the "0" margin is the outline of the Antarctic ice sheet and nearby glaciated islands from BEDMAP2 at 500 m resolution



The RAISED reconstructions are used for the basis of the 20000 year margin, as are the 15000, 10000 and 5000 year margins for West Antarctica and the western parts of East Antarctica. I am using the "scenario B" margin at first, to try and maximize the ice volume. The margins have been adjusted in the intermediate time steps for the following reasons:

- The inferred margins sometimes went over the modern margin
- the intermediate time step margins sometimes overlapped with each other
- the inferred margins sometimes omitted glaciated islands, such as Elephant Island

I assume that the inconsistencies are a result of the margins being drawn on a physical map, then digitized independently, rather than being indicitive of a readvance, just due to the low resolution of digitized margins.

I used the contour interpolation tools to create the margins at other time periods. The work flow went like this:

1) First, I interpolated the margin between 20000 years and 0 years. I used the psudo 15000 year margin as the 15000 year margin for East Antarctica. The margin at Dronning Maud Land was specifically adjusted from the RAISED margin to account for a grounded rise at -74N, -21E

2) I took the resulting 15000 year margin, and again interpolated to the modern margin to a psudo 10000 year margin. Several adjustments to the RAISED margin had to be made on the Antarctic Peninsula.

3) I took the resulting 10000 year margin, and again interpolated to the modern margin to a psudo 5000 year margin. Several adjustments to the RAISED margin had to be made on the Antarctic Peninsula. The 5000 margin from the contour interpolation generally lined up favourably with the RAISED margin where there was overlap, so I used the psudo 5000 year margin, rather one with a more retreated extent. Obviously assuming a linear retreat is not going to work everywhere, but there isn't really a better assumption in areas with almost no data.

4) The margins at 500 year intervals were done between 20000 and 10000 were done at a relatively low resolution (2000 m). This was sufficient in most places, and when there were problems, I manually fixed them.

5) the interpolation between 10000 and 0 were done at a higher resolution because there were problems when the input margins got close together. Lightly edited, especially between 5000 and 0, some places still have inconsistencies and overlaps. Ideally, I want to go back and redo the 5000 to 0 year margins using the original unsimplified polygon, which should fix a lot of the problems. We'll see if I have time.

6) For 30000 to 20000, I used the existing margins from between 20000 and 15000, but going backwards. That means 19500 margin represents 21000, 19000 represents 22000, and so on.


----------
Pre-30000
----------

Not much is directly known.

MIS 3 - Berg et al 2016 - eastern Prydz Bay was ice free during most of MIS 3, possibly warmer than present. Ice overrode the site around 32000 yr BP. Margin was likely similar to present during most of MIS 3, but the ice inland might have been thicker. Exposure date indicates long period of ice free conditions, so this location may not have been glaciated during MIS 4 either.

MIS 3 - Emslie et al 2007 - fossil penguin remains indicate that the Ross Ice Shelf was probably similar in extent to present during MIS 3, and did not advance until after 30000 yr BP. MIS 4 extent is unknown, but there were some non-finite ages for remains.

MIS 3 - Bart and Cone 2012 - in contrast to the Emslie paper, they suggest the western Ross Shelf achieved a maximum extent around 30000 yr BP, based on a plethora of MIS 3 aged forams in a diamict in the grounding zone wedge, and MIS 2 aged formas in the overlying sediment

MIS 5-3 - Bart and Owolana 2012 - they propose that there was a gradual advance through the entire glacial cycle to the LGM limit from some MIS 5e minimum in the West Antarctic sourced part of the Ross Sea ice sheet

MIS 4 - Levy et al 2017 - evidence of an MIS 4 advance for the western Ross Sea region, maybe 70000-80000 yr BP

> MIS 3 - Ingólfsson et al 2003 - West Antarctic maximum predates the Last Glacial Maximum

Framework
---------

It appears that the East Antarctic ice sheet did not really advance beyond modern extent until close to the LGM, perhaps starting around 35000 yr BP. It is probably safe to keep it there through the whole glacial cycle before that.

West Antarctica is a different story, apparently reaching a maximum limit by 30000 yr BP. It may have gradually expanded through the full glacial period from MIS 5.


---------------------------------------------------------------------------
2019/05/02
new margins
---------------------------------------------------------------------------



---------------------
East Antarctica
---------------------

Starting over, because the history of East Antarctica (at least) was wrong

As before, using the 20000 yr BP margin from the RAISED consortium, using scenario B (with further ice extent)

15000, 10000, and 5000 margins for West Antarctica are also from the RAISED consortium, using scenario B

The 0 margin is from the Bedmap grounding line. To smooth it out, I did a -5000 m buffer, followed by a 5000 m buffer, which gets rid of narrow peninsulas and fjords and makes the margin quite smooth. Doing this means I have to readjust the 5000 year margin in west Antarctica

Mackintosh et al 2014 (East Antarctica):
I set the East Antarctica margin at 5000 to be the same as present. There are plenty of data to suggest that margin reached its modern extent by 7000 yr BP. They also state that most of the ice retreat happened after 12000 yr BP

2500 - also set it to be the modern extent
7500 - also set it to be the modern extent
10000 - most places it is close to the modern extent - I took the modern outline and expanded it by 2 km
12500 - put it closer to the LGM extent than modern, but fairly close to halfway
15000 - put it about halfway between the 12500 and maximum extent
17500 - put it about halfway between 15000 and maximum extent
20000 - RAISED margin
22500 - same as RAISED 20000




----------
Pre-20000
----------

Not much is directly known.

MIS 3 - Berg et al 2016 - eastern Prydz Bay was ice free during most of MIS 3, possibly warmer than present. Ice overrode the site around 32000 yr BP. Margin was likely similar to present during most of MIS 3, but the ice inland might have been thicker. Exposure date indicates long period of ice free conditions, so this location may not have been glaciated during MIS 4 either.

MIS 3 - Emslie et al 2007 - fossil penguin remains indicate that the Ross Ice Shelf was probably similar in extent to present during MIS 3, and did not advance until after 30000 yr BP. MIS 4 extent is unknown, but there were some non-finite ages for remains.

MIS 3 - Bart and Cone 2012 - in contrast to the Emslie paper, they suggest the western Ross Shelf achieved a maximum extent around 30000 yr BP, based on a plethora of MIS 3 aged forams in a diamict in the grounding zone wedge, and MIS 2 aged formas in the overlying sediment

MIS 5-3 - Bart and Owolana 2012 - they propose that there was a gradual advance through the entire glacial cycle to the LGM limit from some MIS 5e minimum in the West Antarctic sourced part of the Ross Sea ice sheet

MIS 4 - Levy et al 2017 - evidence of an MIS 4 advance for the western Ross Sea region, maybe 70000-80000 yr BP

> MIS 3 - Ingólfsson et al 2003 - West Antarctic maximum predates the Last Glacial Maximum

Framework
---------

It appears that the East Antarctic ice sheet did not really advance beyond modern extent until close to the LGM, perhaps starting around 35000 yr BP. It is probably safe to keep it there through the whole glacial cycle before that.

West Antarctica is a different story, apparently reaching a maximum limit by 30000 yr BP. It may have gradually expanded through the full glacial period from MIS 5.

East Antarctica

35000 and earlier - use modern margin. Including MIS 4
32500 - make this halfway between modern and 30000
30000 - make this halfway between modern an 27500
27500 - make this time slice be halfway between modern and LGM
25000 - make this halfway between 27500 and LGM


------------------------------
Ross Sea - West Antarctic component
-------------------------------

See review paper by Anderson et al 2014

- Their 20000, 15000, 10000 and 5000 margins are the same as RAISED. Some notes

2500 - grounding line margin was at modern extent about 2000 yr BP, so the 2500 margin is drawn to be close to this.
5000 - RAISED
7500 - grounding line was somewhere north of Ross Island at ~7800 yr BP. The margin at 7500 is close to that
10000 - RAISED
12500 - just put it intermediate of the 15000 and 10000 margins
15000 - RAISED
17500 - put it closer to the LGM limit than the 15000 limit
20000 - RAISED
22500 - same as 20000
25000 - inwards in the west side, close to LGM limit on east side
27500 - same strategy as 25000
30000 - at this point, the west Antartic sourced ice was likely somewhat greater than the LGM limit, but the Emslie et al 2007 paper makes it clear that the ice margin probably was only going over Ross Island at this point.
32500 - west side is close to the maximum, but east side grounding line is south of Ross Island

The rest of the margins follow the idea from Bart and Owolana 2012 that West Antarctic component gradually grew from some minimum in MIS 5, all the way through stage 3.

35000,37500 - for 35000 and 37500, I tried to make them at equal intervals between 32500 and 40000
40000 - put the margin in a location somewhat similar to 10000 (at least in the west side)
42500,45000,47500 - gradual advance between 50000 and 40000
50000 - Put the margin halfway between 40000 and 70000
52500,55000,57500 - gradual advance between 50000 and 60000
60000 - closer to 50000 than 70000
62500 - halfway between 70000 and 60000 - to show a more rapid advance towards the MIS 4 peak
65000, 67500 - gradual advance between 70000 and 62500
70000 - margin somewhere close to the modern extent (though not as detailed).

--------------------------------
Amundsen Sea and Bellingshausen Sea
--------------------------------

See review paper by Larter et al 2014


In this sector, it seems that the maximum extent was achieved prior to the LGM, perhaps best placed before 30000 yr BP (i.e. maybe 32500 yr BP). By 25000, it had retreated a fair amount already.


2500 - If the RAISED margin is to be believed, there was significant margin retreat between 5000 and 0 yr BP. I tried to draw it so that it was sort of between the modern grounding line (which may have retreated a fair amount during the past 50 years) and the RAISED margin
5000 - RAISED 5000 yr BP margin
7500 - intermediate of 5000 and 10000
10000 - RAISED 10000 yr margin
12500 - Larter  suggests that there was still grounded ice on the shelf into the Holocene in the Amundsen Sea, so it makes sence to keep the ice margin on the shelf at 12500.
15000 - RAISED 15000 yr margin
17500 - between the 20000 and 15000 margins
20000 - RAISED 15000 yr margin
22500 - intermediate of 20000 and 25000 margin
25000 - Larter et al 2014 25000 yr margin
30000,27500 - intermediate of 25000 and 32500
32500 - set as maximum extent - the entire margin at the shelf edge. This is the time because there are dates indicating some retreat by 30000, and there are dates older than 35000 that probably predate the advance.

Prior to 30000, there are not a lot of dates, but it seems that at 40000, the margin might not have been far advanced on the shelf. I will set the 40000 yr margin first, and then gradually advance it to the 32500 maximum.

37500, 35000 - advance to 32500 margin
40000 - limited extent, only outside of islands in the region

There is an offshore record by Vautravers et al (2013) which is located just up-peninsula from the region investigated by Larter et al 2014. They suggest that throughout MIS 3 there was open water conditions, indicating a restricted ice sheet extent. During MIS 4, the ice sheet was likely extensive, though not as much as the peak in MIS 2. The exact timing of events in this core is probably not exact, as it is based on magnetic properties and assumptions on accumulation rate. They suggest ice rafted debris increased after 40000 yr BP. Because of this, I set the margin of the ice sheet to be similar to be the 5000 margin for the entirety of MIS 3, from 55000 yr BP to 40000 yr BP

42500, 45000, 47500, 50000, 52500, 55000 - same as 5000 margin

For the MIS 4 advance, I also put 70000 to be the same as the 5000 margin

57500 - closer to 55000 and 60000, signalling rapid retreat
60000 - set as maximum of the MIS 4 advance, not as extensive as MIS 2 maximum, but advances in the major seas
62500 - set to be fairly close to the maximum extent
65000, 67500 - advancing
70000 - same as the 5000 margin

-----------------------
Antarctica Peninsula
-----------------------

See review by O Cofaigh et al 2014

I had questions about the whether the 5000 margin for Larsen C was the grounding line or shelf edge. I emailed Colm O Cofaigh, and he told me that they could not differentiate between whether it was the shelf edge or grounding line, so they set it as the grounding line. Highly uncertain, obviously, and if the global sea level doesn't fit, this would be a target for editing. But for now, I assume it is the grounding line.

2500 - intermediate of the 0 and 5000 margin
5000 - RAISED 5000 margin
7500 - intermediate of 10000 and 5000 - in places with embayments, closer to 5000
10000 - RAISED 10000 margin
12500 - According to the review, most of the retreat happened between 15000 and 10000, I think it would make sense to have more rapid retreat before 12500 than after, from Sea Level considerations
15000 - RAiSED 15000 margin
17500 - intermdiate of the 20000 and 15000 margins, but it makes sence to place the
20000 - raised 20000 margin
25000 - O Cofaigh used the same extent at 25000 as 20000, and I am doing the same.

Much like the Bellingshausen Sea and Amunsen Sea, it is assumed the actual maximum happened prior to 30000 yr BP, and that there was already some retreat by 30000 yr BP. Following what I did for other areas of East Antarctica, I set the maximum to be at 32500, and slightly retreat to the 25000 margin. For this, I am going to use the 20000 margin, but expand it a bit in areas with unglaciated continental shelf.

30000, 27500 - slight reduction in extent from 32500
32500 - extended the ice further along the shelf.


As for the rest of MIS 3, I do the same thing as the Bellingshausen Sea and Amunsen Sea, by setting the margin to be similar to 5000 between 55000 and 42500, then advance to 32500. In contrast, I am going ot link together the nearby islands, and I would assume the ice is a bit more extensive than the mid-Holocene minimum. I assume the main shelves are grounded away from the coast.

35000,37500,40000, gradual advance to maximum
42500,45000,47500,50000,52500,55000 - similar to the 5000 year magin, but shelf regions are grounded

For MIS 4, I follow what I did for the Bellingshausen Sea and Amunsen Sea. Note that there is some evidence that in New Zealand (Schaefer et al 2011), the ice caps were larger in MIS 4 than in MIS 2, and that may have also been true for parts of Antarctica.  Since the peninsula is further north, it was possibly affected by the same oceanic processes lead to an extent that was similar to the LGM extent. Therefor, I am placing the maximum extent, at about 65000 (as per Schaefer), for the northern parts of the Antarctic Peninsula, gradually decreasing closer to the Bellingshausen Sea.

57500 - Schaffer et al suggest that by 58000, the ice sheet should be substantially reduced, so I place it pretty close to the 55000 margin
60000 - halfway between 62500 and 57500
62500 - still close to MIS 4 maximum
65000 - maximum MIS 4 extent, almost the same as late MIS 3 extent
67500 - advance between 70000 and 65000
70000 - set it to be the same as 42500 margin

------------------------
Weddell Sea
------------------------

See review by Hillenbrand et al 2014

There are two proposed reconstructions. Scenario A is based on ice sheet modelling studies, while scenario B is more based on geological constraints, so I am using scenario B. Regardless of the scenario, Hillenbrand et al say that the ice surface profile was likely very shallow, to account for the lack of thickening at terrestrial sites. Much like other West Antartical areas, it appears that a maximum in extent was achieved by 34000, and retreat may have already been underway by 30000. Their 25000 yr BP time slice shows this, that at that time there was more extensive ice than at 20000. Their 25000 time slice has the ice going right to the edge of the continental shelf, except near Coats Land.

2500 - closer to present than 5000
5000 - RAISED scenario B 5000 margin
7500 - between the 5000 and 10000 margins
10000 - RAISED scenario B 10000 margin
12500 - intermediate of 15000 and 10000 margin
15000 - RAISED scenario B 15000 margin
17500 - closer to 20000 and 15000
20000 - RAISED scenario B 20000 margin
22500 - between 20000 and 25000
25000 - Hillenbrand et al scenario B 
30000,27500 - gradual change to 25000 configuration
32500 - make this the maximum time like other places in West Antarctica. Have the margin go right to the shelf edge. Of course, towards East Antarctica, it is not advanced to the shelf yet


There really isn't any evidence about what the Weddell Sea sector was like prior to the LGM. The shear stress of any grounded ice in the area that is now floating shelf was likely very low, so the ice thickness above the floating point was likely not substantial. As a result, it probably isn't important exactly where the margin is. As a result, between 40000 and 55000, I am setting the margin to be equivalent to the 5000 year margin, and the MIS 4 maximum at 60000 to be equal to the 10000 margin. 70000 starts from the 5000 margin

-----------------
2019/06/05
----------------

I want to extend this to 80000, so I simply used 70000 for the other time slices.
