Bay St. Lawrence Hudson Bay -2.9 0.0 -27.3 Canadian Archipelago Beaufort Sea Chukchi Sea 9.5 -24.3 -12.2 Bering Sea Sea of Okhotsk Total Area 0.0 -1.6 -113.7

Wipneus · « **Reply #156 on:** July 14, 2013, 08:35:36 AM »

The action is in the Chukchi and East Siberian Sea's, movement and melting as seen in to days image.

(log in to see attached image)

Wipneus · « **Reply #157 on:** July 14, 2013, 08:51:45 AM »

BTW, new data from Uni Hamburg is comes in normally at about 6:15 UTC. An update is done about 14 hours later, normally without any data actually changing: I only see the file modification time differ.

The update of 20130712 is the first time I noticed a small downward revision.

Extent: -154k3 (from 151k6)
Area: -78k2 (from -72k2)

Wipneus · « **Reply #158 on:** July 15, 2013, 08:19:06 AM »

Update 20130714.

Extent: -108k0 (+263k vs 2012)
Area: -92k8 (+271k vs 2012)

Most of the net decline is in Chukchi, ESS and Kara. Crumbs for CAB, CAA, Beaufort and Laptev.

Updated graphics are in the top post.

The details (in 1000 km²):

Extent: Arctic Basin East Siberian Sea Laptev Sea -5.2 -23.2 -8.6 Kara Sea Barents Sea Greenland Sea -23.0 0.8 -4.6 Baffin/Newfoundland Bay St. Lawrence Hudson Bay -20.5 0.0 12.3 Canadian Archipelago Beaufort Sea Chukchi Sea -10.6 -6.2 -18.6 Bering Sea Sea of Okhotsk Total Extent 0.4 -0.7 -108.0 Area: Arctic Basin East Siberian Sea Laptev Sea 25.0 -21.0 -14.4 Kara Sea Barents Sea Greenland Sea -18.2 0.7 -6.8 Baffin/Newfoundland Bay St. Lawrence Hudson Bay -15.6 0.0 18.0 Canadian Archipelago Beaufort Sea Chukchi Sea -41.6 1.4 -19.8 Bering Sea Sea of Okhotsk Total Area 0.4 -1.2 -92.8

Wipneus · « **Reply #159 on:** July 15, 2013, 08:34:52 AM »

Region of the day is Laptev, slow in the net numbers. The image shows there is movement of the ice, causing increases and decreases in places.

(log in to see attached image)

Bob Wallace · « **Reply #160 on:** July 15, 2013, 10:00:39 AM »

Thanks. You've turned this into a really great horse race for me.

I find it very interesting that 2013 is ahead of 2012 in both extent and area for the CAB. Add in the CA and GS and 2013 is ahead in terms of last year's "didn't fully melt" regions.

LarsBoelen · « **Reply #161 on:** July 15, 2013, 11:37:33 PM »

Wipneus, I got totally inspired by all your graphical work, I'd like to add my 2 cents with a movie of al the frames of AMSR2 (I cleaned the frames up til july)

http://stormglas.wordpress.com/2013/07/15/daily-amsr2-update-movie-of-northpole-melting-in-2013/

Wipneus · « **Reply #162 on:** July 16, 2013, 08:21:56 AM »

Update 20130715.

Extent: -124k9( +225k vs 2012)
Area: -32k7 (+316k vs 2012)

Large extent decline continues in Chukchi, Laptev and Kara, followed by ESS and CAA. The CAB and Beaufort are still "modest".
Area follows mostly the pattern of extent, except for a big increase in the CAB and a more modest one in Chukchi.

The details in 1000 km² :

Extent: Arctic Basin East Siberian Sea Laptev Sea -4.7 -10.1 -22.7 Kara Sea Barents Sea Greenland Sea -20.4 -0.2 -11.2 Baffin/Newfoundland Bay St. Lawrence Hudson Bay -4.9 0.1 -2.9 Canadian Archipelago Beaufort Sea Chukchi Sea -9.2 -5.7 -32.0 Bering Sea Sea of Okhotsk Total Extent 0.2 -1.2 -124.9 Area: Arctic Basin East Siberian Sea Laptev Sea 44.2 -8.5 -28.8 Kara Sea Barents Sea Greenland Sea -14.3 -0.2 -14.8 Baffin/Newfoundland Bay St. Lawrence Hudson Bay -4.4 0.1 -16.1 Canadian Archipelago Beaufort Sea Chukchi Sea -1.5 0.1 12.4 Bering Sea Sea of Okhotsk Total Area 0.2 -1.1 -32.7

Wipneus · « **Reply #163 on:** July 16, 2013, 08:34:14 AM »

Region of today is Kara. The ice flotilla behind Nova Zemlya is nearly gone. The fast Ice behind Severnaya Zemlya looks much tougher.

(log in to see attached image)

werther · « **Reply #164 on:** July 16, 2013, 09:04:29 AM »

Thanks, Wipneus,

Didn't you post the ice thickness anomaly compared to '12 in nice colours in April? That pic illustrated that 'fast ice' against Severnaya zemlya to be tough. It might be a part of Kara Sea ice holding out long enough to contribute 50-100K to remaining minimum SIE.

As for the CAB area increase; in line with my comment on the 'short term' thread, I think it may indicate a stage near the start of complete fading out of large parts of the Chukchi-/ESAS- and Laptev sectors.

Wipneus · « **Reply #165 on:** July 16, 2013, 09:20:16 AM »

Werther,

PIOMAS ice thickness map for May 2013 (latest we have):

And 2012-2013 differences:

The thicker ice seems to be (according to PIOMAS) on the Laptev side of Severnaya Zemlya.

And thanks for reporting what you see on MODIS. It helps interpretating what the different products report as concentration or the derived "area".

werther · « **Reply #166 on:** July 16, 2013, 10:43:32 AM »

Again, thanks for re-issuing that map, Wipneus.

I thought there was an earlier map for April, but I guess it was that one.

Illustrating the different niche-behaviour of the ice formation-growth-melt process in all regions, this had me confusing. Your May map shows the anomalous thickness on the Laptev side, where it is almost depleted now. It does show the anomaly on the South side of the Kara Sea, where indeed there’s still (melting) ice in the Baydaratskaya Bay.

The ‘fast ice’ to the West of Severnaya Zemlya was never a “shelve” like this last year. There was a band of fast ice close to the coast, the rest never even froze completely that season and was almost depleted even as early as day 150.
The difference has been the predominantly South-Eastern winds over the Kara January-April and the associated cold on the continental side of the Arctic Ocean.

I think it shows again never to rely on information from one source. This 100K ice is completely different, structure-wise, than last year. Still, it is FYI, and having another good look at it this morning, I think I suggested too early that part of it might survive…

Wipneus · « **Reply #167 on:** July 17, 2013, 08:25:20 AM »

Update 20130716.

Extent: -94k2 (+198k2 vs 2012)
Area: -38.7 (+239k4 vs 2012)

Kara has the most aggressive melt, followed by Laptev and Chukchi. Beaufort, the region that is lagging most to 2012, is showing a slow but steady increase in melt rates (-3, -6, -9) over the last few days.

Updated graphs can be seen in the top-post.

The details (in 1000 km²)

Extent: Arctic Basin East Siberian Sea Laptev Sea -4.7 -2.6 -12.7 Kara Sea Barents Sea Greenland Sea -31.9 0.2 11.7 Baffin/Newfoundland Bay St. Lawrence Hudson Bay -11.6 -0.1 -16.4 Canadian Archipelago Beaufort Sea Chukchi Sea -3.1 -9.1 -13.8 Bering Sea Sea of Okhotsk Total Extent 0.2 -0.5 -94.2 Area: Arctic Basin East Siberian Sea Laptev Sea 10.8 2.6 -14.0 Kara Sea Barents Sea Greenland Sea -14.7 0.2 13.5 Baffin/Newfoundland Bay St. Lawrence Hudson Bay -14.7 -0.1 -0.9 Canadian Archipelago Beaufort Sea Chukchi Sea -1.5 -9.1 -10.1 Bering Sea Sea of Okhotsk Total Area 0.1 -0.8 -38.7

Wipneus · « **Reply #168 on:** July 17, 2013, 08:43:29 AM »

Region of the day is the Canadian Archipelago. Various darkening caused can be seen that are quite stable indicate probable melting. The NW passage via the Parry Channel and McClure Strait does not look like it is opening soon.

Movement in the Nares strait can be seen as well.

(log in to see attached image)

LarsBoelen · « **Reply #169 on:** July 17, 2013, 11:22:24 PM »

Wipneus, being only a second year MeltSeasonWatcher I wonder if you have any insights in the huge differences between IJIS and your calculations. Ijis has basically stalled the past few days, but if you look at the "movie" of the meltseason (http://stormglas.wordpress.com/2013/07/15/daily-amsr2-update-movie-of-northpole-melting-in-2013/) there does not appear to be any slowdown.

It looks like transport through Fram strait might restart, but does this cause an almost immidiate Increase in extend of almost 100K?

werther · « **Reply #170 on:** July 18, 2013, 12:30:57 AM »

At the risk of talking before my turn...
Yes, it does, Lars...
You mention the Fram, I think more of all fringe regions. The ice is extremely mobile now. Spreading while cracking up and filling the voids with rubble and/or melting foam.
In my opinion, the grid on which the automated algorithm is scaled is too coarse to get this.
The melt plot is not the same in each region.
There are parts where the clear, ice covered, structureless swaths of rounded floes now have transited into seemingly endless low-albedoplates with myriads of holes.
Elsewhere (FI the Chukchi), the ice is fading into foam like end July last year.

Through the last few years most graphs/sensors/numbers were ''on-off" this time of the year on the rapidly melting fringes. I think this year is extreme. Weather, DMI temps...I don't think they are as important as they used to be.
It's the structure... I never suggested temporary ice-free this season. But a record, yes. I think some regions will go "in a flash".
It's just a matter of the right pinch-hitter.

Watch out for ECMWF next week. Circlin' low, strong winds.

Wipneus · « **Reply #171 on:** July 18, 2013, 08:29:25 AM »

Update 20130717.

Extent: -108k0 (+123k vs 2012)
Area: -66k2 ( +260k vs 2012)

No special fast runners today. Chukchi has not moved at all, otherwise all active regions continue to decline.
Curiously that area is now quite slow for a couple of days. It was the same like that last year, so the gap with 2012 is not widening.

Have a look at the updated graphs in the top post.

The details (in 1000 km²):
Extent: Arctic Basin East Siberian Sea Laptev Sea -5.3 -8.6 -14.5 Kara Sea Barents Sea Greenland Sea -17.3 0.3 -14.8 Baffin/Newfoundland Bay St. Lawrence Hudson Bay -17.0 0.1 -10.4 Canadian Archipelago Beaufort Sea Chukchi Sea -11.2 -7.5 0.6 Bering Sea Sea of Okhotsk Total Extent 0.0 -2.4 -108.0 Area: Arctic Basin East Siberian Sea Laptev Sea 25.3 -17.8 -12.3 Kara Sea Barents Sea Greenland Sea -22.2 0.1 -20.0 Baffin/Newfoundland Bay St. Lawrence Hudson Bay -12.4 0.0 -8.0 Canadian Archipelago Beaufort Sea Chukchi Sea 2.7 -2.3 1.9 Bering Sea Sea of Okhotsk Total Area 0.0 -1.3 -66.2

Wipneus · « **Reply #172 on:** July 18, 2013, 08:42:20 AM »

Region of the day is Beaufort. The stronghold that has managed to keep the melting (s)low. The polynya in the middle of the pack that I showed a week ago has also closed again.

(log in to see attached image)

Wipneus · « **Reply #173 on:** July 18, 2013, 09:33:15 AM »

Quote from: LarsBoelen on July 17, 2013, 11:22:24 PM

Wipneus, being only a second year MeltSeasonWatcher I wonder if you have any insights in the huge differences between IJIS and your calculations. Ijis has basically stalled the past few days

I would ignore fluctuations lasting "few days" especially when only one of the timelines is showing them. The declining of NSIDC, IJIS and my AMSR2 are nearly the same, and I expect IJIS either to revise or come with a rather big jump soon.

And I would like to know why exactly my AMSR2 extent is currently about 800k below the others. It could be caused by the source data, it could be in the nitty gritty of the calculation details.

In contrary to my first impression, it probably not the resolution difference. Resolution has been in the past overstated as cause for differences in the various extent/area data.

I will tell you more as soon as find out.

And yes, changes in wind patterns can cause distinctive regional changes. They more or less cancel out on the total.

Rob Dekker · « **Reply #174 on:** July 18, 2013, 11:10:48 AM »

Wipneus,
Awesome work you present here ! Thank you !

You may have explained this before, but considering the interesting high extent/area ratio that your calculations show, could you give again a precise definition of how you calculate extent and how you calculate area on your grid ?

Rob Dekker · « **Reply #175 on:** July 18, 2013, 11:20:02 AM »

I'm sorry. I meant the high area/extent ratio, and specifically the high area compared to Cryosphere Today.

Wipneus · « **Reply #176 on:** July 18, 2013, 12:10:43 PM »

Quote from: Rob Dekker on July 18, 2013, 11:10:48 AM

You may have explained this before, but considering the interesting high extent/area ratio that your calculations show, could you give again a precise definition of how you calculate extent and how you calculate area on your grid ?

Always happy to explain :-).

But first, I outlined the calculation in the top post of this discussion.
Further I went into the CT-nigma yesterday, see this page:
https://forum.arctic-sea-ice.net/index.php/topic,418.50.html

Let me try to be exactly precise about extent and area.

There is some pre-processing to handle 1) pole hole; 2) phantom ice; 3) "coastal" pixels; 4) lakes are excluded; and 5) a mask is used to limit the calculations to the well known(?) CT regions.

Definitions:

The extent is calculated as the sum of areas of all grid cells that have concentration >15%

The area is calculated as the sum of areas of all grid cells that have concentration >15% multiplied with the concentration of those grid cells.

Anyone can convince himself by analyzing the concentration maps in https://sites.google.com/site/apamsr2/home/pngcby32

Made to please A-Team in his desire for scientific standard png files, the have ice concentration 0-100% mapped to 0-255. You will notice that the concentrations values are (on average) bigger than those in the CT maps.

Richard Rathbone · « **Reply #177 on:** July 18, 2013, 12:12:11 PM »

Differences with other methods come in the sensors and algorithms used to get concentration. I think the CT method will report a rather lower concentration than Bremen when concentration is close to 1 (something like CT says 0.8 where Bremen says 0.95 from my eyeballing their concentration maps).

Frivolousz21 · « **Reply #178 on:** July 18, 2013, 12:28:57 PM »

Its because you are using high resolution.

3km res will destroy 12km or 25km when it comes to picking up open water within the ice pack like the Chukchi.

This is why norsex is always so high with there totals vs the others.

And why bremen is typically lower.

this is why in 2011 bremen had a new extent record and jaxa didnt.

even though CT showed 2011 beat 2007 in area.

dree12 · « **Reply #179 on:** July 18, 2013, 03:16:30 PM »

Quote from: Wipneus on July 18, 2013, 08:29:25 AM

Curiously that area is now quite slow for a couple of days. It was the same like that last year, so the gap with 2012 is not widening.

I'm curious about how Arctic Basin area has increased (significantly) in the past 5 days! Along with extent decreases, does this mean the Arctic Basin is now more compacted?

Wipneus · « **Reply #180 on:** July 18, 2013, 03:40:41 PM »

Quote from: dree12 on July 18, 2013, 03:16:30 PM

I'm curious about how Arctic Basin area has increased (significantly) in the past 5 days! Along with extent decreases, does this mean the Arctic Basin is now more compacted?

YYYYMMDD Arctic.Basin.extent Arctic.Basin.area A.B.compactness 190 20130709 4.336055 4.248204 0.9797393 191 20130710 4.326195 4.224234 0.9764317 192 20130711 4.321573 4.159117 0.9624081 193 20130712 4.310013 4.099302 0.9511112 194 20130713 4.307020 4.040629 0.9381496 195 20130714 4.301781 4.065671 0.9451134 196 20130715 4.297041 4.109884 0.9564451 197 20130716 4.292375 4.120708 0.9600065 198 20130717 4.287124 4.146012 0.9670847

Vergent · « **Reply #181 on:** July 18, 2013, 08:29:19 PM »

When ice breaks up, it spreads out. More pixels get counted as ice. This is happening in the eastern CAB.

A lot of the eastern CAB is starting to look like this(from 2012).

Verg

seaice.de · « **Reply #182 on:** July 18, 2013, 10:14:08 PM »

Great work, congratulations, Wipneus!

I am looking forward to a very exciting melt season in the next weeks...

Cheers from Bremen
LK

Peter Ellis · « **Reply #183 on:** July 18, 2013, 11:30:39 PM »

Quote from: Vergent on July 18, 2013, 08:29:19 PM

When ice breaks up, it spreads out. More pixels get counted as ice. This is happening in the eastern CAB.

A lot of the eastern CAB is starting to look like this(from 2012).

Verg

No, the eastern CAB looks nothing like this whatsoever. Those floes are around 10 metres across - i.e. you could fit 25x25 = 625 of them in a single MODIS pixel at maximum resolution. If you go back to MODIS imagery from that location at that date, this is what it looks like.
http://earthdata.nasa.gov/labs/worldview/?map=-555264,766976,-214528,993536&products=baselayers,MODIS_Terra_CorrectedReflectance_TrueColor,!MODIS_Aqua_CorrectedReflectance_TrueColor~overlays,polarview%3AgraticuleN,arctic_coastlines&time=2012-08-28&switch=arctic

The ice in the eastern CAB may be somewhat broken up, but it's still in floes kilometres across, not metres. Honestly, some people on these boards have literally no concept of the different scales involved in microwave imagery, MODIS imagery, and on-the-ground photographs.

ghoti · « **Reply #184 on:** July 19, 2013, 02:18:32 AM »

I think the scale of the MODIS imagery also means the ice could be like the picture presented but we'd not see it. It is such a pity the Polar Stern doesn't have an Arctic cruise this year giving us an actual view of the ice. Doesn't look like the Healy is going anywhere far from shore this year either. We could really use a couple of Global Hawks cruising back and forth over the ice to provide better ground truth. Oh well it isn't going to happen so we're stuck guessing wildly.

Vergent · « **Reply #185 on:** July 19, 2013, 02:34:51 AM »

Quote from: Peter Ellis on July 18, 2013, 11:30:39 PM

The ice in the eastern CAB may be somewhat broken up, but it's still in floes kilometres across, not metres. Honestly, some people on these boards have literally no concept of the different scales involved in microwave imagery, MODIS imagery, and on-the-ground photographs.

Peter

The pixel size for CT area is 15 Km(edit: best recollection). So the principle is that when ice breaks up to a scale smaller than the pixel and consequently spreads out, the area algorithm will see more ice. I chose the picture for purpose of illustration, not for scale. If you can find a picture that illustrates this with kilometer sized chunks, you are welcome to. With a 15 km pixel it makes no difference weather the break-up is at the kilometer scale or the 10 m scale, the geometry is the same.

Honestly, some people on these boards have no concept of when a difference in scale makes no difference to the validity of the argument.

Vergent

jdallen · « **Reply #186 on:** July 19, 2013, 02:52:03 AM »

Quote from: Vergent on July 19, 2013, 02:34:51 AM

Quote from: Peter Ellis on July 18, 2013, 11:30:39 PM

The ice in the eastern CAB may be somewhat broken up, but it's still in floes kilometres across, not metres....

The pixel size for CT area is 15 Km. So the principle is that when ice breaks up to a scale smaller than the pixel and consequently spreads out, the area algorithm will see more ice. I chose the picture for purpose of illustration, not for scale. If you can find a picture that illustrates this with kilometer sized chunks, you are welcome to. With a 15 km pixel it makes no difference weather the break-up is at the kilometer scale or the 10 m scale, the geometry is the same.

I would hasten to add, there is a LOT of far smaller ice, between the floes we can resolve - that also is rather obvious. I suspect there's quite a bit like this about, where we see areas broken up on an obviously larger scale.

Vergent · « **Reply #187 on:** July 19, 2013, 06:52:29 AM »

Quote

A lot of the eastern CAB is starting to look like this(from 2012).

In fairness to Peter, I have to assume that English is his second language.

Vergent

Wipneus · « **Reply #188 on:** July 19, 2013, 07:29:24 AM »

Quote from: seaice.de on July 18, 2013, 10:14:08 PM

Great work, congratulations, Wipneus!

I am looking forward to a very exciting melt season in the next weeks...

Cheers from Bremen
LK

Welcome to the forum!

Thanks for what appears to be excellent data, it surely gives a lot of pleasure to work with.

Wipneus · « **Reply #189 on:** July 19, 2013, 08:20:16 AM »

Update 20130718.

Extent: -97k6 (+120k vs 2012)
Area: -110k2 ( +281k vs 2012)

CAA makes a massive jump downward today. Together with Kara and the last bit of ice in the northern parts of the Hudson that is where the action is, the other regions are on a slow decline.

Updated graphs can be seen in the top post.

The details (in 1000 km²) :

Extent: Arctic Basin East Siberian Sea Laptev Sea -1.0 -1.6 -5.4 Kara Sea Barents Sea Greenland Sea -21.0 -0.3 -5.2 Baffin/Newfoundland Bay St. Lawrence Hudson Bay -2.2 0.4 -16.2 Canadian Archipelago Beaufort Sea Chukchi Sea -37.7 -6.9 0.9 Bering Sea Sea of Okhotsk Total Extent -0.3 -1.1 -97.6 Area: Arctic Basin East Siberian Sea Laptev Sea -6.7 -18.9 -9.5 Kara Sea Barents Sea Greenland Sea -5.9 -0.2 0.5 Baffin/Newfoundland Bay St. Lawrence Hudson Bay 1.4 0.4 -19.3 Canadian Archipelago Beaufort Sea Chukchi Sea -46.9 -1.4 -2.9 Bering Sea Sea of Okhotsk Total Area -0.3 -0.5 -110.2

Wipneus · « **Reply #190 on:** July 19, 2013, 08:36:57 AM »

Region of the day is the Atlantic side of the CAB with what I call the Greenland, Barentsz and Kara sector.
There is quite a lot to notice, the large dark areas in the CAB with concentration between 20-60% looks like it could open real soon now. There is extent expansion: the wind vectors are parallel to the ice edge making the ice move outwards. Slow motion breaking of the fast ice in NE Greenland.

(log in to see attached image)

Anne · « **Reply #191 on:** July 19, 2013, 09:06:01 AM »

Quote from: ghoti on July 19, 2013, 02:18:32 AM

I think the scale of the MODIS imagery also means the ice could be like the picture presented but we'd not see it. It is such a pity the Polar Stern doesn't have an Arctic cruise this year giving us an actual view of the ice. Doesn't look like the Healy is going anywhere far from shore this year either. We could really use a couple of Global Hawks cruising back and forth over the ice to provide better ground truth. Oh well it isn't going to happen so we're stuck guessing wildly.

Polar Stern's sister ship Polar Star is somewhere up there out of Unalaska doing sea trials after her refit. But I haven't been able to find out where, or whether any obs will be made public.

danp · « **Reply #192 on:** July 19, 2013, 09:18:30 AM »

Quote from: Anne on July 19, 2013, 09:06:01 AM

Polar Stern's sister ship Polar Star is somewhere up there out of Unalaska doing sea trials after her refit. But I haven't been able to find out where, or whether any obs will be made public.

Good find Anne! Digging around I came across a little info here: http://www.dvidshub.net/image/973522/coast-guard-conducts-arctic-domain-awareness-flight

There are some photos from July 16th of the Polar Star at the edge of the ice in the Chuckchi sea. There are also some other neat photos taken from various other Arctic overflights, instrument drops, etc.

Wipneus · « **Reply #193 on:** July 19, 2013, 10:42:06 AM »

For fun, see attached images (log in etc.)

First a detail of the AMSR2 image above, showing the black area in the top.

Then the same spot in today's MODIS aqua (250m) image. Mentally rotate 45 degrees anti clockwise and squint eyes to remove cloud blurs and improve resolution.

Neven · « **Reply #194 on:** July 19, 2013, 11:25:19 AM »

Wow, that's really impressive!

LarsBoelen · « **Reply #195 on:** July 19, 2013, 04:57:49 PM »

Wipneus, maybe you explained elsewhere, but couldn't find it:

Is the fact that the Canadian Archipelagao started at .8 for the first half of 2012 and at .7 for the first half of 2013 caused by a definition change of what counts as CA-area?

Pmt111500 · « **Reply #196 on:** July 19, 2013, 05:03:38 PM »

Thanks Wipneus for the comparison between AMSR2 and Modis aqua visible, looks like AMSR2 is proving itself to be a great instrument!

Wipneus · « **Reply #197 on:** July 19, 2013, 05:36:04 PM »

Quote from: LarsBoelen on July 19, 2013, 04:57:49 PM

Wipneus, maybe you explained elsewhere, but couldn't find it:

Is the fact that the Canadian Archipelagao started at .8 for the first half of 2012 and at .7 for the first half of 2013 caused by a definition change of what counts as CA-area?

Yup, I explained that before, but maybe not very well.
Let me try again,

The AMSR2 3.125 km grid data exists only for 2013, so to compare with 2012 I had to use something else. From the same source comes an SSMIS concentration data, available for 2012 and 2013, but in a different grid: 12.5 km. That is 16 times as big, by area.
Yet the SSMIS 12.5 km data and the AMSR2 3.125 km data are quite similar , see this post:
https://forum.arctic-sea-ice.net/index.php/topic,382.msg9041.html#msg9041

So it seems 12.5 km or 3.125 km don't matter very much?
Well for the CAA with its narrow channels and straits it does matter somewhat.
Imagine making a 12.5 km grid, and divide the cells in "land" and "sea" cells. Sea cells must contain almost 100% sea, otherwise it is land. Now make a 3.125 km grid by dividing all those grid cells into 16 smaller cells, some of those marked "land", will now contain only sea -> in total there is more sea surface.

The actual numbers are for the total CAA area:
The amsr2 3.125 km grid counts 0.7917134 Mm²
The ssmis 12.5 km grid counts counts 0.7034836 Mm²

I could compensate for that, perhaps in version 0.0.2

LarsBoelen · « **Reply #198 on:** July 19, 2013, 11:08:00 PM »

Thanks Wipneus,

That totally clarifies it for me! Maybe you can add something to the CAA graph like "* CAA baseline differs due to grid size change".

I think more people wondered why that was, or maybe they didn't :-)

Lars

AartBluestoke · « **Reply #199 on:** July 20, 2013, 01:46:10 AM »

given that there are issues with costal pixels anyway, would there be a significant downside to producing numbers at the high resolution with the low resolution ground mask? ie; just ignore that extra area when computing some of the numbers, for the purpose of comparing the 2 datasets?

News:

Author Topic: Home brew AMSR2 extent & area calculation (Read 1981462 times)

Pmt111500