MIS 4 - Eyles et al 2005 - MIS 4 till is the result of a late advance over proglacial lake sediments, indicating that the ice sheet blocked the St Laurence river drainage before advancing to Ontario

MIS 4 - Johnson et al 1997 - does not show MIS 4 advance into southern Great Lakes

MIS 4 - Karrow et al 2000 - early part of Ontario subepisode is a proglacial lake, later ice advance. They place peak glacial at 70000, but that is highly uncertain and the dates by Berge and Eyles (1994) suggest a later age (40-45 ka). I am splitting the difference and saying 60,000.

MIS 3 - Likely one short lived re-advance into southern Ontario between 50 and 30 ka (Mulligan and Bajc 2018), I have placed this to be equivalent to Heinrich 5 (45000 yr bp)

MIS 3 - Advance over Lake Ontario likely for a brief time (Karig and Miller 2013). They suggested it was coincident to Heinrich 4

MIS 3 - Manley and Jennings 1996 - southern Baffin Island - some shells are in glacio-marine sediment, some are in till, they were hesitant to place them in MIS 3 because of how old they were. If reliable, Hudson Strait was ice free until about 30,000 14C years. Note AA-10252 is likely in-situ and retains a pearly lustre, and has an age of 30,340 +- 450 14C years, calibrates to about 33,000 yr BP

MIS 3 - Munroe et al 2016 - suggests that blockage of the St Laurence valley happened sometime after 35,000 yr BP

MIS 3 - Young and Burr 2006 - incursion into upstate New York, which they dated to Heinrich 4, compatible with the Karig and Miller (2013) data above

MIS 3 - Briner et al 2007 - Lake CF8 on the northeast side of Baffin Island. They called the radiocarbon dates here "non-finite", so that they could ascribe the age of the unit as MIS 5a. I'm not really sure why they had that interpretation, since the ages are finite, the youngest which is 34000 14C yr BP. Could just as easily be MIS 3, separating the MIS4 and MIS 2 glaciations.

MIS 3 - McMartin et al 2019 - Ice free conditions in Rae Isthmus/Repulse Bay region 35000-43500 yr BP

MIS 3 (?) - McMartin and Dredge 2005 - early flow directions were in a south-southeast direction in central coastal Kivalliq, indicating expansion from the north

MIS 3 (?) - Trommelen et al 2012 - southeastward flow directions in northern Manitoba, possibly related to advance? I made it so that the ice margin from 35000-32500 yr BP goes in this direction.

MIS 3 - Belanger et al 2014 - persistent ice free conditions in central Saskatchewan through all of MIS 3, with vast forests, indicating the margin was far to the north

MIS 3 - Paulen et al 2019 - west of Great Slave Lake - ice free conditions in MIS 3. Wood fragment dated to about 40000 yr BP

MIS 3 - Paulen et al 2005 - northern Alberta - wood found in an organic layer between two tills, indicating forests were present in MIS 3

MIS 3-4 - Trommelen and Levson 2008 - non-finite aged radicarbon dates in sediments under late Wisconsin Laurentide diamiction - no Canadian Shield clasts in pre-late Wisconsin sediments indicating a single advance into northeastern British Columbia

MIS 3 - Smith 1992 - redeposited wood in the Mackenzie valley west of Great Bear Lake, indicating ice advance here is not until after 30000 yr BP

MIS 3 - Hughes et al 1993 - Little Bear section, just west of the Smith location indicates only one Laurentide ice advance (late Wisconsin) but several (five) local montane glaciations separated over long periods as there are soils developed after each advance. They dated a piece of spruce wood from non-glacial sediments under the Laurentide ice advance, and it gave an age that is considered to be infinite. One notable thing is that all of the soils indicate that climate conditions here were warmer than the Holocene, and there is no evidence that the soils were subjected to cryoturbation.

MIS 3 - Hughes et al 1981 - only one Laurentide glaciation in Northern Yukon, north of the Mackenzie mountains, younger than 40000 yr BP

> MIS 3 - Catto 1996 - non-finite ages on organics located underneath the late-Wisconsin advance. There is a sequence of non-glacial fluvial sediments overlying another Laurentide till, which is of unknown age (MIS 6?). This paper has a margin reconstruction for the pre-Late Wisconsin magin.

MIS 3 - Catto et al 1996 - Plenty of finite MIS 3 ages that indicate that there was no pre late-Wisconsin advance into the Peace River/Grand Prairie region of Alberta. There was a  pre-MIS 3 Cordilleran Ice SHeet advance (dates not added, but there are a lot)

MIS 3 - England et al 2009 - only a single, late Wisconsin advance of the Laurentide Ice Sheet over Banks Island and Melville Island. Prior to that, it was not glaciated. There are lots of MIS3 aged shells in the late-Wisconsin till.

MIS 3 - Lakeman and England 2014 - on the eastern side of Banks Island, there is evidence of multiple glaciations, but only one of these is in the late Pleistocene, the late Wisconsin glaciation. The other glaciations were prior to the MPT. The ice margin at MIS 4 or MIS 6 should not reach Banks Island.

MIS 3 - England Et al 2006 - The Innuitian ice sheet only started to advance during MIS 2. In MIS 3 and MIS 4 (and possibly earlier), the ice likely did not advance beyond the modern ice caps. I am just using modern ice cap extent for all times.


For Heinrich 4, I am assuming that this was caused by a Foxe Dome advance, because I think the Labrador sector would not have enough time to merge into Hudson Bay.


Cordilera
------------



MIS 3 - Clague and Ward 2011 - they have time slices for 35000 and 30000 yr BP, this is what I am using. There was not much ice at 35000.


The above margins do not include some areas where there is currently ice, I have shifted it to encompass these regions

MIS 4 - Kaufman et al 2011 - In Alaska, the MIS 4 extent is larger than than the MIS 2 (and probably MIS 6) limits. Modifying the Ward and Clague margins so they do not go over the MIS 2 limits. Basically at 30000, I am putting the Alaska limit so that it is not advanced onto the shelf.

MIS 4 - Ward et al 2007 - Northern Cordillera ice sheet was larger in Stage 4 than Stage 2 in some places in the Yukon

MIS 4 - Cosma and Hendy 2008 - increased sediment input prior to 42000 yr BP off the coast of Vancouver Island, probably some expansion during MIS 4

MIS 4 - McDonald et al 2012 - they suggest that there were Missoula floods during MIS 4, indicating the ice sheet was quite extensive within central British Columbia, so that it could advance into Washington

MIS 4 - Mathewes et al 2015 - penultimate till on eastern Haida Gwaii is likely MIS 4 in age, based on an optical date (~57000 yr BP) in sediments overlying that till


MIS 4 - Tulenko et al 2018 - dates in western Alaska confirm that MIS 4 was somewhat more extensive than MIS 2

MIS 4 - Turner et al 2013 - in central Yukon, the MIS 4 limits are somewhat further out than MIS 2 limit (and the MIS 6 limit somewhat further than that)

MIS 4 - Dethier et al 2008 - This paper does some Argon dating on a non-glacial deposit in the Puget lowland, called the Whidbey Formation, which is MIS 5 in age. This is overlain by a till, called the Possession Glaciation, which has to be MIS 4 in age. The Possession Glaciation is a Cordillera Ice Sheet sourced glaciation (Easterbrook 1992)

Outline: extensive, slightly larger than MIS 2 (at least in the north) Cordilleran Ice Sheet, maybe not so large in the south. Set maximum at 55000, somewhat later than the Laurentide (same thing happened in MIS 2)

37500 - 47500 - use present extent

50000 - restricted to mountaious areas, based on 13000 margin (from Dyke 2004)
52500 - set to the same as 60000
55000 - used information on MIS 4 maximum extents, as above. Notably, the Kaufman et al 2011 margin
57500 - set same as 55000
60000 - set limits inwards by 50 km, except in northern Alaska
62500 - set it to between 65000 and 60000
65000 - set to have mountains glaciated, but not advanced to Interior. Largely continuous ice sheet at this point
67500 - set it to be between 70000 and 65000
70000 - set as 35000

Greenland
-----------

Alley et al 2010 (review) - east-central Greenland (Scoresby Sund) - to the mouth of Scoresby Sund at MIS 4 peak, maybe remained near the edge after that. Further south at Kangerdlugssuaq, the ice reached continental shelf edge at about 31 ka. In Thule region (northwestern Greenland), there may have been advances in MIS 5d, 5b and 4. With some retreat in 5c, 4a, and possibly MIS 3.

Simon et al 2014 - advances of central-western Greenland at 110–105, 91–85, 72–63 and 32–16 ka with smaller advances at 58 and 55 ka and around 44 ka. The MIS 4 glaciation might have been close to the same extent as MIS 2. MIS 2 maximum extent probably happened by 32 ka

--------
Framework
--------

Using the LGM extent in Funder et al 2011. It is more extensive than Dyke 2004. I will have to fix Dyke 2004 later. Present day margin is from the grounding line in Rtopo2 (Schaffer et al 2016)

At 70000 yr BP, I will assume that the ice sheet is the same as present. (from Rtopo2)

At 60000 yr BP, I will have the ice sheet at an extent that is on the ice shelf, but not as extensive as the LGM. I did a buffer of the LGM extent, setting it 25 km inwards from it, along with some tweaking to make sure it looked realistic.

Heinrich event times (52500, 45000, 37500), set it to be 25 km inwards of the MIS 4 (60000 extent). For Heinrich event at 32500, set it to MIS 4 extent (60000 extent). Also set 57500 yr BP to be this extent (as per Simon et al 2014)

For times between heinrich events (55000, 50000, 47500, 42500), I used a buffer of 25 km from the modern extent, modified to make it more realistic

For 40000 (implied MIS 3 minimum), I did a -10 km buffer from the above margin (i.e. between heinrich events)

For 35000, did a -5km buffer from the Heinrich event margin

For 62500, used the same as Heinrich Event extent

For 65000, used  the same as 55000

For 65700, used the same as 40000

Iceland

There is a nice review by Patton et al (2017). Suggests that ice advance probably was not significant until close to the LGM. Ice may not have reached the coast until as late as 25000, so it is probably safe to assume the for most of MIS 3, the ice extent was not much further than present. The lack of ice rafted debris from Iceland during MIS 3 supports this assertion (Andrews 2008). Patton's experiments sugested that there was ice buildup between 35000 and 30000 yr BP, but not reaching the shelf, and the western peninsulas were ice free. Between 35000 and 30000, I am drawing this kind of advance.

Obrochta et al (2014) presented a deep sea sediment core record from the mid-Atlantic. They have records of Iceland glass shards, which shows two peaks between 60000 to 50000. They warn that this might not be due to icebergs from Iceland, but from sea ice transport (i.e. volcanos erupted, and the ash deposited on the sea ice). Nevertheless, I think putting an Iceland ice sheet advance during MIS 4 makes sense. However, the synthesis by Eiríksson 2008) does not indicate a MIS 4 glaciation is present, rather a MIS 6 and MIS 2. I think it is safe to put the glacial limits at 60000 and 55000 (to match Obrachta record) to be the same as 30000, with smaller extents from 62500 to 52500.

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

There was a new paper by Gauthier et al (2019) that presents a relative chronological framework for northern Manitoba. As a result of this, I need to make revisions to the margions. One notable thing is that the flowset that may have been MIS 4 indicates almost purely westward flow, indicating a lack of Keewatin influence. As a result, I am reducing the Keewatin extent, and expanding the Labrador extent. This also has a follow-on effect for MIS 3, where the advances must have been focused on Labrador, rather than Keewaitin, at least until after 40000. I really think that Keewaitin must have remained larger throughout MIS 3.

I think at 40000, I will significantly expand Labrador sector, because the flow indicators that show northwards flow from central Quebec might just be the MIS 4 flows. This would make it a lot easier to account for global sea level.



now onto the more recent stuff...

I need to create 2500 year inteval ice margins, which means I have to add a few intermediate steps. Otherwise I am keeping it the same as Gowan et al 2016 and Khosravi 2018. The intermediate steps are simply halfway between. For the moment, I am not touching any of the other margins, which will likely need to be edited to take into account the changes in the Greenland, Cordillera and Iceland margins.

27500 - added. Will need to expand the preceding times later to account for the Ward and Clague margins that were implemented at 30000.

For now, I am using the 30000 extent for Greenland all the way to the LGM

25000 - margin from Clague and Ward (2011). Merged the 30000 margin for greenland with the old 25000 margin, which should give more ice extent

22500 - should be slightly less than the current 22000 margin

20000 - again, merged the Greenland at 30000 here. I am also removing the retreated Hudson Strait that was in Dyke's reconstruction, for maximum ice. They Heinrich event came after.

17500 - intermediate of 18000 and 17000 margins

For Greenland, I am going to take the plunge and replace Dyke's margins with the present margin during the Holocene. The review by Funder et al (2011) says that the margin was behind the present day margin during the Holocene thermal maximum. 

7500 - use present day margin from RTOPO

5000 - for much of Greenland, use the modern boundary, but reduce the extent in western Greenland, which continued to retreat until the mid-Holocene. Funder et al 2011 gave estimates between 10 and 80 km for West Greenland retreat (at least around Disko Bugt), I will put it at 40 km

2500 - put it between 5000 and 0

Before going back to 500 year interval reconstructing, the other times will need to be adjusted as well!

-----------------------------
2019/05/10
-----------------------------

I think that the modelled sea level during stage 3 is too high.... even in a scenario of high sea level, it is currently as high as -15 m. I think the place where the ice could have persisted is the Torngat Mountain region. There is a paper by Marquette et al (2004) where they did a bunch of cosmogenic dates there, and found the bedrock had high levels of inheritance, giving minimum exposures of 100000-200000 years. If the ice sheet was minimally erosive, it could have just sat there after becoming glaciated during the stage 4 glaciation. This is what I am going to do now, so all of the time slices show this. 

60000 - ice reatches the shelf edge

I'm also bringing the Keewaitin sector southwards in the 37500-30000 time slices

Also making the Baffin Island glacial area bigger

Added a persistent ice cap on Newfoundland through all of MIS 3

-----------------------------
2019/05/13
-----------------------------

I received some unpublished (possibly) MIS 3 aged marine shell located in northern Manitoba. If these are truly MIS 3, then it indicates that the MIS 3 sea level in southwestern Hudson Bay was quite high, much higher than what I have (which shows subaerial conditions in MIS 3). I think this means I need to have ice persisting there much longer than I currently have. Binge-purge is back, baby. 

I still think the advance to MIS 4 has to be from a Labrador-centered ice sheet and that Keewaitin was not important until after the MIS 4 maximum. I am basically doing an earlier epansion. After intial retreat, the orientation of the ice sheet should be such that during MIS 3, the ice should be flowing southeastward.

Prior to 60000, I am shifting the Labrador sector backwards by one time step each. Then 60000, I am making it even more extensive.

Since binge/purge is back, the 60000 centered Heinrich even now simply has retreat in Hudson Strait. Got to keep northern Manitoba glaciated.

I really think that I am going to have to shift the timing of the minimum point backwards, so that the 45000 Heinrich event clears out Hudson Bay. Then the 42500 time slice shows an ice free Hudson Bay, and regrowth happening by 40000. Overall, everything is larger in extent.

-----------------------------
2019/06/17
-----------------------------

At this point, it is clear that margins need updating in the east. This is particularly true in the Atlantic side. There is a conceptual model by Shaw et al 2006, which I will use. It shows the maximum extent to be further out than Dyke 2004.

25000 - Shaw doesn't state exactly when the maximum extent happened, but his earliest dated time slice (23.5 cal kyr BP) shows there is already retreat. I am putting the 25000 margin at the shelf edge.

22500 - Shaw showed there was some retreat in ice stream areas by this time. He has margin reconstructions at 23.5 and 21 cal kyr BP, so I set it to be between.

19000 and 21000 - used the margins suggested by Shaw

20000 - intermediate of 19000 and 21000

17000 - based on the 16800 margin by Shaw, slightly expanded

17500 - reading in Shaw's paper, he suggested that ice sheet retreat was very close to the 16800 margin, and it is possible the extent was still quite far out at 17500. I will have this much closer to the 20000 limit to show this.

15000 - this is based on the 14800 margin by Shaw

Occhietti et al (2011) created a revised margin chronology from 15500 to present for the eastern Canadian mainland. Unlike Dyke's reconstruction, they have ice going to the shelf edge in Labrador. This will help the sea level curves a lot, and I will do this for all time steps for the LGM.

15000 - based on Occhietti 2011 15500 margin

12500 - based on Occhietti 2011 12700 margin

10000 - using the 10000 margin from Occhietti 2011

7500 - using the 7500 margin from Occhietti 2011

For LGM, I am going to put the margin at the shelf edge along the Labrador coast. Set maximum at 19000, slightly retreated at 17500. Set to maximum for 20000, 21000, 22500 and 25000

The other intermediate times will need to be edited as well when I need them.


-----------------------------
2019/08/19
-----------------------------

The margin at 15000 has been adjusted in the suture zone between the Laurentide and Cordillera ice sheet. This is because the ice sheet was really low on the Laurentide side, producing a double-ridge that I don't think is realistic considering the location.


-----------------------------
2019/28/19
-----------------------------

Added a second MIS 3 scenario "a1". In this case, Hudson Bay does not become ice free, and instead the margin remains close to the Great Lakes and in southern Manitoba. Overall, the ice sheet extent west of Lake Huron is larger. This should give a better picture of the maximum possible ice sheet configuration in MIS 3 if the data from within Hudson Bay are ignored.
